User login
Regional Variation in Standardized Costs of Care at Children’s Hospitals
With some areas of the country spending close to 3 times more on healthcare than others, regional variation in healthcare spending has been the focus of national attention.1-7 Since 1973, the Dartmouth Institute has studied regional variation in healthcare utilization and spending and concluded that variation is “unwarranted” because it is driven by providers’ practice patterns rather than differences in medical need, patient preferences, or evidence-based medicine.8-11 However, critics of the Dartmouth Institute’s findings argue that their approach does not adequately adjust for community-level income, and that higher costs in some areas reflect greater patient needs that are not reflected in illness acuity alone.12-14
While Medicare data have made it possible to study variations in spending for the senior population, fragmentation of insurance coverage and nonstandardized data structures make studying the pediatric population more difficult. However, the Children’s Hospital Association’s (CHA) Pediatric Health Information System (PHIS) has made large-scale comparisons more feasible. To overcome challenges associated with using charges and nonuniform cost data, PHIS-derived standardized costs provide new opportunities for comparisons.15,16 Initial analyses using PHIS data showed significant interhospital variations in costs of care,15 but they did not adjust for differences in populations and assess the drivers of variation. A more recent study that controlled for payer status, comorbidities, and illness severity found that intensive care unit (ICU) utilization varied significantly for children hospitalized for asthma, suggesting that hospital practice patterns drive differences in cost.17
This study uses PHIS data to analyze regional variations in standardized costs of care for 3 conditions for which children are hospitalized. To assess potential drivers of variation, the study investigates the effects of patient-level demographic and illness-severity variables as well as encounter-level variables on costs of care. It also estimates cost savings from reducing variation.
METHODS
Data Source
This retrospective cohort study uses the PHIS database (CHA, Overland Park, KS), which includes 48 freestanding children’s hospitals located in noncompeting markets across the United States and accounts for approximately 20% of pediatric hospitalizations. PHIS includes patient demographics, International Classification of Diseases, 9th Revision (ICD-9) diagnosis and procedure codes, as well as hospital charges. In addition to total charges, PHIS reports imaging, laboratory, pharmacy, and “other” charges. The “other” category aggregates clinical, supply, room, and nursing charges (including facility fees and ancillary staff services).
Inclusion Criteria
Inpatient- and observation-status hospitalizations for asthma, diabetic ketoacidosis (DKA), and acute gastroenteritis (AGE) at 46 PHIS hospitals from October 2014 to September 2015 were included. Two hospitals were excluded because of missing data. Hospitalizations for patients >18 years were excluded.
Hospitalizations were categorized by using All Patient Refined-Diagnosis Related Groups (APR-DRGs) version 24 (3M Health Information Systems, St. Paul, MN)18 based on the ICD-9 diagnosis and procedure codes assigned during the episode of care. Analyses included APR-DRG 141 (asthma), primary diagnosis ICD-9 codes 250.11 and 250.13 (DKA), and APR-DRG 249 (AGE). ICD-9 codes were used for DKA for increased specificity.19 These conditions were chosen to represent 3 clinical scenarios: (1) a diagnosis for which hospitals differ on whether certain aspects of care are provided in the ICU (asthma), (2) a diagnosis that frequently includes care in an ICU (DKA), and (3) a diagnosis that typically does not include ICU care (AGE).19
Study Design
To focus the analysis on variation in resource utilization across hospitals rather than variations in hospital item charges, each billed resource was assigned a standardized cost.15,16 For each clinical transaction code (CTC), the median unit cost was calculated for each hospital. The median of the hospital medians was defined as the standardized unit cost for that CTC.
The primary outcome variable was the total standardized cost for the hospitalization adjusted for patient-level demographic and illness-severity variables. Patient demographic and illness-severity covariates included age, race, gender, ZIP code-based median annual household income (HHI), rural-urban location, distance from home ZIP code to the hospital, chronic condition indicator (CCI), and severity-of-illness (SOI). When assessing drivers of variation, encounter-level covariates were added, including length of stay (LOS) in hours, ICU utilization, and 7-day readmission (an imprecise measure to account for quality of care during the index visit). The contribution of imaging, laboratory, pharmacy, and “other” costs was also considered.
Median annual HHI for patients’ home ZIP code was obtained from 2010 US Census data. Community-level HHI, a proxy for socioeconomic status (SES),20,21 was classified into categories based on the 2015 US federal poverty level (FPL) for a family of 422: HHI-1 = ≤ 1.5 × FPL; HHI-2 = 1.5 to 2 × FPL; HHI-3 = 2 to 3 × FPL; HHI-4 = ≥ 3 × FPL. Rural-urban commuting area (RUCA) codes were used to determine the rural-urban classification of the patient’s home.23 The distance from home ZIP code to the hospital was included as an additional control for illness severity because patients traveling longer distances are often more sick and require more resources.24
The Agency for Healthcare Research and Quality CCI classification system was used to identify the presence of a chronic condition.25 For asthma, CCI was flagged if the patient had a chronic condition other than asthma; for DKA, CCI was flagged if the patient had a chronic condition other than DKA; and for AGE, CCI was flagged if the patient had any chronic condition.
The APR-DRG system provides a 4-level SOI score with each APR-DRG category. Patient factors, such as comorbid diagnoses, are considered in severity scores generated through 3M’s proprietary algorithms.18
For the first analysis, the 46 hospitals were categorized into 7 geographic regions based on 2010 US Census Divisions.26 To overcome small hospital sample sizes, Mountain and Pacific were combined into West, and Middle Atlantic and New England were combined into North East. Because PHIS hospitals are located in noncompeting geographic regions, for the second analysis, we examined hospital-level variation (considering each hospital as its own region).
Data Analysis
To focus the analysis on “typical” patients and produce more robust estimates of central tendencies, the top and bottom 5% of hospitalizations with the most extreme standardized costs by condition were trimmed.27 Standardized costs were log-transformed because of their nonnormal distribution and analyzed by using linear mixed models. Covariates were added stepwise to assess the proportion of the variance explained by each predictor. Post-hoc tests with conservative single-step stepwise mutation model corrections for multiple testing were used to compare adjusted costs. Statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC). P values < 0.05 were considered significant. The Children’s Hospital of Philadelphia Institutional Review Board did not classify this study as human subjects research.
RESULTS
jhm012100818_t1.jpg
During the study period, there were 26,430 hospitalizations for asthma, 5056 for DKA, and 16,274 for AGE (Table 1).
Variation Across Census Regions
After adjusting for patient-level demographic and illness-severity variables, differences in adjusted total standardized costs remained between regions (P < 0.001). Although no region was an outlier compared to the overall mean for any of the conditions, regions were statistically different in pairwise comparison. The East North Central, South Atlantic, and West South Central regions had the highest adjusted total standardized costs for each of the conditions. The East South Central and West North Central regions had the lowest costs for each of the conditions. Adjusted total standardized costs were 120% higher for asthma ($1920 vs $4227), 46% higher for DKA ($7429 vs $10,881), and 150% higher for AGE ($3316 vs $8292) in the highest-cost region compared with the lowest-cost region (Table 2A).
jhm012100818_t2.jpg
Variation Within Census Regions
After controlling for patient-level demographic and illness-severity variables, standardized costs were different across hospitals in the same region (P < 0.001; panel A in Figure). This was true for all conditions in each region. Differences between the lowest- and highest-cost hospitals within the same region ranged from 111% to 420% for asthma, 101% to 398% for DKA, and 166% to 787% for AGE (Table 3).
jhm012100818_t3.jpg
Variation Across Hospitals (Each Hospital as Its Own Region)
One hospital had the highest adjusted standardized costs for all 3 conditions ($9087 for asthma, $28,564 for DKA, and $23,387 for AGE) and was outside of the 95% confidence interval compared with the overall means. The second highest-cost hospitals for asthma ($5977) and AGE ($18,780) were also outside of the 95% confidence interval. After removing these outliers, the difference between the highest- and lowest-cost hospitals was 549% for asthma ($721 vs $4678), 491% for DKA ($2738 vs $16,192), and 681% for AGE ($1317 vs $10,281; Table 2B).
Drivers of Variation Across Census Regions
Patient-level demographic and illness-severity variables explained very little of the variation in standardized costs across regions. For each of the conditions, age, race, gender, community-level HHI, RUCA, and distance from home to the hospital each accounted for <1.5% of variation, while SOI and CCI each accounted for <5%. Overall, patient-level variables explained 5.5%, 3.7%, and 6.7% of variation for asthma, DKA, and AGE.
Encounter-level variables explained a much larger percentage of the variation in costs. LOS accounted for 17.8% of the variation for asthma, 9.8% for DKA, and 8.7% for AGE. ICU utilization explained 6.9% of the variation for asthma and 12.5% for DKA; ICU use was not a major driver for AGE. Seven-day readmissions accounted for <0.5% for each of the conditions. The combination of patient-level and encounter-level variables explained 27%, 24%, and 15% of the variation for asthma, DKA, and AGE.
Drivers of Variation Across Hospitals
For each of the conditions, patient-level demographic variables each accounted for <2% of variation in costs between hospitals. SOI accounted for 4.5% of the variation for asthma and CCI accounted for 5.2% for AGE. Overall, patient-level variables explained 6.9%, 5.3%, and 7.3% of variation for asthma, DKA, and AGE.
Encounter-level variables accounted for a much larger percentage of the variation in cost. LOS explained 25.4% for asthma, 13.3% for DKA, and 14.2% for AGE. ICU utilization accounted for 13.4% for asthma and 21.9% for DKA; ICU use was not a major driver for AGE. Seven-day readmissions accounted for <0.5% for each of the conditions. Together, patient-level and encounter-level variables explained 40%, 36%, and 22% of variation for asthma, DKA, and AGE.
Imaging, Laboratory, Pharmacy, and “Other” Costs
jhm012100818_f1.jpg
The largest contributor to total costs adjusted for patient-level factors for all conditions was “other,” which aggregates room, nursing, clinical, and supply charges (panel B in Figure). When considering drivers of variation, this category explained >50% for each of the conditions. The next largest contributor to total costs was laboratory charges, which accounted for 15% of the variation across regions for asthma and 11% for DKA. Differences in imaging accounted for 18% of the variation for DKA and 15% for AGE. Differences in pharmacy charges accounted for <4% of the variation for each of the conditions. Adding the 4 cost components to the other patient- and encounter-level covariates, the model explained 81%, 78%, and 72% of the variation across census regions for asthma, DKA, and AGE.
For the hospital-level analysis, differences in “other” remained the largest driver of cost variation. For asthma, “other” explained 61% of variation, while pharmacy, laboratory, and imaging each accounted for <8%. For DKA, differences in imaging accounted for 18% of the variation and laboratory charges accounted for 12%. For AGE, imaging accounted for 15% of the variation. Adding the 4 cost components to the other patient- and encounter-level covariates, the model explained 81%, 72%, and 67% of the variation for asthma, DKA, and AGE.
Cost Savings
If all hospitals in this cohort with adjusted standardized costs above the national PHIS average achieved costs equal to the national PHIS average, estimated annual savings in adjusted standardized costs for these 3 conditions would be $69.1 million. If each hospital with adjusted costs above the average within its census region achieved costs equal to its regional average, estimated annual savings in adjusted standardized costs for these conditions would be $25.2 million.
DISCUSSION
This study reported on the regional variation in costs of care for 3 conditions treated at 46 children’s hospitals across 7 geographic regions, and it demonstrated that variations in costs of care exist in pediatrics. This study used standardized costs to compare utilization patterns across hospitals and adjusted for several patient-level demographic and illness-severity factors, and it found that differences in costs of care for children hospitalized with asthma, DKA, and AGE remained both between and within regions.
These variations are noteworthy, as hospitals strive to improve the value of healthcare. If the higher-cost hospitals in this cohort could achieve costs equal to the national PHIS averages, estimated annual savings in adjusted standardized costs for these conditions alone would equal $69.1 million. If higher-cost hospitals relative to the average in their own region reduced costs to their regional averages, annual standardized cost savings could equal $25.2 million for these conditions.
The differences observed are also significant in that they provide a foundation for exploring whether lower-cost regions or lower-cost hospitals achieve comparable quality outcomes.28 If so, studying what those hospitals do to achieve outcomes more efficiently can serve as the basis for the establishment of best practices.29 Standardizing best practices through protocols, pathways, and care-model redesign can reduce potentially unnecessary spending.30
Our findings showed that patient-level demographic and illness-severity covariates, including community-level HHI and SOI, did not consistently explain cost differences. Instead, LOS and ICU utilization were associated with higher costs.17,19 When considering the effect of the 4 cost components on the variation in total standardized costs between regions and between hospitals, the fact that the “other” category accounted for the largest percent of the variation is not surprising, because the cost of room occupancy and nursing services increases with longer LOS and more time in the ICU. Other individual cost components that were major drivers of variation were laboratory utilization for asthma and imaging for DKA and AGE31 (though they accounted for a much smaller proportion of total adjusted costs).19
To determine if these factors are modifiable, more information is needed to explain why practices differ. Many factors may contribute to varying utilization patterns, including differences in capabilities and resources (in the hospital and in the community) and patient volumes. For example, some hospitals provide continuous albuterol for status asthmaticus only in ICUs, while others provide it on regular units.32 But if certain hospitals do not have adequate resources or volumes to effectively care for certain populations outside of the ICU, their higher-value approach (considering quality and cost) may be to utilize ICU beds, even if some other hospitals care for those patients on non-ICU floors. Another possibility is that family preferences about care delivery (such as how long children stay in the hospital) may vary across regions.33
Other evidence suggests that physician practice and spending patterns are strongly influenced by the practices of the region where they trained.34 Because physicians often practice close to where they trained,35,36 this may partially explain how regional patterns are reinforced.
Even considering all mentioned covariates, our model did not fully explain variation in standardized costs. After adding the cost components as covariates, between one-third and one-fifth of the variation remained unexplained. It is possible that this unexplained variation stemmed from unmeasured patient-level factors.
In addition, while proxies for SES, including community-level HHI, did not significantly predict differences in costs across regions, it is possible that SES affected LOS differently in different regions. Previous studies have suggested that lower SES is associated with longer LOS.37 If this effect is more pronounced in certain regions (potentially because of differences in social service infrastructures), SES may be contributing to variations in cost through LOS.
Our findings were subject to limitations. First, this study only examined 3 diagnoses and did not include surgical or less common conditions. Second, while PHIS includes tertiary care, academic, and freestanding children’s hospitals, it does not include general hospitals, which is where most pediatric patients receive care.38 Third, we used ZIP code-based median annual HHI to account for SES, and we used ZIP codes to determine the distance to the hospital and rural-urban location of patients’ homes. These approximations lack precision because SES and distances vary within ZIP codes.39 Fourth, while adjusted standardized costs allow for comparisons between hospitals, they do not represent actual costs to patients or individual hospitals. Additionally, when determining whether variation remained after controlling for patient-level variables, we included SOI as a reflection of illness-severity at presentation. However, in practice, SOI scores may be assigned partially based on factors determined during the hospitalization.18 Finally, the use of other regional boundaries or the selection of different hospitals may yield different results.
CONCLUSION
This study reveals regional variations in costs of care for 3 inpatient pediatric conditions. Future studies should explore whether lower-cost regions or lower-cost hospitals achieve comparable quality outcomes. To the extent that variation is driven by modifiable factors and lower spending does not compromise outcomes, these data may prompt reviews of care models to reduce unwarranted variation and improve the value of care delivery at local, regional, and national levels.
Disclosure
Internal funds from the CHA and The Children’s Hospital of Philadelphia supported the conduct of this work. The authors have no financial interests, relationships, or affiliations relevant to the subject matter or materials discussed in the manuscript to disclose. The authors have no potential conflicts of interest relevant to the subject matter or materials discussed in the manuscript to disclose
1. Fisher E, Skinner J. Making Sense of Geographic Variations in Health Care: The New IOM Report. 2013; http://healthaffairs.org/blog/2013/07/24/making-sense-of-geographic-variations-in-health-care-the-new-iom-report/. Accessed on April 11, 2014.
With some areas of the country spending close to 3 times more on healthcare than others, regional variation in healthcare spending has been the focus of national attention.1-7 Since 1973, the Dartmouth Institute has studied regional variation in healthcare utilization and spending and concluded that variation is “unwarranted” because it is driven by providers’ practice patterns rather than differences in medical need, patient preferences, or evidence-based medicine.8-11 However, critics of the Dartmouth Institute’s findings argue that their approach does not adequately adjust for community-level income, and that higher costs in some areas reflect greater patient needs that are not reflected in illness acuity alone.12-14
While Medicare data have made it possible to study variations in spending for the senior population, fragmentation of insurance coverage and nonstandardized data structures make studying the pediatric population more difficult. However, the Children’s Hospital Association’s (CHA) Pediatric Health Information System (PHIS) has made large-scale comparisons more feasible. To overcome challenges associated with using charges and nonuniform cost data, PHIS-derived standardized costs provide new opportunities for comparisons.15,16 Initial analyses using PHIS data showed significant interhospital variations in costs of care,15 but they did not adjust for differences in populations and assess the drivers of variation. A more recent study that controlled for payer status, comorbidities, and illness severity found that intensive care unit (ICU) utilization varied significantly for children hospitalized for asthma, suggesting that hospital practice patterns drive differences in cost.17
This study uses PHIS data to analyze regional variations in standardized costs of care for 3 conditions for which children are hospitalized. To assess potential drivers of variation, the study investigates the effects of patient-level demographic and illness-severity variables as well as encounter-level variables on costs of care. It also estimates cost savings from reducing variation.
METHODS
Data Source
This retrospective cohort study uses the PHIS database (CHA, Overland Park, KS), which includes 48 freestanding children’s hospitals located in noncompeting markets across the United States and accounts for approximately 20% of pediatric hospitalizations. PHIS includes patient demographics, International Classification of Diseases, 9th Revision (ICD-9) diagnosis and procedure codes, as well as hospital charges. In addition to total charges, PHIS reports imaging, laboratory, pharmacy, and “other” charges. The “other” category aggregates clinical, supply, room, and nursing charges (including facility fees and ancillary staff services).
Inclusion Criteria
Inpatient- and observation-status hospitalizations for asthma, diabetic ketoacidosis (DKA), and acute gastroenteritis (AGE) at 46 PHIS hospitals from October 2014 to September 2015 were included. Two hospitals were excluded because of missing data. Hospitalizations for patients >18 years were excluded.
Hospitalizations were categorized by using All Patient Refined-Diagnosis Related Groups (APR-DRGs) version 24 (3M Health Information Systems, St. Paul, MN)18 based on the ICD-9 diagnosis and procedure codes assigned during the episode of care. Analyses included APR-DRG 141 (asthma), primary diagnosis ICD-9 codes 250.11 and 250.13 (DKA), and APR-DRG 249 (AGE). ICD-9 codes were used for DKA for increased specificity.19 These conditions were chosen to represent 3 clinical scenarios: (1) a diagnosis for which hospitals differ on whether certain aspects of care are provided in the ICU (asthma), (2) a diagnosis that frequently includes care in an ICU (DKA), and (3) a diagnosis that typically does not include ICU care (AGE).19
Study Design
To focus the analysis on variation in resource utilization across hospitals rather than variations in hospital item charges, each billed resource was assigned a standardized cost.15,16 For each clinical transaction code (CTC), the median unit cost was calculated for each hospital. The median of the hospital medians was defined as the standardized unit cost for that CTC.
The primary outcome variable was the total standardized cost for the hospitalization adjusted for patient-level demographic and illness-severity variables. Patient demographic and illness-severity covariates included age, race, gender, ZIP code-based median annual household income (HHI), rural-urban location, distance from home ZIP code to the hospital, chronic condition indicator (CCI), and severity-of-illness (SOI). When assessing drivers of variation, encounter-level covariates were added, including length of stay (LOS) in hours, ICU utilization, and 7-day readmission (an imprecise measure to account for quality of care during the index visit). The contribution of imaging, laboratory, pharmacy, and “other” costs was also considered.
Median annual HHI for patients’ home ZIP code was obtained from 2010 US Census data. Community-level HHI, a proxy for socioeconomic status (SES),20,21 was classified into categories based on the 2015 US federal poverty level (FPL) for a family of 422: HHI-1 = ≤ 1.5 × FPL; HHI-2 = 1.5 to 2 × FPL; HHI-3 = 2 to 3 × FPL; HHI-4 = ≥ 3 × FPL. Rural-urban commuting area (RUCA) codes were used to determine the rural-urban classification of the patient’s home.23 The distance from home ZIP code to the hospital was included as an additional control for illness severity because patients traveling longer distances are often more sick and require more resources.24
The Agency for Healthcare Research and Quality CCI classification system was used to identify the presence of a chronic condition.25 For asthma, CCI was flagged if the patient had a chronic condition other than asthma; for DKA, CCI was flagged if the patient had a chronic condition other than DKA; and for AGE, CCI was flagged if the patient had any chronic condition.
The APR-DRG system provides a 4-level SOI score with each APR-DRG category. Patient factors, such as comorbid diagnoses, are considered in severity scores generated through 3M’s proprietary algorithms.18
For the first analysis, the 46 hospitals were categorized into 7 geographic regions based on 2010 US Census Divisions.26 To overcome small hospital sample sizes, Mountain and Pacific were combined into West, and Middle Atlantic and New England were combined into North East. Because PHIS hospitals are located in noncompeting geographic regions, for the second analysis, we examined hospital-level variation (considering each hospital as its own region).
Data Analysis
To focus the analysis on “typical” patients and produce more robust estimates of central tendencies, the top and bottom 5% of hospitalizations with the most extreme standardized costs by condition were trimmed.27 Standardized costs were log-transformed because of their nonnormal distribution and analyzed by using linear mixed models. Covariates were added stepwise to assess the proportion of the variance explained by each predictor. Post-hoc tests with conservative single-step stepwise mutation model corrections for multiple testing were used to compare adjusted costs. Statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC). P values < 0.05 were considered significant. The Children’s Hospital of Philadelphia Institutional Review Board did not classify this study as human subjects research.
RESULTS
jhm012100818_t1.jpg
During the study period, there were 26,430 hospitalizations for asthma, 5056 for DKA, and 16,274 for AGE (Table 1).
Variation Across Census Regions
After adjusting for patient-level demographic and illness-severity variables, differences in adjusted total standardized costs remained between regions (P < 0.001). Although no region was an outlier compared to the overall mean for any of the conditions, regions were statistically different in pairwise comparison. The East North Central, South Atlantic, and West South Central regions had the highest adjusted total standardized costs for each of the conditions. The East South Central and West North Central regions had the lowest costs for each of the conditions. Adjusted total standardized costs were 120% higher for asthma ($1920 vs $4227), 46% higher for DKA ($7429 vs $10,881), and 150% higher for AGE ($3316 vs $8292) in the highest-cost region compared with the lowest-cost region (Table 2A).
jhm012100818_t2.jpg
Variation Within Census Regions
After controlling for patient-level demographic and illness-severity variables, standardized costs were different across hospitals in the same region (P < 0.001; panel A in Figure). This was true for all conditions in each region. Differences between the lowest- and highest-cost hospitals within the same region ranged from 111% to 420% for asthma, 101% to 398% for DKA, and 166% to 787% for AGE (Table 3).
jhm012100818_t3.jpg
Variation Across Hospitals (Each Hospital as Its Own Region)
One hospital had the highest adjusted standardized costs for all 3 conditions ($9087 for asthma, $28,564 for DKA, and $23,387 for AGE) and was outside of the 95% confidence interval compared with the overall means. The second highest-cost hospitals for asthma ($5977) and AGE ($18,780) were also outside of the 95% confidence interval. After removing these outliers, the difference between the highest- and lowest-cost hospitals was 549% for asthma ($721 vs $4678), 491% for DKA ($2738 vs $16,192), and 681% for AGE ($1317 vs $10,281; Table 2B).
Drivers of Variation Across Census Regions
Patient-level demographic and illness-severity variables explained very little of the variation in standardized costs across regions. For each of the conditions, age, race, gender, community-level HHI, RUCA, and distance from home to the hospital each accounted for <1.5% of variation, while SOI and CCI each accounted for <5%. Overall, patient-level variables explained 5.5%, 3.7%, and 6.7% of variation for asthma, DKA, and AGE.
Encounter-level variables explained a much larger percentage of the variation in costs. LOS accounted for 17.8% of the variation for asthma, 9.8% for DKA, and 8.7% for AGE. ICU utilization explained 6.9% of the variation for asthma and 12.5% for DKA; ICU use was not a major driver for AGE. Seven-day readmissions accounted for <0.5% for each of the conditions. The combination of patient-level and encounter-level variables explained 27%, 24%, and 15% of the variation for asthma, DKA, and AGE.
Drivers of Variation Across Hospitals
For each of the conditions, patient-level demographic variables each accounted for <2% of variation in costs between hospitals. SOI accounted for 4.5% of the variation for asthma and CCI accounted for 5.2% for AGE. Overall, patient-level variables explained 6.9%, 5.3%, and 7.3% of variation for asthma, DKA, and AGE.
Encounter-level variables accounted for a much larger percentage of the variation in cost. LOS explained 25.4% for asthma, 13.3% for DKA, and 14.2% for AGE. ICU utilization accounted for 13.4% for asthma and 21.9% for DKA; ICU use was not a major driver for AGE. Seven-day readmissions accounted for <0.5% for each of the conditions. Together, patient-level and encounter-level variables explained 40%, 36%, and 22% of variation for asthma, DKA, and AGE.
Imaging, Laboratory, Pharmacy, and “Other” Costs
jhm012100818_f1.jpg
The largest contributor to total costs adjusted for patient-level factors for all conditions was “other,” which aggregates room, nursing, clinical, and supply charges (panel B in Figure). When considering drivers of variation, this category explained >50% for each of the conditions. The next largest contributor to total costs was laboratory charges, which accounted for 15% of the variation across regions for asthma and 11% for DKA. Differences in imaging accounted for 18% of the variation for DKA and 15% for AGE. Differences in pharmacy charges accounted for <4% of the variation for each of the conditions. Adding the 4 cost components to the other patient- and encounter-level covariates, the model explained 81%, 78%, and 72% of the variation across census regions for asthma, DKA, and AGE.
For the hospital-level analysis, differences in “other” remained the largest driver of cost variation. For asthma, “other” explained 61% of variation, while pharmacy, laboratory, and imaging each accounted for <8%. For DKA, differences in imaging accounted for 18% of the variation and laboratory charges accounted for 12%. For AGE, imaging accounted for 15% of the variation. Adding the 4 cost components to the other patient- and encounter-level covariates, the model explained 81%, 72%, and 67% of the variation for asthma, DKA, and AGE.
Cost Savings
If all hospitals in this cohort with adjusted standardized costs above the national PHIS average achieved costs equal to the national PHIS average, estimated annual savings in adjusted standardized costs for these 3 conditions would be $69.1 million. If each hospital with adjusted costs above the average within its census region achieved costs equal to its regional average, estimated annual savings in adjusted standardized costs for these conditions would be $25.2 million.
DISCUSSION
This study reported on the regional variation in costs of care for 3 conditions treated at 46 children’s hospitals across 7 geographic regions, and it demonstrated that variations in costs of care exist in pediatrics. This study used standardized costs to compare utilization patterns across hospitals and adjusted for several patient-level demographic and illness-severity factors, and it found that differences in costs of care for children hospitalized with asthma, DKA, and AGE remained both between and within regions.
These variations are noteworthy, as hospitals strive to improve the value of healthcare. If the higher-cost hospitals in this cohort could achieve costs equal to the national PHIS averages, estimated annual savings in adjusted standardized costs for these conditions alone would equal $69.1 million. If higher-cost hospitals relative to the average in their own region reduced costs to their regional averages, annual standardized cost savings could equal $25.2 million for these conditions.
The differences observed are also significant in that they provide a foundation for exploring whether lower-cost regions or lower-cost hospitals achieve comparable quality outcomes.28 If so, studying what those hospitals do to achieve outcomes more efficiently can serve as the basis for the establishment of best practices.29 Standardizing best practices through protocols, pathways, and care-model redesign can reduce potentially unnecessary spending.30
Our findings showed that patient-level demographic and illness-severity covariates, including community-level HHI and SOI, did not consistently explain cost differences. Instead, LOS and ICU utilization were associated with higher costs.17,19 When considering the effect of the 4 cost components on the variation in total standardized costs between regions and between hospitals, the fact that the “other” category accounted for the largest percent of the variation is not surprising, because the cost of room occupancy and nursing services increases with longer LOS and more time in the ICU. Other individual cost components that were major drivers of variation were laboratory utilization for asthma and imaging for DKA and AGE31 (though they accounted for a much smaller proportion of total adjusted costs).19
To determine if these factors are modifiable, more information is needed to explain why practices differ. Many factors may contribute to varying utilization patterns, including differences in capabilities and resources (in the hospital and in the community) and patient volumes. For example, some hospitals provide continuous albuterol for status asthmaticus only in ICUs, while others provide it on regular units.32 But if certain hospitals do not have adequate resources or volumes to effectively care for certain populations outside of the ICU, their higher-value approach (considering quality and cost) may be to utilize ICU beds, even if some other hospitals care for those patients on non-ICU floors. Another possibility is that family preferences about care delivery (such as how long children stay in the hospital) may vary across regions.33
Other evidence suggests that physician practice and spending patterns are strongly influenced by the practices of the region where they trained.34 Because physicians often practice close to where they trained,35,36 this may partially explain how regional patterns are reinforced.
Even considering all mentioned covariates, our model did not fully explain variation in standardized costs. After adding the cost components as covariates, between one-third and one-fifth of the variation remained unexplained. It is possible that this unexplained variation stemmed from unmeasured patient-level factors.
In addition, while proxies for SES, including community-level HHI, did not significantly predict differences in costs across regions, it is possible that SES affected LOS differently in different regions. Previous studies have suggested that lower SES is associated with longer LOS.37 If this effect is more pronounced in certain regions (potentially because of differences in social service infrastructures), SES may be contributing to variations in cost through LOS.
Our findings were subject to limitations. First, this study only examined 3 diagnoses and did not include surgical or less common conditions. Second, while PHIS includes tertiary care, academic, and freestanding children’s hospitals, it does not include general hospitals, which is where most pediatric patients receive care.38 Third, we used ZIP code-based median annual HHI to account for SES, and we used ZIP codes to determine the distance to the hospital and rural-urban location of patients’ homes. These approximations lack precision because SES and distances vary within ZIP codes.39 Fourth, while adjusted standardized costs allow for comparisons between hospitals, they do not represent actual costs to patients or individual hospitals. Additionally, when determining whether variation remained after controlling for patient-level variables, we included SOI as a reflection of illness-severity at presentation. However, in practice, SOI scores may be assigned partially based on factors determined during the hospitalization.18 Finally, the use of other regional boundaries or the selection of different hospitals may yield different results.
CONCLUSION
This study reveals regional variations in costs of care for 3 inpatient pediatric conditions. Future studies should explore whether lower-cost regions or lower-cost hospitals achieve comparable quality outcomes. To the extent that variation is driven by modifiable factors and lower spending does not compromise outcomes, these data may prompt reviews of care models to reduce unwarranted variation and improve the value of care delivery at local, regional, and national levels.
Disclosure
Internal funds from the CHA and The Children’s Hospital of Philadelphia supported the conduct of this work. The authors have no financial interests, relationships, or affiliations relevant to the subject matter or materials discussed in the manuscript to disclose. The authors have no potential conflicts of interest relevant to the subject matter or materials discussed in the manuscript to disclose
With some areas of the country spending close to 3 times more on healthcare than others, regional variation in healthcare spending has been the focus of national attention.1-7 Since 1973, the Dartmouth Institute has studied regional variation in healthcare utilization and spending and concluded that variation is “unwarranted” because it is driven by providers’ practice patterns rather than differences in medical need, patient preferences, or evidence-based medicine.8-11 However, critics of the Dartmouth Institute’s findings argue that their approach does not adequately adjust for community-level income, and that higher costs in some areas reflect greater patient needs that are not reflected in illness acuity alone.12-14
While Medicare data have made it possible to study variations in spending for the senior population, fragmentation of insurance coverage and nonstandardized data structures make studying the pediatric population more difficult. However, the Children’s Hospital Association’s (CHA) Pediatric Health Information System (PHIS) has made large-scale comparisons more feasible. To overcome challenges associated with using charges and nonuniform cost data, PHIS-derived standardized costs provide new opportunities for comparisons.15,16 Initial analyses using PHIS data showed significant interhospital variations in costs of care,15 but they did not adjust for differences in populations and assess the drivers of variation. A more recent study that controlled for payer status, comorbidities, and illness severity found that intensive care unit (ICU) utilization varied significantly for children hospitalized for asthma, suggesting that hospital practice patterns drive differences in cost.17
This study uses PHIS data to analyze regional variations in standardized costs of care for 3 conditions for which children are hospitalized. To assess potential drivers of variation, the study investigates the effects of patient-level demographic and illness-severity variables as well as encounter-level variables on costs of care. It also estimates cost savings from reducing variation.
METHODS
Data Source
This retrospective cohort study uses the PHIS database (CHA, Overland Park, KS), which includes 48 freestanding children’s hospitals located in noncompeting markets across the United States and accounts for approximately 20% of pediatric hospitalizations. PHIS includes patient demographics, International Classification of Diseases, 9th Revision (ICD-9) diagnosis and procedure codes, as well as hospital charges. In addition to total charges, PHIS reports imaging, laboratory, pharmacy, and “other” charges. The “other” category aggregates clinical, supply, room, and nursing charges (including facility fees and ancillary staff services).
Inclusion Criteria
Inpatient- and observation-status hospitalizations for asthma, diabetic ketoacidosis (DKA), and acute gastroenteritis (AGE) at 46 PHIS hospitals from October 2014 to September 2015 were included. Two hospitals were excluded because of missing data. Hospitalizations for patients >18 years were excluded.
Hospitalizations were categorized by using All Patient Refined-Diagnosis Related Groups (APR-DRGs) version 24 (3M Health Information Systems, St. Paul, MN)18 based on the ICD-9 diagnosis and procedure codes assigned during the episode of care. Analyses included APR-DRG 141 (asthma), primary diagnosis ICD-9 codes 250.11 and 250.13 (DKA), and APR-DRG 249 (AGE). ICD-9 codes were used for DKA for increased specificity.19 These conditions were chosen to represent 3 clinical scenarios: (1) a diagnosis for which hospitals differ on whether certain aspects of care are provided in the ICU (asthma), (2) a diagnosis that frequently includes care in an ICU (DKA), and (3) a diagnosis that typically does not include ICU care (AGE).19
Study Design
To focus the analysis on variation in resource utilization across hospitals rather than variations in hospital item charges, each billed resource was assigned a standardized cost.15,16 For each clinical transaction code (CTC), the median unit cost was calculated for each hospital. The median of the hospital medians was defined as the standardized unit cost for that CTC.
The primary outcome variable was the total standardized cost for the hospitalization adjusted for patient-level demographic and illness-severity variables. Patient demographic and illness-severity covariates included age, race, gender, ZIP code-based median annual household income (HHI), rural-urban location, distance from home ZIP code to the hospital, chronic condition indicator (CCI), and severity-of-illness (SOI). When assessing drivers of variation, encounter-level covariates were added, including length of stay (LOS) in hours, ICU utilization, and 7-day readmission (an imprecise measure to account for quality of care during the index visit). The contribution of imaging, laboratory, pharmacy, and “other” costs was also considered.
Median annual HHI for patients’ home ZIP code was obtained from 2010 US Census data. Community-level HHI, a proxy for socioeconomic status (SES),20,21 was classified into categories based on the 2015 US federal poverty level (FPL) for a family of 422: HHI-1 = ≤ 1.5 × FPL; HHI-2 = 1.5 to 2 × FPL; HHI-3 = 2 to 3 × FPL; HHI-4 = ≥ 3 × FPL. Rural-urban commuting area (RUCA) codes were used to determine the rural-urban classification of the patient’s home.23 The distance from home ZIP code to the hospital was included as an additional control for illness severity because patients traveling longer distances are often more sick and require more resources.24
The Agency for Healthcare Research and Quality CCI classification system was used to identify the presence of a chronic condition.25 For asthma, CCI was flagged if the patient had a chronic condition other than asthma; for DKA, CCI was flagged if the patient had a chronic condition other than DKA; and for AGE, CCI was flagged if the patient had any chronic condition.
The APR-DRG system provides a 4-level SOI score with each APR-DRG category. Patient factors, such as comorbid diagnoses, are considered in severity scores generated through 3M’s proprietary algorithms.18
For the first analysis, the 46 hospitals were categorized into 7 geographic regions based on 2010 US Census Divisions.26 To overcome small hospital sample sizes, Mountain and Pacific were combined into West, and Middle Atlantic and New England were combined into North East. Because PHIS hospitals are located in noncompeting geographic regions, for the second analysis, we examined hospital-level variation (considering each hospital as its own region).
Data Analysis
To focus the analysis on “typical” patients and produce more robust estimates of central tendencies, the top and bottom 5% of hospitalizations with the most extreme standardized costs by condition were trimmed.27 Standardized costs were log-transformed because of their nonnormal distribution and analyzed by using linear mixed models. Covariates were added stepwise to assess the proportion of the variance explained by each predictor. Post-hoc tests with conservative single-step stepwise mutation model corrections for multiple testing were used to compare adjusted costs. Statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC). P values < 0.05 were considered significant. The Children’s Hospital of Philadelphia Institutional Review Board did not classify this study as human subjects research.
RESULTS
jhm012100818_t1.jpg
During the study period, there were 26,430 hospitalizations for asthma, 5056 for DKA, and 16,274 for AGE (Table 1).
Variation Across Census Regions
After adjusting for patient-level demographic and illness-severity variables, differences in adjusted total standardized costs remained between regions (P < 0.001). Although no region was an outlier compared to the overall mean for any of the conditions, regions were statistically different in pairwise comparison. The East North Central, South Atlantic, and West South Central regions had the highest adjusted total standardized costs for each of the conditions. The East South Central and West North Central regions had the lowest costs for each of the conditions. Adjusted total standardized costs were 120% higher for asthma ($1920 vs $4227), 46% higher for DKA ($7429 vs $10,881), and 150% higher for AGE ($3316 vs $8292) in the highest-cost region compared with the lowest-cost region (Table 2A).
jhm012100818_t2.jpg
Variation Within Census Regions
After controlling for patient-level demographic and illness-severity variables, standardized costs were different across hospitals in the same region (P < 0.001; panel A in Figure). This was true for all conditions in each region. Differences between the lowest- and highest-cost hospitals within the same region ranged from 111% to 420% for asthma, 101% to 398% for DKA, and 166% to 787% for AGE (Table 3).
jhm012100818_t3.jpg
Variation Across Hospitals (Each Hospital as Its Own Region)
One hospital had the highest adjusted standardized costs for all 3 conditions ($9087 for asthma, $28,564 for DKA, and $23,387 for AGE) and was outside of the 95% confidence interval compared with the overall means. The second highest-cost hospitals for asthma ($5977) and AGE ($18,780) were also outside of the 95% confidence interval. After removing these outliers, the difference between the highest- and lowest-cost hospitals was 549% for asthma ($721 vs $4678), 491% for DKA ($2738 vs $16,192), and 681% for AGE ($1317 vs $10,281; Table 2B).
Drivers of Variation Across Census Regions
Patient-level demographic and illness-severity variables explained very little of the variation in standardized costs across regions. For each of the conditions, age, race, gender, community-level HHI, RUCA, and distance from home to the hospital each accounted for <1.5% of variation, while SOI and CCI each accounted for <5%. Overall, patient-level variables explained 5.5%, 3.7%, and 6.7% of variation for asthma, DKA, and AGE.
Encounter-level variables explained a much larger percentage of the variation in costs. LOS accounted for 17.8% of the variation for asthma, 9.8% for DKA, and 8.7% for AGE. ICU utilization explained 6.9% of the variation for asthma and 12.5% for DKA; ICU use was not a major driver for AGE. Seven-day readmissions accounted for <0.5% for each of the conditions. The combination of patient-level and encounter-level variables explained 27%, 24%, and 15% of the variation for asthma, DKA, and AGE.
Drivers of Variation Across Hospitals
For each of the conditions, patient-level demographic variables each accounted for <2% of variation in costs between hospitals. SOI accounted for 4.5% of the variation for asthma and CCI accounted for 5.2% for AGE. Overall, patient-level variables explained 6.9%, 5.3%, and 7.3% of variation for asthma, DKA, and AGE.
Encounter-level variables accounted for a much larger percentage of the variation in cost. LOS explained 25.4% for asthma, 13.3% for DKA, and 14.2% for AGE. ICU utilization accounted for 13.4% for asthma and 21.9% for DKA; ICU use was not a major driver for AGE. Seven-day readmissions accounted for <0.5% for each of the conditions. Together, patient-level and encounter-level variables explained 40%, 36%, and 22% of variation for asthma, DKA, and AGE.
Imaging, Laboratory, Pharmacy, and “Other” Costs
jhm012100818_f1.jpg
The largest contributor to total costs adjusted for patient-level factors for all conditions was “other,” which aggregates room, nursing, clinical, and supply charges (panel B in Figure). When considering drivers of variation, this category explained >50% for each of the conditions. The next largest contributor to total costs was laboratory charges, which accounted for 15% of the variation across regions for asthma and 11% for DKA. Differences in imaging accounted for 18% of the variation for DKA and 15% for AGE. Differences in pharmacy charges accounted for <4% of the variation for each of the conditions. Adding the 4 cost components to the other patient- and encounter-level covariates, the model explained 81%, 78%, and 72% of the variation across census regions for asthma, DKA, and AGE.
For the hospital-level analysis, differences in “other” remained the largest driver of cost variation. For asthma, “other” explained 61% of variation, while pharmacy, laboratory, and imaging each accounted for <8%. For DKA, differences in imaging accounted for 18% of the variation and laboratory charges accounted for 12%. For AGE, imaging accounted for 15% of the variation. Adding the 4 cost components to the other patient- and encounter-level covariates, the model explained 81%, 72%, and 67% of the variation for asthma, DKA, and AGE.
Cost Savings
If all hospitals in this cohort with adjusted standardized costs above the national PHIS average achieved costs equal to the national PHIS average, estimated annual savings in adjusted standardized costs for these 3 conditions would be $69.1 million. If each hospital with adjusted costs above the average within its census region achieved costs equal to its regional average, estimated annual savings in adjusted standardized costs for these conditions would be $25.2 million.
DISCUSSION
This study reported on the regional variation in costs of care for 3 conditions treated at 46 children’s hospitals across 7 geographic regions, and it demonstrated that variations in costs of care exist in pediatrics. This study used standardized costs to compare utilization patterns across hospitals and adjusted for several patient-level demographic and illness-severity factors, and it found that differences in costs of care for children hospitalized with asthma, DKA, and AGE remained both between and within regions.
These variations are noteworthy, as hospitals strive to improve the value of healthcare. If the higher-cost hospitals in this cohort could achieve costs equal to the national PHIS averages, estimated annual savings in adjusted standardized costs for these conditions alone would equal $69.1 million. If higher-cost hospitals relative to the average in their own region reduced costs to their regional averages, annual standardized cost savings could equal $25.2 million for these conditions.
The differences observed are also significant in that they provide a foundation for exploring whether lower-cost regions or lower-cost hospitals achieve comparable quality outcomes.28 If so, studying what those hospitals do to achieve outcomes more efficiently can serve as the basis for the establishment of best practices.29 Standardizing best practices through protocols, pathways, and care-model redesign can reduce potentially unnecessary spending.30
Our findings showed that patient-level demographic and illness-severity covariates, including community-level HHI and SOI, did not consistently explain cost differences. Instead, LOS and ICU utilization were associated with higher costs.17,19 When considering the effect of the 4 cost components on the variation in total standardized costs between regions and between hospitals, the fact that the “other” category accounted for the largest percent of the variation is not surprising, because the cost of room occupancy and nursing services increases with longer LOS and more time in the ICU. Other individual cost components that were major drivers of variation were laboratory utilization for asthma and imaging for DKA and AGE31 (though they accounted for a much smaller proportion of total adjusted costs).19
To determine if these factors are modifiable, more information is needed to explain why practices differ. Many factors may contribute to varying utilization patterns, including differences in capabilities and resources (in the hospital and in the community) and patient volumes. For example, some hospitals provide continuous albuterol for status asthmaticus only in ICUs, while others provide it on regular units.32 But if certain hospitals do not have adequate resources or volumes to effectively care for certain populations outside of the ICU, their higher-value approach (considering quality and cost) may be to utilize ICU beds, even if some other hospitals care for those patients on non-ICU floors. Another possibility is that family preferences about care delivery (such as how long children stay in the hospital) may vary across regions.33
Other evidence suggests that physician practice and spending patterns are strongly influenced by the practices of the region where they trained.34 Because physicians often practice close to where they trained,35,36 this may partially explain how regional patterns are reinforced.
Even considering all mentioned covariates, our model did not fully explain variation in standardized costs. After adding the cost components as covariates, between one-third and one-fifth of the variation remained unexplained. It is possible that this unexplained variation stemmed from unmeasured patient-level factors.
In addition, while proxies for SES, including community-level HHI, did not significantly predict differences in costs across regions, it is possible that SES affected LOS differently in different regions. Previous studies have suggested that lower SES is associated with longer LOS.37 If this effect is more pronounced in certain regions (potentially because of differences in social service infrastructures), SES may be contributing to variations in cost through LOS.
Our findings were subject to limitations. First, this study only examined 3 diagnoses and did not include surgical or less common conditions. Second, while PHIS includes tertiary care, academic, and freestanding children’s hospitals, it does not include general hospitals, which is where most pediatric patients receive care.38 Third, we used ZIP code-based median annual HHI to account for SES, and we used ZIP codes to determine the distance to the hospital and rural-urban location of patients’ homes. These approximations lack precision because SES and distances vary within ZIP codes.39 Fourth, while adjusted standardized costs allow for comparisons between hospitals, they do not represent actual costs to patients or individual hospitals. Additionally, when determining whether variation remained after controlling for patient-level variables, we included SOI as a reflection of illness-severity at presentation. However, in practice, SOI scores may be assigned partially based on factors determined during the hospitalization.18 Finally, the use of other regional boundaries or the selection of different hospitals may yield different results.
CONCLUSION
This study reveals regional variations in costs of care for 3 inpatient pediatric conditions. Future studies should explore whether lower-cost regions or lower-cost hospitals achieve comparable quality outcomes. To the extent that variation is driven by modifiable factors and lower spending does not compromise outcomes, these data may prompt reviews of care models to reduce unwarranted variation and improve the value of care delivery at local, regional, and national levels.
Disclosure
Internal funds from the CHA and The Children’s Hospital of Philadelphia supported the conduct of this work. The authors have no financial interests, relationships, or affiliations relevant to the subject matter or materials discussed in the manuscript to disclose. The authors have no potential conflicts of interest relevant to the subject matter or materials discussed in the manuscript to disclose
1. Fisher E, Skinner J. Making Sense of Geographic Variations in Health Care: The New IOM Report. 2013; http://healthaffairs.org/blog/2013/07/24/making-sense-of-geographic-variations-in-health-care-the-new-iom-report/. Accessed on April 11, 2014.
1. Fisher E, Skinner J. Making Sense of Geographic Variations in Health Care: The New IOM Report. 2013; http://healthaffairs.org/blog/2013/07/24/making-sense-of-geographic-variations-in-health-care-the-new-iom-report/. Accessed on April 11, 2014.
© 2017 Society of Hospital Medicine
Readmission Analysis Using Fault Tree
As physicians strive to increase the value of healthcare delivery, there has been increased focus on improving the quality of care that patients receive while lowering per capita costs. A provision of the Affordable Care Act implemented in 2012 identified all‐cause 30‐day readmission rates as a measure of hospital quality, and as part of the Act's Hospital Readmission and Reduction Program, Medicare now penalizes hospitals with higher than expected all‐cause readmissions rates for adult patients with certain conditions by lowering reimbursements.[1] Although readmissions are not yet commonly used to determine reimbursements for pediatric hospitals, several states are penalizing higher than expected readmission rates for Medicaid enrollees,[2, 3] using an imprecise algorithm to determine which readmissions resulted from low‐quality care during the index admission.[4, 5, 6]
There is growing concern, however, that readmission rates are not an accurate gauge of the quality of care patients receive while in the hospital or during the discharge process to prepare them for their transition home.[7, 8, 9, 10] This is especially true in pediatric settings, where overall readmission rates are much lower than in adult settings, many readmissions are expected as part of a patient's planned course of care, and variation in readmission rates between hospitals is correlated with the percentage of patients with certain complex chronic conditions.[1, 7, 11] Thus, there is increasing agreement that hospitals and external evaluators need to shift the focus from all‐cause readmissions to a reliable, consistent, and fair measure of potentially preventable readmissions.[12, 13] In addition to being a more useful quality metric, analyzing preventable readmissions will help hospitals focus resources on patients with potentially modifiable risk factors and develop meaningful quality‐improvement initiatives to improve inpatient care as well as the discharge process to prepare families for their transition to home.[14]
Although previous studies have attempted to distinguish preventable from nonpreventable readmissions, many reported significant challenges in completing reviews efficiently, achieving consistency in how readmissions were classified, and attaining consensus on final determinations.[12, 13, 14] Studies have also demonstrated that the algorithms some states are using to streamline preventability reviews and determine reimbursements overestimate the rate of potentially preventable readmissions.[4, 5, 6]
To increase the efficiency of preventability reviews and reduce the subjectivity involved in reaching final determinations, while still accounting for the nuances necessary to conduct a fair review, a quality‐improvement team from the Division of General Pediatrics at The Children's Hospital of Philadelphia (CHOP) implemented a fault tree analysis tool based on a framework developed by Howard Parker at Intermountain Primary Children's Hospital. The CHOP team coded this framework into a secure Web‐based data‐collection tool in the form of a decision tree to guide reviewers through a logical progression of questions that result in 1 of 18 root causes of readmissions, 8 of which are considered potentially preventable. We hypothesized that this method would help reviewers efficiently reach consensus on the root causes of hospital readmissions, and thus help the division and the hospital focus efforts on developing relevant quality‐improvement initiatives.
METHODS
Inclusion Criteria and Study Design
This study was conducted at CHOP, a 535‐bed urban, tertiary‐care, freestanding children's hospital with approximately 29,000 annual discharges. Of those discharges, 7000 to 8000 are from the general pediatrics service, meaning that the attending of record was a general pediatrician. Patients were included in the study if (1) they were discharged from the general pediatrics service between January 2014 and December 2014, and (2) they were readmitted to the hospital, for any reason, within 15 days of discharge. Because this analysis was done as part of a quality‐improvement initiative, it focuses on 15‐day, early readmissions to target cases with a higher probability of being potentially preventable from the perspective of the hospital care team.[10, 12, 13] Patients under observation status during the index admission or the readmission were included. However, patients who returned to the emergency department but were not admitted to an inpatient unit were excluded. Objective details about each case, including the patient's name, demographics, chart number, and diagnosis code, were pre‐loaded from EPIC (Epic Systems Corp., Verona, WI) into REDCap (Research Electronic Data Capture;
A panel of 10 general pediatricians divided up the cases to perform retrospective chart reviews. For each case, REDCap guided reviewers through the fault tree analysis. Reviewers met monthly to discuss difficult cases and reach consensus on any identified ambiguities in the process. After all cases were reviewed once, 3 panel members independently reviewed a random selection of cases to measure inter‐rater reliability and confirm reproducibility of final determinations. The inter‐rater reliability statistic was calculated using Stata 12.1 (StataCorp LP, College Station, TX). During chart reviews, panel members were not blinded to the identity of physicians and other staff members caring for the patients under review. CHOP's institutional review board determined this study to be exempt from ongoing review.
Fault Tree Analysis
Using the decision tree framework for analyzing readmissions that was developed at Intermountain Primary Children's Hospital, the REDCap tool prompted reviewers with a series of sequential questions, each with mutually exclusive options. Using imbedded branching logic to select follow‐up questions, the tool guided reviewers to 1 of 18 terminal nodes, each representing a potential root cause of the readmission. Of those 18 potential causes, 8 were considered potentially preventable. A diagram of the fault tree framework, color coded to indicate which nodes were considered potentially preventable, is shown in Figure 1.
RESULTS
In 2014, 7252 patients were discharged from the general pediatrics service at CHOP. Of those patients, 248 were readmitted within 15 days for an overall general pediatrics 15‐day readmission rate of 3.4%.
Preventability Analysis
Of the 248 readmissions, 233 (94.0%) were considered not preventable. The most common cause for readmission, which accounted for 145 cases (58.5%), was a patient developing an unpredictable problem related to the index diagnosis or a natural progression of the disease that required readmission. The second most common cause, which accounted for 53 cases (21.4%), was a patient developing a new condition unrelated to the index diagnosis or a readmission unrelated to the quality of care received during the index stay. The third most frequent cause, which accounted for 11 cases (4.4%), was a legitimate nonclinical readmission due to lack of alternative resources, psychosocial or economic factors, or case‐specific factors. Other nonpreventable causes of readmission, including scheduled readmissions, each accounted for 7 or fewer cases and <3% of total readmissions.
The 15 readmissions considered potentially preventable accounted for 6.0% of total readmissions and 0.2% of total discharges from the general pediatrics service in 2014. The most common cause of preventable readmissions, which accounted for 6 cases, was premature discharge. The second most common cause, which accounted for 4 cases, was a problem resulting from nosocomial or iatrogenic factors. Other potentially preventable causes included delayed detection of problem (3 cases), inappropriate readmission (1 case), and inadequate postdischarge care planning (1 case).
A breakdown of fault tree results, including examples of cases associated with each terminal node, is shown in Table 1. Information about general pediatrics patients and readmitted patients is included in Tables 2 and 3. A breakdown of determinations for each reviewer is included in Supporting Table 1 in the online version of this article.
| Fault Tree Terminal Node | Root Cause of Readmission | No. of Cases | % of Total Readmissions | % Within Preventability Category | % of Total Discharges |
|---|---|---|---|---|---|
| |||||
| 2 (Potentially Preventable) | Problematic condition on discharge. Example:* Index admission: Infant with history of prematurity admitted with RSV and rhinovirus bronchiolitis. Had some waxing and waning symptoms. Just prior to discharge, noted to have increased work of breathing related to feeds. Readmission: 12 hours later with tachypnea, retractions, and hypoxia. | 6 | 2.4% | 40.0% | 0.08% |
| 3 (Potentially Preventable) | Nosocomial/Iatrogenic factors. Example*: Index admission: Toddler admitted with fever and neutropenia. Treated with antibiotics 24 hours. Diagnosed with viral illness and discharged home. Readmission: symptomatic Clostridum difficile infection. | 4 | 1.6% | 26.7% | 0.06% |
| 8 (Potentially Preventable) | Detection/treatment of problem was delayed and not appropriately facilitated. Example:* Index admission: Preteen admitted with abdominal pain, concern for appendicitis. Ultrasound and abdominal MRI negative for appendicitis. Symptoms improved. Tolerated PO. Readmission: 3 days later with similar abdominal pain. Diagnosed with constipation with significant improvement following clean‐out. | 3 | 1.2% | 20.0% | 0.04% |
| 1 (Potentially Preventable) | Inappropriate readmission. Example:* Index admission: Infant with laryngomalacia admitted with bronchiolitis. Readmission: Continued mild bronchiolitis symptoms but did not require oxygen or suctioning, normal CXR. | 1 | 0.4% | 6.7% | 0.01% |
| 5 (Potentially Preventable) | Resulted from inadequate postdischarge care planning. Example:* Index diagnosis: Infant with vomiting, prior admissions, and extensive evaluation, diagnosed with milk protein allergy and GERD. PPI increased. Readmission: Persistent symptoms, required NGT feeds supplementation. | 1 | 0.4% | 6.7% | 0.01% |
| 4 (Potentially Preventable) | Resulted from a preventable complication and hospital/physician did not take the appropriate steps to minimize likelihood of complication. | ||||
| 6 (Potentially Preventable) | Resulted from improper care by patient/family and effort by hospital/physician to ensure correct postdischarge care was inadequate. | ||||
| 7 (Potentially Preventable) | Resulted from inadequate care by community services and effort by hospital/physician to ensure correct postdischarge care was inadequate. | ||||
| 15 | 6.0% | 100% | 0.2% | ||
| 12 (Not Preventable) | Problem was unpredictable. Example:* Index admission: Infant admitted with gastroenteritis and dehydration with an anion gap metabolic acidosis. Vomiting and diarrhea improved, rehydrated, acidosis improved. Readmission: 1 day later, presented with emesis and fussiness. Readmitted for metabolic acidosis. | 145 | 58.5% | 62.2% | 2.00% |
| 10 (Not Preventable) | Patient developed new condition unrelated to index diagnosis or quality of care. Example:* Index admission: Toddler admitted with cellulitis. Readmission: Bronchiolitis (did not meet CDC guidelines for nosocomial infection). | 53 | 21.4% | 22.7% | 0.73% |
| 9 (Not Preventable) | Legitimate nonclinical readmission. Example:* Index admission: Infant admitted with second episode of bronchiolitis. Readmission: 4 days later with mild diarrhea. Tolerated PO challenge in emergency department. Admitted due to parental anxiety. | 11 | 4.4% | 4.7% | 0.15% |
| 17 (Not Preventable) | Problem resulted from improper care by patient/family but effort by hospital/physician to ensure correct postdischarge care was appropriate. Example:* Index admission: Infant admitted with diarrhea, diagnosed with milk protein allergy. Discharged on soy formula. Readmission: Developed vomiting and diarrhea with cow milk formula. | 7 | 2.8% | 3.0% | 0.10% |
| 11 (Not Preventable) | Scheduled readmission. Example:* Index admission: Infant with conjunctivitis and preseptal cellulitis with nasolacrimal duct obstruction. Readmission: Postoperatively following scheduled nasolacrimal duct repair. | 7 | 2.8% | 3.0% | 0.10% |
| 14 (Not Preventable) | Detection/treatment of problem was delayed, but earlier detection was not feasible. Example:* Index admission: Preteen admitted with fever, abdominal pain, and elevated inflammatory markers. Fever resolved and symptoms improved. Diagnosed with unspecified viral infection. Readmission: 4 days later with lower extremity pyomyositis and possible osteomyelitis. | 4 | 1.6% | 1.7% | 0.06% |
| 15 (Not Preventable) | Detection/treatment of problem was delayed, earlier detection was feasible, but detection was appropriately facilitated. Example:* Index admission: Infant with history of laryngomalacia and GER admitted with an ALTE. No events during hospitalization. Appropriate workup and cleared by consultants for discharge. Zantac increased. Readmission: Infant had similar ALTE events within a week after discharge. Ultimately underwent supraglottoplasty. | 2 | 0.8% | 0.9% | 0.03% |
| 13 (Not Preventable) | Resulted from preventable complication but efforts to minimize likelihood were appropriate. Example:* Index admission: Patient on GJ feeds admitted for dislodged GJ. Extensive conversations between primary team and multiple consulting services regarding best type of tube. Determined that no other tube options were appropriate. Temporizing measures were initiated. Readmission: GJ tube dislodged again. | 2 | 0.8% | 0.9% | 0.03% |
| 18 (Not Preventable) | Resulted from medication side effect (after watch period). Example:* Index admission: Preteen with MSSA bacteremia spread to other organs. Sent home on appropriate IV antibiotics. Readmission: Fever, rash, increased LFTs. Blood cultures negative. Presumed drug reaction. Fevers resolved with alternate medication. | 2 | 0.8% | 0.9% | 0.03% |
| 16 (Not Preventable) | Resulted from inadequate care by community services, but effort by hospital/physician to ensure correct postdischarge care was appropriate. | ||||
| 233 | 94.0% | 100% | 3.2% | ||
| Fault Tree Terminal Node | Root Cause of Potentially Preventable Readmission with Case Descriptions* |
|---|---|
| |
| 2 (Potentially Preventable) | Problematic condition on discharge |
| Case 1: Index admission: Infant with history of prematurity admitted with RSV and rhinovirus bronchiolitis. Had some waxing and waning symptoms. Just prior to discharge, noted to have increased work of breathing related to feeds. Readmission: 12 hours later with tachypnea, retractions, and hypoxia. | |
| Case 2: Index admission: Toddler admitted with febrile seizure in setting of gastroenteritis. Poor PO intake during hospitalization. Readmission: 1 day later with dehydration. | |
| Case 3: Index admission: Infant admitted with a prolonged complex febrile seizure. Workup included an unremarkable lumbar puncture. No additional seizures. No inpatient imaging obtained. Readmission: Abnormal outpatient MRI requiring intervention. | |
| Case 4: Index admission: Teenager with wheezing and history of chronic daily symptoms. Discharged <24 hours later on albuterol every 4 hours and prednisone. Readmission: 1 day later, seen by primary care physician with persistent asthma flare. | |
| Case 5: Index admission: Exfull‐term infant admitted with bronchiolitis, early in course. At time of discharge, had been off oxygen for 24 hours, but last recorded respiratory rate was >70. Readmission: 1 day later due to continued tachypnea and increased work of breathing. No hypoxia. CXR normal. | |
| Case 6: Exfull‐term infant admitted with bilious emesis, diarrhea, and dehydration. Ultrasound of pylorus, UGI, and BMP all normal. Tolerated oral intake but had emesis and loose stools prior to discharge. Readmission: <48 hours later with severe metabolic acidosis. | |
| 3 (Potentially Preventable) | Nosocomial/ematrogenic factors |
| Case 1: Index admission: Toddler admitted with fever and neutropenia. Treated with antibiotics 24 hours. Diagnosed with viral illness and discharged home. Readmission: Symptomatic Clostridum difficile infection. | |
| Case 2: Index admission: Patient with autism admitted with viral gastroenteritis. Readmission: Presumed nosocominal upper respiratory infection. | |
| Case 3: Index admission: Infant admitted with bronchiolitis. Recovered from initial infection. Readmission: New upper respiratory infection and presumed nosocomial infection. | |
| Case 4: Index admission: <28‐day‐old full‐term neonate presenting with neonatal fever and rash. Full septic workup performed and all cultures negative at 24 hours. Readmission: CSF culture positive at 36 hours and readmitted while awaiting speciation. Discharged once culture grew out a contaminant. | |
| 8 (Potentially Preventable) | Detection/treatment of problem was delayed and/or not appropriately facilitated |
| Case 1: Index admission: Preteen admitted with abdominal pain, concern for appendicitis. Ultrasound and MRI abdomen negative for appendicitis. Symptoms improved. Tolerated PO. Readmission: 3 days later with similar abdominal pain. Diagnosed with constipation with significant improvement following clean‐out. | |
| Case 2: Index admission: Infant with history of macrocephaly presented with fever and full fontanelle. Head CT showed mild prominence of the extra‐axial space, and lumbar puncture was normal. Readmission: Patient developed torticollis. MRI demonstrated a malignant lesion. | |
| Case 3: Index admission: School‐age child with RLQ abdominal pain, fever, leukocytosis, and indeterminate RLQ abdominal ultrasound. Twelve‐hour observation with no further fevers. Pain and appetite improved. Readmission: 1 day later with fever, anorexia, and abdominal pain. RLQ ultrasound unchanged. Appendectomy performed with inflamed appendix. | |
| 1 (Potentially Preventable) | Inappropriate readmission |
| Case 1: Index admission: Infant with laryngomalacia admitted with bronchiolitis. Readmission: Continued mild bronchiolitis symptoms but did not require oxygen or suctioning. Normal CXR. | |
| 5 (Potentially Preventable) | Resulted from inadequate postdischarge care planning |
| Case 1: Index diagnosis: Infant with vomiting, prior admissions, and extensive evaluation, diagnosed with milk protein allergy and GERD. PPI increased. Readmission: Persistent symptoms, required NGT feeds supplementation. | |
| All General Pediatrics Patients in 2014 | General Pediatric Readmitted Patients in 2014 | ||||
|---|---|---|---|---|---|
| Major Diagnosis Category at Index Admission | No. | % | Major Diagnosis Category at Index Admission | No. | % |
| |||||
| Respiratory | 2,723 | 37.5% | Respiratory | 79 | 31.9% |
| Digestive | 748 | 10.3% | Digestive | 41 | 16.5% |
| Ear, nose, mouth, throat | 675 | 9.3% | Ear, nose, mouth, throat | 24 | 9.7% |
| Skin, subcutaneous tissue | 480 | 6.6% | Musculoskeletal and connective tissue | 14 | 5.6% |
| Infectious, parasitic, systemic | 455 | 6.3% | Nervous | 13 | 5.2% |
| Factors influencing health status | 359 | 5.0% | Endocrine, nutritional, metabolic | 13 | 5.2% |
| Endocrine, nutritional, metabolic | 339 | 4.7% | Infectious, parasitic, systemic | 12 | 4.8% |
| Nervous | 239 | 3.3% | Newborn, neonate, perinatal period | 11 | 4.4% |
| Musculoskeletal and connective tissue | 228 | 3.1% | Hepatobiliary system and pancreas | 8 | 3.2% |
| Newborn, neonate, perinatal period | 206 | 2.8% | Skin, subcutaneous tissue | 8 | 3.2% |
| Other* | 800 | 11.0% | Other | 25 | 10.1% |
| Total | 7,252 | 100% | Total | 248 | 100% |
Inter‐Rater Reliability Analysis
A random selection of 50 cases (20% of total readmissions) was selected for a second review to test the tool's inter‐rater reliability. The second review resulted in the same terminal node for 44 (86%) of the cross‐checked files ( = 0.79; 95% confidence interval: 0.60‐0.98). Of the 6 cross‐checked files that ended at different nodes, 5 resulted in the same final determination about preventability. Only 1 of the cross‐checks (2% of total cross‐checked files) resulted in a different conclusion about preventability.
Efficiency Analysis
Reviewers reported that using the tool to reach a determination about preventability took approximately 20 minutes per case. Thus, initial reviews on the 248 cases required approximately 82.6 reviewer hours. Divided across 10 reviewers, this resulted in 8 to 9 hours of review time per reviewer over the year.
DISCUSSION
As part of an effort to direct quality‐improvement initiatives, this project used a Web‐based fault tree tool to identify root causes of general pediatrics readmissions at a freestanding children's hospital and classify them as either preventable or not preventable. The project also investigated the efficiency and inter‐rater reliability of the tool, which was designed to systematically guide physicians through the chart review process to a final determination about preventability. The project confirmed that using the tool helped reviewers reach final determinations about preventability efficiently with a high degree of consistency. It also confirmed that only a very small percentage of general pediatrics 15‐day readmissions are potentially preventable. Specifically, potentially preventable readmissions accounted for only 6.0% of total readmissions and 0.2% of general pediatrics discharges in 2014. Although our analysis focused on 15‐day readmissions, the fault tree methodology can be applied to any timeframe.
Previous studies attempting to distinguish preventable from nonpreventable readmissions, which used a range of methodologies to reach final determinations, reported that their review process was both time intensive and highly subjective. One study, which had 4 reviewers independently review charts and assign each case a preventability score on a 5‐point Likert scale, reported that reviewers disagreed on the final determination in 62.5% of cases.[12] Another study had 2 physicians independently review a selection of cases and assign a preventability score on a scale from 0 to 3. Scores for the 2 reviewers were added together, and cases above a certain composite threshold were classified as preventable. Despite being time‐intensive, this method resulted in only moderate agreement among physicians about the likelihood of preventability (weighted statistic of 0.44).[14] A more recent study, in which 2 physicians independently classified readmissions into 1 of 4 predefined categories, also reported only moderate agreement between reviewers ( = 0.44).[13] Other methods that have been reported include classifying readmissions as preventable only if multiple reviewers independently agreed, and using a third reviewer as a tie‐breaker.[14]
In an attempt to identify potentially preventable readmissions without using chart reviews, 3M (St. Paul, MN) developed its Potentially Preventable Readmissions software (3M‐PPR), which uses administrative data to identify which readmissions were potentially preventable. Although this automated approach is less time intensive, evidence suggests that due to a lack of nuance, the algorithm significantly overestimates the percentage of readmissions that are potentially preventable.[4, 5] A study that used 3M‐PPR to assess 1.7 million hospitalizations across 58 children's hospitals found that the algorithm classified 81% of sickle cell crisis and asthma readmissions, and 83% of bronchiolitis readmissions as potentially preventable.[10, 11] However, many readmissions for asthma and bronchiolitis are due to social factors that are outside of a hospital's direct control,[4, 5] and at many hospitals, readmissions for sickle cell crisis are part of a high‐value care model that weighs length of stay against potential readmissions. In addition, when assessing readmissions 7, 15, and 30 days after discharge, the algorithm classified almost the same percentage as potentially preventable, which is inconsistent with the notion that readmissions are more likely to have been preventable if they occurred closer to the initial discharge.[4, 13] Another study that assessed the performance of the software in the adult population reported that the algorithm performed with 85% sensitivity, but only 28% specificity.[5, 6]
The results of this quality‐improvement project indicate that using the fault tree tool to guide physicians through the chart review process helped address some of the shortcomings of methods reported in previous studies, by increasing the efficiency and reducing the subjectivity of final determinations, while still accounting for the nuances necessary to conduct a fair review. Because the tool provided a systematic framework for reviews, each case was completed in approximately 20 minutes, and because the process was the same for all reviewers, inter‐rater reliability was extremely high. In 86% of cross‐checked cases, the second reviewer ended at the same terminal node in the decision tree as the original reviewer, and in 98% of cross‐checked cases the second reviewer reached the same conclusion about preventability, even if they did not end at the same terminal node. Even accounting for agreement due to chance, the statistic of 0.79 confirmed that there was substantial agreement among reviewers about final determinations. Because the tool is easily adaptable, other hospitals can adopt this framework for their own preventability reviews and quality‐improvement initiatives.
Using the fault tree tool to access root causes of all 15‐day general pediatric readmissions helped the division focus quality‐improvement efforts on the most common causes of potentially preventable readmissions. Because 40% of potentially preventable readmissions were due to premature discharges, this prompted quality‐improvement teams to focus efforts on improving and clarifying the division's discharge criteria and clinical pathways. The division also initiated processes to improve discharge planning, including improved teaching of discharge instructions and having families pick up prescriptions prior to discharge.
Although these results did help the division identify a few areas of focus to potentially reduce readmissions, the fact that the overall 15‐day readmission rate for general pediatrics, as well as the percentage of readmissions and total discharges that were deemed potentially preventable, were so low (3.4%, 6.0%, and 0.2%, respectively), supports those who question whether prioritizing pediatric readmissions is the best place for hospitals to focus quality‐improvement efforts.[10, 12, 15, 16] As these results indicate, most pediatric readmissions are not preventable, and thus consistent with an efficient, effective, timely, patient‐centered, and equitable health system. Other studies have also shown that because overall and condition‐specific readmissions at pediatric hospitals are low, few pediatric hospitals are high or low performing for readmissions, and thus readmission rates are likely not a good measure of hospital quality.[8]
However, other condition‐specific studies of readmissions in pediatrics have indicated that there are some areas of opportunity to identify populations at high risk for readmission. One study found that although pneumonia‐specific 30‐day readmission rates in a national cohort of children hospitalized with pneumonia was only 3.1%, the chances of readmission were higher for children <1 year old, children with chronic comorbidities or complicated pneumonia, and children cared for in hospitals with lower volumes of pneumonia admissions.[17] Another study found that 17.1% of adolescents in a statewide database were readmitted post‐tonsillectomy for pain, nausea, and dehydration.[18] Thus, adapting the tool to identify root causes of condition‐specific or procedure‐specific readmissions, especially for surgical patients, may be an area of opportunity for future quality‐improvement efforts.[5] However, for general pediatrics, shifting the focus from reducing readmissions to improving the quality of care patients receive in the hospital, improving the discharge process, and adopting a population health approach to mitigate external risk factors, may be appropriate.
This project was subject to limitations. First, because it was conducted at a single site and only on general pediatrics patients, results may not be generalizable to other hospitals or other pediatric divisions. Thus, future studies might use the fault tree framework to assess preventability of pediatric readmissions in other divisions or specialties. Second, because readmissions to other hospitals were not included in the sample, the overall readmissions rate is likely underestimated.[19] However, it is unclear how this would affect the rate of potentially preventable readmissions. Third, although the fault tree framework reduced the subjectivity of the review process, there is still a degree of subjectivity inherent at each decision node. To minimize this, reviewers should try to discuss and come to consensus on how they are making determinations at each juncture in the decision tree. Similarly, because reviewers' answers to decision‐tree questions rely heavily on chart documentation, reviews may be compromised by unclear or incomplete documentation. For example, if information about steps the hospital team took to prepare a family for discharge were not properly documented, it would be difficult to determine whether appropriate steps were taken to minimize the likelihood of a complication. In the case of insufficient documentation of relevant social concerns, cases may be incorrectly classified as preventable, because addressing social issues is often not within a hospital's direct control. Finally, because reviewers were not blinded to the original discharging physician, there may have been some unconscious bias of unknown direction in the reviews.
CONCLUSION
Using the Web‐based fault tree tool helped physicians to identify the root causes of hospital readmissions and classify them as preventable or not preventable in a standardized, efficient, and consistent way, while still accounting for the nuances necessary to conduct a fair review. Thus, other hospitals should consider adopting this framework for their own preventability reviews and quality‐improvement initiatives. However, this project also confirmed that only a very small percentage of general pediatrics 15‐day readmissions are potentially preventable, suggesting that general pediatrics readmissions are not an appropriate measure of hospital quality. Instead, adapting the tool to identify root causes of condition‐specific or procedure‐specific readmission rates may be an area of opportunity for future quality‐improvement efforts.
Disclosures: This work was supported through internal funds from The Children's Hospital of Philadelphia. The authors have no financial interests, relationships or affiliations relevant to the subject matter or materials discussed in the article to disclose. The authors have no potential conflicts of interest relevant to the subject matter or materials discussed in the article to disclose.
- , . Pediatric readmissions as a hospital quality measure. JAMA. 2013;309(4):396–398.
- Texas Health and Human Services Commission. Potentially preventable readmissions in the Texas Medicaid population, state fiscal year 2012. Available at: http://www.hhsc.state.tx.us/reports/2013/ppr‐report.pdf. Published November 2013. Accessed August 16, 2015.
- Illinois Department of Healthcare and Family Services. Quality initiative to reduce hospital potentially preventable readmissions (PPR): Status update. Available at: http://www.illinois.gov/hfs/SiteCollectionDocuments/PPRPolicyStatusUpdate.pdf. Published September 3, 2014. Accessed August 16, 2015.
- , , , et al. Rates and impact of potentially preventable readmissions at children's hospitals. J Pediatr. 2015;166(3):613–619.e615.
- , . Preventing pediatric readmissions: which ones and how? J Pediatr. 2015;166(3):519–520.
- , , , , , . Manual and automated methods for identifying potentially preventable readmissions: a comparison in a large healthcare system. BMC Med Inform Decis Mak. 2014;14:28.
- , . Section on hospital medicine leadership and staff. Hosp Pediatr. 2013;3(4):390–393.
- , , , et al. Measuring hospital quality using pediatric readmission and revisit rates. Pediatrics. 2013;132(3):429–436.
- , . Hospital readmissions—not just a measure of quality. JAMA. 2011;306(16):1796–1797.
- , . Preventing readmissions in children: how do we do that? Hosp Pediatr. 2015;5(11):602–604.
- , , , et al. Pediatric readmission prevalence and variability across hospitals. JAMA. 2013;309(4):372–380.
- , , , , , . Preventability of early readmissions at a children's hospital. Pediatrics. 2013;131(1):e171–e181.
- , , , et al. An examination of physician‐, caregiver‐, and disease‐related factors associated with readmission from a pediatric hospital medicine service. Hosp Pediatr. 2015;5(11):566–573.
- , , , et al. Clinical preventability of 30‐day readmission after percutaneous coronary intervention. J Am Heart Assoc. 2014;3(5):e001290.
- . 3M algorithm overestimates preventable pediatric readmissions. Hospitalist News website. Available at: http://www.ehospitalistnews.com/specialty‐focus/pediatrics/single‐article‐page/3m‐algorithm‐overestimates‐preventable‐pediatric‐readmissions.html. Published August 16, 2013. Accessed August 16, 2015.
- . The 30‐day readmission rate: not a quality measure but an accountability measure. An Ounce of Evidence: Health Policy blog. Available at: https://blogs.sph.harvard.edu/ashish‐jha/?s=30‐day+readmission+rate. Published February 14, 2013. Accessed August 16, 2015.
- , , , et al. Readmissions among children previously hospitalized with pneumonia. Pediatrics. 2014;134(1):100–109.
- , , . A population‐based study of acute care revisits following tonsillectomy. J Pediatr. 2015;166(3):607–612.e605.
- , , , et al. Same‐hospital readmission rates as a measure of pediatric quality of care. JAMA Pediatr. 2015;169(10):905–912.
As physicians strive to increase the value of healthcare delivery, there has been increased focus on improving the quality of care that patients receive while lowering per capita costs. A provision of the Affordable Care Act implemented in 2012 identified all‐cause 30‐day readmission rates as a measure of hospital quality, and as part of the Act's Hospital Readmission and Reduction Program, Medicare now penalizes hospitals with higher than expected all‐cause readmissions rates for adult patients with certain conditions by lowering reimbursements.[1] Although readmissions are not yet commonly used to determine reimbursements for pediatric hospitals, several states are penalizing higher than expected readmission rates for Medicaid enrollees,[2, 3] using an imprecise algorithm to determine which readmissions resulted from low‐quality care during the index admission.[4, 5, 6]
There is growing concern, however, that readmission rates are not an accurate gauge of the quality of care patients receive while in the hospital or during the discharge process to prepare them for their transition home.[7, 8, 9, 10] This is especially true in pediatric settings, where overall readmission rates are much lower than in adult settings, many readmissions are expected as part of a patient's planned course of care, and variation in readmission rates between hospitals is correlated with the percentage of patients with certain complex chronic conditions.[1, 7, 11] Thus, there is increasing agreement that hospitals and external evaluators need to shift the focus from all‐cause readmissions to a reliable, consistent, and fair measure of potentially preventable readmissions.[12, 13] In addition to being a more useful quality metric, analyzing preventable readmissions will help hospitals focus resources on patients with potentially modifiable risk factors and develop meaningful quality‐improvement initiatives to improve inpatient care as well as the discharge process to prepare families for their transition to home.[14]
Although previous studies have attempted to distinguish preventable from nonpreventable readmissions, many reported significant challenges in completing reviews efficiently, achieving consistency in how readmissions were classified, and attaining consensus on final determinations.[12, 13, 14] Studies have also demonstrated that the algorithms some states are using to streamline preventability reviews and determine reimbursements overestimate the rate of potentially preventable readmissions.[4, 5, 6]
To increase the efficiency of preventability reviews and reduce the subjectivity involved in reaching final determinations, while still accounting for the nuances necessary to conduct a fair review, a quality‐improvement team from the Division of General Pediatrics at The Children's Hospital of Philadelphia (CHOP) implemented a fault tree analysis tool based on a framework developed by Howard Parker at Intermountain Primary Children's Hospital. The CHOP team coded this framework into a secure Web‐based data‐collection tool in the form of a decision tree to guide reviewers through a logical progression of questions that result in 1 of 18 root causes of readmissions, 8 of which are considered potentially preventable. We hypothesized that this method would help reviewers efficiently reach consensus on the root causes of hospital readmissions, and thus help the division and the hospital focus efforts on developing relevant quality‐improvement initiatives.
METHODS
Inclusion Criteria and Study Design
This study was conducted at CHOP, a 535‐bed urban, tertiary‐care, freestanding children's hospital with approximately 29,000 annual discharges. Of those discharges, 7000 to 8000 are from the general pediatrics service, meaning that the attending of record was a general pediatrician. Patients were included in the study if (1) they were discharged from the general pediatrics service between January 2014 and December 2014, and (2) they were readmitted to the hospital, for any reason, within 15 days of discharge. Because this analysis was done as part of a quality‐improvement initiative, it focuses on 15‐day, early readmissions to target cases with a higher probability of being potentially preventable from the perspective of the hospital care team.[10, 12, 13] Patients under observation status during the index admission or the readmission were included. However, patients who returned to the emergency department but were not admitted to an inpatient unit were excluded. Objective details about each case, including the patient's name, demographics, chart number, and diagnosis code, were pre‐loaded from EPIC (Epic Systems Corp., Verona, WI) into REDCap (Research Electronic Data Capture;
A panel of 10 general pediatricians divided up the cases to perform retrospective chart reviews. For each case, REDCap guided reviewers through the fault tree analysis. Reviewers met monthly to discuss difficult cases and reach consensus on any identified ambiguities in the process. After all cases were reviewed once, 3 panel members independently reviewed a random selection of cases to measure inter‐rater reliability and confirm reproducibility of final determinations. The inter‐rater reliability statistic was calculated using Stata 12.1 (StataCorp LP, College Station, TX). During chart reviews, panel members were not blinded to the identity of physicians and other staff members caring for the patients under review. CHOP's institutional review board determined this study to be exempt from ongoing review.
Fault Tree Analysis
Using the decision tree framework for analyzing readmissions that was developed at Intermountain Primary Children's Hospital, the REDCap tool prompted reviewers with a series of sequential questions, each with mutually exclusive options. Using imbedded branching logic to select follow‐up questions, the tool guided reviewers to 1 of 18 terminal nodes, each representing a potential root cause of the readmission. Of those 18 potential causes, 8 were considered potentially preventable. A diagram of the fault tree framework, color coded to indicate which nodes were considered potentially preventable, is shown in Figure 1.
RESULTS
In 2014, 7252 patients were discharged from the general pediatrics service at CHOP. Of those patients, 248 were readmitted within 15 days for an overall general pediatrics 15‐day readmission rate of 3.4%.
Preventability Analysis
Of the 248 readmissions, 233 (94.0%) were considered not preventable. The most common cause for readmission, which accounted for 145 cases (58.5%), was a patient developing an unpredictable problem related to the index diagnosis or a natural progression of the disease that required readmission. The second most common cause, which accounted for 53 cases (21.4%), was a patient developing a new condition unrelated to the index diagnosis or a readmission unrelated to the quality of care received during the index stay. The third most frequent cause, which accounted for 11 cases (4.4%), was a legitimate nonclinical readmission due to lack of alternative resources, psychosocial or economic factors, or case‐specific factors. Other nonpreventable causes of readmission, including scheduled readmissions, each accounted for 7 or fewer cases and <3% of total readmissions.
The 15 readmissions considered potentially preventable accounted for 6.0% of total readmissions and 0.2% of total discharges from the general pediatrics service in 2014. The most common cause of preventable readmissions, which accounted for 6 cases, was premature discharge. The second most common cause, which accounted for 4 cases, was a problem resulting from nosocomial or iatrogenic factors. Other potentially preventable causes included delayed detection of problem (3 cases), inappropriate readmission (1 case), and inadequate postdischarge care planning (1 case).
A breakdown of fault tree results, including examples of cases associated with each terminal node, is shown in Table 1. Information about general pediatrics patients and readmitted patients is included in Tables 2 and 3. A breakdown of determinations for each reviewer is included in Supporting Table 1 in the online version of this article.
| Fault Tree Terminal Node | Root Cause of Readmission | No. of Cases | % of Total Readmissions | % Within Preventability Category | % of Total Discharges |
|---|---|---|---|---|---|
| |||||
| 2 (Potentially Preventable) | Problematic condition on discharge. Example:* Index admission: Infant with history of prematurity admitted with RSV and rhinovirus bronchiolitis. Had some waxing and waning symptoms. Just prior to discharge, noted to have increased work of breathing related to feeds. Readmission: 12 hours later with tachypnea, retractions, and hypoxia. | 6 | 2.4% | 40.0% | 0.08% |
| 3 (Potentially Preventable) | Nosocomial/Iatrogenic factors. Example*: Index admission: Toddler admitted with fever and neutropenia. Treated with antibiotics 24 hours. Diagnosed with viral illness and discharged home. Readmission: symptomatic Clostridum difficile infection. | 4 | 1.6% | 26.7% | 0.06% |
| 8 (Potentially Preventable) | Detection/treatment of problem was delayed and not appropriately facilitated. Example:* Index admission: Preteen admitted with abdominal pain, concern for appendicitis. Ultrasound and abdominal MRI negative for appendicitis. Symptoms improved. Tolerated PO. Readmission: 3 days later with similar abdominal pain. Diagnosed with constipation with significant improvement following clean‐out. | 3 | 1.2% | 20.0% | 0.04% |
| 1 (Potentially Preventable) | Inappropriate readmission. Example:* Index admission: Infant with laryngomalacia admitted with bronchiolitis. Readmission: Continued mild bronchiolitis symptoms but did not require oxygen or suctioning, normal CXR. | 1 | 0.4% | 6.7% | 0.01% |
| 5 (Potentially Preventable) | Resulted from inadequate postdischarge care planning. Example:* Index diagnosis: Infant with vomiting, prior admissions, and extensive evaluation, diagnosed with milk protein allergy and GERD. PPI increased. Readmission: Persistent symptoms, required NGT feeds supplementation. | 1 | 0.4% | 6.7% | 0.01% |
| 4 (Potentially Preventable) | Resulted from a preventable complication and hospital/physician did not take the appropriate steps to minimize likelihood of complication. | ||||
| 6 (Potentially Preventable) | Resulted from improper care by patient/family and effort by hospital/physician to ensure correct postdischarge care was inadequate. | ||||
| 7 (Potentially Preventable) | Resulted from inadequate care by community services and effort by hospital/physician to ensure correct postdischarge care was inadequate. | ||||
| 15 | 6.0% | 100% | 0.2% | ||
| 12 (Not Preventable) | Problem was unpredictable. Example:* Index admission: Infant admitted with gastroenteritis and dehydration with an anion gap metabolic acidosis. Vomiting and diarrhea improved, rehydrated, acidosis improved. Readmission: 1 day later, presented with emesis and fussiness. Readmitted for metabolic acidosis. | 145 | 58.5% | 62.2% | 2.00% |
| 10 (Not Preventable) | Patient developed new condition unrelated to index diagnosis or quality of care. Example:* Index admission: Toddler admitted with cellulitis. Readmission: Bronchiolitis (did not meet CDC guidelines for nosocomial infection). | 53 | 21.4% | 22.7% | 0.73% |
| 9 (Not Preventable) | Legitimate nonclinical readmission. Example:* Index admission: Infant admitted with second episode of bronchiolitis. Readmission: 4 days later with mild diarrhea. Tolerated PO challenge in emergency department. Admitted due to parental anxiety. | 11 | 4.4% | 4.7% | 0.15% |
| 17 (Not Preventable) | Problem resulted from improper care by patient/family but effort by hospital/physician to ensure correct postdischarge care was appropriate. Example:* Index admission: Infant admitted with diarrhea, diagnosed with milk protein allergy. Discharged on soy formula. Readmission: Developed vomiting and diarrhea with cow milk formula. | 7 | 2.8% | 3.0% | 0.10% |
| 11 (Not Preventable) | Scheduled readmission. Example:* Index admission: Infant with conjunctivitis and preseptal cellulitis with nasolacrimal duct obstruction. Readmission: Postoperatively following scheduled nasolacrimal duct repair. | 7 | 2.8% | 3.0% | 0.10% |
| 14 (Not Preventable) | Detection/treatment of problem was delayed, but earlier detection was not feasible. Example:* Index admission: Preteen admitted with fever, abdominal pain, and elevated inflammatory markers. Fever resolved and symptoms improved. Diagnosed with unspecified viral infection. Readmission: 4 days later with lower extremity pyomyositis and possible osteomyelitis. | 4 | 1.6% | 1.7% | 0.06% |
| 15 (Not Preventable) | Detection/treatment of problem was delayed, earlier detection was feasible, but detection was appropriately facilitated. Example:* Index admission: Infant with history of laryngomalacia and GER admitted with an ALTE. No events during hospitalization. Appropriate workup and cleared by consultants for discharge. Zantac increased. Readmission: Infant had similar ALTE events within a week after discharge. Ultimately underwent supraglottoplasty. | 2 | 0.8% | 0.9% | 0.03% |
| 13 (Not Preventable) | Resulted from preventable complication but efforts to minimize likelihood were appropriate. Example:* Index admission: Patient on GJ feeds admitted for dislodged GJ. Extensive conversations between primary team and multiple consulting services regarding best type of tube. Determined that no other tube options were appropriate. Temporizing measures were initiated. Readmission: GJ tube dislodged again. | 2 | 0.8% | 0.9% | 0.03% |
| 18 (Not Preventable) | Resulted from medication side effect (after watch period). Example:* Index admission: Preteen with MSSA bacteremia spread to other organs. Sent home on appropriate IV antibiotics. Readmission: Fever, rash, increased LFTs. Blood cultures negative. Presumed drug reaction. Fevers resolved with alternate medication. | 2 | 0.8% | 0.9% | 0.03% |
| 16 (Not Preventable) | Resulted from inadequate care by community services, but effort by hospital/physician to ensure correct postdischarge care was appropriate. | ||||
| 233 | 94.0% | 100% | 3.2% | ||
| Fault Tree Terminal Node | Root Cause of Potentially Preventable Readmission with Case Descriptions* |
|---|---|
| |
| 2 (Potentially Preventable) | Problematic condition on discharge |
| Case 1: Index admission: Infant with history of prematurity admitted with RSV and rhinovirus bronchiolitis. Had some waxing and waning symptoms. Just prior to discharge, noted to have increased work of breathing related to feeds. Readmission: 12 hours later with tachypnea, retractions, and hypoxia. | |
| Case 2: Index admission: Toddler admitted with febrile seizure in setting of gastroenteritis. Poor PO intake during hospitalization. Readmission: 1 day later with dehydration. | |
| Case 3: Index admission: Infant admitted with a prolonged complex febrile seizure. Workup included an unremarkable lumbar puncture. No additional seizures. No inpatient imaging obtained. Readmission: Abnormal outpatient MRI requiring intervention. | |
| Case 4: Index admission: Teenager with wheezing and history of chronic daily symptoms. Discharged <24 hours later on albuterol every 4 hours and prednisone. Readmission: 1 day later, seen by primary care physician with persistent asthma flare. | |
| Case 5: Index admission: Exfull‐term infant admitted with bronchiolitis, early in course. At time of discharge, had been off oxygen for 24 hours, but last recorded respiratory rate was >70. Readmission: 1 day later due to continued tachypnea and increased work of breathing. No hypoxia. CXR normal. | |
| Case 6: Exfull‐term infant admitted with bilious emesis, diarrhea, and dehydration. Ultrasound of pylorus, UGI, and BMP all normal. Tolerated oral intake but had emesis and loose stools prior to discharge. Readmission: <48 hours later with severe metabolic acidosis. | |
| 3 (Potentially Preventable) | Nosocomial/ematrogenic factors |
| Case 1: Index admission: Toddler admitted with fever and neutropenia. Treated with antibiotics 24 hours. Diagnosed with viral illness and discharged home. Readmission: Symptomatic Clostridum difficile infection. | |
| Case 2: Index admission: Patient with autism admitted with viral gastroenteritis. Readmission: Presumed nosocominal upper respiratory infection. | |
| Case 3: Index admission: Infant admitted with bronchiolitis. Recovered from initial infection. Readmission: New upper respiratory infection and presumed nosocomial infection. | |
| Case 4: Index admission: <28‐day‐old full‐term neonate presenting with neonatal fever and rash. Full septic workup performed and all cultures negative at 24 hours. Readmission: CSF culture positive at 36 hours and readmitted while awaiting speciation. Discharged once culture grew out a contaminant. | |
| 8 (Potentially Preventable) | Detection/treatment of problem was delayed and/or not appropriately facilitated |
| Case 1: Index admission: Preteen admitted with abdominal pain, concern for appendicitis. Ultrasound and MRI abdomen negative for appendicitis. Symptoms improved. Tolerated PO. Readmission: 3 days later with similar abdominal pain. Diagnosed with constipation with significant improvement following clean‐out. | |
| Case 2: Index admission: Infant with history of macrocephaly presented with fever and full fontanelle. Head CT showed mild prominence of the extra‐axial space, and lumbar puncture was normal. Readmission: Patient developed torticollis. MRI demonstrated a malignant lesion. | |
| Case 3: Index admission: School‐age child with RLQ abdominal pain, fever, leukocytosis, and indeterminate RLQ abdominal ultrasound. Twelve‐hour observation with no further fevers. Pain and appetite improved. Readmission: 1 day later with fever, anorexia, and abdominal pain. RLQ ultrasound unchanged. Appendectomy performed with inflamed appendix. | |
| 1 (Potentially Preventable) | Inappropriate readmission |
| Case 1: Index admission: Infant with laryngomalacia admitted with bronchiolitis. Readmission: Continued mild bronchiolitis symptoms but did not require oxygen or suctioning. Normal CXR. | |
| 5 (Potentially Preventable) | Resulted from inadequate postdischarge care planning |
| Case 1: Index diagnosis: Infant with vomiting, prior admissions, and extensive evaluation, diagnosed with milk protein allergy and GERD. PPI increased. Readmission: Persistent symptoms, required NGT feeds supplementation. | |
| All General Pediatrics Patients in 2014 | General Pediatric Readmitted Patients in 2014 | ||||
|---|---|---|---|---|---|
| Major Diagnosis Category at Index Admission | No. | % | Major Diagnosis Category at Index Admission | No. | % |
| |||||
| Respiratory | 2,723 | 37.5% | Respiratory | 79 | 31.9% |
| Digestive | 748 | 10.3% | Digestive | 41 | 16.5% |
| Ear, nose, mouth, throat | 675 | 9.3% | Ear, nose, mouth, throat | 24 | 9.7% |
| Skin, subcutaneous tissue | 480 | 6.6% | Musculoskeletal and connective tissue | 14 | 5.6% |
| Infectious, parasitic, systemic | 455 | 6.3% | Nervous | 13 | 5.2% |
| Factors influencing health status | 359 | 5.0% | Endocrine, nutritional, metabolic | 13 | 5.2% |
| Endocrine, nutritional, metabolic | 339 | 4.7% | Infectious, parasitic, systemic | 12 | 4.8% |
| Nervous | 239 | 3.3% | Newborn, neonate, perinatal period | 11 | 4.4% |
| Musculoskeletal and connective tissue | 228 | 3.1% | Hepatobiliary system and pancreas | 8 | 3.2% |
| Newborn, neonate, perinatal period | 206 | 2.8% | Skin, subcutaneous tissue | 8 | 3.2% |
| Other* | 800 | 11.0% | Other | 25 | 10.1% |
| Total | 7,252 | 100% | Total | 248 | 100% |
Inter‐Rater Reliability Analysis
A random selection of 50 cases (20% of total readmissions) was selected for a second review to test the tool's inter‐rater reliability. The second review resulted in the same terminal node for 44 (86%) of the cross‐checked files ( = 0.79; 95% confidence interval: 0.60‐0.98). Of the 6 cross‐checked files that ended at different nodes, 5 resulted in the same final determination about preventability. Only 1 of the cross‐checks (2% of total cross‐checked files) resulted in a different conclusion about preventability.
Efficiency Analysis
Reviewers reported that using the tool to reach a determination about preventability took approximately 20 minutes per case. Thus, initial reviews on the 248 cases required approximately 82.6 reviewer hours. Divided across 10 reviewers, this resulted in 8 to 9 hours of review time per reviewer over the year.
DISCUSSION
As part of an effort to direct quality‐improvement initiatives, this project used a Web‐based fault tree tool to identify root causes of general pediatrics readmissions at a freestanding children's hospital and classify them as either preventable or not preventable. The project also investigated the efficiency and inter‐rater reliability of the tool, which was designed to systematically guide physicians through the chart review process to a final determination about preventability. The project confirmed that using the tool helped reviewers reach final determinations about preventability efficiently with a high degree of consistency. It also confirmed that only a very small percentage of general pediatrics 15‐day readmissions are potentially preventable. Specifically, potentially preventable readmissions accounted for only 6.0% of total readmissions and 0.2% of general pediatrics discharges in 2014. Although our analysis focused on 15‐day readmissions, the fault tree methodology can be applied to any timeframe.
Previous studies attempting to distinguish preventable from nonpreventable readmissions, which used a range of methodologies to reach final determinations, reported that their review process was both time intensive and highly subjective. One study, which had 4 reviewers independently review charts and assign each case a preventability score on a 5‐point Likert scale, reported that reviewers disagreed on the final determination in 62.5% of cases.[12] Another study had 2 physicians independently review a selection of cases and assign a preventability score on a scale from 0 to 3. Scores for the 2 reviewers were added together, and cases above a certain composite threshold were classified as preventable. Despite being time‐intensive, this method resulted in only moderate agreement among physicians about the likelihood of preventability (weighted statistic of 0.44).[14] A more recent study, in which 2 physicians independently classified readmissions into 1 of 4 predefined categories, also reported only moderate agreement between reviewers ( = 0.44).[13] Other methods that have been reported include classifying readmissions as preventable only if multiple reviewers independently agreed, and using a third reviewer as a tie‐breaker.[14]
In an attempt to identify potentially preventable readmissions without using chart reviews, 3M (St. Paul, MN) developed its Potentially Preventable Readmissions software (3M‐PPR), which uses administrative data to identify which readmissions were potentially preventable. Although this automated approach is less time intensive, evidence suggests that due to a lack of nuance, the algorithm significantly overestimates the percentage of readmissions that are potentially preventable.[4, 5] A study that used 3M‐PPR to assess 1.7 million hospitalizations across 58 children's hospitals found that the algorithm classified 81% of sickle cell crisis and asthma readmissions, and 83% of bronchiolitis readmissions as potentially preventable.[10, 11] However, many readmissions for asthma and bronchiolitis are due to social factors that are outside of a hospital's direct control,[4, 5] and at many hospitals, readmissions for sickle cell crisis are part of a high‐value care model that weighs length of stay against potential readmissions. In addition, when assessing readmissions 7, 15, and 30 days after discharge, the algorithm classified almost the same percentage as potentially preventable, which is inconsistent with the notion that readmissions are more likely to have been preventable if they occurred closer to the initial discharge.[4, 13] Another study that assessed the performance of the software in the adult population reported that the algorithm performed with 85% sensitivity, but only 28% specificity.[5, 6]
The results of this quality‐improvement project indicate that using the fault tree tool to guide physicians through the chart review process helped address some of the shortcomings of methods reported in previous studies, by increasing the efficiency and reducing the subjectivity of final determinations, while still accounting for the nuances necessary to conduct a fair review. Because the tool provided a systematic framework for reviews, each case was completed in approximately 20 minutes, and because the process was the same for all reviewers, inter‐rater reliability was extremely high. In 86% of cross‐checked cases, the second reviewer ended at the same terminal node in the decision tree as the original reviewer, and in 98% of cross‐checked cases the second reviewer reached the same conclusion about preventability, even if they did not end at the same terminal node. Even accounting for agreement due to chance, the statistic of 0.79 confirmed that there was substantial agreement among reviewers about final determinations. Because the tool is easily adaptable, other hospitals can adopt this framework for their own preventability reviews and quality‐improvement initiatives.
Using the fault tree tool to access root causes of all 15‐day general pediatric readmissions helped the division focus quality‐improvement efforts on the most common causes of potentially preventable readmissions. Because 40% of potentially preventable readmissions were due to premature discharges, this prompted quality‐improvement teams to focus efforts on improving and clarifying the division's discharge criteria and clinical pathways. The division also initiated processes to improve discharge planning, including improved teaching of discharge instructions and having families pick up prescriptions prior to discharge.
Although these results did help the division identify a few areas of focus to potentially reduce readmissions, the fact that the overall 15‐day readmission rate for general pediatrics, as well as the percentage of readmissions and total discharges that were deemed potentially preventable, were so low (3.4%, 6.0%, and 0.2%, respectively), supports those who question whether prioritizing pediatric readmissions is the best place for hospitals to focus quality‐improvement efforts.[10, 12, 15, 16] As these results indicate, most pediatric readmissions are not preventable, and thus consistent with an efficient, effective, timely, patient‐centered, and equitable health system. Other studies have also shown that because overall and condition‐specific readmissions at pediatric hospitals are low, few pediatric hospitals are high or low performing for readmissions, and thus readmission rates are likely not a good measure of hospital quality.[8]
However, other condition‐specific studies of readmissions in pediatrics have indicated that there are some areas of opportunity to identify populations at high risk for readmission. One study found that although pneumonia‐specific 30‐day readmission rates in a national cohort of children hospitalized with pneumonia was only 3.1%, the chances of readmission were higher for children <1 year old, children with chronic comorbidities or complicated pneumonia, and children cared for in hospitals with lower volumes of pneumonia admissions.[17] Another study found that 17.1% of adolescents in a statewide database were readmitted post‐tonsillectomy for pain, nausea, and dehydration.[18] Thus, adapting the tool to identify root causes of condition‐specific or procedure‐specific readmissions, especially for surgical patients, may be an area of opportunity for future quality‐improvement efforts.[5] However, for general pediatrics, shifting the focus from reducing readmissions to improving the quality of care patients receive in the hospital, improving the discharge process, and adopting a population health approach to mitigate external risk factors, may be appropriate.
This project was subject to limitations. First, because it was conducted at a single site and only on general pediatrics patients, results may not be generalizable to other hospitals or other pediatric divisions. Thus, future studies might use the fault tree framework to assess preventability of pediatric readmissions in other divisions or specialties. Second, because readmissions to other hospitals were not included in the sample, the overall readmissions rate is likely underestimated.[19] However, it is unclear how this would affect the rate of potentially preventable readmissions. Third, although the fault tree framework reduced the subjectivity of the review process, there is still a degree of subjectivity inherent at each decision node. To minimize this, reviewers should try to discuss and come to consensus on how they are making determinations at each juncture in the decision tree. Similarly, because reviewers' answers to decision‐tree questions rely heavily on chart documentation, reviews may be compromised by unclear or incomplete documentation. For example, if information about steps the hospital team took to prepare a family for discharge were not properly documented, it would be difficult to determine whether appropriate steps were taken to minimize the likelihood of a complication. In the case of insufficient documentation of relevant social concerns, cases may be incorrectly classified as preventable, because addressing social issues is often not within a hospital's direct control. Finally, because reviewers were not blinded to the original discharging physician, there may have been some unconscious bias of unknown direction in the reviews.
CONCLUSION
Using the Web‐based fault tree tool helped physicians to identify the root causes of hospital readmissions and classify them as preventable or not preventable in a standardized, efficient, and consistent way, while still accounting for the nuances necessary to conduct a fair review. Thus, other hospitals should consider adopting this framework for their own preventability reviews and quality‐improvement initiatives. However, this project also confirmed that only a very small percentage of general pediatrics 15‐day readmissions are potentially preventable, suggesting that general pediatrics readmissions are not an appropriate measure of hospital quality. Instead, adapting the tool to identify root causes of condition‐specific or procedure‐specific readmission rates may be an area of opportunity for future quality‐improvement efforts.
Disclosures: This work was supported through internal funds from The Children's Hospital of Philadelphia. The authors have no financial interests, relationships or affiliations relevant to the subject matter or materials discussed in the article to disclose. The authors have no potential conflicts of interest relevant to the subject matter or materials discussed in the article to disclose.
As physicians strive to increase the value of healthcare delivery, there has been increased focus on improving the quality of care that patients receive while lowering per capita costs. A provision of the Affordable Care Act implemented in 2012 identified all‐cause 30‐day readmission rates as a measure of hospital quality, and as part of the Act's Hospital Readmission and Reduction Program, Medicare now penalizes hospitals with higher than expected all‐cause readmissions rates for adult patients with certain conditions by lowering reimbursements.[1] Although readmissions are not yet commonly used to determine reimbursements for pediatric hospitals, several states are penalizing higher than expected readmission rates for Medicaid enrollees,[2, 3] using an imprecise algorithm to determine which readmissions resulted from low‐quality care during the index admission.[4, 5, 6]
There is growing concern, however, that readmission rates are not an accurate gauge of the quality of care patients receive while in the hospital or during the discharge process to prepare them for their transition home.[7, 8, 9, 10] This is especially true in pediatric settings, where overall readmission rates are much lower than in adult settings, many readmissions are expected as part of a patient's planned course of care, and variation in readmission rates between hospitals is correlated with the percentage of patients with certain complex chronic conditions.[1, 7, 11] Thus, there is increasing agreement that hospitals and external evaluators need to shift the focus from all‐cause readmissions to a reliable, consistent, and fair measure of potentially preventable readmissions.[12, 13] In addition to being a more useful quality metric, analyzing preventable readmissions will help hospitals focus resources on patients with potentially modifiable risk factors and develop meaningful quality‐improvement initiatives to improve inpatient care as well as the discharge process to prepare families for their transition to home.[14]
Although previous studies have attempted to distinguish preventable from nonpreventable readmissions, many reported significant challenges in completing reviews efficiently, achieving consistency in how readmissions were classified, and attaining consensus on final determinations.[12, 13, 14] Studies have also demonstrated that the algorithms some states are using to streamline preventability reviews and determine reimbursements overestimate the rate of potentially preventable readmissions.[4, 5, 6]
To increase the efficiency of preventability reviews and reduce the subjectivity involved in reaching final determinations, while still accounting for the nuances necessary to conduct a fair review, a quality‐improvement team from the Division of General Pediatrics at The Children's Hospital of Philadelphia (CHOP) implemented a fault tree analysis tool based on a framework developed by Howard Parker at Intermountain Primary Children's Hospital. The CHOP team coded this framework into a secure Web‐based data‐collection tool in the form of a decision tree to guide reviewers through a logical progression of questions that result in 1 of 18 root causes of readmissions, 8 of which are considered potentially preventable. We hypothesized that this method would help reviewers efficiently reach consensus on the root causes of hospital readmissions, and thus help the division and the hospital focus efforts on developing relevant quality‐improvement initiatives.
METHODS
Inclusion Criteria and Study Design
This study was conducted at CHOP, a 535‐bed urban, tertiary‐care, freestanding children's hospital with approximately 29,000 annual discharges. Of those discharges, 7000 to 8000 are from the general pediatrics service, meaning that the attending of record was a general pediatrician. Patients were included in the study if (1) they were discharged from the general pediatrics service between January 2014 and December 2014, and (2) they were readmitted to the hospital, for any reason, within 15 days of discharge. Because this analysis was done as part of a quality‐improvement initiative, it focuses on 15‐day, early readmissions to target cases with a higher probability of being potentially preventable from the perspective of the hospital care team.[10, 12, 13] Patients under observation status during the index admission or the readmission were included. However, patients who returned to the emergency department but were not admitted to an inpatient unit were excluded. Objective details about each case, including the patient's name, demographics, chart number, and diagnosis code, were pre‐loaded from EPIC (Epic Systems Corp., Verona, WI) into REDCap (Research Electronic Data Capture;
A panel of 10 general pediatricians divided up the cases to perform retrospective chart reviews. For each case, REDCap guided reviewers through the fault tree analysis. Reviewers met monthly to discuss difficult cases and reach consensus on any identified ambiguities in the process. After all cases were reviewed once, 3 panel members independently reviewed a random selection of cases to measure inter‐rater reliability and confirm reproducibility of final determinations. The inter‐rater reliability statistic was calculated using Stata 12.1 (StataCorp LP, College Station, TX). During chart reviews, panel members were not blinded to the identity of physicians and other staff members caring for the patients under review. CHOP's institutional review board determined this study to be exempt from ongoing review.
Fault Tree Analysis
Using the decision tree framework for analyzing readmissions that was developed at Intermountain Primary Children's Hospital, the REDCap tool prompted reviewers with a series of sequential questions, each with mutually exclusive options. Using imbedded branching logic to select follow‐up questions, the tool guided reviewers to 1 of 18 terminal nodes, each representing a potential root cause of the readmission. Of those 18 potential causes, 8 were considered potentially preventable. A diagram of the fault tree framework, color coded to indicate which nodes were considered potentially preventable, is shown in Figure 1.
RESULTS
In 2014, 7252 patients were discharged from the general pediatrics service at CHOP. Of those patients, 248 were readmitted within 15 days for an overall general pediatrics 15‐day readmission rate of 3.4%.
Preventability Analysis
Of the 248 readmissions, 233 (94.0%) were considered not preventable. The most common cause for readmission, which accounted for 145 cases (58.5%), was a patient developing an unpredictable problem related to the index diagnosis or a natural progression of the disease that required readmission. The second most common cause, which accounted for 53 cases (21.4%), was a patient developing a new condition unrelated to the index diagnosis or a readmission unrelated to the quality of care received during the index stay. The third most frequent cause, which accounted for 11 cases (4.4%), was a legitimate nonclinical readmission due to lack of alternative resources, psychosocial or economic factors, or case‐specific factors. Other nonpreventable causes of readmission, including scheduled readmissions, each accounted for 7 or fewer cases and <3% of total readmissions.
The 15 readmissions considered potentially preventable accounted for 6.0% of total readmissions and 0.2% of total discharges from the general pediatrics service in 2014. The most common cause of preventable readmissions, which accounted for 6 cases, was premature discharge. The second most common cause, which accounted for 4 cases, was a problem resulting from nosocomial or iatrogenic factors. Other potentially preventable causes included delayed detection of problem (3 cases), inappropriate readmission (1 case), and inadequate postdischarge care planning (1 case).
A breakdown of fault tree results, including examples of cases associated with each terminal node, is shown in Table 1. Information about general pediatrics patients and readmitted patients is included in Tables 2 and 3. A breakdown of determinations for each reviewer is included in Supporting Table 1 in the online version of this article.
| Fault Tree Terminal Node | Root Cause of Readmission | No. of Cases | % of Total Readmissions | % Within Preventability Category | % of Total Discharges |
|---|---|---|---|---|---|
| |||||
| 2 (Potentially Preventable) | Problematic condition on discharge. Example:* Index admission: Infant with history of prematurity admitted with RSV and rhinovirus bronchiolitis. Had some waxing and waning symptoms. Just prior to discharge, noted to have increased work of breathing related to feeds. Readmission: 12 hours later with tachypnea, retractions, and hypoxia. | 6 | 2.4% | 40.0% | 0.08% |
| 3 (Potentially Preventable) | Nosocomial/Iatrogenic factors. Example*: Index admission: Toddler admitted with fever and neutropenia. Treated with antibiotics 24 hours. Diagnosed with viral illness and discharged home. Readmission: symptomatic Clostridum difficile infection. | 4 | 1.6% | 26.7% | 0.06% |
| 8 (Potentially Preventable) | Detection/treatment of problem was delayed and not appropriately facilitated. Example:* Index admission: Preteen admitted with abdominal pain, concern for appendicitis. Ultrasound and abdominal MRI negative for appendicitis. Symptoms improved. Tolerated PO. Readmission: 3 days later with similar abdominal pain. Diagnosed with constipation with significant improvement following clean‐out. | 3 | 1.2% | 20.0% | 0.04% |
| 1 (Potentially Preventable) | Inappropriate readmission. Example:* Index admission: Infant with laryngomalacia admitted with bronchiolitis. Readmission: Continued mild bronchiolitis symptoms but did not require oxygen or suctioning, normal CXR. | 1 | 0.4% | 6.7% | 0.01% |
| 5 (Potentially Preventable) | Resulted from inadequate postdischarge care planning. Example:* Index diagnosis: Infant with vomiting, prior admissions, and extensive evaluation, diagnosed with milk protein allergy and GERD. PPI increased. Readmission: Persistent symptoms, required NGT feeds supplementation. | 1 | 0.4% | 6.7% | 0.01% |
| 4 (Potentially Preventable) | Resulted from a preventable complication and hospital/physician did not take the appropriate steps to minimize likelihood of complication. | ||||
| 6 (Potentially Preventable) | Resulted from improper care by patient/family and effort by hospital/physician to ensure correct postdischarge care was inadequate. | ||||
| 7 (Potentially Preventable) | Resulted from inadequate care by community services and effort by hospital/physician to ensure correct postdischarge care was inadequate. | ||||
| 15 | 6.0% | 100% | 0.2% | ||
| 12 (Not Preventable) | Problem was unpredictable. Example:* Index admission: Infant admitted with gastroenteritis and dehydration with an anion gap metabolic acidosis. Vomiting and diarrhea improved, rehydrated, acidosis improved. Readmission: 1 day later, presented with emesis and fussiness. Readmitted for metabolic acidosis. | 145 | 58.5% | 62.2% | 2.00% |
| 10 (Not Preventable) | Patient developed new condition unrelated to index diagnosis or quality of care. Example:* Index admission: Toddler admitted with cellulitis. Readmission: Bronchiolitis (did not meet CDC guidelines for nosocomial infection). | 53 | 21.4% | 22.7% | 0.73% |
| 9 (Not Preventable) | Legitimate nonclinical readmission. Example:* Index admission: Infant admitted with second episode of bronchiolitis. Readmission: 4 days later with mild diarrhea. Tolerated PO challenge in emergency department. Admitted due to parental anxiety. | 11 | 4.4% | 4.7% | 0.15% |
| 17 (Not Preventable) | Problem resulted from improper care by patient/family but effort by hospital/physician to ensure correct postdischarge care was appropriate. Example:* Index admission: Infant admitted with diarrhea, diagnosed with milk protein allergy. Discharged on soy formula. Readmission: Developed vomiting and diarrhea with cow milk formula. | 7 | 2.8% | 3.0% | 0.10% |
| 11 (Not Preventable) | Scheduled readmission. Example:* Index admission: Infant with conjunctivitis and preseptal cellulitis with nasolacrimal duct obstruction. Readmission: Postoperatively following scheduled nasolacrimal duct repair. | 7 | 2.8% | 3.0% | 0.10% |
| 14 (Not Preventable) | Detection/treatment of problem was delayed, but earlier detection was not feasible. Example:* Index admission: Preteen admitted with fever, abdominal pain, and elevated inflammatory markers. Fever resolved and symptoms improved. Diagnosed with unspecified viral infection. Readmission: 4 days later with lower extremity pyomyositis and possible osteomyelitis. | 4 | 1.6% | 1.7% | 0.06% |
| 15 (Not Preventable) | Detection/treatment of problem was delayed, earlier detection was feasible, but detection was appropriately facilitated. Example:* Index admission: Infant with history of laryngomalacia and GER admitted with an ALTE. No events during hospitalization. Appropriate workup and cleared by consultants for discharge. Zantac increased. Readmission: Infant had similar ALTE events within a week after discharge. Ultimately underwent supraglottoplasty. | 2 | 0.8% | 0.9% | 0.03% |
| 13 (Not Preventable) | Resulted from preventable complication but efforts to minimize likelihood were appropriate. Example:* Index admission: Patient on GJ feeds admitted for dislodged GJ. Extensive conversations between primary team and multiple consulting services regarding best type of tube. Determined that no other tube options were appropriate. Temporizing measures were initiated. Readmission: GJ tube dislodged again. | 2 | 0.8% | 0.9% | 0.03% |
| 18 (Not Preventable) | Resulted from medication side effect (after watch period). Example:* Index admission: Preteen with MSSA bacteremia spread to other organs. Sent home on appropriate IV antibiotics. Readmission: Fever, rash, increased LFTs. Blood cultures negative. Presumed drug reaction. Fevers resolved with alternate medication. | 2 | 0.8% | 0.9% | 0.03% |
| 16 (Not Preventable) | Resulted from inadequate care by community services, but effort by hospital/physician to ensure correct postdischarge care was appropriate. | ||||
| 233 | 94.0% | 100% | 3.2% | ||
| Fault Tree Terminal Node | Root Cause of Potentially Preventable Readmission with Case Descriptions* |
|---|---|
| |
| 2 (Potentially Preventable) | Problematic condition on discharge |
| Case 1: Index admission: Infant with history of prematurity admitted with RSV and rhinovirus bronchiolitis. Had some waxing and waning symptoms. Just prior to discharge, noted to have increased work of breathing related to feeds. Readmission: 12 hours later with tachypnea, retractions, and hypoxia. | |
| Case 2: Index admission: Toddler admitted with febrile seizure in setting of gastroenteritis. Poor PO intake during hospitalization. Readmission: 1 day later with dehydration. | |
| Case 3: Index admission: Infant admitted with a prolonged complex febrile seizure. Workup included an unremarkable lumbar puncture. No additional seizures. No inpatient imaging obtained. Readmission: Abnormal outpatient MRI requiring intervention. | |
| Case 4: Index admission: Teenager with wheezing and history of chronic daily symptoms. Discharged <24 hours later on albuterol every 4 hours and prednisone. Readmission: 1 day later, seen by primary care physician with persistent asthma flare. | |
| Case 5: Index admission: Exfull‐term infant admitted with bronchiolitis, early in course. At time of discharge, had been off oxygen for 24 hours, but last recorded respiratory rate was >70. Readmission: 1 day later due to continued tachypnea and increased work of breathing. No hypoxia. CXR normal. | |
| Case 6: Exfull‐term infant admitted with bilious emesis, diarrhea, and dehydration. Ultrasound of pylorus, UGI, and BMP all normal. Tolerated oral intake but had emesis and loose stools prior to discharge. Readmission: <48 hours later with severe metabolic acidosis. | |
| 3 (Potentially Preventable) | Nosocomial/ematrogenic factors |
| Case 1: Index admission: Toddler admitted with fever and neutropenia. Treated with antibiotics 24 hours. Diagnosed with viral illness and discharged home. Readmission: Symptomatic Clostridum difficile infection. | |
| Case 2: Index admission: Patient with autism admitted with viral gastroenteritis. Readmission: Presumed nosocominal upper respiratory infection. | |
| Case 3: Index admission: Infant admitted with bronchiolitis. Recovered from initial infection. Readmission: New upper respiratory infection and presumed nosocomial infection. | |
| Case 4: Index admission: <28‐day‐old full‐term neonate presenting with neonatal fever and rash. Full septic workup performed and all cultures negative at 24 hours. Readmission: CSF culture positive at 36 hours and readmitted while awaiting speciation. Discharged once culture grew out a contaminant. | |
| 8 (Potentially Preventable) | Detection/treatment of problem was delayed and/or not appropriately facilitated |
| Case 1: Index admission: Preteen admitted with abdominal pain, concern for appendicitis. Ultrasound and MRI abdomen negative for appendicitis. Symptoms improved. Tolerated PO. Readmission: 3 days later with similar abdominal pain. Diagnosed with constipation with significant improvement following clean‐out. | |
| Case 2: Index admission: Infant with history of macrocephaly presented with fever and full fontanelle. Head CT showed mild prominence of the extra‐axial space, and lumbar puncture was normal. Readmission: Patient developed torticollis. MRI demonstrated a malignant lesion. | |
| Case 3: Index admission: School‐age child with RLQ abdominal pain, fever, leukocytosis, and indeterminate RLQ abdominal ultrasound. Twelve‐hour observation with no further fevers. Pain and appetite improved. Readmission: 1 day later with fever, anorexia, and abdominal pain. RLQ ultrasound unchanged. Appendectomy performed with inflamed appendix. | |
| 1 (Potentially Preventable) | Inappropriate readmission |
| Case 1: Index admission: Infant with laryngomalacia admitted with bronchiolitis. Readmission: Continued mild bronchiolitis symptoms but did not require oxygen or suctioning. Normal CXR. | |
| 5 (Potentially Preventable) | Resulted from inadequate postdischarge care planning |
| Case 1: Index diagnosis: Infant with vomiting, prior admissions, and extensive evaluation, diagnosed with milk protein allergy and GERD. PPI increased. Readmission: Persistent symptoms, required NGT feeds supplementation. | |
| All General Pediatrics Patients in 2014 | General Pediatric Readmitted Patients in 2014 | ||||
|---|---|---|---|---|---|
| Major Diagnosis Category at Index Admission | No. | % | Major Diagnosis Category at Index Admission | No. | % |
| |||||
| Respiratory | 2,723 | 37.5% | Respiratory | 79 | 31.9% |
| Digestive | 748 | 10.3% | Digestive | 41 | 16.5% |
| Ear, nose, mouth, throat | 675 | 9.3% | Ear, nose, mouth, throat | 24 | 9.7% |
| Skin, subcutaneous tissue | 480 | 6.6% | Musculoskeletal and connective tissue | 14 | 5.6% |
| Infectious, parasitic, systemic | 455 | 6.3% | Nervous | 13 | 5.2% |
| Factors influencing health status | 359 | 5.0% | Endocrine, nutritional, metabolic | 13 | 5.2% |
| Endocrine, nutritional, metabolic | 339 | 4.7% | Infectious, parasitic, systemic | 12 | 4.8% |
| Nervous | 239 | 3.3% | Newborn, neonate, perinatal period | 11 | 4.4% |
| Musculoskeletal and connective tissue | 228 | 3.1% | Hepatobiliary system and pancreas | 8 | 3.2% |
| Newborn, neonate, perinatal period | 206 | 2.8% | Skin, subcutaneous tissue | 8 | 3.2% |
| Other* | 800 | 11.0% | Other | 25 | 10.1% |
| Total | 7,252 | 100% | Total | 248 | 100% |
Inter‐Rater Reliability Analysis
A random selection of 50 cases (20% of total readmissions) was selected for a second review to test the tool's inter‐rater reliability. The second review resulted in the same terminal node for 44 (86%) of the cross‐checked files ( = 0.79; 95% confidence interval: 0.60‐0.98). Of the 6 cross‐checked files that ended at different nodes, 5 resulted in the same final determination about preventability. Only 1 of the cross‐checks (2% of total cross‐checked files) resulted in a different conclusion about preventability.
Efficiency Analysis
Reviewers reported that using the tool to reach a determination about preventability took approximately 20 minutes per case. Thus, initial reviews on the 248 cases required approximately 82.6 reviewer hours. Divided across 10 reviewers, this resulted in 8 to 9 hours of review time per reviewer over the year.
DISCUSSION
As part of an effort to direct quality‐improvement initiatives, this project used a Web‐based fault tree tool to identify root causes of general pediatrics readmissions at a freestanding children's hospital and classify them as either preventable or not preventable. The project also investigated the efficiency and inter‐rater reliability of the tool, which was designed to systematically guide physicians through the chart review process to a final determination about preventability. The project confirmed that using the tool helped reviewers reach final determinations about preventability efficiently with a high degree of consistency. It also confirmed that only a very small percentage of general pediatrics 15‐day readmissions are potentially preventable. Specifically, potentially preventable readmissions accounted for only 6.0% of total readmissions and 0.2% of general pediatrics discharges in 2014. Although our analysis focused on 15‐day readmissions, the fault tree methodology can be applied to any timeframe.
Previous studies attempting to distinguish preventable from nonpreventable readmissions, which used a range of methodologies to reach final determinations, reported that their review process was both time intensive and highly subjective. One study, which had 4 reviewers independently review charts and assign each case a preventability score on a 5‐point Likert scale, reported that reviewers disagreed on the final determination in 62.5% of cases.[12] Another study had 2 physicians independently review a selection of cases and assign a preventability score on a scale from 0 to 3. Scores for the 2 reviewers were added together, and cases above a certain composite threshold were classified as preventable. Despite being time‐intensive, this method resulted in only moderate agreement among physicians about the likelihood of preventability (weighted statistic of 0.44).[14] A more recent study, in which 2 physicians independently classified readmissions into 1 of 4 predefined categories, also reported only moderate agreement between reviewers ( = 0.44).[13] Other methods that have been reported include classifying readmissions as preventable only if multiple reviewers independently agreed, and using a third reviewer as a tie‐breaker.[14]
In an attempt to identify potentially preventable readmissions without using chart reviews, 3M (St. Paul, MN) developed its Potentially Preventable Readmissions software (3M‐PPR), which uses administrative data to identify which readmissions were potentially preventable. Although this automated approach is less time intensive, evidence suggests that due to a lack of nuance, the algorithm significantly overestimates the percentage of readmissions that are potentially preventable.[4, 5] A study that used 3M‐PPR to assess 1.7 million hospitalizations across 58 children's hospitals found that the algorithm classified 81% of sickle cell crisis and asthma readmissions, and 83% of bronchiolitis readmissions as potentially preventable.[10, 11] However, many readmissions for asthma and bronchiolitis are due to social factors that are outside of a hospital's direct control,[4, 5] and at many hospitals, readmissions for sickle cell crisis are part of a high‐value care model that weighs length of stay against potential readmissions. In addition, when assessing readmissions 7, 15, and 30 days after discharge, the algorithm classified almost the same percentage as potentially preventable, which is inconsistent with the notion that readmissions are more likely to have been preventable if they occurred closer to the initial discharge.[4, 13] Another study that assessed the performance of the software in the adult population reported that the algorithm performed with 85% sensitivity, but only 28% specificity.[5, 6]
The results of this quality‐improvement project indicate that using the fault tree tool to guide physicians through the chart review process helped address some of the shortcomings of methods reported in previous studies, by increasing the efficiency and reducing the subjectivity of final determinations, while still accounting for the nuances necessary to conduct a fair review. Because the tool provided a systematic framework for reviews, each case was completed in approximately 20 minutes, and because the process was the same for all reviewers, inter‐rater reliability was extremely high. In 86% of cross‐checked cases, the second reviewer ended at the same terminal node in the decision tree as the original reviewer, and in 98% of cross‐checked cases the second reviewer reached the same conclusion about preventability, even if they did not end at the same terminal node. Even accounting for agreement due to chance, the statistic of 0.79 confirmed that there was substantial agreement among reviewers about final determinations. Because the tool is easily adaptable, other hospitals can adopt this framework for their own preventability reviews and quality‐improvement initiatives.
Using the fault tree tool to access root causes of all 15‐day general pediatric readmissions helped the division focus quality‐improvement efforts on the most common causes of potentially preventable readmissions. Because 40% of potentially preventable readmissions were due to premature discharges, this prompted quality‐improvement teams to focus efforts on improving and clarifying the division's discharge criteria and clinical pathways. The division also initiated processes to improve discharge planning, including improved teaching of discharge instructions and having families pick up prescriptions prior to discharge.
Although these results did help the division identify a few areas of focus to potentially reduce readmissions, the fact that the overall 15‐day readmission rate for general pediatrics, as well as the percentage of readmissions and total discharges that were deemed potentially preventable, were so low (3.4%, 6.0%, and 0.2%, respectively), supports those who question whether prioritizing pediatric readmissions is the best place for hospitals to focus quality‐improvement efforts.[10, 12, 15, 16] As these results indicate, most pediatric readmissions are not preventable, and thus consistent with an efficient, effective, timely, patient‐centered, and equitable health system. Other studies have also shown that because overall and condition‐specific readmissions at pediatric hospitals are low, few pediatric hospitals are high or low performing for readmissions, and thus readmission rates are likely not a good measure of hospital quality.[8]
However, other condition‐specific studies of readmissions in pediatrics have indicated that there are some areas of opportunity to identify populations at high risk for readmission. One study found that although pneumonia‐specific 30‐day readmission rates in a national cohort of children hospitalized with pneumonia was only 3.1%, the chances of readmission were higher for children <1 year old, children with chronic comorbidities or complicated pneumonia, and children cared for in hospitals with lower volumes of pneumonia admissions.[17] Another study found that 17.1% of adolescents in a statewide database were readmitted post‐tonsillectomy for pain, nausea, and dehydration.[18] Thus, adapting the tool to identify root causes of condition‐specific or procedure‐specific readmissions, especially for surgical patients, may be an area of opportunity for future quality‐improvement efforts.[5] However, for general pediatrics, shifting the focus from reducing readmissions to improving the quality of care patients receive in the hospital, improving the discharge process, and adopting a population health approach to mitigate external risk factors, may be appropriate.
This project was subject to limitations. First, because it was conducted at a single site and only on general pediatrics patients, results may not be generalizable to other hospitals or other pediatric divisions. Thus, future studies might use the fault tree framework to assess preventability of pediatric readmissions in other divisions or specialties. Second, because readmissions to other hospitals were not included in the sample, the overall readmissions rate is likely underestimated.[19] However, it is unclear how this would affect the rate of potentially preventable readmissions. Third, although the fault tree framework reduced the subjectivity of the review process, there is still a degree of subjectivity inherent at each decision node. To minimize this, reviewers should try to discuss and come to consensus on how they are making determinations at each juncture in the decision tree. Similarly, because reviewers' answers to decision‐tree questions rely heavily on chart documentation, reviews may be compromised by unclear or incomplete documentation. For example, if information about steps the hospital team took to prepare a family for discharge were not properly documented, it would be difficult to determine whether appropriate steps were taken to minimize the likelihood of a complication. In the case of insufficient documentation of relevant social concerns, cases may be incorrectly classified as preventable, because addressing social issues is often not within a hospital's direct control. Finally, because reviewers were not blinded to the original discharging physician, there may have been some unconscious bias of unknown direction in the reviews.
CONCLUSION
Using the Web‐based fault tree tool helped physicians to identify the root causes of hospital readmissions and classify them as preventable or not preventable in a standardized, efficient, and consistent way, while still accounting for the nuances necessary to conduct a fair review. Thus, other hospitals should consider adopting this framework for their own preventability reviews and quality‐improvement initiatives. However, this project also confirmed that only a very small percentage of general pediatrics 15‐day readmissions are potentially preventable, suggesting that general pediatrics readmissions are not an appropriate measure of hospital quality. Instead, adapting the tool to identify root causes of condition‐specific or procedure‐specific readmission rates may be an area of opportunity for future quality‐improvement efforts.
Disclosures: This work was supported through internal funds from The Children's Hospital of Philadelphia. The authors have no financial interests, relationships or affiliations relevant to the subject matter or materials discussed in the article to disclose. The authors have no potential conflicts of interest relevant to the subject matter or materials discussed in the article to disclose.
- , . Pediatric readmissions as a hospital quality measure. JAMA. 2013;309(4):396–398.
- Texas Health and Human Services Commission. Potentially preventable readmissions in the Texas Medicaid population, state fiscal year 2012. Available at: http://www.hhsc.state.tx.us/reports/2013/ppr‐report.pdf. Published November 2013. Accessed August 16, 2015.
- Illinois Department of Healthcare and Family Services. Quality initiative to reduce hospital potentially preventable readmissions (PPR): Status update. Available at: http://www.illinois.gov/hfs/SiteCollectionDocuments/PPRPolicyStatusUpdate.pdf. Published September 3, 2014. Accessed August 16, 2015.
- , , , et al. Rates and impact of potentially preventable readmissions at children's hospitals. J Pediatr. 2015;166(3):613–619.e615.
- , . Preventing pediatric readmissions: which ones and how? J Pediatr. 2015;166(3):519–520.
- , , , , , . Manual and automated methods for identifying potentially preventable readmissions: a comparison in a large healthcare system. BMC Med Inform Decis Mak. 2014;14:28.
- , . Section on hospital medicine leadership and staff. Hosp Pediatr. 2013;3(4):390–393.
- , , , et al. Measuring hospital quality using pediatric readmission and revisit rates. Pediatrics. 2013;132(3):429–436.
- , . Hospital readmissions—not just a measure of quality. JAMA. 2011;306(16):1796–1797.
- , . Preventing readmissions in children: how do we do that? Hosp Pediatr. 2015;5(11):602–604.
- , , , et al. Pediatric readmission prevalence and variability across hospitals. JAMA. 2013;309(4):372–380.
- , , , , , . Preventability of early readmissions at a children's hospital. Pediatrics. 2013;131(1):e171–e181.
- , , , et al. An examination of physician‐, caregiver‐, and disease‐related factors associated with readmission from a pediatric hospital medicine service. Hosp Pediatr. 2015;5(11):566–573.
- , , , et al. Clinical preventability of 30‐day readmission after percutaneous coronary intervention. J Am Heart Assoc. 2014;3(5):e001290.
- . 3M algorithm overestimates preventable pediatric readmissions. Hospitalist News website. Available at: http://www.ehospitalistnews.com/specialty‐focus/pediatrics/single‐article‐page/3m‐algorithm‐overestimates‐preventable‐pediatric‐readmissions.html. Published August 16, 2013. Accessed August 16, 2015.
- . The 30‐day readmission rate: not a quality measure but an accountability measure. An Ounce of Evidence: Health Policy blog. Available at: https://blogs.sph.harvard.edu/ashish‐jha/?s=30‐day+readmission+rate. Published February 14, 2013. Accessed August 16, 2015.
- , , , et al. Readmissions among children previously hospitalized with pneumonia. Pediatrics. 2014;134(1):100–109.
- , , . A population‐based study of acute care revisits following tonsillectomy. J Pediatr. 2015;166(3):607–612.e605.
- , , , et al. Same‐hospital readmission rates as a measure of pediatric quality of care. JAMA Pediatr. 2015;169(10):905–912.
- , . Pediatric readmissions as a hospital quality measure. JAMA. 2013;309(4):396–398.
- Texas Health and Human Services Commission. Potentially preventable readmissions in the Texas Medicaid population, state fiscal year 2012. Available at: http://www.hhsc.state.tx.us/reports/2013/ppr‐report.pdf. Published November 2013. Accessed August 16, 2015.
- Illinois Department of Healthcare and Family Services. Quality initiative to reduce hospital potentially preventable readmissions (PPR): Status update. Available at: http://www.illinois.gov/hfs/SiteCollectionDocuments/PPRPolicyStatusUpdate.pdf. Published September 3, 2014. Accessed August 16, 2015.
- , , , et al. Rates and impact of potentially preventable readmissions at children's hospitals. J Pediatr. 2015;166(3):613–619.e615.
- , . Preventing pediatric readmissions: which ones and how? J Pediatr. 2015;166(3):519–520.
- , , , , , . Manual and automated methods for identifying potentially preventable readmissions: a comparison in a large healthcare system. BMC Med Inform Decis Mak. 2014;14:28.
- , . Section on hospital medicine leadership and staff. Hosp Pediatr. 2013;3(4):390–393.
- , , , et al. Measuring hospital quality using pediatric readmission and revisit rates. Pediatrics. 2013;132(3):429–436.
- , . Hospital readmissions—not just a measure of quality. JAMA. 2011;306(16):1796–1797.
- , . Preventing readmissions in children: how do we do that? Hosp Pediatr. 2015;5(11):602–604.
- , , , et al. Pediatric readmission prevalence and variability across hospitals. JAMA. 2013;309(4):372–380.
- , , , , , . Preventability of early readmissions at a children's hospital. Pediatrics. 2013;131(1):e171–e181.
- , , , et al. An examination of physician‐, caregiver‐, and disease‐related factors associated with readmission from a pediatric hospital medicine service. Hosp Pediatr. 2015;5(11):566–573.
- , , , et al. Clinical preventability of 30‐day readmission after percutaneous coronary intervention. J Am Heart Assoc. 2014;3(5):e001290.
- . 3M algorithm overestimates preventable pediatric readmissions. Hospitalist News website. Available at: http://www.ehospitalistnews.com/specialty‐focus/pediatrics/single‐article‐page/3m‐algorithm‐overestimates‐preventable‐pediatric‐readmissions.html. Published August 16, 2013. Accessed August 16, 2015.
- . The 30‐day readmission rate: not a quality measure but an accountability measure. An Ounce of Evidence: Health Policy blog. Available at: https://blogs.sph.harvard.edu/ashish‐jha/?s=30‐day+readmission+rate. Published February 14, 2013. Accessed August 16, 2015.
- , , , et al. Readmissions among children previously hospitalized with pneumonia. Pediatrics. 2014;134(1):100–109.
- , , . A population‐based study of acute care revisits following tonsillectomy. J Pediatr. 2015;166(3):607–612.e605.
- , , , et al. Same‐hospital readmission rates as a measure of pediatric quality of care. JAMA Pediatr. 2015;169(10):905–912.
© 2016 Society of Hospital Medicine
OUs and Patient Outcomes
Many pediatric hospitalizations are of short duration, and more than half of short‐stay hospitalizations are designated as observation status.[1, 2] Observation status is an administrative label assigned to patients who do not meet hospital or payer criteria for inpatient‐status care. Short‐stay observation‐status patients do not fit in traditional models of emergency department (ED) or inpatient care. EDs often focus on discharging or admitting patients within a matter of hours, whereas inpatient units tend to measure length of stay (LOS) in terms of days[3] and may not have systems in place to facilitate rapid discharge of short‐stay patients.[4] Observation units (OUs) have been established in some hospitals to address the unique care needs of short‐stay patients.[5, 6, 7]
Single‐site reports from children's hospitals with successful OUs have demonstrated shorter LOS and lower costs compared with inpatient settings.[6, 8, 9, 10, 11, 12, 13, 14] No prior study has examined hospital‐level effects of an OU on observation‐status patient outcomes. The Pediatric Health Information System (PHIS) database provides a unique opportunity to explore this question, because unlike other national hospital administrative databases,[15, 16] the PHIS dataset contains information about children under observation status. In addition, we know which PHIS hospitals had a dedicated OU in 2011.7
We hypothesized that overall observation‐status stays in hospitals with a dedicated OU would be of shorter duration with earlier discharges at lower cost than observation‐status stays in hospitals without a dedicated OU. We compared hospitals with and without a dedicated OU on secondary outcomes including rates of conversion to inpatient status and return care for any reason.
METHODS
We conducted a cross‐sectional analysis of hospital administrative data using the 2011 PHIS databasea national administrative database that contains resource utilization data from 43 participating hospitals located in 26 states plus the District of Columbia. These hospitals account for approximately 20% of pediatric hospitalizations in the United States.
For each hospital encounter, PHIS includes patient demographics, up to 41 International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) diagnoses, up to 41 ICD‐9‐CM procedures, and hospital charges for services. Data are deidentified prior to inclusion, but unique identifiers allow for determination of return visits and readmissions following an index visit for an individual patient. Data quality and reliability are assured jointly by the Children's Hospital Association (formerly Child Health Corporation of America, Overland Park, KS), participating hospitals, and Truven Health Analytics (New York, NY). This study, using administrative data, was not considered human subjects research by the policies of the Cincinnati Children's Hospital Medical Center Institutional Review Board.
Hospital Selection and Hospital Characteristics
The study sample was drawn from the 31 hospitals that reported observation‐status patient data to PHIS in 2011. Analyses were conducted in 2013, at which time 2011 was the most recent year of data. We categorized 14 hospitals as having a dedicated OU during 2011 based on information collected in 2013.7 To summarize briefly, we interviewed by telephone representatives of hospitals responding to an email query as to the presence of a geographically distinct OU for the care of unscheduled patients from the ED. Three of the 14 representatives reported their hospital had 2 OUs, 1 of which was a separate surgical OU. Ten OUs cared for both ED patients and patients with scheduled procedures; 8 units received patients from non‐ED sources. Hospitalists provided staffing in more than half of the OUs.
We attempted to identify administrative data that would signal care delivered in a dedicated OU using hospital charge codes reported to PHIS, but learned this was not possible due to between‐hospital variation in the specificity of the charge codes. Therefore, we were unable to determine if patient care was delivered in a dedicated OU or another setting, such as a general inpatient unit or the ED. Other hospital characteristics available from the PHIS dataset included the number of inpatient beds, ED visits, inpatient admissions, observation‐status stays, and payer mix. We calculated the percentage of ED visits resulting in admission by dividing the number of ED visits with associated inpatient or observation status by the total number of ED visits and the percentage of admissions under observation status by dividing the number of observation‐status stays by the total number of admissions under observation or inpatient status.
Visit Selection and Patient Characteristics
All observation‐status stays regardless of the point of entry into the hospital were eligible for this study. We excluded stays that were birth‐related, included intensive care, or resulted in transfer or death. Patient demographic characteristics used to describe the cohort included age, gender, race/ethnicity, and primary payer. Stays that began in the ED were identified by an emergency room charge within PHIS. Eligible stays were categorized using All Patient Refined Diagnosis Related Groups (APR‐DRGs) version 24 using the ICD‐9‐CM code‐based proprietary 3M software (3M Health Information Systems, St. Paul, MN). We determined the 15 top‐ranking APR‐DRGs among observation‐status stays in hospitals with a dedicated OU and hospitals without. Procedural stays were identified based on procedural APR‐DRGs (eg, tonsil and adenoid procedures) or the presence of an ICD‐9‐CM procedure code (eg, 331 spinal tap).
Measured Outcomes
Outcomes of observation‐status stays were determined within 4 categories: (1) LOS, (2) standardized costs, (3) conversion to inpatient status, and (4) return visits and readmissions. LOS was calculated in terms of nights spent in hospital for all stays by subtracting the discharge date from the admission date and in terms of hours for stays in the 28 hospitals that report admission and discharge hour to the PHIS database. Discharge timing was examined in 4, 6‐hour blocks starting at midnight. Standardized costs were derived from a charge master index that was created by taking the median costs from all PHIS hospitals for each charged service.[17] Standardized costs represent the estimated cost of providing any particular clinical activity but are not the cost to patients, nor do they represent the actual cost to any given hospital. This approach allows for cost comparisons across hospitals, without biases arising from using charges or from deriving costs using hospitals' ratios of costs to charges.[18] Conversion from observation to inpatient status was calculated by dividing the number of inpatient‐status stays with observation codes by the number of observation‐statusonly stays plus the number of inpatient‐status stays with observation codes. All‐cause 3‐day ED return visits and 30‐day readmissions to the same hospital were assessed using patient‐specific identifiers that allowed for tracking of ED return visits and readmissions following the index observation stay.
Data Analysis
Descriptive statistics were calculated for hospital and patient characteristics using medians and interquartile ranges (IQRs) for continuous factors and frequencies with percentages for categorical factors. Comparisons of these factors between hospitals with dedicated OUs and without were made using [2] and Wilcoxon rank sum tests as appropriate. Multivariable regression was performed using generalized linear mixed models treating hospital as a random effect and used patient age, the case‐mix index based on the APR‐DRG severity of illness, ED visit, and procedures associated with the index observation‐status stay. For continuous outcomes, we performed a log transformation on the outcome, confirmed the normality assumption, and back transformed the results. Sensitivity analyses were conducted to compare LOS, standardized costs, and conversation rates by hospital type for 10 of the 15 top‐ranking APR‐DRGs commonly cared for by pediatric hospitalists and to compare hospitals that reported the presence of an OU that was consistently open (24 hours per day, 7 days per week) and operating during the entire 2011 calendar year, and those without. Based on information gathered from the telephone interviews, hospitals with partially open OUs were similar to hospitals with continuously open OUs, such that they were included in our main analyses. All statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC). P values <0.05 were considered statistically significant.
RESULTS
Hospital Characteristics
Dedicated OUs were present in 14 of the 31 hospitals that reported observation‐status patient data to PHIS (Figure 1). Three of these hospitals had OUs that were open for 5 months or less in 2011; 1 unit opened, 1 unit closed, and 1 hospital operated a seasonal unit. The remaining 17 hospitals reported no OU that admitted unscheduled patients from the ED during 2011. Hospitals with a dedicated OU had more inpatient beds and higher median number of inpatient admissions than those without (Table 1). Hospitals were statistically similar in terms of total volume of ED visits, percentage of ED visits resulting in admission, total number of observation‐status stays, percentage of admissions under observation status, and payer mix.
| Overall, Median (IQR) | Hospitals With a Dedicated Observation Unit, Median (IQR) | Hospitals Without a Dedicated Observation Unit, Median (IQR) | P Value | |
|---|---|---|---|---|
| ||||
| No. of hospitals | 31 | 14 | 17 | |
| Total no. of inpatient beds | 273 (213311) | 304 (269425) | 246 (175293) | 0.006 |
| Total no. ED visits | 62971 (47,50497,723) | 87,892 (55,102117,119) | 53,151 (4750470,882) | 0.21 |
| ED visits resulting in admission, % | 13.1 (9.715.0) | 13.8 (10.5, 19.1) | 12.5 (9.714.5) | 0.31 |
| Total no. of inpatient admissions | 11,537 (9,26814,568) | 13,206 (11,32517,869) | 10,207 (8,64013,363) | 0.04 |
| Admissions under observation status, % | 25.7 (19.733.8) | 25.5 (21.431.4) | 26.0 (16.935.1) | 0.98 |
| Total no. of observation stays | 3,820 (27935672) | 4,850 (3,309 6,196) | 3,141 (2,3654,616) | 0.07 |
| Government payer, % | 60.2 (53.371.2) | 62.1 (54.9, 65.9) | 59.2 (53.373.7) | 0.89 |
Observation‐Status Patients by Hospital Type
In 2011, there were a total of 136,239 observation‐status stays69,983 (51.4%) within the 14 hospitals with a dedicated OU and 66,256 (48.6%) within the 17 hospitals without. Patient care originated in the ED for 57.8% observation‐status stays in hospitals with an OU compared with 53.0% of observation‐status stays in hospitals without (P<0.001). Compared with hospitals with a dedicated OU, those without a dedicated OU had higher percentages of observation‐status patients older than 12 years and non‐Hispanic and a higher percentage of observation‐status patients with private payer type (Table 2). The 15 top‐ranking APR‐DRGs accounted for roughly half of all observation‐status stays and were relatively consistent between hospitals with and without a dedicated OU (Table 3). Procedural care was frequently associated with observation‐status stays.
| Overall, No. (%) | Hospitals With a Dedicated Observation Unit, No. (%)* | Hospitals Without a Dedicated Observation Unit, No. (%) | P Value | |
|---|---|---|---|---|
| ||||
| Age | ||||
| <1 year | 23,845 (17.5) | 12,101 (17.3) | 11,744 (17.7) | <0.001 |
| 15 years | 53,405 (38.5) | 28,052 (40.1) | 24,353 (36.8) | |
| 612 years | 33,674 (24.7) | 17,215 (24.6) | 16,459 (24.8) | |
| 1318 years | 23,607 (17.3) | 11,472 (16.4) | 12,135 (18.3) | |
| >18 years | 2,708 (2) | 1,143 (1.6) | 1,565 (2.4) | |
| Gender | ||||
| Male | 76,142 (55.9) | 39,178 (56) | 36,964 (55.8) | 0.43 |
| Female | 60,025 (44.1) | 30,756 (44) | 29,269 (44.2) | |
| Race/ethnicity | ||||
| Non‐Hispanic white | 72,183 (53.0) | 30,653 (43.8) | 41,530 (62.7) | <0.001 |
| Non‐Hispanic black | 30,995 (22.8) | 16,314 (23.3) | 14,681 (22.2) | |
| Hispanic | 21,255 (15.6) | 16,583 (23.7) | 4,672 (7.1) | |
| Asian | 2,075 (1.5) | 1,313 (1.9) | 762 (1.2) | |
| Non‐Hispanic other | 9,731 (7.1) | 5,120 (7.3) | 4,611 (7.0) | |
| Payer | ||||
| Government | 68,725 (50.4) | 36,967 (52.8) | 31,758 (47.9) | <0.001 |
| Private | 48,416 (35.5) | 21,112 (30.2) | 27,304 (41.2) | |
| Other | 19,098 (14.0) | 11,904 (17) | 7,194 (10.9) | |
| Observation‐Status Patients in Hospitals With a Dedicated Observation Unit* | Observation‐Status Patients in Hospitals Without a Dedicated Observation Unit | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Rank | APR‐DRG | No. | % of All Observation Status Stays | % Began in ED | Rank | APR‐DRG | No. | % of All Observation Status Stays | % Began in ED |
| |||||||||
| 1 | Tonsil and adenoid procedures | 4,621 | 6.6 | 1.3 | 1 | Tonsil and adenoid procedures | 3,806 | 5.7 | 1.6 |
| 2 | Asthma | 4,246 | 6.1 | 85.3 | 2 | Asthma | 3,756 | 5.7 | 79.0 |
| 3 | Seizure | 3,516 | 5.0 | 52.0 | 3 | Seizure | 2,846 | 4.3 | 54.9 |
| 4 | Nonbacterial gastroenteritis | 3,286 | 4.7 | 85.8 | 4 | Upper respiratory infections | 2,733 | 4.1 | 69.6 |
| 5 | Bronchiolitis, RSV pneumonia | 3,093 | 4.4 | 78.5 | 5 | Nonbacterial gastroenteritis | 2,682 | 4.0 | 74.5 |
| 6 | Upper respiratory infections | 2,923 | 4.2 | 80.0 | 6 | Other digestive system diagnoses | 2,545 | 3.8 | 66.3 |
| 7 | Other digestive system diagnoses | 2,064 | 2.9 | 74.0 | 7 | Bronchiolitis, RSV pneumonia | 2,544 | 3.8 | 69.2 |
| 8 | Respiratory signs, symptoms, diagnoses | 2,052 | 2.9 | 81.6 | 8 | Shoulder and arm procedures | 1,862 | 2.8 | 72.6 |
| 9 | Other ENT/cranial/facial diagnoses | 1,684 | 2.4 | 43.6 | 9 | Appendectomy | 1,785 | 2.7 | 79.2 |
| 10 | Shoulder and arm procedures | 1,624 | 2.3 | 79.1 | 10 | Other ENT/cranial/facial diagnoses | 1,624 | 2.5 | 29.9 |
| 11 | Abdominal pain | 1,612 | 2.3 | 86.2 | 11 | Abdominal pain | 1,461 | 2.2 | 82.3 |
| 12 | Fever | 1,494 | 2.1 | 85.1 | 12 | Other factors influencing health status | 1,461 | 2.2 | 66.3 |
| 13 | Appendectomy | 1,465 | 2.1 | 66.4 | 13 | Cellulitis/other bacterial skin infections | 1,383 | 2.1 | 84.2 |
| 14 | Cellulitis/other bacterial skin infections | 1,393 | 2.0 | 86.4 | 14 | Respiratory signs, symptoms, diagnoses | 1,308 | 2.0 | 39.1 |
| 15 | Pneumonia NEC | 1,356 | 1.9 | 79.1 | 15 | Pneumonia NEC | 1,245 | 1.9 | 73.1 |
| Total | 36,429 | 52.0 | 57.8 | Total | 33,041 | 49.87 | 53.0 | ||
Outcomes of Observation‐Status Stays
A greater percentage of observation‐status stays in hospitals with a dedicated OU experienced a same‐day discharge (Table 4). In addition, a higher percentage of discharges occurred between midnight and 11 am in hospitals with a dedicated OU. However, overall risk‐adjusted LOS in hours (12.8 vs 12.2 hours, P=0.90) and risk‐adjusted total standardized costs ($2551 vs $2433, P=0.75) were similar between hospital types. These findings were consistent within the 1 APR‐DRGs commonly cared for by pediatric hospitalists (see Supporting Information, Appendix 1, in the online version of this article). Overall, conversion from observation to inpatient status was significantly higher in hospitals with a dedicated OU compared with hospitals without; however, this pattern was not consistent across the 10 APR‐DRGs commonly cared for by pediatric hospitalists (see Supporting Information, Appendix 1, in the online version of this article). Adjusted odds of 3‐day ED return visits and 30‐day readmissions were comparable between hospital groups.
| Observation‐Status Patients in Hospitals With a Dedicated Observation Unit | Observation‐Status Patients in Hospitals Without a Dedicated Observation Unit | P Value | |
|---|---|---|---|
| |||
| No. of hospitals | 14 | 17 | |
| Length of stay, h, median (IQR) | 12.8 (6.923.7) | 12.2 (721.3) | 0.90 |
| 0 midnights, no. (%) | 16,678 (23.8) | 14,648 (22.1) | <.001 |
| 1 midnight, no. (%) | 46,144 (65.9) | 44,559 (67.3) | |
| 2 midnights or more, no. (%) | 7,161 (10.2) | 7,049 (10.6) | |
| Discharge timing, no. (%) | |||
| Midnight5 am | 1,223 (1.9) | 408 (0.7) | <0.001 |
| 6 am11 am | 18,916 (29.3) | 15,914 (27.1) | |
| Noon5 pm | 32,699 (50.7) | 31,619 (53.9) | |
| 6 pm11 pm | 11,718 (18.2) | 10,718 (18.3) | |
| Total standardized costs, $, median (IQR) | 2,551.3 (2,053.93,169.1) | 2,433.4 (1,998.42,963) | 0.75 |
| Conversion to inpatient status | 11.06% | 9.63% | <0.01 |
| Return care, AOR (95% CI) | |||
| 3‐day ED return visit | 0.93 (0.77‐1.12) | Referent | 0.46 |
| 30‐day readmission | 0.88 (0.67‐1.15) | Referent | 0.36 |
We found similar results in sensitivity analyses comparing observation‐status stays in hospitals with a continuously open OU (open 24 hours per day, 7 days per week, for all of 2011 [n=10 hospitals]) to those without(see Supporting Information, Appendix 2, in the online version of this article). However, there were, on average, more observation‐status stays in hospitals with a continuously open OU (median 5605, IQR 42077089) than hospitals without (median 3309, IQR 26784616) (P=0.04). In contrast to our main results, conversion to inpatient status was lower in hospitals with a continuously open OU compared with hospitals without (8.52% vs 11.57%, P<0.01).
DISCUSSION
Counter to our hypothesis, we did not find hospital‐level differences in length of stay or costs for observation‐status patients cared for in hospitals with and without a dedicated OU, though hospitals with dedicated OUs did have more same‐day discharges and more morning discharges. The lack of observed differences in LOS and costs may reflect the fact that many children under observation status are treated throughout the hospital, even in facilities with a dedicated OU. Access to a dedicated OU is limited by factors including small numbers of OU beds and specific low acuity/low complexity OU admission criteria.[7] The inclusion of all children admitted under observation status in our analyses may have diluted any effect of dedicated OUs at the hospital level, but was necessary due to the inability to identify location of care for children admitted under observation status. Location of care is an important variable that should be incorporated into administrative databases to allow for comparative effectiveness research designs. Until such data are available, chart review at individual hospitals would be necessary to determine which patients received care in an OU.
We did find that discharges for observation‐status patients occurred earlier in the day in hospitals with a dedicated OU when compared with observation‐status patients in hospitals without a dedicated OU. In addition, the percentage of same‐day discharges was higher among observation‐status patients treated in hospitals with a dedicated OU. These differences may stem from policies and procedures that encourage rapid discharge in dedicated OUs, and those practices may affect other care areas. For example, OUs may enforce policies requiring family presence at the bedside or utilize staffing models where doctors and nurses are in frequent communication, both of which would facilitate discharge as soon as a patient no longer required hospital‐based care.[7] A retrospective chart review study design could be used to identify discharge processes and other key characteristics of highly performing OUs.
We found conflicting results in our main and sensitivity analyses related to conversion to inpatient status. Lower percentages of observation‐status patients converting to inpatient status indicates greater success in the delivery of observation care based on established performance metrics.[19] Lower rates of conversion to inpatient status may be the result of stricter admission criteria for some diagnosis and in hospitals with a continuously open dedicate OU, more refined processes for utilization review that allow for patients to be placed into the correct status (observation vs inpatient) at the time of admission, or efforts to educate providers about the designation of observation status.[7] It is also possible that fewer observation‐status patients convert to inpatient status in hospitals with a continuously open dedicated OU because such a change would require movement of the patient to an inpatient bed.
These analyses were more comprehensive than our prior studies[2, 20] in that we included both patients who were treated first in the ED and those who were not. In addition to the APR‐DRGs representative of conditions that have been successfully treated in ED‐based pediatric OUs (eg, asthma, seizures, gastroenteritis, cellulitis),[8, 9, 21, 22] we found observation‐status was commonly associated with procedural care. This population of patients may be relevant to hospitalists who staff OUs that provide both unscheduled and postprocedural care. The colocation of medical and postprocedural patients has been described by others[8, 23] and was reported to occur in over half of the OUs included in this study.[7] The extent to which postprocedure observation care is provided in general OUs staffed by hospitalists represents another opportunity for further study.
Hospitals face many considerations when determining if and how they will provide observation services to patients expected to experience short stays.[7] Some hospitals may be unable to justify an OU for all or part of the year based on the volume of admissions or the costs to staff an OU.[24, 25] Other hospitals may open an OU to promote patient flow and reduce ED crowding.[26] Hospitals may also be influenced by reimbursement policies related to observation‐status stays. Although we did not observe differences in overall payer mix, we did find higher percentages of observation‐status patients in hospitals with dedicated OUs to have public insurance. Although hospital contracts with payers around observation status patients are complex and beyond the scope of this analysis, it is possible that hospitals have established OUs because of increasingly stringent rules or criteria to meet inpatient status or experiences with high volumes of observation‐status patients covered by a particular payer. Nevertheless, the brief nature of many pediatric hospitalizations and the scarcity of pediatric OU beds must be considered in policy changes that result from national discussions about the appropriateness of inpatient stays shorter than 2 nights in duration.[27]
Limitations
The primary limitation to our analyses is the lack of ability to identify patients who were treated in a dedicated OU because few hospitals provided data to PHIS that allowed for the identification of the unit or location of care. Second, it is possible that some hospitals were misclassified as not having a dedicated OU based on our survey, which initially inquired about OUs that provided care to patients first treated in the ED. Therefore, OUs that exclusively care for postoperative patients or patients with scheduled treatments may be present in hospitals that we have labeled as not having a dedicated OU. This potential misclassification would bias our results toward finding no differences. Third, in any study of administrative data there is potential that diagnosis codes are incomplete or inaccurately capture the underlying reason for the episode of care. Fourth, the experiences of the free‐standing children's hospitals that contribute data to PHIS may not be generalizable to other hospitals that provide observation care to children. Finally, return care may be underestimated, as children could receive treatment at another hospital following discharge from a PHIS hospital. Care outside of PHIS hospitals would not be captured, but we do not expect this to differ for hospitals with and without dedicated OUs. It is possible that health information exchanges will permit more comprehensive analyses of care across different hospitals in the future.
CONCLUSION
Observation status patients are similar in hospitals with and without dedicated observation units that admit children from the ED. The presence of a dedicated OU appears to have an influence on same‐day and morning discharges across all observation‐status stays without impacting other hospital‐level outcomes. Inclusion of location of care (eg, geographically distinct dedicated OU vs general inpatient unit vs ED) in hospital administrative datasets would allow for meaningful comparisons of different models of care for short‐stay observation‐status patients.
Acknowledgements
The authors thank John P. Harding, MBA, FACHE, Children's Hospital of the King's Daughters, Norfolk, Virginia for his input on the study design.
Disclosures: Dr. Hall had full access to the data and takes responsibility for the integrity of the data and the accuracy of the data analysis. Internal funds from the Children's Hospital Association supported the conduct of this work. The authors have no financial relationships or conflicts of interest to disclose.
Many pediatric hospitalizations are of short duration, and more than half of short‐stay hospitalizations are designated as observation status.[1, 2] Observation status is an administrative label assigned to patients who do not meet hospital or payer criteria for inpatient‐status care. Short‐stay observation‐status patients do not fit in traditional models of emergency department (ED) or inpatient care. EDs often focus on discharging or admitting patients within a matter of hours, whereas inpatient units tend to measure length of stay (LOS) in terms of days[3] and may not have systems in place to facilitate rapid discharge of short‐stay patients.[4] Observation units (OUs) have been established in some hospitals to address the unique care needs of short‐stay patients.[5, 6, 7]
Single‐site reports from children's hospitals with successful OUs have demonstrated shorter LOS and lower costs compared with inpatient settings.[6, 8, 9, 10, 11, 12, 13, 14] No prior study has examined hospital‐level effects of an OU on observation‐status patient outcomes. The Pediatric Health Information System (PHIS) database provides a unique opportunity to explore this question, because unlike other national hospital administrative databases,[15, 16] the PHIS dataset contains information about children under observation status. In addition, we know which PHIS hospitals had a dedicated OU in 2011.7
We hypothesized that overall observation‐status stays in hospitals with a dedicated OU would be of shorter duration with earlier discharges at lower cost than observation‐status stays in hospitals without a dedicated OU. We compared hospitals with and without a dedicated OU on secondary outcomes including rates of conversion to inpatient status and return care for any reason.
METHODS
We conducted a cross‐sectional analysis of hospital administrative data using the 2011 PHIS databasea national administrative database that contains resource utilization data from 43 participating hospitals located in 26 states plus the District of Columbia. These hospitals account for approximately 20% of pediatric hospitalizations in the United States.
For each hospital encounter, PHIS includes patient demographics, up to 41 International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) diagnoses, up to 41 ICD‐9‐CM procedures, and hospital charges for services. Data are deidentified prior to inclusion, but unique identifiers allow for determination of return visits and readmissions following an index visit for an individual patient. Data quality and reliability are assured jointly by the Children's Hospital Association (formerly Child Health Corporation of America, Overland Park, KS), participating hospitals, and Truven Health Analytics (New York, NY). This study, using administrative data, was not considered human subjects research by the policies of the Cincinnati Children's Hospital Medical Center Institutional Review Board.
Hospital Selection and Hospital Characteristics
The study sample was drawn from the 31 hospitals that reported observation‐status patient data to PHIS in 2011. Analyses were conducted in 2013, at which time 2011 was the most recent year of data. We categorized 14 hospitals as having a dedicated OU during 2011 based on information collected in 2013.7 To summarize briefly, we interviewed by telephone representatives of hospitals responding to an email query as to the presence of a geographically distinct OU for the care of unscheduled patients from the ED. Three of the 14 representatives reported their hospital had 2 OUs, 1 of which was a separate surgical OU. Ten OUs cared for both ED patients and patients with scheduled procedures; 8 units received patients from non‐ED sources. Hospitalists provided staffing in more than half of the OUs.
We attempted to identify administrative data that would signal care delivered in a dedicated OU using hospital charge codes reported to PHIS, but learned this was not possible due to between‐hospital variation in the specificity of the charge codes. Therefore, we were unable to determine if patient care was delivered in a dedicated OU or another setting, such as a general inpatient unit or the ED. Other hospital characteristics available from the PHIS dataset included the number of inpatient beds, ED visits, inpatient admissions, observation‐status stays, and payer mix. We calculated the percentage of ED visits resulting in admission by dividing the number of ED visits with associated inpatient or observation status by the total number of ED visits and the percentage of admissions under observation status by dividing the number of observation‐status stays by the total number of admissions under observation or inpatient status.
Visit Selection and Patient Characteristics
All observation‐status stays regardless of the point of entry into the hospital were eligible for this study. We excluded stays that were birth‐related, included intensive care, or resulted in transfer or death. Patient demographic characteristics used to describe the cohort included age, gender, race/ethnicity, and primary payer. Stays that began in the ED were identified by an emergency room charge within PHIS. Eligible stays were categorized using All Patient Refined Diagnosis Related Groups (APR‐DRGs) version 24 using the ICD‐9‐CM code‐based proprietary 3M software (3M Health Information Systems, St. Paul, MN). We determined the 15 top‐ranking APR‐DRGs among observation‐status stays in hospitals with a dedicated OU and hospitals without. Procedural stays were identified based on procedural APR‐DRGs (eg, tonsil and adenoid procedures) or the presence of an ICD‐9‐CM procedure code (eg, 331 spinal tap).
Measured Outcomes
Outcomes of observation‐status stays were determined within 4 categories: (1) LOS, (2) standardized costs, (3) conversion to inpatient status, and (4) return visits and readmissions. LOS was calculated in terms of nights spent in hospital for all stays by subtracting the discharge date from the admission date and in terms of hours for stays in the 28 hospitals that report admission and discharge hour to the PHIS database. Discharge timing was examined in 4, 6‐hour blocks starting at midnight. Standardized costs were derived from a charge master index that was created by taking the median costs from all PHIS hospitals for each charged service.[17] Standardized costs represent the estimated cost of providing any particular clinical activity but are not the cost to patients, nor do they represent the actual cost to any given hospital. This approach allows for cost comparisons across hospitals, without biases arising from using charges or from deriving costs using hospitals' ratios of costs to charges.[18] Conversion from observation to inpatient status was calculated by dividing the number of inpatient‐status stays with observation codes by the number of observation‐statusonly stays plus the number of inpatient‐status stays with observation codes. All‐cause 3‐day ED return visits and 30‐day readmissions to the same hospital were assessed using patient‐specific identifiers that allowed for tracking of ED return visits and readmissions following the index observation stay.
Data Analysis
Descriptive statistics were calculated for hospital and patient characteristics using medians and interquartile ranges (IQRs) for continuous factors and frequencies with percentages for categorical factors. Comparisons of these factors between hospitals with dedicated OUs and without were made using [2] and Wilcoxon rank sum tests as appropriate. Multivariable regression was performed using generalized linear mixed models treating hospital as a random effect and used patient age, the case‐mix index based on the APR‐DRG severity of illness, ED visit, and procedures associated with the index observation‐status stay. For continuous outcomes, we performed a log transformation on the outcome, confirmed the normality assumption, and back transformed the results. Sensitivity analyses were conducted to compare LOS, standardized costs, and conversation rates by hospital type for 10 of the 15 top‐ranking APR‐DRGs commonly cared for by pediatric hospitalists and to compare hospitals that reported the presence of an OU that was consistently open (24 hours per day, 7 days per week) and operating during the entire 2011 calendar year, and those without. Based on information gathered from the telephone interviews, hospitals with partially open OUs were similar to hospitals with continuously open OUs, such that they were included in our main analyses. All statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC). P values <0.05 were considered statistically significant.
RESULTS
Hospital Characteristics
Dedicated OUs were present in 14 of the 31 hospitals that reported observation‐status patient data to PHIS (Figure 1). Three of these hospitals had OUs that were open for 5 months or less in 2011; 1 unit opened, 1 unit closed, and 1 hospital operated a seasonal unit. The remaining 17 hospitals reported no OU that admitted unscheduled patients from the ED during 2011. Hospitals with a dedicated OU had more inpatient beds and higher median number of inpatient admissions than those without (Table 1). Hospitals were statistically similar in terms of total volume of ED visits, percentage of ED visits resulting in admission, total number of observation‐status stays, percentage of admissions under observation status, and payer mix.
| Overall, Median (IQR) | Hospitals With a Dedicated Observation Unit, Median (IQR) | Hospitals Without a Dedicated Observation Unit, Median (IQR) | P Value | |
|---|---|---|---|---|
| ||||
| No. of hospitals | 31 | 14 | 17 | |
| Total no. of inpatient beds | 273 (213311) | 304 (269425) | 246 (175293) | 0.006 |
| Total no. ED visits | 62971 (47,50497,723) | 87,892 (55,102117,119) | 53,151 (4750470,882) | 0.21 |
| ED visits resulting in admission, % | 13.1 (9.715.0) | 13.8 (10.5, 19.1) | 12.5 (9.714.5) | 0.31 |
| Total no. of inpatient admissions | 11,537 (9,26814,568) | 13,206 (11,32517,869) | 10,207 (8,64013,363) | 0.04 |
| Admissions under observation status, % | 25.7 (19.733.8) | 25.5 (21.431.4) | 26.0 (16.935.1) | 0.98 |
| Total no. of observation stays | 3,820 (27935672) | 4,850 (3,309 6,196) | 3,141 (2,3654,616) | 0.07 |
| Government payer, % | 60.2 (53.371.2) | 62.1 (54.9, 65.9) | 59.2 (53.373.7) | 0.89 |
Observation‐Status Patients by Hospital Type
In 2011, there were a total of 136,239 observation‐status stays69,983 (51.4%) within the 14 hospitals with a dedicated OU and 66,256 (48.6%) within the 17 hospitals without. Patient care originated in the ED for 57.8% observation‐status stays in hospitals with an OU compared with 53.0% of observation‐status stays in hospitals without (P<0.001). Compared with hospitals with a dedicated OU, those without a dedicated OU had higher percentages of observation‐status patients older than 12 years and non‐Hispanic and a higher percentage of observation‐status patients with private payer type (Table 2). The 15 top‐ranking APR‐DRGs accounted for roughly half of all observation‐status stays and were relatively consistent between hospitals with and without a dedicated OU (Table 3). Procedural care was frequently associated with observation‐status stays.
| Overall, No. (%) | Hospitals With a Dedicated Observation Unit, No. (%)* | Hospitals Without a Dedicated Observation Unit, No. (%) | P Value | |
|---|---|---|---|---|
| ||||
| Age | ||||
| <1 year | 23,845 (17.5) | 12,101 (17.3) | 11,744 (17.7) | <0.001 |
| 15 years | 53,405 (38.5) | 28,052 (40.1) | 24,353 (36.8) | |
| 612 years | 33,674 (24.7) | 17,215 (24.6) | 16,459 (24.8) | |
| 1318 years | 23,607 (17.3) | 11,472 (16.4) | 12,135 (18.3) | |
| >18 years | 2,708 (2) | 1,143 (1.6) | 1,565 (2.4) | |
| Gender | ||||
| Male | 76,142 (55.9) | 39,178 (56) | 36,964 (55.8) | 0.43 |
| Female | 60,025 (44.1) | 30,756 (44) | 29,269 (44.2) | |
| Race/ethnicity | ||||
| Non‐Hispanic white | 72,183 (53.0) | 30,653 (43.8) | 41,530 (62.7) | <0.001 |
| Non‐Hispanic black | 30,995 (22.8) | 16,314 (23.3) | 14,681 (22.2) | |
| Hispanic | 21,255 (15.6) | 16,583 (23.7) | 4,672 (7.1) | |
| Asian | 2,075 (1.5) | 1,313 (1.9) | 762 (1.2) | |
| Non‐Hispanic other | 9,731 (7.1) | 5,120 (7.3) | 4,611 (7.0) | |
| Payer | ||||
| Government | 68,725 (50.4) | 36,967 (52.8) | 31,758 (47.9) | <0.001 |
| Private | 48,416 (35.5) | 21,112 (30.2) | 27,304 (41.2) | |
| Other | 19,098 (14.0) | 11,904 (17) | 7,194 (10.9) | |
| Observation‐Status Patients in Hospitals With a Dedicated Observation Unit* | Observation‐Status Patients in Hospitals Without a Dedicated Observation Unit | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Rank | APR‐DRG | No. | % of All Observation Status Stays | % Began in ED | Rank | APR‐DRG | No. | % of All Observation Status Stays | % Began in ED |
| |||||||||
| 1 | Tonsil and adenoid procedures | 4,621 | 6.6 | 1.3 | 1 | Tonsil and adenoid procedures | 3,806 | 5.7 | 1.6 |
| 2 | Asthma | 4,246 | 6.1 | 85.3 | 2 | Asthma | 3,756 | 5.7 | 79.0 |
| 3 | Seizure | 3,516 | 5.0 | 52.0 | 3 | Seizure | 2,846 | 4.3 | 54.9 |
| 4 | Nonbacterial gastroenteritis | 3,286 | 4.7 | 85.8 | 4 | Upper respiratory infections | 2,733 | 4.1 | 69.6 |
| 5 | Bronchiolitis, RSV pneumonia | 3,093 | 4.4 | 78.5 | 5 | Nonbacterial gastroenteritis | 2,682 | 4.0 | 74.5 |
| 6 | Upper respiratory infections | 2,923 | 4.2 | 80.0 | 6 | Other digestive system diagnoses | 2,545 | 3.8 | 66.3 |
| 7 | Other digestive system diagnoses | 2,064 | 2.9 | 74.0 | 7 | Bronchiolitis, RSV pneumonia | 2,544 | 3.8 | 69.2 |
| 8 | Respiratory signs, symptoms, diagnoses | 2,052 | 2.9 | 81.6 | 8 | Shoulder and arm procedures | 1,862 | 2.8 | 72.6 |
| 9 | Other ENT/cranial/facial diagnoses | 1,684 | 2.4 | 43.6 | 9 | Appendectomy | 1,785 | 2.7 | 79.2 |
| 10 | Shoulder and arm procedures | 1,624 | 2.3 | 79.1 | 10 | Other ENT/cranial/facial diagnoses | 1,624 | 2.5 | 29.9 |
| 11 | Abdominal pain | 1,612 | 2.3 | 86.2 | 11 | Abdominal pain | 1,461 | 2.2 | 82.3 |
| 12 | Fever | 1,494 | 2.1 | 85.1 | 12 | Other factors influencing health status | 1,461 | 2.2 | 66.3 |
| 13 | Appendectomy | 1,465 | 2.1 | 66.4 | 13 | Cellulitis/other bacterial skin infections | 1,383 | 2.1 | 84.2 |
| 14 | Cellulitis/other bacterial skin infections | 1,393 | 2.0 | 86.4 | 14 | Respiratory signs, symptoms, diagnoses | 1,308 | 2.0 | 39.1 |
| 15 | Pneumonia NEC | 1,356 | 1.9 | 79.1 | 15 | Pneumonia NEC | 1,245 | 1.9 | 73.1 |
| Total | 36,429 | 52.0 | 57.8 | Total | 33,041 | 49.87 | 53.0 | ||
Outcomes of Observation‐Status Stays
A greater percentage of observation‐status stays in hospitals with a dedicated OU experienced a same‐day discharge (Table 4). In addition, a higher percentage of discharges occurred between midnight and 11 am in hospitals with a dedicated OU. However, overall risk‐adjusted LOS in hours (12.8 vs 12.2 hours, P=0.90) and risk‐adjusted total standardized costs ($2551 vs $2433, P=0.75) were similar between hospital types. These findings were consistent within the 1 APR‐DRGs commonly cared for by pediatric hospitalists (see Supporting Information, Appendix 1, in the online version of this article). Overall, conversion from observation to inpatient status was significantly higher in hospitals with a dedicated OU compared with hospitals without; however, this pattern was not consistent across the 10 APR‐DRGs commonly cared for by pediatric hospitalists (see Supporting Information, Appendix 1, in the online version of this article). Adjusted odds of 3‐day ED return visits and 30‐day readmissions were comparable between hospital groups.
| Observation‐Status Patients in Hospitals With a Dedicated Observation Unit | Observation‐Status Patients in Hospitals Without a Dedicated Observation Unit | P Value | |
|---|---|---|---|
| |||
| No. of hospitals | 14 | 17 | |
| Length of stay, h, median (IQR) | 12.8 (6.923.7) | 12.2 (721.3) | 0.90 |
| 0 midnights, no. (%) | 16,678 (23.8) | 14,648 (22.1) | <.001 |
| 1 midnight, no. (%) | 46,144 (65.9) | 44,559 (67.3) | |
| 2 midnights or more, no. (%) | 7,161 (10.2) | 7,049 (10.6) | |
| Discharge timing, no. (%) | |||
| Midnight5 am | 1,223 (1.9) | 408 (0.7) | <0.001 |
| 6 am11 am | 18,916 (29.3) | 15,914 (27.1) | |
| Noon5 pm | 32,699 (50.7) | 31,619 (53.9) | |
| 6 pm11 pm | 11,718 (18.2) | 10,718 (18.3) | |
| Total standardized costs, $, median (IQR) | 2,551.3 (2,053.93,169.1) | 2,433.4 (1,998.42,963) | 0.75 |
| Conversion to inpatient status | 11.06% | 9.63% | <0.01 |
| Return care, AOR (95% CI) | |||
| 3‐day ED return visit | 0.93 (0.77‐1.12) | Referent | 0.46 |
| 30‐day readmission | 0.88 (0.67‐1.15) | Referent | 0.36 |
We found similar results in sensitivity analyses comparing observation‐status stays in hospitals with a continuously open OU (open 24 hours per day, 7 days per week, for all of 2011 [n=10 hospitals]) to those without(see Supporting Information, Appendix 2, in the online version of this article). However, there were, on average, more observation‐status stays in hospitals with a continuously open OU (median 5605, IQR 42077089) than hospitals without (median 3309, IQR 26784616) (P=0.04). In contrast to our main results, conversion to inpatient status was lower in hospitals with a continuously open OU compared with hospitals without (8.52% vs 11.57%, P<0.01).
DISCUSSION
Counter to our hypothesis, we did not find hospital‐level differences in length of stay or costs for observation‐status patients cared for in hospitals with and without a dedicated OU, though hospitals with dedicated OUs did have more same‐day discharges and more morning discharges. The lack of observed differences in LOS and costs may reflect the fact that many children under observation status are treated throughout the hospital, even in facilities with a dedicated OU. Access to a dedicated OU is limited by factors including small numbers of OU beds and specific low acuity/low complexity OU admission criteria.[7] The inclusion of all children admitted under observation status in our analyses may have diluted any effect of dedicated OUs at the hospital level, but was necessary due to the inability to identify location of care for children admitted under observation status. Location of care is an important variable that should be incorporated into administrative databases to allow for comparative effectiveness research designs. Until such data are available, chart review at individual hospitals would be necessary to determine which patients received care in an OU.
We did find that discharges for observation‐status patients occurred earlier in the day in hospitals with a dedicated OU when compared with observation‐status patients in hospitals without a dedicated OU. In addition, the percentage of same‐day discharges was higher among observation‐status patients treated in hospitals with a dedicated OU. These differences may stem from policies and procedures that encourage rapid discharge in dedicated OUs, and those practices may affect other care areas. For example, OUs may enforce policies requiring family presence at the bedside or utilize staffing models where doctors and nurses are in frequent communication, both of which would facilitate discharge as soon as a patient no longer required hospital‐based care.[7] A retrospective chart review study design could be used to identify discharge processes and other key characteristics of highly performing OUs.
We found conflicting results in our main and sensitivity analyses related to conversion to inpatient status. Lower percentages of observation‐status patients converting to inpatient status indicates greater success in the delivery of observation care based on established performance metrics.[19] Lower rates of conversion to inpatient status may be the result of stricter admission criteria for some diagnosis and in hospitals with a continuously open dedicate OU, more refined processes for utilization review that allow for patients to be placed into the correct status (observation vs inpatient) at the time of admission, or efforts to educate providers about the designation of observation status.[7] It is also possible that fewer observation‐status patients convert to inpatient status in hospitals with a continuously open dedicated OU because such a change would require movement of the patient to an inpatient bed.
These analyses were more comprehensive than our prior studies[2, 20] in that we included both patients who were treated first in the ED and those who were not. In addition to the APR‐DRGs representative of conditions that have been successfully treated in ED‐based pediatric OUs (eg, asthma, seizures, gastroenteritis, cellulitis),[8, 9, 21, 22] we found observation‐status was commonly associated with procedural care. This population of patients may be relevant to hospitalists who staff OUs that provide both unscheduled and postprocedural care. The colocation of medical and postprocedural patients has been described by others[8, 23] and was reported to occur in over half of the OUs included in this study.[7] The extent to which postprocedure observation care is provided in general OUs staffed by hospitalists represents another opportunity for further study.
Hospitals face many considerations when determining if and how they will provide observation services to patients expected to experience short stays.[7] Some hospitals may be unable to justify an OU for all or part of the year based on the volume of admissions or the costs to staff an OU.[24, 25] Other hospitals may open an OU to promote patient flow and reduce ED crowding.[26] Hospitals may also be influenced by reimbursement policies related to observation‐status stays. Although we did not observe differences in overall payer mix, we did find higher percentages of observation‐status patients in hospitals with dedicated OUs to have public insurance. Although hospital contracts with payers around observation status patients are complex and beyond the scope of this analysis, it is possible that hospitals have established OUs because of increasingly stringent rules or criteria to meet inpatient status or experiences with high volumes of observation‐status patients covered by a particular payer. Nevertheless, the brief nature of many pediatric hospitalizations and the scarcity of pediatric OU beds must be considered in policy changes that result from national discussions about the appropriateness of inpatient stays shorter than 2 nights in duration.[27]
Limitations
The primary limitation to our analyses is the lack of ability to identify patients who were treated in a dedicated OU because few hospitals provided data to PHIS that allowed for the identification of the unit or location of care. Second, it is possible that some hospitals were misclassified as not having a dedicated OU based on our survey, which initially inquired about OUs that provided care to patients first treated in the ED. Therefore, OUs that exclusively care for postoperative patients or patients with scheduled treatments may be present in hospitals that we have labeled as not having a dedicated OU. This potential misclassification would bias our results toward finding no differences. Third, in any study of administrative data there is potential that diagnosis codes are incomplete or inaccurately capture the underlying reason for the episode of care. Fourth, the experiences of the free‐standing children's hospitals that contribute data to PHIS may not be generalizable to other hospitals that provide observation care to children. Finally, return care may be underestimated, as children could receive treatment at another hospital following discharge from a PHIS hospital. Care outside of PHIS hospitals would not be captured, but we do not expect this to differ for hospitals with and without dedicated OUs. It is possible that health information exchanges will permit more comprehensive analyses of care across different hospitals in the future.
CONCLUSION
Observation status patients are similar in hospitals with and without dedicated observation units that admit children from the ED. The presence of a dedicated OU appears to have an influence on same‐day and morning discharges across all observation‐status stays without impacting other hospital‐level outcomes. Inclusion of location of care (eg, geographically distinct dedicated OU vs general inpatient unit vs ED) in hospital administrative datasets would allow for meaningful comparisons of different models of care for short‐stay observation‐status patients.
Acknowledgements
The authors thank John P. Harding, MBA, FACHE, Children's Hospital of the King's Daughters, Norfolk, Virginia for his input on the study design.
Disclosures: Dr. Hall had full access to the data and takes responsibility for the integrity of the data and the accuracy of the data analysis. Internal funds from the Children's Hospital Association supported the conduct of this work. The authors have no financial relationships or conflicts of interest to disclose.
Many pediatric hospitalizations are of short duration, and more than half of short‐stay hospitalizations are designated as observation status.[1, 2] Observation status is an administrative label assigned to patients who do not meet hospital or payer criteria for inpatient‐status care. Short‐stay observation‐status patients do not fit in traditional models of emergency department (ED) or inpatient care. EDs often focus on discharging or admitting patients within a matter of hours, whereas inpatient units tend to measure length of stay (LOS) in terms of days[3] and may not have systems in place to facilitate rapid discharge of short‐stay patients.[4] Observation units (OUs) have been established in some hospitals to address the unique care needs of short‐stay patients.[5, 6, 7]
Single‐site reports from children's hospitals with successful OUs have demonstrated shorter LOS and lower costs compared with inpatient settings.[6, 8, 9, 10, 11, 12, 13, 14] No prior study has examined hospital‐level effects of an OU on observation‐status patient outcomes. The Pediatric Health Information System (PHIS) database provides a unique opportunity to explore this question, because unlike other national hospital administrative databases,[15, 16] the PHIS dataset contains information about children under observation status. In addition, we know which PHIS hospitals had a dedicated OU in 2011.7
We hypothesized that overall observation‐status stays in hospitals with a dedicated OU would be of shorter duration with earlier discharges at lower cost than observation‐status stays in hospitals without a dedicated OU. We compared hospitals with and without a dedicated OU on secondary outcomes including rates of conversion to inpatient status and return care for any reason.
METHODS
We conducted a cross‐sectional analysis of hospital administrative data using the 2011 PHIS databasea national administrative database that contains resource utilization data from 43 participating hospitals located in 26 states plus the District of Columbia. These hospitals account for approximately 20% of pediatric hospitalizations in the United States.
For each hospital encounter, PHIS includes patient demographics, up to 41 International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) diagnoses, up to 41 ICD‐9‐CM procedures, and hospital charges for services. Data are deidentified prior to inclusion, but unique identifiers allow for determination of return visits and readmissions following an index visit for an individual patient. Data quality and reliability are assured jointly by the Children's Hospital Association (formerly Child Health Corporation of America, Overland Park, KS), participating hospitals, and Truven Health Analytics (New York, NY). This study, using administrative data, was not considered human subjects research by the policies of the Cincinnati Children's Hospital Medical Center Institutional Review Board.
Hospital Selection and Hospital Characteristics
The study sample was drawn from the 31 hospitals that reported observation‐status patient data to PHIS in 2011. Analyses were conducted in 2013, at which time 2011 was the most recent year of data. We categorized 14 hospitals as having a dedicated OU during 2011 based on information collected in 2013.7 To summarize briefly, we interviewed by telephone representatives of hospitals responding to an email query as to the presence of a geographically distinct OU for the care of unscheduled patients from the ED. Three of the 14 representatives reported their hospital had 2 OUs, 1 of which was a separate surgical OU. Ten OUs cared for both ED patients and patients with scheduled procedures; 8 units received patients from non‐ED sources. Hospitalists provided staffing in more than half of the OUs.
We attempted to identify administrative data that would signal care delivered in a dedicated OU using hospital charge codes reported to PHIS, but learned this was not possible due to between‐hospital variation in the specificity of the charge codes. Therefore, we were unable to determine if patient care was delivered in a dedicated OU or another setting, such as a general inpatient unit or the ED. Other hospital characteristics available from the PHIS dataset included the number of inpatient beds, ED visits, inpatient admissions, observation‐status stays, and payer mix. We calculated the percentage of ED visits resulting in admission by dividing the number of ED visits with associated inpatient or observation status by the total number of ED visits and the percentage of admissions under observation status by dividing the number of observation‐status stays by the total number of admissions under observation or inpatient status.
Visit Selection and Patient Characteristics
All observation‐status stays regardless of the point of entry into the hospital were eligible for this study. We excluded stays that were birth‐related, included intensive care, or resulted in transfer or death. Patient demographic characteristics used to describe the cohort included age, gender, race/ethnicity, and primary payer. Stays that began in the ED were identified by an emergency room charge within PHIS. Eligible stays were categorized using All Patient Refined Diagnosis Related Groups (APR‐DRGs) version 24 using the ICD‐9‐CM code‐based proprietary 3M software (3M Health Information Systems, St. Paul, MN). We determined the 15 top‐ranking APR‐DRGs among observation‐status stays in hospitals with a dedicated OU and hospitals without. Procedural stays were identified based on procedural APR‐DRGs (eg, tonsil and adenoid procedures) or the presence of an ICD‐9‐CM procedure code (eg, 331 spinal tap).
Measured Outcomes
Outcomes of observation‐status stays were determined within 4 categories: (1) LOS, (2) standardized costs, (3) conversion to inpatient status, and (4) return visits and readmissions. LOS was calculated in terms of nights spent in hospital for all stays by subtracting the discharge date from the admission date and in terms of hours for stays in the 28 hospitals that report admission and discharge hour to the PHIS database. Discharge timing was examined in 4, 6‐hour blocks starting at midnight. Standardized costs were derived from a charge master index that was created by taking the median costs from all PHIS hospitals for each charged service.[17] Standardized costs represent the estimated cost of providing any particular clinical activity but are not the cost to patients, nor do they represent the actual cost to any given hospital. This approach allows for cost comparisons across hospitals, without biases arising from using charges or from deriving costs using hospitals' ratios of costs to charges.[18] Conversion from observation to inpatient status was calculated by dividing the number of inpatient‐status stays with observation codes by the number of observation‐statusonly stays plus the number of inpatient‐status stays with observation codes. All‐cause 3‐day ED return visits and 30‐day readmissions to the same hospital were assessed using patient‐specific identifiers that allowed for tracking of ED return visits and readmissions following the index observation stay.
Data Analysis
Descriptive statistics were calculated for hospital and patient characteristics using medians and interquartile ranges (IQRs) for continuous factors and frequencies with percentages for categorical factors. Comparisons of these factors between hospitals with dedicated OUs and without were made using [2] and Wilcoxon rank sum tests as appropriate. Multivariable regression was performed using generalized linear mixed models treating hospital as a random effect and used patient age, the case‐mix index based on the APR‐DRG severity of illness, ED visit, and procedures associated with the index observation‐status stay. For continuous outcomes, we performed a log transformation on the outcome, confirmed the normality assumption, and back transformed the results. Sensitivity analyses were conducted to compare LOS, standardized costs, and conversation rates by hospital type for 10 of the 15 top‐ranking APR‐DRGs commonly cared for by pediatric hospitalists and to compare hospitals that reported the presence of an OU that was consistently open (24 hours per day, 7 days per week) and operating during the entire 2011 calendar year, and those without. Based on information gathered from the telephone interviews, hospitals with partially open OUs were similar to hospitals with continuously open OUs, such that they were included in our main analyses. All statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC). P values <0.05 were considered statistically significant.
RESULTS
Hospital Characteristics
Dedicated OUs were present in 14 of the 31 hospitals that reported observation‐status patient data to PHIS (Figure 1). Three of these hospitals had OUs that were open for 5 months or less in 2011; 1 unit opened, 1 unit closed, and 1 hospital operated a seasonal unit. The remaining 17 hospitals reported no OU that admitted unscheduled patients from the ED during 2011. Hospitals with a dedicated OU had more inpatient beds and higher median number of inpatient admissions than those without (Table 1). Hospitals were statistically similar in terms of total volume of ED visits, percentage of ED visits resulting in admission, total number of observation‐status stays, percentage of admissions under observation status, and payer mix.
| Overall, Median (IQR) | Hospitals With a Dedicated Observation Unit, Median (IQR) | Hospitals Without a Dedicated Observation Unit, Median (IQR) | P Value | |
|---|---|---|---|---|
| ||||
| No. of hospitals | 31 | 14 | 17 | |
| Total no. of inpatient beds | 273 (213311) | 304 (269425) | 246 (175293) | 0.006 |
| Total no. ED visits | 62971 (47,50497,723) | 87,892 (55,102117,119) | 53,151 (4750470,882) | 0.21 |
| ED visits resulting in admission, % | 13.1 (9.715.0) | 13.8 (10.5, 19.1) | 12.5 (9.714.5) | 0.31 |
| Total no. of inpatient admissions | 11,537 (9,26814,568) | 13,206 (11,32517,869) | 10,207 (8,64013,363) | 0.04 |
| Admissions under observation status, % | 25.7 (19.733.8) | 25.5 (21.431.4) | 26.0 (16.935.1) | 0.98 |
| Total no. of observation stays | 3,820 (27935672) | 4,850 (3,309 6,196) | 3,141 (2,3654,616) | 0.07 |
| Government payer, % | 60.2 (53.371.2) | 62.1 (54.9, 65.9) | 59.2 (53.373.7) | 0.89 |
Observation‐Status Patients by Hospital Type
In 2011, there were a total of 136,239 observation‐status stays69,983 (51.4%) within the 14 hospitals with a dedicated OU and 66,256 (48.6%) within the 17 hospitals without. Patient care originated in the ED for 57.8% observation‐status stays in hospitals with an OU compared with 53.0% of observation‐status stays in hospitals without (P<0.001). Compared with hospitals with a dedicated OU, those without a dedicated OU had higher percentages of observation‐status patients older than 12 years and non‐Hispanic and a higher percentage of observation‐status patients with private payer type (Table 2). The 15 top‐ranking APR‐DRGs accounted for roughly half of all observation‐status stays and were relatively consistent between hospitals with and without a dedicated OU (Table 3). Procedural care was frequently associated with observation‐status stays.
| Overall, No. (%) | Hospitals With a Dedicated Observation Unit, No. (%)* | Hospitals Without a Dedicated Observation Unit, No. (%) | P Value | |
|---|---|---|---|---|
| ||||
| Age | ||||
| <1 year | 23,845 (17.5) | 12,101 (17.3) | 11,744 (17.7) | <0.001 |
| 15 years | 53,405 (38.5) | 28,052 (40.1) | 24,353 (36.8) | |
| 612 years | 33,674 (24.7) | 17,215 (24.6) | 16,459 (24.8) | |
| 1318 years | 23,607 (17.3) | 11,472 (16.4) | 12,135 (18.3) | |
| >18 years | 2,708 (2) | 1,143 (1.6) | 1,565 (2.4) | |
| Gender | ||||
| Male | 76,142 (55.9) | 39,178 (56) | 36,964 (55.8) | 0.43 |
| Female | 60,025 (44.1) | 30,756 (44) | 29,269 (44.2) | |
| Race/ethnicity | ||||
| Non‐Hispanic white | 72,183 (53.0) | 30,653 (43.8) | 41,530 (62.7) | <0.001 |
| Non‐Hispanic black | 30,995 (22.8) | 16,314 (23.3) | 14,681 (22.2) | |
| Hispanic | 21,255 (15.6) | 16,583 (23.7) | 4,672 (7.1) | |
| Asian | 2,075 (1.5) | 1,313 (1.9) | 762 (1.2) | |
| Non‐Hispanic other | 9,731 (7.1) | 5,120 (7.3) | 4,611 (7.0) | |
| Payer | ||||
| Government | 68,725 (50.4) | 36,967 (52.8) | 31,758 (47.9) | <0.001 |
| Private | 48,416 (35.5) | 21,112 (30.2) | 27,304 (41.2) | |
| Other | 19,098 (14.0) | 11,904 (17) | 7,194 (10.9) | |
| Observation‐Status Patients in Hospitals With a Dedicated Observation Unit* | Observation‐Status Patients in Hospitals Without a Dedicated Observation Unit | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Rank | APR‐DRG | No. | % of All Observation Status Stays | % Began in ED | Rank | APR‐DRG | No. | % of All Observation Status Stays | % Began in ED |
| |||||||||
| 1 | Tonsil and adenoid procedures | 4,621 | 6.6 | 1.3 | 1 | Tonsil and adenoid procedures | 3,806 | 5.7 | 1.6 |
| 2 | Asthma | 4,246 | 6.1 | 85.3 | 2 | Asthma | 3,756 | 5.7 | 79.0 |
| 3 | Seizure | 3,516 | 5.0 | 52.0 | 3 | Seizure | 2,846 | 4.3 | 54.9 |
| 4 | Nonbacterial gastroenteritis | 3,286 | 4.7 | 85.8 | 4 | Upper respiratory infections | 2,733 | 4.1 | 69.6 |
| 5 | Bronchiolitis, RSV pneumonia | 3,093 | 4.4 | 78.5 | 5 | Nonbacterial gastroenteritis | 2,682 | 4.0 | 74.5 |
| 6 | Upper respiratory infections | 2,923 | 4.2 | 80.0 | 6 | Other digestive system diagnoses | 2,545 | 3.8 | 66.3 |
| 7 | Other digestive system diagnoses | 2,064 | 2.9 | 74.0 | 7 | Bronchiolitis, RSV pneumonia | 2,544 | 3.8 | 69.2 |
| 8 | Respiratory signs, symptoms, diagnoses | 2,052 | 2.9 | 81.6 | 8 | Shoulder and arm procedures | 1,862 | 2.8 | 72.6 |
| 9 | Other ENT/cranial/facial diagnoses | 1,684 | 2.4 | 43.6 | 9 | Appendectomy | 1,785 | 2.7 | 79.2 |
| 10 | Shoulder and arm procedures | 1,624 | 2.3 | 79.1 | 10 | Other ENT/cranial/facial diagnoses | 1,624 | 2.5 | 29.9 |
| 11 | Abdominal pain | 1,612 | 2.3 | 86.2 | 11 | Abdominal pain | 1,461 | 2.2 | 82.3 |
| 12 | Fever | 1,494 | 2.1 | 85.1 | 12 | Other factors influencing health status | 1,461 | 2.2 | 66.3 |
| 13 | Appendectomy | 1,465 | 2.1 | 66.4 | 13 | Cellulitis/other bacterial skin infections | 1,383 | 2.1 | 84.2 |
| 14 | Cellulitis/other bacterial skin infections | 1,393 | 2.0 | 86.4 | 14 | Respiratory signs, symptoms, diagnoses | 1,308 | 2.0 | 39.1 |
| 15 | Pneumonia NEC | 1,356 | 1.9 | 79.1 | 15 | Pneumonia NEC | 1,245 | 1.9 | 73.1 |
| Total | 36,429 | 52.0 | 57.8 | Total | 33,041 | 49.87 | 53.0 | ||
Outcomes of Observation‐Status Stays
A greater percentage of observation‐status stays in hospitals with a dedicated OU experienced a same‐day discharge (Table 4). In addition, a higher percentage of discharges occurred between midnight and 11 am in hospitals with a dedicated OU. However, overall risk‐adjusted LOS in hours (12.8 vs 12.2 hours, P=0.90) and risk‐adjusted total standardized costs ($2551 vs $2433, P=0.75) were similar between hospital types. These findings were consistent within the 1 APR‐DRGs commonly cared for by pediatric hospitalists (see Supporting Information, Appendix 1, in the online version of this article). Overall, conversion from observation to inpatient status was significantly higher in hospitals with a dedicated OU compared with hospitals without; however, this pattern was not consistent across the 10 APR‐DRGs commonly cared for by pediatric hospitalists (see Supporting Information, Appendix 1, in the online version of this article). Adjusted odds of 3‐day ED return visits and 30‐day readmissions were comparable between hospital groups.
| Observation‐Status Patients in Hospitals With a Dedicated Observation Unit | Observation‐Status Patients in Hospitals Without a Dedicated Observation Unit | P Value | |
|---|---|---|---|
| |||
| No. of hospitals | 14 | 17 | |
| Length of stay, h, median (IQR) | 12.8 (6.923.7) | 12.2 (721.3) | 0.90 |
| 0 midnights, no. (%) | 16,678 (23.8) | 14,648 (22.1) | <.001 |
| 1 midnight, no. (%) | 46,144 (65.9) | 44,559 (67.3) | |
| 2 midnights or more, no. (%) | 7,161 (10.2) | 7,049 (10.6) | |
| Discharge timing, no. (%) | |||
| Midnight5 am | 1,223 (1.9) | 408 (0.7) | <0.001 |
| 6 am11 am | 18,916 (29.3) | 15,914 (27.1) | |
| Noon5 pm | 32,699 (50.7) | 31,619 (53.9) | |
| 6 pm11 pm | 11,718 (18.2) | 10,718 (18.3) | |
| Total standardized costs, $, median (IQR) | 2,551.3 (2,053.93,169.1) | 2,433.4 (1,998.42,963) | 0.75 |
| Conversion to inpatient status | 11.06% | 9.63% | <0.01 |
| Return care, AOR (95% CI) | |||
| 3‐day ED return visit | 0.93 (0.77‐1.12) | Referent | 0.46 |
| 30‐day readmission | 0.88 (0.67‐1.15) | Referent | 0.36 |
We found similar results in sensitivity analyses comparing observation‐status stays in hospitals with a continuously open OU (open 24 hours per day, 7 days per week, for all of 2011 [n=10 hospitals]) to those without(see Supporting Information, Appendix 2, in the online version of this article). However, there were, on average, more observation‐status stays in hospitals with a continuously open OU (median 5605, IQR 42077089) than hospitals without (median 3309, IQR 26784616) (P=0.04). In contrast to our main results, conversion to inpatient status was lower in hospitals with a continuously open OU compared with hospitals without (8.52% vs 11.57%, P<0.01).
DISCUSSION
Counter to our hypothesis, we did not find hospital‐level differences in length of stay or costs for observation‐status patients cared for in hospitals with and without a dedicated OU, though hospitals with dedicated OUs did have more same‐day discharges and more morning discharges. The lack of observed differences in LOS and costs may reflect the fact that many children under observation status are treated throughout the hospital, even in facilities with a dedicated OU. Access to a dedicated OU is limited by factors including small numbers of OU beds and specific low acuity/low complexity OU admission criteria.[7] The inclusion of all children admitted under observation status in our analyses may have diluted any effect of dedicated OUs at the hospital level, but was necessary due to the inability to identify location of care for children admitted under observation status. Location of care is an important variable that should be incorporated into administrative databases to allow for comparative effectiveness research designs. Until such data are available, chart review at individual hospitals would be necessary to determine which patients received care in an OU.
We did find that discharges for observation‐status patients occurred earlier in the day in hospitals with a dedicated OU when compared with observation‐status patients in hospitals without a dedicated OU. In addition, the percentage of same‐day discharges was higher among observation‐status patients treated in hospitals with a dedicated OU. These differences may stem from policies and procedures that encourage rapid discharge in dedicated OUs, and those practices may affect other care areas. For example, OUs may enforce policies requiring family presence at the bedside or utilize staffing models where doctors and nurses are in frequent communication, both of which would facilitate discharge as soon as a patient no longer required hospital‐based care.[7] A retrospective chart review study design could be used to identify discharge processes and other key characteristics of highly performing OUs.
We found conflicting results in our main and sensitivity analyses related to conversion to inpatient status. Lower percentages of observation‐status patients converting to inpatient status indicates greater success in the delivery of observation care based on established performance metrics.[19] Lower rates of conversion to inpatient status may be the result of stricter admission criteria for some diagnosis and in hospitals with a continuously open dedicate OU, more refined processes for utilization review that allow for patients to be placed into the correct status (observation vs inpatient) at the time of admission, or efforts to educate providers about the designation of observation status.[7] It is also possible that fewer observation‐status patients convert to inpatient status in hospitals with a continuously open dedicated OU because such a change would require movement of the patient to an inpatient bed.
These analyses were more comprehensive than our prior studies[2, 20] in that we included both patients who were treated first in the ED and those who were not. In addition to the APR‐DRGs representative of conditions that have been successfully treated in ED‐based pediatric OUs (eg, asthma, seizures, gastroenteritis, cellulitis),[8, 9, 21, 22] we found observation‐status was commonly associated with procedural care. This population of patients may be relevant to hospitalists who staff OUs that provide both unscheduled and postprocedural care. The colocation of medical and postprocedural patients has been described by others[8, 23] and was reported to occur in over half of the OUs included in this study.[7] The extent to which postprocedure observation care is provided in general OUs staffed by hospitalists represents another opportunity for further study.
Hospitals face many considerations when determining if and how they will provide observation services to patients expected to experience short stays.[7] Some hospitals may be unable to justify an OU for all or part of the year based on the volume of admissions or the costs to staff an OU.[24, 25] Other hospitals may open an OU to promote patient flow and reduce ED crowding.[26] Hospitals may also be influenced by reimbursement policies related to observation‐status stays. Although we did not observe differences in overall payer mix, we did find higher percentages of observation‐status patients in hospitals with dedicated OUs to have public insurance. Although hospital contracts with payers around observation status patients are complex and beyond the scope of this analysis, it is possible that hospitals have established OUs because of increasingly stringent rules or criteria to meet inpatient status or experiences with high volumes of observation‐status patients covered by a particular payer. Nevertheless, the brief nature of many pediatric hospitalizations and the scarcity of pediatric OU beds must be considered in policy changes that result from national discussions about the appropriateness of inpatient stays shorter than 2 nights in duration.[27]
Limitations
The primary limitation to our analyses is the lack of ability to identify patients who were treated in a dedicated OU because few hospitals provided data to PHIS that allowed for the identification of the unit or location of care. Second, it is possible that some hospitals were misclassified as not having a dedicated OU based on our survey, which initially inquired about OUs that provided care to patients first treated in the ED. Therefore, OUs that exclusively care for postoperative patients or patients with scheduled treatments may be present in hospitals that we have labeled as not having a dedicated OU. This potential misclassification would bias our results toward finding no differences. Third, in any study of administrative data there is potential that diagnosis codes are incomplete or inaccurately capture the underlying reason for the episode of care. Fourth, the experiences of the free‐standing children's hospitals that contribute data to PHIS may not be generalizable to other hospitals that provide observation care to children. Finally, return care may be underestimated, as children could receive treatment at another hospital following discharge from a PHIS hospital. Care outside of PHIS hospitals would not be captured, but we do not expect this to differ for hospitals with and without dedicated OUs. It is possible that health information exchanges will permit more comprehensive analyses of care across different hospitals in the future.
CONCLUSION
Observation status patients are similar in hospitals with and without dedicated observation units that admit children from the ED. The presence of a dedicated OU appears to have an influence on same‐day and morning discharges across all observation‐status stays without impacting other hospital‐level outcomes. Inclusion of location of care (eg, geographically distinct dedicated OU vs general inpatient unit vs ED) in hospital administrative datasets would allow for meaningful comparisons of different models of care for short‐stay observation‐status patients.
Acknowledgements
The authors thank John P. Harding, MBA, FACHE, Children's Hospital of the King's Daughters, Norfolk, Virginia for his input on the study design.
Disclosures: Dr. Hall had full access to the data and takes responsibility for the integrity of the data and the accuracy of the data analysis. Internal funds from the Children's Hospital Association supported the conduct of this work. The authors have no financial relationships or conflicts of interest to disclose.
© 2015 Society of Hospital Medicine
Patient Flow Composite Measurement
Patient flow refers to the management and movement of patients in a healthcare facility. Healthcare institutions utilize patient flow analyses to evaluate and improve aspects of the patient experience including safety, effectiveness, efficiency, timeliness, patient centeredness, and equity.[1, 2, 3, 4, 5, 6, 7, 8] Hospitals can evaluate patient flow using specific metrics, such as time in emergency department (ED) or percent of discharges completed by a certain time of day. However, no single metric can represent the full spectrum of processes inherent to patient flow. For example, ED length of stay (LOS) is dependent on inpatient occupancy, which is dependent on discharge timeliness. Each of these activities depends on various smaller activities, such as cleaning rooms or identifying available beds.
Evaluating the quality that healthcare organizations deliver is growing in importance.[9] Composite scores are being used increasingly to assess clinical processes and outcomes for professionals and institutions.[10, 11] Where various aspects of performance coexist, composite measures can incorporate multiple metrics into a comprehensive summary.[12, 13, 14, 15, 16] They also allow organizations to track a range of metrics for more holistic, comprehensive evaluations.[9, 13]
This article describes a balanced scorecard with composite scoring used at a large urban children's hospital to evaluate patient flow and direct improvement resources where they are needed most.
METHODS
The Children's Hospital of Philadelphia identified patient flow improvement as an operating plan initiative. Previously, performance was measured with a series of independent measures including time from ED arrival to transfer to the inpatient floor, and time from discharge order to room vacancy. These metrics were dismissed as sole measures of flow because they did not reflect the complexity and interdependence of processes or improvement efforts. There were also concerns that efforts to improve a measure caused unintended consequences for others, which at best lead to little overall improvement, and at worst reduced performance elsewhere in the value chain. For example, to meet a goal time for entering discharge orders, physicians could enter orders earlier. But, if patients were not actually ready to leave, their beds were not made available any earlier. Similarly, bed management staff could rush to meet a goal for speed of unit assignment, but this could cause an increase in patients admitted to the wrong specialty floor.
To address these concerns, a group of physicians, nurses, quality improvement specialists, and researchers designed a patient flow scorecard with composite measurement. Five domains of patient flow were identified: (1) ED and ED‐to‐inpatient transition, (2) bed management, (3) discharge process, (4) room turnover and environmental services department (ESD) activities, and (5) scheduling and utilization. Component measures for each domain were selected for 1 of 3 purposes: (1) to correspond to processes of importance to flow and improvement work, (2) to act as adjusters for factors that affect performance, or (3) to act as balancing measures so that progress in a measure would not result in the degradation of another. Each domain was assigned 20 points, which were distributed across the domain's components based on a consensus of the component's relative importance to overall domain performance (Figure 1). Data from the previous year were used as guidelines for setting performance percentile goals. For example, a goal of 80% in 60 minutes for arrival to physician evaluation meant that 80% of patients should see a physician within 1 hour of arriving at the ED.
Scores were also categorized to correspond to commonly used color descriptors.[17] For each component measure, performance meeting or exceeding the goal fell into the green category. Performances <10 percentage points below the goal fell into the yellow category, and performances below that level fell into the red category. Domain‐level scores and overall composite scores were also assigned colors. Performance at or above 80% (16 on the 20‐point domain scale, or 80 on the 100‐point overall scale) were designated green, scores between 70% and 79% were yellow, and scores below 70% were red.
DOMAINS OF THE PATIENT FLOW COMPOSITE SCORE
ED and ED‐to‐Inpatient Transition
Patient progression from the ED to an inpatient unit was separated into 4 steps (Figure 1A): (1) arrival to physician evaluation, (2) ED physician evaluation to decision to admit, (3) decision to admit to medical doctor (MD) report complete, and (4) registered nurse (RN) report to patient to floor. Four additional metrics included: (5) ED LOS for nonadmitted patients, (6) leaving without being seen (LWBS) rate, (7) ED admission rate, and (8) ED volume.
Arrival to physician evaluation measures time between patient arrival in the ED and self‐assignment by the first doctor or nurse practitioner in the electronic record, with a goal of 80% of patients seen within 60 minutes. The component score is calculated as percent of patients meeting this goal (ie, seen within 60 minutes) component weight. ED physician evaluation to decision to admit measures time from the start of the physician evaluation to the decision to admit, using bed request as a proxy; the goal was 80% within 4 hours. Decision to admit to MD report complete measures time from bed request to patient sign‐out to the inpatient floor, with a goal of 80% within 2 hours. RN report to patient to floor measures time from sign‐out to the patient leaving the ED, with a goal of 80% within 1 hour. ED LOS for nonadmitted patients measures time in the ED for patients who are not admitted, and the goal was 80% in <5 hours. The domain also tracks the LWBS rate, with a goal of keeping it below 3%. Its component score is calculated as percent patients seen component weight. ED admission rate is an adjusting factor for the severity of patients visiting the ED. Its component score is calculated as (percent of patients visiting the ED who are admitted to the hospital 5) component weight. Because the average admission rate is around 20%, the percent admitted is multiplied by 5 to more effectively adjust for high‐severity patients. ED volume is an adjusting factor that accounts for high volume. Its component score is calculated as percent of days in a month with more than 250 visits (a threshold chosen by the ED team) component weight. If these days exceed 50%, that percent would be added to the component score as an additional adjustment for excessive volume.
Bed Management
The bed management domain measures how efficiently and effectively patients are assigned to units and beds using 4 metrics (Figure 1B): (1) bed request to unit assignment, (2) unit assignment to bed assignment, (3) percentage of patients placed on right unit for service, and (4) percent of days with peak occupancy >95%.
Bed request to unit assignment measures time from the ED request for a bed in the electronic system to patient being assigned to a unit, with a goal of 80% of assignments made within 20 minutes. Unit assignment to bed assignment measures time from unit assignment to bed assignment, with a goal of 75% within 25 minutes. Because this goal was set to 75% rather than 80%, this component score was multiplied by 80/75 so that all component scores could be compared on the same scale. Percentage of patients placed on right unit for service is a balancing measure for speed of assignment. Because the goal was set to 90% rather than 80%, this component score was also multiplied by an adjusting factor (80/90) so that all components could be compared on the same scale. Percent of days with peak occupancy >95% is an adjusting measure that reflects that locating an appropriate bed takes longer when the hospital is approaching full occupancy. Its component score is calculated as (percent of days with peak occupancy >95% + 1) component weight. The was added to more effectively adjust for high occupancy. If more than 20% of days had peak occupancy greater than 95%, that percent would be added to the component score as an additional adjustment for excessive capacity.
Discharge Process
The discharge process domain measures the efficiency of patient discharge using 2 metrics (Figure 1C): (1) decision to discharge and (2) homeward bound time.
Decision to discharge tracks when clinicians enter electronic discharge orders. The goal was 50% by 1:30 pm for medical services and 10:30 am for surgical services. This encourages physicians to enter discharge orders early to enable downstream discharge work to begin. The component score is calculated as percent entered by goal time component weight (80/50) to adjust the 50% goal up to 80% so all component scores could be compared on the same scale. Homeward bound time measures the time between the discharge order and room vacancy as entered by the unit clerk, with a goal of 80% of patients leaving within 110 minutes for medical services and 240 minutes for surgical services. This balancing measure captures the fact that entering discharge orders early does not facilitate flow if the patients do not actually leave the hospital.
Room Turnover and Environmental Services Department
The room turnover and ESD domain measures the quality of the room turnover processes using 4 metrics (Figure 1D): (1) discharge to in progress time, (2) in progress to complete time, (3) total discharge to clean time, and (4) room cleanliness.
Discharge to in progress time measures time from patient vacancy until ESD staff enters the room, with a goal of 75% within 35 minutes. Because the goal was set to 75% rather than 80%, this component score was multiplied by 80/75 so all component scores could be compared on the same scale. In progress to complete time measures time as entered in the electronic health record from ESD staff entering the room to the room being clean, with a goal of 75% within 55 minutes. The component score is calculated identically to the previous metric. Total discharge to clean time measures the length of the total process, with a goal of 75% within 90 minutes. This component score was also multiplied by 80/75 so that all component scores could be compared on the same scale. Although this repeats the first 2 measures, given workflow and interface issues with our electronic health record (Epic, Epic Systems Corporation, Verona Wisconsin), it is necessary to include a total end‐to‐end measure in addition to the subparts. Patient and family ratings of room cleanliness serve as balancing measures, with the component score calculated as percent satisfaction component weight (80/85) to adjust the 85% satisfaction goal to 80% so all component scores could be compared on the same scale.
Scheduling and Utilization
The scheduling and utilization domain measures hospital operations and variations in bed utilization using 7 metrics including (Figure 1E): (1) coefficient of variation (CV): scheduled admissions, (2) CV: scheduled admissions for weekdays only, (3) CV: emergent admissions, (4) CV: scheduled occupancy, (5) CV: emergent occupancy, (6) percent emergent admissions with LOS >1 day, and (7) percent of days with peak occupancy <95%.
The CV, standard deviation divided by the mean of a distribution, is a measure of dispersion. Because it is a normalized value reported as a percentage, CV can be used to compare variability when sample sizes differ. CV: scheduled admissions captures the variability in admissions coded as an elective across all days in a month. The raw CV score is the standard deviation of the elective admissions for each day divided by the mean. The component score is (1 CV) component weight. A higher CV indicates greater variability, and yields a lower component score. CV on scheduled and emergent occupancy is derived from peak daily occupancy. Percent emergent admissions with LOS >1 day captures the efficiency of bed use, because high volumes of short‐stay patients increases turnover work. Its component score is calculated as the percent of emergent admissions in a month with LOS >1 day component weight. Percent of days with peak occupancy <95% incentivizes the hospital to avoid full occupancy, because effective flow requires that some beds remain open.[18, 19] Its component score is calculated as the percent of days in the month with peak occupancy <95% component weight. Although a similar measure, percent of days with peak occupancy >95%, was an adjusting factor in the bed management domain, it is included again here, because this factor has a unique effect on both domains.
RESULTS
The balanced scorecard with composite measures provided improvement teams and administrators with a picture of patient flow (Figure 2). The overall score provided a global perspective on patient flow over time and captured trends in performance during various states of hospital occupancy. One trend that it captured was an association between high volume and poor composite scores (Figure 3). Notably, the H1N1 influenza pandemic in the fall of 2009 and the turnover of computer systems in January 2011 can be linked to dips in performance. The changes between fiscal years reflect a shift in baseline metrics.
In addition to the overall composite score, the domain level and individual component scores allowed for more specific evaluation of variables affecting quality of care and enabled targeted improvement activities (Figure 4). For example, in December 2010 and January 2011, room turnover and ESD domain scores dropped, especially in the total discharge to clean time component. In response, the ESD made staffing adjustments, and starting in February 2011, component scores and the domain score improved. Feedback from the scheduling and utilization domain scores also initiated positive change. In August 2010, the CV: scheduled occupancy component score started to drop. In response, certain elective admissions were shifted to weekends to distribute hospital occupancy more evenly throughout the week. By February 2011, the component returned to its goal level. This continual evaluation of performance motivates continual improvement.
DISCUSSION
The use of a patient flow balanced scorecard with composite measurement overcomes pitfalls associated with a single or unaggregated measure. Aggregate scores alone mask important differences and relationships among components.[13] For example, 2 domains may be inversely related, or a provider with an overall average score might score above average in 1 domain but below in another. The composite scorecard, however, shows individual component and domain scores in addition to an aggregate score. The individual component and domain level scores highlight specific areas that need improvement and allow attention to be directed to those areas.
Additionally, a composite score is more likely to engage the range of staff involved in patient flow. Scaling out of 100 points and the red‐yellow‐green model are familiar for operations performance and can be easily understood.[17] Moreover, a composite score allows for dynamic performance goals while maintaining a stable measurement structure. For example, standardized LOS ratios, readmission rates, and denied hospital days can be added to the scorecard to provide more information and balancing measures.
Although balanced scorecards with composites can make holistic performance visible across multiple operational domains, they have some disadvantages. First, because there is a degree of complexity associated with a measure that incorporates multiple aspects of flow, certain elements, such as the relationship between a metric and its balancing measure, may not be readily apparent. Second, composite measures may not provide actionable information if the measure is not clearly related to a process that can be improved.[13, 14] Third, individual metrics may not be replicable between locations, so composites may need to be individualized to each setting.[10, 20]
Improving patient flow is a goal at many hospitals. Although measurement is crucial to identifying and mitigating variations, measuring the multidimensional aspects of flow and their impact on quality is difficult. Our scorecard, with composite measurement, addresses the need for an improved method to assess patient flow and improve quality by tracking care processes simultaneously.
Acknowledgements
The authors thank Bhuvaneswari Jayaraman for her contributions to the original calculations for the first version of the composite score.
Disclosures: Internal funds from The Children's Hospital of Philadelphia supported the conduct of this work. The authors report no conflicts of interest.
- AHA Solutions. Patient Flow Challenges Assessment 2009. Chicago, IL: American Hospital Association; 2009.
- , , , et al. The impact of emergency department crowding measures on time to antibiotics for patients with community‐acquired pneumonia. Ann Emerg Med. 2007;50(5):510–516.
- . Practice variation: implications for our health care system. Manag Care. 2004;13(9 suppl):3–7.
- . Managing variability in patient flow is the key to improving access to care, nursing staffing, quality of care, and reducing its cost. Paper presented at: Institute of Medicine; June 24, 2004; Washington, DC.
- , , . Developing models for patient flow and daily surge capacity research. Acad Emerg Med. 2006;13(11):1109–1113.
- , , , , . Patient flow variability and unplanned readmissions to an intensive care unit. Crit Care Med. 2009;37(11):2882–2887.
- , , , , , . Scheduled admissions and high occupancy at a children's hospital. J Hosp Med. 2011;6(2):81–87.
- , , . Frequent overcrowding in US emergency departments. Acad Emerg Med. 2001;8(2):151–155.
- Institute of Medicine. Performance measurement: accelerating improvement. Available at: http://www.iom.edu/Reports/2005/Performance‐Measurement‐Accelerating‐Improvement.aspx. Published December 1, 2005. Accessed December 5, 2012.
- , , , . Emergency department performance measures and benchmarking summit. Acad Emerg Med. 2006;13(10):1074–1080.
- . The Surgical Infection Prevention and Surgical Care Improvement Projects: promises and pitfalls. Am Surg. 2006;72(11):1010–1016; discussion 1021–1030, 1133–1048.
- , , . Patient safety quality indicators. Composite measures workgroup. Final report. Rockville, MD; Agency for Healthcare Research and Quality; 2008.
- , , , et al. ACCF/AHA 2010 position statement on composite measures for healthcare performance assessment: a report of the American College of Cardiology Foundation/American Heart Association Task Force on performance measures (Writing Committee to develop a position statement on composite measures). Circulation. 2010;121(15):1780–1791.
- , . A five‐point checklist to help performance reports incentivize improvement and effectively guide patients. Health Aff (Millwood). 2012;31(3):612–618.
- , , , , . Composite measures for profiling hospitals on surgical morbidity. Ann Surg. 2013;257(1):67–72.
- , . All‐or‐none measurement raises the bar on performance. JAMA. 2006;295(10):1168–1170.
- , , , et al. Quality improvement. Red light‐green light: from kids' game to discharge tool. Healthc Q. 2011;14:77–81.
- , , , . Myths of ideal hospital occupancy. Med J Aust. 2010;192(1):42–43.
- , . Emergency department overcrowding in the United States: an emerging threat to patient safety and public health. Emerg Med J. 2003;20(5):402–405.
- , , , . Emergency department crowding: consensus development of potential measures. Ann Emerg Med. 2003;42(6):824–834.
Patient flow refers to the management and movement of patients in a healthcare facility. Healthcare institutions utilize patient flow analyses to evaluate and improve aspects of the patient experience including safety, effectiveness, efficiency, timeliness, patient centeredness, and equity.[1, 2, 3, 4, 5, 6, 7, 8] Hospitals can evaluate patient flow using specific metrics, such as time in emergency department (ED) or percent of discharges completed by a certain time of day. However, no single metric can represent the full spectrum of processes inherent to patient flow. For example, ED length of stay (LOS) is dependent on inpatient occupancy, which is dependent on discharge timeliness. Each of these activities depends on various smaller activities, such as cleaning rooms or identifying available beds.
Evaluating the quality that healthcare organizations deliver is growing in importance.[9] Composite scores are being used increasingly to assess clinical processes and outcomes for professionals and institutions.[10, 11] Where various aspects of performance coexist, composite measures can incorporate multiple metrics into a comprehensive summary.[12, 13, 14, 15, 16] They also allow organizations to track a range of metrics for more holistic, comprehensive evaluations.[9, 13]
This article describes a balanced scorecard with composite scoring used at a large urban children's hospital to evaluate patient flow and direct improvement resources where they are needed most.
METHODS
The Children's Hospital of Philadelphia identified patient flow improvement as an operating plan initiative. Previously, performance was measured with a series of independent measures including time from ED arrival to transfer to the inpatient floor, and time from discharge order to room vacancy. These metrics were dismissed as sole measures of flow because they did not reflect the complexity and interdependence of processes or improvement efforts. There were also concerns that efforts to improve a measure caused unintended consequences for others, which at best lead to little overall improvement, and at worst reduced performance elsewhere in the value chain. For example, to meet a goal time for entering discharge orders, physicians could enter orders earlier. But, if patients were not actually ready to leave, their beds were not made available any earlier. Similarly, bed management staff could rush to meet a goal for speed of unit assignment, but this could cause an increase in patients admitted to the wrong specialty floor.
To address these concerns, a group of physicians, nurses, quality improvement specialists, and researchers designed a patient flow scorecard with composite measurement. Five domains of patient flow were identified: (1) ED and ED‐to‐inpatient transition, (2) bed management, (3) discharge process, (4) room turnover and environmental services department (ESD) activities, and (5) scheduling and utilization. Component measures for each domain were selected for 1 of 3 purposes: (1) to correspond to processes of importance to flow and improvement work, (2) to act as adjusters for factors that affect performance, or (3) to act as balancing measures so that progress in a measure would not result in the degradation of another. Each domain was assigned 20 points, which were distributed across the domain's components based on a consensus of the component's relative importance to overall domain performance (Figure 1). Data from the previous year were used as guidelines for setting performance percentile goals. For example, a goal of 80% in 60 minutes for arrival to physician evaluation meant that 80% of patients should see a physician within 1 hour of arriving at the ED.
Scores were also categorized to correspond to commonly used color descriptors.[17] For each component measure, performance meeting or exceeding the goal fell into the green category. Performances <10 percentage points below the goal fell into the yellow category, and performances below that level fell into the red category. Domain‐level scores and overall composite scores were also assigned colors. Performance at or above 80% (16 on the 20‐point domain scale, or 80 on the 100‐point overall scale) were designated green, scores between 70% and 79% were yellow, and scores below 70% were red.
DOMAINS OF THE PATIENT FLOW COMPOSITE SCORE
ED and ED‐to‐Inpatient Transition
Patient progression from the ED to an inpatient unit was separated into 4 steps (Figure 1A): (1) arrival to physician evaluation, (2) ED physician evaluation to decision to admit, (3) decision to admit to medical doctor (MD) report complete, and (4) registered nurse (RN) report to patient to floor. Four additional metrics included: (5) ED LOS for nonadmitted patients, (6) leaving without being seen (LWBS) rate, (7) ED admission rate, and (8) ED volume.
Arrival to physician evaluation measures time between patient arrival in the ED and self‐assignment by the first doctor or nurse practitioner in the electronic record, with a goal of 80% of patients seen within 60 minutes. The component score is calculated as percent of patients meeting this goal (ie, seen within 60 minutes) component weight. ED physician evaluation to decision to admit measures time from the start of the physician evaluation to the decision to admit, using bed request as a proxy; the goal was 80% within 4 hours. Decision to admit to MD report complete measures time from bed request to patient sign‐out to the inpatient floor, with a goal of 80% within 2 hours. RN report to patient to floor measures time from sign‐out to the patient leaving the ED, with a goal of 80% within 1 hour. ED LOS for nonadmitted patients measures time in the ED for patients who are not admitted, and the goal was 80% in <5 hours. The domain also tracks the LWBS rate, with a goal of keeping it below 3%. Its component score is calculated as percent patients seen component weight. ED admission rate is an adjusting factor for the severity of patients visiting the ED. Its component score is calculated as (percent of patients visiting the ED who are admitted to the hospital 5) component weight. Because the average admission rate is around 20%, the percent admitted is multiplied by 5 to more effectively adjust for high‐severity patients. ED volume is an adjusting factor that accounts for high volume. Its component score is calculated as percent of days in a month with more than 250 visits (a threshold chosen by the ED team) component weight. If these days exceed 50%, that percent would be added to the component score as an additional adjustment for excessive volume.
Bed Management
The bed management domain measures how efficiently and effectively patients are assigned to units and beds using 4 metrics (Figure 1B): (1) bed request to unit assignment, (2) unit assignment to bed assignment, (3) percentage of patients placed on right unit for service, and (4) percent of days with peak occupancy >95%.
Bed request to unit assignment measures time from the ED request for a bed in the electronic system to patient being assigned to a unit, with a goal of 80% of assignments made within 20 minutes. Unit assignment to bed assignment measures time from unit assignment to bed assignment, with a goal of 75% within 25 minutes. Because this goal was set to 75% rather than 80%, this component score was multiplied by 80/75 so that all component scores could be compared on the same scale. Percentage of patients placed on right unit for service is a balancing measure for speed of assignment. Because the goal was set to 90% rather than 80%, this component score was also multiplied by an adjusting factor (80/90) so that all components could be compared on the same scale. Percent of days with peak occupancy >95% is an adjusting measure that reflects that locating an appropriate bed takes longer when the hospital is approaching full occupancy. Its component score is calculated as (percent of days with peak occupancy >95% + 1) component weight. The was added to more effectively adjust for high occupancy. If more than 20% of days had peak occupancy greater than 95%, that percent would be added to the component score as an additional adjustment for excessive capacity.
Discharge Process
The discharge process domain measures the efficiency of patient discharge using 2 metrics (Figure 1C): (1) decision to discharge and (2) homeward bound time.
Decision to discharge tracks when clinicians enter electronic discharge orders. The goal was 50% by 1:30 pm for medical services and 10:30 am for surgical services. This encourages physicians to enter discharge orders early to enable downstream discharge work to begin. The component score is calculated as percent entered by goal time component weight (80/50) to adjust the 50% goal up to 80% so all component scores could be compared on the same scale. Homeward bound time measures the time between the discharge order and room vacancy as entered by the unit clerk, with a goal of 80% of patients leaving within 110 minutes for medical services and 240 minutes for surgical services. This balancing measure captures the fact that entering discharge orders early does not facilitate flow if the patients do not actually leave the hospital.
Room Turnover and Environmental Services Department
The room turnover and ESD domain measures the quality of the room turnover processes using 4 metrics (Figure 1D): (1) discharge to in progress time, (2) in progress to complete time, (3) total discharge to clean time, and (4) room cleanliness.
Discharge to in progress time measures time from patient vacancy until ESD staff enters the room, with a goal of 75% within 35 minutes. Because the goal was set to 75% rather than 80%, this component score was multiplied by 80/75 so all component scores could be compared on the same scale. In progress to complete time measures time as entered in the electronic health record from ESD staff entering the room to the room being clean, with a goal of 75% within 55 minutes. The component score is calculated identically to the previous metric. Total discharge to clean time measures the length of the total process, with a goal of 75% within 90 minutes. This component score was also multiplied by 80/75 so that all component scores could be compared on the same scale. Although this repeats the first 2 measures, given workflow and interface issues with our electronic health record (Epic, Epic Systems Corporation, Verona Wisconsin), it is necessary to include a total end‐to‐end measure in addition to the subparts. Patient and family ratings of room cleanliness serve as balancing measures, with the component score calculated as percent satisfaction component weight (80/85) to adjust the 85% satisfaction goal to 80% so all component scores could be compared on the same scale.
Scheduling and Utilization
The scheduling and utilization domain measures hospital operations and variations in bed utilization using 7 metrics including (Figure 1E): (1) coefficient of variation (CV): scheduled admissions, (2) CV: scheduled admissions for weekdays only, (3) CV: emergent admissions, (4) CV: scheduled occupancy, (5) CV: emergent occupancy, (6) percent emergent admissions with LOS >1 day, and (7) percent of days with peak occupancy <95%.
The CV, standard deviation divided by the mean of a distribution, is a measure of dispersion. Because it is a normalized value reported as a percentage, CV can be used to compare variability when sample sizes differ. CV: scheduled admissions captures the variability in admissions coded as an elective across all days in a month. The raw CV score is the standard deviation of the elective admissions for each day divided by the mean. The component score is (1 CV) component weight. A higher CV indicates greater variability, and yields a lower component score. CV on scheduled and emergent occupancy is derived from peak daily occupancy. Percent emergent admissions with LOS >1 day captures the efficiency of bed use, because high volumes of short‐stay patients increases turnover work. Its component score is calculated as the percent of emergent admissions in a month with LOS >1 day component weight. Percent of days with peak occupancy <95% incentivizes the hospital to avoid full occupancy, because effective flow requires that some beds remain open.[18, 19] Its component score is calculated as the percent of days in the month with peak occupancy <95% component weight. Although a similar measure, percent of days with peak occupancy >95%, was an adjusting factor in the bed management domain, it is included again here, because this factor has a unique effect on both domains.
RESULTS
The balanced scorecard with composite measures provided improvement teams and administrators with a picture of patient flow (Figure 2). The overall score provided a global perspective on patient flow over time and captured trends in performance during various states of hospital occupancy. One trend that it captured was an association between high volume and poor composite scores (Figure 3). Notably, the H1N1 influenza pandemic in the fall of 2009 and the turnover of computer systems in January 2011 can be linked to dips in performance. The changes between fiscal years reflect a shift in baseline metrics.
In addition to the overall composite score, the domain level and individual component scores allowed for more specific evaluation of variables affecting quality of care and enabled targeted improvement activities (Figure 4). For example, in December 2010 and January 2011, room turnover and ESD domain scores dropped, especially in the total discharge to clean time component. In response, the ESD made staffing adjustments, and starting in February 2011, component scores and the domain score improved. Feedback from the scheduling and utilization domain scores also initiated positive change. In August 2010, the CV: scheduled occupancy component score started to drop. In response, certain elective admissions were shifted to weekends to distribute hospital occupancy more evenly throughout the week. By February 2011, the component returned to its goal level. This continual evaluation of performance motivates continual improvement.
DISCUSSION
The use of a patient flow balanced scorecard with composite measurement overcomes pitfalls associated with a single or unaggregated measure. Aggregate scores alone mask important differences and relationships among components.[13] For example, 2 domains may be inversely related, or a provider with an overall average score might score above average in 1 domain but below in another. The composite scorecard, however, shows individual component and domain scores in addition to an aggregate score. The individual component and domain level scores highlight specific areas that need improvement and allow attention to be directed to those areas.
Additionally, a composite score is more likely to engage the range of staff involved in patient flow. Scaling out of 100 points and the red‐yellow‐green model are familiar for operations performance and can be easily understood.[17] Moreover, a composite score allows for dynamic performance goals while maintaining a stable measurement structure. For example, standardized LOS ratios, readmission rates, and denied hospital days can be added to the scorecard to provide more information and balancing measures.
Although balanced scorecards with composites can make holistic performance visible across multiple operational domains, they have some disadvantages. First, because there is a degree of complexity associated with a measure that incorporates multiple aspects of flow, certain elements, such as the relationship between a metric and its balancing measure, may not be readily apparent. Second, composite measures may not provide actionable information if the measure is not clearly related to a process that can be improved.[13, 14] Third, individual metrics may not be replicable between locations, so composites may need to be individualized to each setting.[10, 20]
Improving patient flow is a goal at many hospitals. Although measurement is crucial to identifying and mitigating variations, measuring the multidimensional aspects of flow and their impact on quality is difficult. Our scorecard, with composite measurement, addresses the need for an improved method to assess patient flow and improve quality by tracking care processes simultaneously.
Acknowledgements
The authors thank Bhuvaneswari Jayaraman for her contributions to the original calculations for the first version of the composite score.
Disclosures: Internal funds from The Children's Hospital of Philadelphia supported the conduct of this work. The authors report no conflicts of interest.
Patient flow refers to the management and movement of patients in a healthcare facility. Healthcare institutions utilize patient flow analyses to evaluate and improve aspects of the patient experience including safety, effectiveness, efficiency, timeliness, patient centeredness, and equity.[1, 2, 3, 4, 5, 6, 7, 8] Hospitals can evaluate patient flow using specific metrics, such as time in emergency department (ED) or percent of discharges completed by a certain time of day. However, no single metric can represent the full spectrum of processes inherent to patient flow. For example, ED length of stay (LOS) is dependent on inpatient occupancy, which is dependent on discharge timeliness. Each of these activities depends on various smaller activities, such as cleaning rooms or identifying available beds.
Evaluating the quality that healthcare organizations deliver is growing in importance.[9] Composite scores are being used increasingly to assess clinical processes and outcomes for professionals and institutions.[10, 11] Where various aspects of performance coexist, composite measures can incorporate multiple metrics into a comprehensive summary.[12, 13, 14, 15, 16] They also allow organizations to track a range of metrics for more holistic, comprehensive evaluations.[9, 13]
This article describes a balanced scorecard with composite scoring used at a large urban children's hospital to evaluate patient flow and direct improvement resources where they are needed most.
METHODS
The Children's Hospital of Philadelphia identified patient flow improvement as an operating plan initiative. Previously, performance was measured with a series of independent measures including time from ED arrival to transfer to the inpatient floor, and time from discharge order to room vacancy. These metrics were dismissed as sole measures of flow because they did not reflect the complexity and interdependence of processes or improvement efforts. There were also concerns that efforts to improve a measure caused unintended consequences for others, which at best lead to little overall improvement, and at worst reduced performance elsewhere in the value chain. For example, to meet a goal time for entering discharge orders, physicians could enter orders earlier. But, if patients were not actually ready to leave, their beds were not made available any earlier. Similarly, bed management staff could rush to meet a goal for speed of unit assignment, but this could cause an increase in patients admitted to the wrong specialty floor.
To address these concerns, a group of physicians, nurses, quality improvement specialists, and researchers designed a patient flow scorecard with composite measurement. Five domains of patient flow were identified: (1) ED and ED‐to‐inpatient transition, (2) bed management, (3) discharge process, (4) room turnover and environmental services department (ESD) activities, and (5) scheduling and utilization. Component measures for each domain were selected for 1 of 3 purposes: (1) to correspond to processes of importance to flow and improvement work, (2) to act as adjusters for factors that affect performance, or (3) to act as balancing measures so that progress in a measure would not result in the degradation of another. Each domain was assigned 20 points, which were distributed across the domain's components based on a consensus of the component's relative importance to overall domain performance (Figure 1). Data from the previous year were used as guidelines for setting performance percentile goals. For example, a goal of 80% in 60 minutes for arrival to physician evaluation meant that 80% of patients should see a physician within 1 hour of arriving at the ED.
Scores were also categorized to correspond to commonly used color descriptors.[17] For each component measure, performance meeting or exceeding the goal fell into the green category. Performances <10 percentage points below the goal fell into the yellow category, and performances below that level fell into the red category. Domain‐level scores and overall composite scores were also assigned colors. Performance at or above 80% (16 on the 20‐point domain scale, or 80 on the 100‐point overall scale) were designated green, scores between 70% and 79% were yellow, and scores below 70% were red.
DOMAINS OF THE PATIENT FLOW COMPOSITE SCORE
ED and ED‐to‐Inpatient Transition
Patient progression from the ED to an inpatient unit was separated into 4 steps (Figure 1A): (1) arrival to physician evaluation, (2) ED physician evaluation to decision to admit, (3) decision to admit to medical doctor (MD) report complete, and (4) registered nurse (RN) report to patient to floor. Four additional metrics included: (5) ED LOS for nonadmitted patients, (6) leaving without being seen (LWBS) rate, (7) ED admission rate, and (8) ED volume.
Arrival to physician evaluation measures time between patient arrival in the ED and self‐assignment by the first doctor or nurse practitioner in the electronic record, with a goal of 80% of patients seen within 60 minutes. The component score is calculated as percent of patients meeting this goal (ie, seen within 60 minutes) component weight. ED physician evaluation to decision to admit measures time from the start of the physician evaluation to the decision to admit, using bed request as a proxy; the goal was 80% within 4 hours. Decision to admit to MD report complete measures time from bed request to patient sign‐out to the inpatient floor, with a goal of 80% within 2 hours. RN report to patient to floor measures time from sign‐out to the patient leaving the ED, with a goal of 80% within 1 hour. ED LOS for nonadmitted patients measures time in the ED for patients who are not admitted, and the goal was 80% in <5 hours. The domain also tracks the LWBS rate, with a goal of keeping it below 3%. Its component score is calculated as percent patients seen component weight. ED admission rate is an adjusting factor for the severity of patients visiting the ED. Its component score is calculated as (percent of patients visiting the ED who are admitted to the hospital 5) component weight. Because the average admission rate is around 20%, the percent admitted is multiplied by 5 to more effectively adjust for high‐severity patients. ED volume is an adjusting factor that accounts for high volume. Its component score is calculated as percent of days in a month with more than 250 visits (a threshold chosen by the ED team) component weight. If these days exceed 50%, that percent would be added to the component score as an additional adjustment for excessive volume.
Bed Management
The bed management domain measures how efficiently and effectively patients are assigned to units and beds using 4 metrics (Figure 1B): (1) bed request to unit assignment, (2) unit assignment to bed assignment, (3) percentage of patients placed on right unit for service, and (4) percent of days with peak occupancy >95%.
Bed request to unit assignment measures time from the ED request for a bed in the electronic system to patient being assigned to a unit, with a goal of 80% of assignments made within 20 minutes. Unit assignment to bed assignment measures time from unit assignment to bed assignment, with a goal of 75% within 25 minutes. Because this goal was set to 75% rather than 80%, this component score was multiplied by 80/75 so that all component scores could be compared on the same scale. Percentage of patients placed on right unit for service is a balancing measure for speed of assignment. Because the goal was set to 90% rather than 80%, this component score was also multiplied by an adjusting factor (80/90) so that all components could be compared on the same scale. Percent of days with peak occupancy >95% is an adjusting measure that reflects that locating an appropriate bed takes longer when the hospital is approaching full occupancy. Its component score is calculated as (percent of days with peak occupancy >95% + 1) component weight. The was added to more effectively adjust for high occupancy. If more than 20% of days had peak occupancy greater than 95%, that percent would be added to the component score as an additional adjustment for excessive capacity.
Discharge Process
The discharge process domain measures the efficiency of patient discharge using 2 metrics (Figure 1C): (1) decision to discharge and (2) homeward bound time.
Decision to discharge tracks when clinicians enter electronic discharge orders. The goal was 50% by 1:30 pm for medical services and 10:30 am for surgical services. This encourages physicians to enter discharge orders early to enable downstream discharge work to begin. The component score is calculated as percent entered by goal time component weight (80/50) to adjust the 50% goal up to 80% so all component scores could be compared on the same scale. Homeward bound time measures the time between the discharge order and room vacancy as entered by the unit clerk, with a goal of 80% of patients leaving within 110 minutes for medical services and 240 minutes for surgical services. This balancing measure captures the fact that entering discharge orders early does not facilitate flow if the patients do not actually leave the hospital.
Room Turnover and Environmental Services Department
The room turnover and ESD domain measures the quality of the room turnover processes using 4 metrics (Figure 1D): (1) discharge to in progress time, (2) in progress to complete time, (3) total discharge to clean time, and (4) room cleanliness.
Discharge to in progress time measures time from patient vacancy until ESD staff enters the room, with a goal of 75% within 35 minutes. Because the goal was set to 75% rather than 80%, this component score was multiplied by 80/75 so all component scores could be compared on the same scale. In progress to complete time measures time as entered in the electronic health record from ESD staff entering the room to the room being clean, with a goal of 75% within 55 minutes. The component score is calculated identically to the previous metric. Total discharge to clean time measures the length of the total process, with a goal of 75% within 90 minutes. This component score was also multiplied by 80/75 so that all component scores could be compared on the same scale. Although this repeats the first 2 measures, given workflow and interface issues with our electronic health record (Epic, Epic Systems Corporation, Verona Wisconsin), it is necessary to include a total end‐to‐end measure in addition to the subparts. Patient and family ratings of room cleanliness serve as balancing measures, with the component score calculated as percent satisfaction component weight (80/85) to adjust the 85% satisfaction goal to 80% so all component scores could be compared on the same scale.
Scheduling and Utilization
The scheduling and utilization domain measures hospital operations and variations in bed utilization using 7 metrics including (Figure 1E): (1) coefficient of variation (CV): scheduled admissions, (2) CV: scheduled admissions for weekdays only, (3) CV: emergent admissions, (4) CV: scheduled occupancy, (5) CV: emergent occupancy, (6) percent emergent admissions with LOS >1 day, and (7) percent of days with peak occupancy <95%.
The CV, standard deviation divided by the mean of a distribution, is a measure of dispersion. Because it is a normalized value reported as a percentage, CV can be used to compare variability when sample sizes differ. CV: scheduled admissions captures the variability in admissions coded as an elective across all days in a month. The raw CV score is the standard deviation of the elective admissions for each day divided by the mean. The component score is (1 CV) component weight. A higher CV indicates greater variability, and yields a lower component score. CV on scheduled and emergent occupancy is derived from peak daily occupancy. Percent emergent admissions with LOS >1 day captures the efficiency of bed use, because high volumes of short‐stay patients increases turnover work. Its component score is calculated as the percent of emergent admissions in a month with LOS >1 day component weight. Percent of days with peak occupancy <95% incentivizes the hospital to avoid full occupancy, because effective flow requires that some beds remain open.[18, 19] Its component score is calculated as the percent of days in the month with peak occupancy <95% component weight. Although a similar measure, percent of days with peak occupancy >95%, was an adjusting factor in the bed management domain, it is included again here, because this factor has a unique effect on both domains.
RESULTS
The balanced scorecard with composite measures provided improvement teams and administrators with a picture of patient flow (Figure 2). The overall score provided a global perspective on patient flow over time and captured trends in performance during various states of hospital occupancy. One trend that it captured was an association between high volume and poor composite scores (Figure 3). Notably, the H1N1 influenza pandemic in the fall of 2009 and the turnover of computer systems in January 2011 can be linked to dips in performance. The changes between fiscal years reflect a shift in baseline metrics.
In addition to the overall composite score, the domain level and individual component scores allowed for more specific evaluation of variables affecting quality of care and enabled targeted improvement activities (Figure 4). For example, in December 2010 and January 2011, room turnover and ESD domain scores dropped, especially in the total discharge to clean time component. In response, the ESD made staffing adjustments, and starting in February 2011, component scores and the domain score improved. Feedback from the scheduling and utilization domain scores also initiated positive change. In August 2010, the CV: scheduled occupancy component score started to drop. In response, certain elective admissions were shifted to weekends to distribute hospital occupancy more evenly throughout the week. By February 2011, the component returned to its goal level. This continual evaluation of performance motivates continual improvement.
DISCUSSION
The use of a patient flow balanced scorecard with composite measurement overcomes pitfalls associated with a single or unaggregated measure. Aggregate scores alone mask important differences and relationships among components.[13] For example, 2 domains may be inversely related, or a provider with an overall average score might score above average in 1 domain but below in another. The composite scorecard, however, shows individual component and domain scores in addition to an aggregate score. The individual component and domain level scores highlight specific areas that need improvement and allow attention to be directed to those areas.
Additionally, a composite score is more likely to engage the range of staff involved in patient flow. Scaling out of 100 points and the red‐yellow‐green model are familiar for operations performance and can be easily understood.[17] Moreover, a composite score allows for dynamic performance goals while maintaining a stable measurement structure. For example, standardized LOS ratios, readmission rates, and denied hospital days can be added to the scorecard to provide more information and balancing measures.
Although balanced scorecards with composites can make holistic performance visible across multiple operational domains, they have some disadvantages. First, because there is a degree of complexity associated with a measure that incorporates multiple aspects of flow, certain elements, such as the relationship between a metric and its balancing measure, may not be readily apparent. Second, composite measures may not provide actionable information if the measure is not clearly related to a process that can be improved.[13, 14] Third, individual metrics may not be replicable between locations, so composites may need to be individualized to each setting.[10, 20]
Improving patient flow is a goal at many hospitals. Although measurement is crucial to identifying and mitigating variations, measuring the multidimensional aspects of flow and their impact on quality is difficult. Our scorecard, with composite measurement, addresses the need for an improved method to assess patient flow and improve quality by tracking care processes simultaneously.
Acknowledgements
The authors thank Bhuvaneswari Jayaraman for her contributions to the original calculations for the first version of the composite score.
Disclosures: Internal funds from The Children's Hospital of Philadelphia supported the conduct of this work. The authors report no conflicts of interest.
- AHA Solutions. Patient Flow Challenges Assessment 2009. Chicago, IL: American Hospital Association; 2009.
- , , , et al. The impact of emergency department crowding measures on time to antibiotics for patients with community‐acquired pneumonia. Ann Emerg Med. 2007;50(5):510–516.
- . Practice variation: implications for our health care system. Manag Care. 2004;13(9 suppl):3–7.
- . Managing variability in patient flow is the key to improving access to care, nursing staffing, quality of care, and reducing its cost. Paper presented at: Institute of Medicine; June 24, 2004; Washington, DC.
- , , . Developing models for patient flow and daily surge capacity research. Acad Emerg Med. 2006;13(11):1109–1113.
- , , , , . Patient flow variability and unplanned readmissions to an intensive care unit. Crit Care Med. 2009;37(11):2882–2887.
- , , , , , . Scheduled admissions and high occupancy at a children's hospital. J Hosp Med. 2011;6(2):81–87.
- , , . Frequent overcrowding in US emergency departments. Acad Emerg Med. 2001;8(2):151–155.
- Institute of Medicine. Performance measurement: accelerating improvement. Available at: http://www.iom.edu/Reports/2005/Performance‐Measurement‐Accelerating‐Improvement.aspx. Published December 1, 2005. Accessed December 5, 2012.
- , , , . Emergency department performance measures and benchmarking summit. Acad Emerg Med. 2006;13(10):1074–1080.
- . The Surgical Infection Prevention and Surgical Care Improvement Projects: promises and pitfalls. Am Surg. 2006;72(11):1010–1016; discussion 1021–1030, 1133–1048.
- , , . Patient safety quality indicators. Composite measures workgroup. Final report. Rockville, MD; Agency for Healthcare Research and Quality; 2008.
- , , , et al. ACCF/AHA 2010 position statement on composite measures for healthcare performance assessment: a report of the American College of Cardiology Foundation/American Heart Association Task Force on performance measures (Writing Committee to develop a position statement on composite measures). Circulation. 2010;121(15):1780–1791.
- , . A five‐point checklist to help performance reports incentivize improvement and effectively guide patients. Health Aff (Millwood). 2012;31(3):612–618.
- , , , , . Composite measures for profiling hospitals on surgical morbidity. Ann Surg. 2013;257(1):67–72.
- , . All‐or‐none measurement raises the bar on performance. JAMA. 2006;295(10):1168–1170.
- , , , et al. Quality improvement. Red light‐green light: from kids' game to discharge tool. Healthc Q. 2011;14:77–81.
- , , , . Myths of ideal hospital occupancy. Med J Aust. 2010;192(1):42–43.
- , . Emergency department overcrowding in the United States: an emerging threat to patient safety and public health. Emerg Med J. 2003;20(5):402–405.
- , , , . Emergency department crowding: consensus development of potential measures. Ann Emerg Med. 2003;42(6):824–834.
- AHA Solutions. Patient Flow Challenges Assessment 2009. Chicago, IL: American Hospital Association; 2009.
- , , , et al. The impact of emergency department crowding measures on time to antibiotics for patients with community‐acquired pneumonia. Ann Emerg Med. 2007;50(5):510–516.
- . Practice variation: implications for our health care system. Manag Care. 2004;13(9 suppl):3–7.
- . Managing variability in patient flow is the key to improving access to care, nursing staffing, quality of care, and reducing its cost. Paper presented at: Institute of Medicine; June 24, 2004; Washington, DC.
- , , . Developing models for patient flow and daily surge capacity research. Acad Emerg Med. 2006;13(11):1109–1113.
- , , , , . Patient flow variability and unplanned readmissions to an intensive care unit. Crit Care Med. 2009;37(11):2882–2887.
- , , , , , . Scheduled admissions and high occupancy at a children's hospital. J Hosp Med. 2011;6(2):81–87.
- , , . Frequent overcrowding in US emergency departments. Acad Emerg Med. 2001;8(2):151–155.
- Institute of Medicine. Performance measurement: accelerating improvement. Available at: http://www.iom.edu/Reports/2005/Performance‐Measurement‐Accelerating‐Improvement.aspx. Published December 1, 2005. Accessed December 5, 2012.
- , , , . Emergency department performance measures and benchmarking summit. Acad Emerg Med. 2006;13(10):1074–1080.
- . The Surgical Infection Prevention and Surgical Care Improvement Projects: promises and pitfalls. Am Surg. 2006;72(11):1010–1016; discussion 1021–1030, 1133–1048.
- , , . Patient safety quality indicators. Composite measures workgroup. Final report. Rockville, MD; Agency for Healthcare Research and Quality; 2008.
- , , , et al. ACCF/AHA 2010 position statement on composite measures for healthcare performance assessment: a report of the American College of Cardiology Foundation/American Heart Association Task Force on performance measures (Writing Committee to develop a position statement on composite measures). Circulation. 2010;121(15):1780–1791.
- , . A five‐point checklist to help performance reports incentivize improvement and effectively guide patients. Health Aff (Millwood). 2012;31(3):612–618.
- , , , , . Composite measures for profiling hospitals on surgical morbidity. Ann Surg. 2013;257(1):67–72.
- , . All‐or‐none measurement raises the bar on performance. JAMA. 2006;295(10):1168–1170.
- , , , et al. Quality improvement. Red light‐green light: from kids' game to discharge tool. Healthc Q. 2011;14:77–81.
- , , , . Myths of ideal hospital occupancy. Med J Aust. 2010;192(1):42–43.
- , . Emergency department overcrowding in the United States: an emerging threat to patient safety and public health. Emerg Med J. 2003;20(5):402–405.
- , , , . Emergency department crowding: consensus development of potential measures. Ann Emerg Med. 2003;42(6):824–834.
Bed Utilization in the PICU
Patient flow refers to the management and movement of patients in health care settings and is linked to quality, safety, and cost.16 The intensive care unit (ICU) is crucial in patient flow.7, 8 The limited number of beds and the resource‐intensive services and staffing associated with them require that hospitals optimize their utilization, as is increasingly true of all hospital resources. To maximize delivery of services to patients who need them and minimize real and opportunity losses (eg, postponed surgery, diverted transports, or inability to accept patients), patients in ICU beds should receive critical care medicine/nursing services while there and be transferred or discharged when appropriate.
The time between arrival and departure from any area of the hospital, including the ICU, is considered the time when a patient is receiving needed clinical carethe value‐added portion of health care operationsand time waiting to move on to the next step.911 This period includes both necessary logistics (eg, signing out a patient or waiting a reasonable amount of time for room cleaning) and nonvalue‐added time (eg, an excessively long amount of time for room cleaning). Operations management labels nonvalue‐added time as waste, and its reduction is vital for high‐quality health care.9, 12, 13 As in other industries, one important way to understand value versus waste is through direct observation.11, 14 Although operating rooms have been the subject of several published process improvement projects to improve efficiency,1518 inpatient beds have not been the subject of such scrutiny. The objectives of this study were to generate a direct observation method and use it to describe pediatric ICU (PICU) bed utilization from a value‐added perspective.
METHODS
An interdisciplinary work group of physicians, nurses, quality improvement specialists, and 1 operations management expert developed an Excel spreadsheet to categorize hour‐by‐hour status of PICU beds. The clinicians generated a list of 27 activities. A critical care nurse trained in quality improvement piloted the list for 3 separate 4‐hour blocks over 2 weeks adding 18 activities; 2 additional activities were added during the 5 weeks of observation (Table 1). (The recording tool is provided in the Supporting Information Appendix.) Three observers with knowledge of medical terminology (2 third‐year medical students and 1 premedical student with years of experience as an emergency medical technician) were trained over 12 hours to conduct the observations. Prior to the observations, the 3 observers also spent time in the PICU, and terminology used for recordings was reviewed. Interobserver reliability was checked during 3 sets of observation circuits by all 3 observers and the principal investigator, as well as by spot checks during the study.
| Activity Description | Activity Code | Total Hours Over 5 Weeks | % Total Hours Over 5 Weeks* | Mean Hours per Week* |
|---|---|---|---|---|
| ||||
| Ventilated patient | Vent | 8996 | 45 | 1799 |
| CCSs not otherwise specified | NOS | 2982 | 15 | 596 |
| Neurosurgery patient with ICU needs | NeurosurgICU | 1534 | 8 | 307 |
| Room empty and unassigned | Empty‐unassigned | 1511 | 8 | 302 |
| Patient on continuous infusion | ContinInfus | 958 | 5 | 192 |
| Awaiting floor bed assignment | Floorbedassign | 919 | 5 | 184 |
| Patient with arterial line | ArtLine | 508 | 3 | 102 |
| Patient on high‐flow nasal cannula | HFNC | 475 | 2 | 95 |
| Room cleaning | EVS | 318 | 2 | 64 |
| Patient <12 hours after extubation | PostVent | 226 | 1 | 45 |
| Patient in OR, bed being held | OR | 210 | 1 | 42 |
| Neurosurgery patient, post‐ICU needs | NeurosurgPostICU | 164 | 0.8 | 33 |
| No clear ICU need, but no other accepting floor or service | Unclear | 163 | 0.8 | 33 |
| Patient at procedure, bed being held | Proced | 133 | 0.7 | 27 |
| Patient awaiting a rehabilitation bed | Rehab | 99 | 0.5 | 20 |
| Patient with ventriculostomy | Ventriculostomy | 82 | 0.4 | 16 |
| Patient eligible to be in NICU | NICU | 76 | 0.4 | 15 |
| Patient awaiting social work, case management, prescriptions before discharge | AwaitingOtherServ | 66 | 0.3 | 13 |
| Empty bed, assigned to ED patient | Empty‐ED | 40 | 0.2 | 8 |
| Empty bed, assigned to incoming transport patient | Empty‐Transport | 37 | 0.2 | 7 |
| Patient awaiting transport to another facility | Transport | 37 | 0.2 | 7 |
| Patient awaiting consult to determine transfer | Consult | 33 | 0.2 | 7 |
| Patient awaiting physician or NP sign‐out to floor before transfer | CallMDNP | 30 | 0.2 | 6 |
| PICU room needs a bed for next patient | Bed | 26 | 0.1 | 5 |
| Patient eligible to be in CCU | CCU | 24 | 0.1 | 5 |
| Patient eligible to be in CICU | CICU | 24 | 0.1 | 5 |
| Patient awaiting laboratory result to determine transfer or discharge | LabResult | 21 | 0.1 | 4 |
| Patient awaiting a ride home | Ride | 21 | 0.1 | 4 |
| Empty bed, assigned to floor patient | Empty‐floor | 19 | 0.1 | 4 |
| Patient awaiting nursing report to floor for transfer | Callnurse | 18 | 0.1 | 4 |
| Patient eligible to be in PCU | PCU | 18 | 0.1 | 4 |
| Patient on cardiac pressor | Pressor | 16 | 0.1 | 3 |
| Patient actively coding | Code | 15 | 0.1 | 3 |
| Patient on continuous veno‐venous hemofiltration | CVVH | 15 | 0.1 | 3 |
| Nursing work needed to enable transfer out | Nursing | 11 | 0.1 | 2 |
| Patient awaiting order for transfer to floor | Order | 11 | 0.1 | 2 |
| Patient in interventional radiology, bed being held | IR | 10 | 0.1 | 2 |
| Patient deceased in PICU room | Deceased | 9 | 0.1 | 2 |
| Awaiting radiology result to clear transfer or discharge | RadResult | 9 | 0.1 | 2 |
| Patient awaiting a floor bed to be cleaned for transfer out | Floorbedclean | 7 | <0.1 | 1 |
| Other logistical need for an empty room | Logistics | 7 | <0.1 | 1 |
| Disagreement among services for disposition | Disagreement | 4 | <0.1 | 1 |
| Family request to stay in PICU | Family | 3 | <0.1 | 1 |
| Awaiting accepting attending/fellow for transfer out | Accept | 1 | <0.1 | <1 |
| PICU room needs a crib for next patient | Crib | 1 | <0.1 | <1 |
| Patient with preventable reason for being in PICU | Prev | 0 | 0 | 0 |
| PICU room needs specialty bed for next patient | SpecialBed | 0 | 0 | 0 |
| Total | 19,887 | 100 | ||
The targeted area included 24 single‐patient rooms. The activity of each bed was recorded hourly. Real‐time recording in to the Excel spreadsheet on a dedicated laptop occurred from 8:00 AM until 11:00 PM. The most visible or critical event was recorded. Although some activities were not mutually exclusive (eg, a patient could be ventilated and on a continuous infusion simultaneously), the objective was to identify when a room was being used for any critical care service, not enumerate all of them. The observers noted overnight events that occurred from 11:00 PM to 8:00 AM in the morning by reviewing the bedside record and talking to the staff to complete each day's 24‐hour recording. The observers also recorded the hospital‐wide census and the census for the other half of the PICU every 4 hours. The observations occurred over 5 noncontiguous weeks between January 2009 and April 2009.
After all observations were complete, activities were classified as critical care services (CCS) or noncritical‐care services (NCCS). NCCSs were further divided into necessary logistics (defined for analysis purposes as the first hour of any NCCS activity) or nonvalue‐added (the second or greater hour of NCCS). A time limit of 1 hour was chosen to define necessary logistics based on a consensus that nonclinical activities optimally would not take more than 1 hour each. We also analyzed results with 2 hours as the cutoff for necessary logistics. Admission, discharge, and transfer records were reviewed to check for returns to the PICU or hospital within 48 hours of transfer or discharge from the PICU.
Analyses were conducted using Microsoft Excel (Microsoft, Redmond, WA) and Stata 10.0 (StataCorp, College Station, TX). The study was approved by the Children's Hospital of Philadelphia Institutional Review Board with waiver of consent.
RESULTS
A total of 824 hours of recordings included 19,887 bed‐hours with 219 unique patients; among them, 2 remained from the first day of recording in January to the last day in April (sample recording in Figure 1). A total of 50 patients (range, 812 per week) stayed for the entirety of each 1‐week observation period. Of the 47 possible activities, 45 of them were recorded for at least 1 hour in the 5 weeks. Overall, 14 activities accounted for 95% of the observed bed‐hours and 31 activities accounted for the remaining 5%. CCS accounted for 82% of observed bed‐hours, NCCS accounted for 10.4%, and empty unassigned accounted for 8% (Figure 2). Using the 1‐hour cutoff for necessary services, 77% of NCCS time was nonvalue‐added, whereas 23% of it was necessary logistics; using the 2‐hour cutoff, 54% was nonvalue‐added, and 46% was necessary logistics.
During the observation period, <1% of bed‐hours were used for CCS for overflow patients from the neonatal ICU (NICU), cardiac care unit (CCU), cardiac ICU (CICU), or progressive care unit (PCU; tracheostomy/ventilator unit). Although only 4 patients required transport to a rehabilitation facility, their wait time comprised 99 hours (<1%) of total recordings. Eight patients waited a mean of 2.6 hours for transportation home (maximum, 10 hours).
To demonstrate the cycle of room use, activities were divided into 4 categories: room preparation, critical care services, disposition pending, and postcritical care services (Figure 3). As an example of detailed data revealed by direct observation, we identified 102 instances totaling 919 hours when a patient was waiting for a bed assignment on another floor (5% of all bed‐hours). The mean wait time was 9 hours (range, 188 hours) and the median time was 5.5 hours. There were only 15 instances when floor bed assignment took 1 hour or less, and only 9 instances when it took 12 hours. Similarly, considerable time was spent on cleaning rooms between patients: only 66 of 146 instances of cleaning took 1 hour or less. The mean time for cleaning was 2.2 hours (range, 115), and the median was 2 hours. (There were 136 recorded instances of room cleaning and 10 additional episodes that were not recorded but had to be completed for the room to turnover from one patient to the next, yielding a total of 146 instances of cleaning.)
From the 824 hours of recording, we identified 200 hours (25% of time) when there were zero empty unassigned beds available in the section of the PICU being observed. Episodes of full occupancy occurred mostly on weekdays, with 23% of hours of full capacity on Thursdays, 21% on Mondays, and 21% on Wednesdays; only 8% were on Saturdays and <1% on Sundays. These 200 hours fell into 36 separate episodes of complete occupancy, each lasting 122 hours. Each patient, on average, received 3.1 hours of NCCS during each episode of full occupancy (range, 111 hours). Within these 200 hours at capacity, we identified only 15 hours (8%) when all 24 beds were used for CCS. For 72% of the time, there was at least 1 bed with NCCS, and for 37% at least 2 beds. A small portion of the time (7%), the lack of beds was affected by occupancy by patients who should have been in the NICU, CICU, CCU, or PCU.
Data collected through direct observation can be used to understand aggregated and averaged experiences, but also more specific time periods. For example, we identified 1 week with the highest consistent level of occupancy and turnover: March 915 had empty unassigned beds for only 4% of the week. Of the 168 hours in the week, 68 (40%) had full capacity. However, for 90% of the time, at least 1 bed was used for a NCCS. Other analytic options included varying the assumptions around time needed for logistics. Overall, NCCS time on necessary logistics changes from 23% to 46% using 1 hour versus 2 hours as the cutoff. For floor bed assignments, assuming that the first hour of this activity is necessary logistics and any hour thereafter is not, 817 hours were wasted. Even after assuming 2 hours of necessary logistical time (which may also include steps such as nursing and physician sign‐out to the receiving team, often not recorded in the observations), this left 715 hours of NCCS time in which patients waited to be placed elsewhere in the hospital. For room cleaning, because recordings were hourly, but room cleaning could take less time, we performed a sensitivity analysis, converting all 1‐hour recordings to half‐hour recordings to half‐hour recordings (an exaggerated shortening since industry‐standard cleaning may take longer).
Of the 219 patients directly observed, 15 were noted to be waiting for a transfer out of the PICU but experienced a change in disposition before the transfer. On average, these patients waited 8 hours for a floor bed assignment (range, 221) before reverting to a CCS, which then lasted an average of 16.5 hours (range, 149). (Included in this group are 2 patients who experienced this change in disposition twice.) In post hoc review across the 5 weeks, no patients were transferred back to the PICU within 48 hours after being transferred out. During the study period, 19 patients were discharged directly from the PICU (8 to home, 7 by transport to another facility, and 4 to rehabilitation). One patient returned to the hospital (but not the PICU) within 48 hours of being discharged home from the PICU.
During the study period, using the highest census value for recorded for each 24‐hour period and the number of beds available that day, median hospital‐wide occupancy was 93% (interquartile range, 90%96%). During the 35 days of observation, 71% of the days had occupancy >90%, 29% of days had occupancy >95%, and 3% of days had occupancy >100%.
DISCUSSION
In this direct observation of a PICU, we found high usage of beds for delivery of CCS. We identified many episodes in which the half of the PICU we observed was fully occupied (200 of 824 hours), but not necessarily delivering PICU‐level care to all patients. In fact, 75% of the full‐capacity hours had at least 1 patient receiving NCCS and 37% had at least 2. Patients waiting for a floor bed assignment represented nearly 5% of bed‐hours observed (mean 9 hours per patient). That full occupancy was not random, but rather clustered on weekdays, is consistent with other work showing that hospitals are at greater risk for midweek crowding due to the way in which scheduled admissions enter and leave.1925
Our methods provide the basis for operational analysis and improvement to patient flow, such as value stream mapping.9, 26 Process improvement work could be directed to areas of delay uncovered through this analysis and inform clinical and nonclinical management. For example, one of the key problems faced by the PICU was finding floor bed assignments for patients leaving the unit. Simply building more beds in the PICU will not solve this problemand at an estimated cost of $2 million to add a bed, it is likely not an efficient means of responding to poor flow. In these cases, the problem seems to lie downstream, and could suggest shortage of regular floor beds or inefficient bed assignment procedures within the hospital. The output also suggests that variation in nonclinical processes should be a target for improvement, such as time to clean rooms, because variation is known to be a source of nonvalue‐added time in many operations.9, 26 High occupancy on weekdays but low occupancy on weekends also emphasizes the potential for smoothing occupancy to reduce the risk of midweek crowding and to better manage bed utilization and staffing.24, 25
When seeking to reduce nonvalue‐added time, one must weigh the risks of increased efficiency against clinical outcomes. For example, if patients could be transferred out of the PICU faster, would the risk of returns to the PICU be higher? In this study, 15 patients (7%) had a change in disposition from awaiting transfer back to a CCS. The fact that transfers did not happen instantaneously may serve as a safety check to reduce rapid returns, but it is not possible for us to evaluate the reasons why patients did not actually complete the pending transfers. Specifically, we cannot determine whether the patient's clinical status objectively deteriorated, the ICU team made a judgment call to hold the patient, or the floor team refused to accept the transfer. Given this fact, although it appears in this study (and in the health care system more broadly) that there are opportunities to increase efficiency and reduce nonvalue‐added time, it is not realistic (nor advisable) that such time be reduced to zero. Along this line, one must consider separately purely nonclinical functions such as room cleaning and those that include some clinical element, such as time waiting for a patient to be transferred.
Beyond the direct findings of this study, the method should be replicable in other settings and can reveal important information about health care efficiency, capacity, and flexibility. The bottlenecks identified would have been difficult to identify through administrative record review. The exact amount of time to spend on observation may vary from place to place and would depend on the expected variation over time and the level of detail sought. In general, the more common the event and the less variation, the less time needed to observe it.
This study has several limitations that should be considered in terms of interpreting the results and in seeking to reproduce the approach. First, hourly recordings may not be discrete enough for events that took less than 1 hour. To assess the degree to which this would affect our results, we reanalyzed all NCCS by subtracting 30 minutes (0.5 hour) from all recordings, which increased total CCS from 82% to 87% and decreased NCCS by the same 5 percentage points. In a related fashion, our recordings were truncated at the start and end of each 1‐week period, so we could only observe a maximum of 168 hours for any given activity and did not record how long an activity was happening before or after the recordings started or stopped, respectively. Second, each recording could only be for 1 activity per hour. Separate from the level of granularity already noted, this also limits interpretation of critical care activities that may have been simultaneous. However, because the goal of the study was not to describe the provision of critical care services, but rather the times when they were not being delivered, this does not influence our conclusions. For movement of patients, however, we missed instances of physician and nursing calling sign‐out on patients to receiving units, as these events last less than 1 hour (and in the case of surgical patients, generally do not occur as the team provides continuous coverage). The time for such events is then included in other activities. To the extent that this may influence the results, it would increase the perceived time for nonvalue‐added services, but to a limited degree, and never by more than 59 minutes. Third, the overnight hours (11:00 PM to 8:00 AM) were not directly observed, but retrospectively recorded each morning by reviewing the records and discussing the overnight events with the clinical staff. For example, if a patient was intubated at 11:00 PM and at 8:00 AM, the observer would confirm this and record that status for the intervening hours. This is unlikely to result in a substantial impact on the findings, because the overnight hours have a relative degree of stability even for unstable patients in terms of their status of needing or not needing a CCS. Fourth, we did not evaluate the appropriateness of CCS delivered (eg, how long a patient was ventilated). Our definitions for CCS and NCSS were based on Children's Hospital of Philadelphia practices, which may not be the same as those of other facilities. The categorization of CCS was objective for activities such as ventilation or continuous infusion, but was less clear for the not otherwise specified recordings, which represented patients with a complex illness or projected organ, respiratory, cardiac, or neurological failure. These patients were not receiving a specific critical care intervention, but were deemed to need to be in the PICU as opposed to a regular floor (eg, for frequent monitoring of potential respiratory failure). It would also include patients receiving combinations of therapies more efficiently delivered in the PICU. For that, the observers relied on the judgment of clinicians (primarily nurses) to determine whether the patient needed to be in the PICU or not; if no specific reason could be provided, not otherwise specified was applied. These 192 instances accounted for 2982 aggregate bed‐hours (15% of total). It is difficult to judge the direction of bias, because overestimation of need to be in the PICU may be as likely to occur as underestimation. Fifth, the very presence of the observers may have changed behavior. Knowing that they were being observed staff may have acted with greater efficiency than otherwise. We expect that such a finding would lead to less time appearing as necessary logistics or NCCS. Finally, results may not be generalizable to other hospitals or hospital settings. There are clearly important contextual factors, not only for the location but also for the duration. For example, staffing was never an issue during the 5 weeks of observation, but there are locations where an empty bed is not necessarily usable due to lack of staffing. Nonetheless, we believe the results provide a generalizable approach and methodology for other settings (and staffing could be a reason for an empty bed).
In terms of the setting, as noted, we observed one discrete 24‐bed unit, which comprises half of the total PICU. Thus, statements that the PICU was at full capacity must be interpreted in the context that additional rooms may have been available on the other side. Patients are generally admitted alternately to each unit, so the occupancies should parallel each other. We recorded the census every 4 hours for both sides from the electronic system (Sunrise Clinical Manager [SCM]). However, this only accounts for patients physically in beds, not beds held for patients in other locations. Thus, we would expect a discrepancy between direct observation and the SCM value. Through analysis of the entire pediatric intensive care unit,* that part which observed directly, and that which we did not observe directly using census data, we think it reasonable to assert that both units of the total PICU had constrained capacity during the times we directly observed and recorded such constraint on one side.
This study demonstrates the use of direct observation for inpatient settings to learn about resource utilization and identification of value‐added services. PubMed searches for the terms efficiency, flow, process redesign, and time management bring up many more references for operating rooms than for ICUs or inpatient beds. Some examples of ICU‐directed work include videography of an ICU in Australia27 and human factor analysis in ICU nursing.5 Time‐motion studies have also been conducted on clinical staff, such as physicians.28, 29
In conclusion, we found that direct observation provided important insights into the utilization of patient rooms in an important inpatient setting. Data such as these are valuable for clinical and process improvement work, as well as understanding how best to match capacity to patient need. Finally, the methodology is reproducible for other settings and would be an additional tool to measuring and improving the efficiency and value of the health system. When appropriate, this approach can also evaluate the effectiveness of process improvement, help identify and reduce waste,13 and contribute to the growing field that merges operations management with hospital administration and clinical care: in other words, evidence‐based management.30
Acknowledgements
The authors thank Paula Agosto, Patricia Hubbs, Heidi Martin, and Annette Bollig for contributions to the study design.
In comparing direct observation to the SCM count, we found perfect concordance for 110 hours (55%) during which 0 beds were available. For the other 90 hours, SCM reported 1 bed being available in 46 hours (23%), 2 beds being available in 24 hours (12%), 3 beds being available in 17 hours (9%), and 4 beds being available in 3 hours (2%)all while we directly observed 0 beds being available. Thus, cumulatively, 90% of the hours observed with no beds had an SCM report availability of 02 beds; 99% of the time that was 03 beds. Applying this rate of mismatch to the unit that we did not observe directly, SCM reported 0 beds for 46 (23%) of the 200 hours the observation unit was full; SCM reported 1 bed available in 70 hours (35%), 2 beds open in 42 hours (21%), 3 beds open in 26 hours (13%), and 4 beds open in 16 hours (8%). Cumulatively, that is 79% of the time with 02 beds and 92% at 03 beds. From this, we conclude that the combined PICU for both sides was likely functionally full at least 158 of the 200 hours that the side we observed was full (79% 200 hours) and likely had very constrained capacity during the other 42 hours.
- ,,,,.The effect of hospital occupancy on emergency department length of stay and patient disposition.Acad Emerg Med.2003;10:127–133.
- ,,,.The effect of hospital bed occupancy on throughput in the pediatric emergency department.Ann Emerg Med.2009;53:767–776.
- ,,,.A Comparison of in‐hospital mortality risk conferred by high hospital occupancy, differences in nurse staffing levels, weekend admission, and seasonal influenza.Medical Care.2010;48:224–232.
- .Intensive care unit occupancy: making room for more patients.Crit Care Med.2009;37:1794–1795.
- ,.A human factors engineering conceptual framework of nursing workload and patient safety in intensive care units.Intensive Crit Care Nurs.2005;21:284–301.
- ,,, et al.High level of burnout in intensivists: prevalence and associated factors.Am J Respir Crit Care Med.2007;175:686–692.
- ,,.Length of stay and efficiency in pediatric intensive care units.J Pediatr.1998;133:79–85.
- ,.Variability in duration of stay in pediatric intensive care units: a multiinstitutional study.J Pediatr.1996;128:35–44.
- ,.Matching Supply with Demand: An Introduction to Operations Management.New York, NY:McGraw‐Hill;2006.
- ,.Impact of workload on service time and patient safety: an econometric analysis of hospital operations.Management Science.2009;55:1486–1498.
- .OPIM 631: Operations Management.Philadelphia, PA:Wharton School, University of Pennsylvania;2008.
- ,,.From waste to value in health care.JAMA.2008;299:568–571.
- .Eliminating “waste” in health care.JAMA.2009;302:2481–2482.
- .Toyota Production System: Beyond Large‐scale Production.London, UK:Productivity Press;1995.
- ,.Interdisciplinary work flow assessment and redesign decreases operating room turnover time and allows for additional caseload.Arch Surg.2006;141:65–69.
- ,,,.Improving operating room efficiency through process redesign.Surgery.2006;140:509–514.
- ,,,.Successful strategies for improving operating room efficiency at academic institutions.Anesth Analg.1998year="1998"1998;86:896–906.
- ,,.Efficiency of the operating room suite.Am J Surg.2003;185:244–250.
- ,,, et al.Children's hospitals do not acutely respond to high occupancy.Pediatrics.2010;125:974–981.
- Institute for Healthcare Improvement. Smoothing elective surgical admissions. Available at: http://www.ihi.org/IHI/Topics/Flow/Patient Flow/EmergingContent/SmoothingElectiveSurgicalAdmissions.htm. Accessed October 24,2008.
- Boston Hospital Sees Big Impact from Smoothing Elective Schedule.OR Manager. Volume 20, no. 12,2004.
- ,.Rethinking rapid response teams.JAMA.2010;304:1375–1376.
- Litvak E, ed.Managing Patient Flow in Hospitals: Strategies and Solutions.2nd ed.Oak Brook, IL:Joint Commission Resources;2009.
- ,,,,,.Scheduled admissions and high occupancy at a children's hospital.J Hosp Med.2011;6:81–87.
- ,,, et al.Addressing inpatient crowding by smoothing occupancy at children's hospitals.J Hosp Med.2011;6:466–473.
- ,.Learning to See: Value Stream Mapping to Add Value and Eliminate MUDA.Cambridge, MA:Lean Enterprise Institute;1999.
- ,,.Reshaping ICU ward round practices using video‐reflexive ethnography.Qual Health Res.2008;18:380–390.
- ,,.How hospitalists spend their time: insights on efficiency and safety.J Hosp Med.2006;1:88–93.
- ,,, et al.Where did the day go? A time‐motion study of hospitalists.J Hosp Med2010;5:323–238.
- ,,.Improving patient care by linking evidence‐based medicine and evidence‐based management.JAMA.2007;298:673–676.
Patient flow refers to the management and movement of patients in health care settings and is linked to quality, safety, and cost.16 The intensive care unit (ICU) is crucial in patient flow.7, 8 The limited number of beds and the resource‐intensive services and staffing associated with them require that hospitals optimize their utilization, as is increasingly true of all hospital resources. To maximize delivery of services to patients who need them and minimize real and opportunity losses (eg, postponed surgery, diverted transports, or inability to accept patients), patients in ICU beds should receive critical care medicine/nursing services while there and be transferred or discharged when appropriate.
The time between arrival and departure from any area of the hospital, including the ICU, is considered the time when a patient is receiving needed clinical carethe value‐added portion of health care operationsand time waiting to move on to the next step.911 This period includes both necessary logistics (eg, signing out a patient or waiting a reasonable amount of time for room cleaning) and nonvalue‐added time (eg, an excessively long amount of time for room cleaning). Operations management labels nonvalue‐added time as waste, and its reduction is vital for high‐quality health care.9, 12, 13 As in other industries, one important way to understand value versus waste is through direct observation.11, 14 Although operating rooms have been the subject of several published process improvement projects to improve efficiency,1518 inpatient beds have not been the subject of such scrutiny. The objectives of this study were to generate a direct observation method and use it to describe pediatric ICU (PICU) bed utilization from a value‐added perspective.
METHODS
An interdisciplinary work group of physicians, nurses, quality improvement specialists, and 1 operations management expert developed an Excel spreadsheet to categorize hour‐by‐hour status of PICU beds. The clinicians generated a list of 27 activities. A critical care nurse trained in quality improvement piloted the list for 3 separate 4‐hour blocks over 2 weeks adding 18 activities; 2 additional activities were added during the 5 weeks of observation (Table 1). (The recording tool is provided in the Supporting Information Appendix.) Three observers with knowledge of medical terminology (2 third‐year medical students and 1 premedical student with years of experience as an emergency medical technician) were trained over 12 hours to conduct the observations. Prior to the observations, the 3 observers also spent time in the PICU, and terminology used for recordings was reviewed. Interobserver reliability was checked during 3 sets of observation circuits by all 3 observers and the principal investigator, as well as by spot checks during the study.
| Activity Description | Activity Code | Total Hours Over 5 Weeks | % Total Hours Over 5 Weeks* | Mean Hours per Week* |
|---|---|---|---|---|
| ||||
| Ventilated patient | Vent | 8996 | 45 | 1799 |
| CCSs not otherwise specified | NOS | 2982 | 15 | 596 |
| Neurosurgery patient with ICU needs | NeurosurgICU | 1534 | 8 | 307 |
| Room empty and unassigned | Empty‐unassigned | 1511 | 8 | 302 |
| Patient on continuous infusion | ContinInfus | 958 | 5 | 192 |
| Awaiting floor bed assignment | Floorbedassign | 919 | 5 | 184 |
| Patient with arterial line | ArtLine | 508 | 3 | 102 |
| Patient on high‐flow nasal cannula | HFNC | 475 | 2 | 95 |
| Room cleaning | EVS | 318 | 2 | 64 |
| Patient <12 hours after extubation | PostVent | 226 | 1 | 45 |
| Patient in OR, bed being held | OR | 210 | 1 | 42 |
| Neurosurgery patient, post‐ICU needs | NeurosurgPostICU | 164 | 0.8 | 33 |
| No clear ICU need, but no other accepting floor or service | Unclear | 163 | 0.8 | 33 |
| Patient at procedure, bed being held | Proced | 133 | 0.7 | 27 |
| Patient awaiting a rehabilitation bed | Rehab | 99 | 0.5 | 20 |
| Patient with ventriculostomy | Ventriculostomy | 82 | 0.4 | 16 |
| Patient eligible to be in NICU | NICU | 76 | 0.4 | 15 |
| Patient awaiting social work, case management, prescriptions before discharge | AwaitingOtherServ | 66 | 0.3 | 13 |
| Empty bed, assigned to ED patient | Empty‐ED | 40 | 0.2 | 8 |
| Empty bed, assigned to incoming transport patient | Empty‐Transport | 37 | 0.2 | 7 |
| Patient awaiting transport to another facility | Transport | 37 | 0.2 | 7 |
| Patient awaiting consult to determine transfer | Consult | 33 | 0.2 | 7 |
| Patient awaiting physician or NP sign‐out to floor before transfer | CallMDNP | 30 | 0.2 | 6 |
| PICU room needs a bed for next patient | Bed | 26 | 0.1 | 5 |
| Patient eligible to be in CCU | CCU | 24 | 0.1 | 5 |
| Patient eligible to be in CICU | CICU | 24 | 0.1 | 5 |
| Patient awaiting laboratory result to determine transfer or discharge | LabResult | 21 | 0.1 | 4 |
| Patient awaiting a ride home | Ride | 21 | 0.1 | 4 |
| Empty bed, assigned to floor patient | Empty‐floor | 19 | 0.1 | 4 |
| Patient awaiting nursing report to floor for transfer | Callnurse | 18 | 0.1 | 4 |
| Patient eligible to be in PCU | PCU | 18 | 0.1 | 4 |
| Patient on cardiac pressor | Pressor | 16 | 0.1 | 3 |
| Patient actively coding | Code | 15 | 0.1 | 3 |
| Patient on continuous veno‐venous hemofiltration | CVVH | 15 | 0.1 | 3 |
| Nursing work needed to enable transfer out | Nursing | 11 | 0.1 | 2 |
| Patient awaiting order for transfer to floor | Order | 11 | 0.1 | 2 |
| Patient in interventional radiology, bed being held | IR | 10 | 0.1 | 2 |
| Patient deceased in PICU room | Deceased | 9 | 0.1 | 2 |
| Awaiting radiology result to clear transfer or discharge | RadResult | 9 | 0.1 | 2 |
| Patient awaiting a floor bed to be cleaned for transfer out | Floorbedclean | 7 | <0.1 | 1 |
| Other logistical need for an empty room | Logistics | 7 | <0.1 | 1 |
| Disagreement among services for disposition | Disagreement | 4 | <0.1 | 1 |
| Family request to stay in PICU | Family | 3 | <0.1 | 1 |
| Awaiting accepting attending/fellow for transfer out | Accept | 1 | <0.1 | <1 |
| PICU room needs a crib for next patient | Crib | 1 | <0.1 | <1 |
| Patient with preventable reason for being in PICU | Prev | 0 | 0 | 0 |
| PICU room needs specialty bed for next patient | SpecialBed | 0 | 0 | 0 |
| Total | 19,887 | 100 | ||
The targeted area included 24 single‐patient rooms. The activity of each bed was recorded hourly. Real‐time recording in to the Excel spreadsheet on a dedicated laptop occurred from 8:00 AM until 11:00 PM. The most visible or critical event was recorded. Although some activities were not mutually exclusive (eg, a patient could be ventilated and on a continuous infusion simultaneously), the objective was to identify when a room was being used for any critical care service, not enumerate all of them. The observers noted overnight events that occurred from 11:00 PM to 8:00 AM in the morning by reviewing the bedside record and talking to the staff to complete each day's 24‐hour recording. The observers also recorded the hospital‐wide census and the census for the other half of the PICU every 4 hours. The observations occurred over 5 noncontiguous weeks between January 2009 and April 2009.
After all observations were complete, activities were classified as critical care services (CCS) or noncritical‐care services (NCCS). NCCSs were further divided into necessary logistics (defined for analysis purposes as the first hour of any NCCS activity) or nonvalue‐added (the second or greater hour of NCCS). A time limit of 1 hour was chosen to define necessary logistics based on a consensus that nonclinical activities optimally would not take more than 1 hour each. We also analyzed results with 2 hours as the cutoff for necessary logistics. Admission, discharge, and transfer records were reviewed to check for returns to the PICU or hospital within 48 hours of transfer or discharge from the PICU.
Analyses were conducted using Microsoft Excel (Microsoft, Redmond, WA) and Stata 10.0 (StataCorp, College Station, TX). The study was approved by the Children's Hospital of Philadelphia Institutional Review Board with waiver of consent.
RESULTS
A total of 824 hours of recordings included 19,887 bed‐hours with 219 unique patients; among them, 2 remained from the first day of recording in January to the last day in April (sample recording in Figure 1). A total of 50 patients (range, 812 per week) stayed for the entirety of each 1‐week observation period. Of the 47 possible activities, 45 of them were recorded for at least 1 hour in the 5 weeks. Overall, 14 activities accounted for 95% of the observed bed‐hours and 31 activities accounted for the remaining 5%. CCS accounted for 82% of observed bed‐hours, NCCS accounted for 10.4%, and empty unassigned accounted for 8% (Figure 2). Using the 1‐hour cutoff for necessary services, 77% of NCCS time was nonvalue‐added, whereas 23% of it was necessary logistics; using the 2‐hour cutoff, 54% was nonvalue‐added, and 46% was necessary logistics.
During the observation period, <1% of bed‐hours were used for CCS for overflow patients from the neonatal ICU (NICU), cardiac care unit (CCU), cardiac ICU (CICU), or progressive care unit (PCU; tracheostomy/ventilator unit). Although only 4 patients required transport to a rehabilitation facility, their wait time comprised 99 hours (<1%) of total recordings. Eight patients waited a mean of 2.6 hours for transportation home (maximum, 10 hours).
To demonstrate the cycle of room use, activities were divided into 4 categories: room preparation, critical care services, disposition pending, and postcritical care services (Figure 3). As an example of detailed data revealed by direct observation, we identified 102 instances totaling 919 hours when a patient was waiting for a bed assignment on another floor (5% of all bed‐hours). The mean wait time was 9 hours (range, 188 hours) and the median time was 5.5 hours. There were only 15 instances when floor bed assignment took 1 hour or less, and only 9 instances when it took 12 hours. Similarly, considerable time was spent on cleaning rooms between patients: only 66 of 146 instances of cleaning took 1 hour or less. The mean time for cleaning was 2.2 hours (range, 115), and the median was 2 hours. (There were 136 recorded instances of room cleaning and 10 additional episodes that were not recorded but had to be completed for the room to turnover from one patient to the next, yielding a total of 146 instances of cleaning.)
From the 824 hours of recording, we identified 200 hours (25% of time) when there were zero empty unassigned beds available in the section of the PICU being observed. Episodes of full occupancy occurred mostly on weekdays, with 23% of hours of full capacity on Thursdays, 21% on Mondays, and 21% on Wednesdays; only 8% were on Saturdays and <1% on Sundays. These 200 hours fell into 36 separate episodes of complete occupancy, each lasting 122 hours. Each patient, on average, received 3.1 hours of NCCS during each episode of full occupancy (range, 111 hours). Within these 200 hours at capacity, we identified only 15 hours (8%) when all 24 beds were used for CCS. For 72% of the time, there was at least 1 bed with NCCS, and for 37% at least 2 beds. A small portion of the time (7%), the lack of beds was affected by occupancy by patients who should have been in the NICU, CICU, CCU, or PCU.
Data collected through direct observation can be used to understand aggregated and averaged experiences, but also more specific time periods. For example, we identified 1 week with the highest consistent level of occupancy and turnover: March 915 had empty unassigned beds for only 4% of the week. Of the 168 hours in the week, 68 (40%) had full capacity. However, for 90% of the time, at least 1 bed was used for a NCCS. Other analytic options included varying the assumptions around time needed for logistics. Overall, NCCS time on necessary logistics changes from 23% to 46% using 1 hour versus 2 hours as the cutoff. For floor bed assignments, assuming that the first hour of this activity is necessary logistics and any hour thereafter is not, 817 hours were wasted. Even after assuming 2 hours of necessary logistical time (which may also include steps such as nursing and physician sign‐out to the receiving team, often not recorded in the observations), this left 715 hours of NCCS time in which patients waited to be placed elsewhere in the hospital. For room cleaning, because recordings were hourly, but room cleaning could take less time, we performed a sensitivity analysis, converting all 1‐hour recordings to half‐hour recordings to half‐hour recordings (an exaggerated shortening since industry‐standard cleaning may take longer).
Of the 219 patients directly observed, 15 were noted to be waiting for a transfer out of the PICU but experienced a change in disposition before the transfer. On average, these patients waited 8 hours for a floor bed assignment (range, 221) before reverting to a CCS, which then lasted an average of 16.5 hours (range, 149). (Included in this group are 2 patients who experienced this change in disposition twice.) In post hoc review across the 5 weeks, no patients were transferred back to the PICU within 48 hours after being transferred out. During the study period, 19 patients were discharged directly from the PICU (8 to home, 7 by transport to another facility, and 4 to rehabilitation). One patient returned to the hospital (but not the PICU) within 48 hours of being discharged home from the PICU.
During the study period, using the highest census value for recorded for each 24‐hour period and the number of beds available that day, median hospital‐wide occupancy was 93% (interquartile range, 90%96%). During the 35 days of observation, 71% of the days had occupancy >90%, 29% of days had occupancy >95%, and 3% of days had occupancy >100%.
DISCUSSION
In this direct observation of a PICU, we found high usage of beds for delivery of CCS. We identified many episodes in which the half of the PICU we observed was fully occupied (200 of 824 hours), but not necessarily delivering PICU‐level care to all patients. In fact, 75% of the full‐capacity hours had at least 1 patient receiving NCCS and 37% had at least 2. Patients waiting for a floor bed assignment represented nearly 5% of bed‐hours observed (mean 9 hours per patient). That full occupancy was not random, but rather clustered on weekdays, is consistent with other work showing that hospitals are at greater risk for midweek crowding due to the way in which scheduled admissions enter and leave.1925
Our methods provide the basis for operational analysis and improvement to patient flow, such as value stream mapping.9, 26 Process improvement work could be directed to areas of delay uncovered through this analysis and inform clinical and nonclinical management. For example, one of the key problems faced by the PICU was finding floor bed assignments for patients leaving the unit. Simply building more beds in the PICU will not solve this problemand at an estimated cost of $2 million to add a bed, it is likely not an efficient means of responding to poor flow. In these cases, the problem seems to lie downstream, and could suggest shortage of regular floor beds or inefficient bed assignment procedures within the hospital. The output also suggests that variation in nonclinical processes should be a target for improvement, such as time to clean rooms, because variation is known to be a source of nonvalue‐added time in many operations.9, 26 High occupancy on weekdays but low occupancy on weekends also emphasizes the potential for smoothing occupancy to reduce the risk of midweek crowding and to better manage bed utilization and staffing.24, 25
When seeking to reduce nonvalue‐added time, one must weigh the risks of increased efficiency against clinical outcomes. For example, if patients could be transferred out of the PICU faster, would the risk of returns to the PICU be higher? In this study, 15 patients (7%) had a change in disposition from awaiting transfer back to a CCS. The fact that transfers did not happen instantaneously may serve as a safety check to reduce rapid returns, but it is not possible for us to evaluate the reasons why patients did not actually complete the pending transfers. Specifically, we cannot determine whether the patient's clinical status objectively deteriorated, the ICU team made a judgment call to hold the patient, or the floor team refused to accept the transfer. Given this fact, although it appears in this study (and in the health care system more broadly) that there are opportunities to increase efficiency and reduce nonvalue‐added time, it is not realistic (nor advisable) that such time be reduced to zero. Along this line, one must consider separately purely nonclinical functions such as room cleaning and those that include some clinical element, such as time waiting for a patient to be transferred.
Beyond the direct findings of this study, the method should be replicable in other settings and can reveal important information about health care efficiency, capacity, and flexibility. The bottlenecks identified would have been difficult to identify through administrative record review. The exact amount of time to spend on observation may vary from place to place and would depend on the expected variation over time and the level of detail sought. In general, the more common the event and the less variation, the less time needed to observe it.
This study has several limitations that should be considered in terms of interpreting the results and in seeking to reproduce the approach. First, hourly recordings may not be discrete enough for events that took less than 1 hour. To assess the degree to which this would affect our results, we reanalyzed all NCCS by subtracting 30 minutes (0.5 hour) from all recordings, which increased total CCS from 82% to 87% and decreased NCCS by the same 5 percentage points. In a related fashion, our recordings were truncated at the start and end of each 1‐week period, so we could only observe a maximum of 168 hours for any given activity and did not record how long an activity was happening before or after the recordings started or stopped, respectively. Second, each recording could only be for 1 activity per hour. Separate from the level of granularity already noted, this also limits interpretation of critical care activities that may have been simultaneous. However, because the goal of the study was not to describe the provision of critical care services, but rather the times when they were not being delivered, this does not influence our conclusions. For movement of patients, however, we missed instances of physician and nursing calling sign‐out on patients to receiving units, as these events last less than 1 hour (and in the case of surgical patients, generally do not occur as the team provides continuous coverage). The time for such events is then included in other activities. To the extent that this may influence the results, it would increase the perceived time for nonvalue‐added services, but to a limited degree, and never by more than 59 minutes. Third, the overnight hours (11:00 PM to 8:00 AM) were not directly observed, but retrospectively recorded each morning by reviewing the records and discussing the overnight events with the clinical staff. For example, if a patient was intubated at 11:00 PM and at 8:00 AM, the observer would confirm this and record that status for the intervening hours. This is unlikely to result in a substantial impact on the findings, because the overnight hours have a relative degree of stability even for unstable patients in terms of their status of needing or not needing a CCS. Fourth, we did not evaluate the appropriateness of CCS delivered (eg, how long a patient was ventilated). Our definitions for CCS and NCSS were based on Children's Hospital of Philadelphia practices, which may not be the same as those of other facilities. The categorization of CCS was objective for activities such as ventilation or continuous infusion, but was less clear for the not otherwise specified recordings, which represented patients with a complex illness or projected organ, respiratory, cardiac, or neurological failure. These patients were not receiving a specific critical care intervention, but were deemed to need to be in the PICU as opposed to a regular floor (eg, for frequent monitoring of potential respiratory failure). It would also include patients receiving combinations of therapies more efficiently delivered in the PICU. For that, the observers relied on the judgment of clinicians (primarily nurses) to determine whether the patient needed to be in the PICU or not; if no specific reason could be provided, not otherwise specified was applied. These 192 instances accounted for 2982 aggregate bed‐hours (15% of total). It is difficult to judge the direction of bias, because overestimation of need to be in the PICU may be as likely to occur as underestimation. Fifth, the very presence of the observers may have changed behavior. Knowing that they were being observed staff may have acted with greater efficiency than otherwise. We expect that such a finding would lead to less time appearing as necessary logistics or NCCS. Finally, results may not be generalizable to other hospitals or hospital settings. There are clearly important contextual factors, not only for the location but also for the duration. For example, staffing was never an issue during the 5 weeks of observation, but there are locations where an empty bed is not necessarily usable due to lack of staffing. Nonetheless, we believe the results provide a generalizable approach and methodology for other settings (and staffing could be a reason for an empty bed).
In terms of the setting, as noted, we observed one discrete 24‐bed unit, which comprises half of the total PICU. Thus, statements that the PICU was at full capacity must be interpreted in the context that additional rooms may have been available on the other side. Patients are generally admitted alternately to each unit, so the occupancies should parallel each other. We recorded the census every 4 hours for both sides from the electronic system (Sunrise Clinical Manager [SCM]). However, this only accounts for patients physically in beds, not beds held for patients in other locations. Thus, we would expect a discrepancy between direct observation and the SCM value. Through analysis of the entire pediatric intensive care unit,* that part which observed directly, and that which we did not observe directly using census data, we think it reasonable to assert that both units of the total PICU had constrained capacity during the times we directly observed and recorded such constraint on one side.
This study demonstrates the use of direct observation for inpatient settings to learn about resource utilization and identification of value‐added services. PubMed searches for the terms efficiency, flow, process redesign, and time management bring up many more references for operating rooms than for ICUs or inpatient beds. Some examples of ICU‐directed work include videography of an ICU in Australia27 and human factor analysis in ICU nursing.5 Time‐motion studies have also been conducted on clinical staff, such as physicians.28, 29
In conclusion, we found that direct observation provided important insights into the utilization of patient rooms in an important inpatient setting. Data such as these are valuable for clinical and process improvement work, as well as understanding how best to match capacity to patient need. Finally, the methodology is reproducible for other settings and would be an additional tool to measuring and improving the efficiency and value of the health system. When appropriate, this approach can also evaluate the effectiveness of process improvement, help identify and reduce waste,13 and contribute to the growing field that merges operations management with hospital administration and clinical care: in other words, evidence‐based management.30
Acknowledgements
The authors thank Paula Agosto, Patricia Hubbs, Heidi Martin, and Annette Bollig for contributions to the study design.
In comparing direct observation to the SCM count, we found perfect concordance for 110 hours (55%) during which 0 beds were available. For the other 90 hours, SCM reported 1 bed being available in 46 hours (23%), 2 beds being available in 24 hours (12%), 3 beds being available in 17 hours (9%), and 4 beds being available in 3 hours (2%)all while we directly observed 0 beds being available. Thus, cumulatively, 90% of the hours observed with no beds had an SCM report availability of 02 beds; 99% of the time that was 03 beds. Applying this rate of mismatch to the unit that we did not observe directly, SCM reported 0 beds for 46 (23%) of the 200 hours the observation unit was full; SCM reported 1 bed available in 70 hours (35%), 2 beds open in 42 hours (21%), 3 beds open in 26 hours (13%), and 4 beds open in 16 hours (8%). Cumulatively, that is 79% of the time with 02 beds and 92% at 03 beds. From this, we conclude that the combined PICU for both sides was likely functionally full at least 158 of the 200 hours that the side we observed was full (79% 200 hours) and likely had very constrained capacity during the other 42 hours.
Patient flow refers to the management and movement of patients in health care settings and is linked to quality, safety, and cost.16 The intensive care unit (ICU) is crucial in patient flow.7, 8 The limited number of beds and the resource‐intensive services and staffing associated with them require that hospitals optimize their utilization, as is increasingly true of all hospital resources. To maximize delivery of services to patients who need them and minimize real and opportunity losses (eg, postponed surgery, diverted transports, or inability to accept patients), patients in ICU beds should receive critical care medicine/nursing services while there and be transferred or discharged when appropriate.
The time between arrival and departure from any area of the hospital, including the ICU, is considered the time when a patient is receiving needed clinical carethe value‐added portion of health care operationsand time waiting to move on to the next step.911 This period includes both necessary logistics (eg, signing out a patient or waiting a reasonable amount of time for room cleaning) and nonvalue‐added time (eg, an excessively long amount of time for room cleaning). Operations management labels nonvalue‐added time as waste, and its reduction is vital for high‐quality health care.9, 12, 13 As in other industries, one important way to understand value versus waste is through direct observation.11, 14 Although operating rooms have been the subject of several published process improvement projects to improve efficiency,1518 inpatient beds have not been the subject of such scrutiny. The objectives of this study were to generate a direct observation method and use it to describe pediatric ICU (PICU) bed utilization from a value‐added perspective.
METHODS
An interdisciplinary work group of physicians, nurses, quality improvement specialists, and 1 operations management expert developed an Excel spreadsheet to categorize hour‐by‐hour status of PICU beds. The clinicians generated a list of 27 activities. A critical care nurse trained in quality improvement piloted the list for 3 separate 4‐hour blocks over 2 weeks adding 18 activities; 2 additional activities were added during the 5 weeks of observation (Table 1). (The recording tool is provided in the Supporting Information Appendix.) Three observers with knowledge of medical terminology (2 third‐year medical students and 1 premedical student with years of experience as an emergency medical technician) were trained over 12 hours to conduct the observations. Prior to the observations, the 3 observers also spent time in the PICU, and terminology used for recordings was reviewed. Interobserver reliability was checked during 3 sets of observation circuits by all 3 observers and the principal investigator, as well as by spot checks during the study.
| Activity Description | Activity Code | Total Hours Over 5 Weeks | % Total Hours Over 5 Weeks* | Mean Hours per Week* |
|---|---|---|---|---|
| ||||
| Ventilated patient | Vent | 8996 | 45 | 1799 |
| CCSs not otherwise specified | NOS | 2982 | 15 | 596 |
| Neurosurgery patient with ICU needs | NeurosurgICU | 1534 | 8 | 307 |
| Room empty and unassigned | Empty‐unassigned | 1511 | 8 | 302 |
| Patient on continuous infusion | ContinInfus | 958 | 5 | 192 |
| Awaiting floor bed assignment | Floorbedassign | 919 | 5 | 184 |
| Patient with arterial line | ArtLine | 508 | 3 | 102 |
| Patient on high‐flow nasal cannula | HFNC | 475 | 2 | 95 |
| Room cleaning | EVS | 318 | 2 | 64 |
| Patient <12 hours after extubation | PostVent | 226 | 1 | 45 |
| Patient in OR, bed being held | OR | 210 | 1 | 42 |
| Neurosurgery patient, post‐ICU needs | NeurosurgPostICU | 164 | 0.8 | 33 |
| No clear ICU need, but no other accepting floor or service | Unclear | 163 | 0.8 | 33 |
| Patient at procedure, bed being held | Proced | 133 | 0.7 | 27 |
| Patient awaiting a rehabilitation bed | Rehab | 99 | 0.5 | 20 |
| Patient with ventriculostomy | Ventriculostomy | 82 | 0.4 | 16 |
| Patient eligible to be in NICU | NICU | 76 | 0.4 | 15 |
| Patient awaiting social work, case management, prescriptions before discharge | AwaitingOtherServ | 66 | 0.3 | 13 |
| Empty bed, assigned to ED patient | Empty‐ED | 40 | 0.2 | 8 |
| Empty bed, assigned to incoming transport patient | Empty‐Transport | 37 | 0.2 | 7 |
| Patient awaiting transport to another facility | Transport | 37 | 0.2 | 7 |
| Patient awaiting consult to determine transfer | Consult | 33 | 0.2 | 7 |
| Patient awaiting physician or NP sign‐out to floor before transfer | CallMDNP | 30 | 0.2 | 6 |
| PICU room needs a bed for next patient | Bed | 26 | 0.1 | 5 |
| Patient eligible to be in CCU | CCU | 24 | 0.1 | 5 |
| Patient eligible to be in CICU | CICU | 24 | 0.1 | 5 |
| Patient awaiting laboratory result to determine transfer or discharge | LabResult | 21 | 0.1 | 4 |
| Patient awaiting a ride home | Ride | 21 | 0.1 | 4 |
| Empty bed, assigned to floor patient | Empty‐floor | 19 | 0.1 | 4 |
| Patient awaiting nursing report to floor for transfer | Callnurse | 18 | 0.1 | 4 |
| Patient eligible to be in PCU | PCU | 18 | 0.1 | 4 |
| Patient on cardiac pressor | Pressor | 16 | 0.1 | 3 |
| Patient actively coding | Code | 15 | 0.1 | 3 |
| Patient on continuous veno‐venous hemofiltration | CVVH | 15 | 0.1 | 3 |
| Nursing work needed to enable transfer out | Nursing | 11 | 0.1 | 2 |
| Patient awaiting order for transfer to floor | Order | 11 | 0.1 | 2 |
| Patient in interventional radiology, bed being held | IR | 10 | 0.1 | 2 |
| Patient deceased in PICU room | Deceased | 9 | 0.1 | 2 |
| Awaiting radiology result to clear transfer or discharge | RadResult | 9 | 0.1 | 2 |
| Patient awaiting a floor bed to be cleaned for transfer out | Floorbedclean | 7 | <0.1 | 1 |
| Other logistical need for an empty room | Logistics | 7 | <0.1 | 1 |
| Disagreement among services for disposition | Disagreement | 4 | <0.1 | 1 |
| Family request to stay in PICU | Family | 3 | <0.1 | 1 |
| Awaiting accepting attending/fellow for transfer out | Accept | 1 | <0.1 | <1 |
| PICU room needs a crib for next patient | Crib | 1 | <0.1 | <1 |
| Patient with preventable reason for being in PICU | Prev | 0 | 0 | 0 |
| PICU room needs specialty bed for next patient | SpecialBed | 0 | 0 | 0 |
| Total | 19,887 | 100 | ||
The targeted area included 24 single‐patient rooms. The activity of each bed was recorded hourly. Real‐time recording in to the Excel spreadsheet on a dedicated laptop occurred from 8:00 AM until 11:00 PM. The most visible or critical event was recorded. Although some activities were not mutually exclusive (eg, a patient could be ventilated and on a continuous infusion simultaneously), the objective was to identify when a room was being used for any critical care service, not enumerate all of them. The observers noted overnight events that occurred from 11:00 PM to 8:00 AM in the morning by reviewing the bedside record and talking to the staff to complete each day's 24‐hour recording. The observers also recorded the hospital‐wide census and the census for the other half of the PICU every 4 hours. The observations occurred over 5 noncontiguous weeks between January 2009 and April 2009.
After all observations were complete, activities were classified as critical care services (CCS) or noncritical‐care services (NCCS). NCCSs were further divided into necessary logistics (defined for analysis purposes as the first hour of any NCCS activity) or nonvalue‐added (the second or greater hour of NCCS). A time limit of 1 hour was chosen to define necessary logistics based on a consensus that nonclinical activities optimally would not take more than 1 hour each. We also analyzed results with 2 hours as the cutoff for necessary logistics. Admission, discharge, and transfer records were reviewed to check for returns to the PICU or hospital within 48 hours of transfer or discharge from the PICU.
Analyses were conducted using Microsoft Excel (Microsoft, Redmond, WA) and Stata 10.0 (StataCorp, College Station, TX). The study was approved by the Children's Hospital of Philadelphia Institutional Review Board with waiver of consent.
RESULTS
A total of 824 hours of recordings included 19,887 bed‐hours with 219 unique patients; among them, 2 remained from the first day of recording in January to the last day in April (sample recording in Figure 1). A total of 50 patients (range, 812 per week) stayed for the entirety of each 1‐week observation period. Of the 47 possible activities, 45 of them were recorded for at least 1 hour in the 5 weeks. Overall, 14 activities accounted for 95% of the observed bed‐hours and 31 activities accounted for the remaining 5%. CCS accounted for 82% of observed bed‐hours, NCCS accounted for 10.4%, and empty unassigned accounted for 8% (Figure 2). Using the 1‐hour cutoff for necessary services, 77% of NCCS time was nonvalue‐added, whereas 23% of it was necessary logistics; using the 2‐hour cutoff, 54% was nonvalue‐added, and 46% was necessary logistics.
During the observation period, <1% of bed‐hours were used for CCS for overflow patients from the neonatal ICU (NICU), cardiac care unit (CCU), cardiac ICU (CICU), or progressive care unit (PCU; tracheostomy/ventilator unit). Although only 4 patients required transport to a rehabilitation facility, their wait time comprised 99 hours (<1%) of total recordings. Eight patients waited a mean of 2.6 hours for transportation home (maximum, 10 hours).
To demonstrate the cycle of room use, activities were divided into 4 categories: room preparation, critical care services, disposition pending, and postcritical care services (Figure 3). As an example of detailed data revealed by direct observation, we identified 102 instances totaling 919 hours when a patient was waiting for a bed assignment on another floor (5% of all bed‐hours). The mean wait time was 9 hours (range, 188 hours) and the median time was 5.5 hours. There were only 15 instances when floor bed assignment took 1 hour or less, and only 9 instances when it took 12 hours. Similarly, considerable time was spent on cleaning rooms between patients: only 66 of 146 instances of cleaning took 1 hour or less. The mean time for cleaning was 2.2 hours (range, 115), and the median was 2 hours. (There were 136 recorded instances of room cleaning and 10 additional episodes that were not recorded but had to be completed for the room to turnover from one patient to the next, yielding a total of 146 instances of cleaning.)
From the 824 hours of recording, we identified 200 hours (25% of time) when there were zero empty unassigned beds available in the section of the PICU being observed. Episodes of full occupancy occurred mostly on weekdays, with 23% of hours of full capacity on Thursdays, 21% on Mondays, and 21% on Wednesdays; only 8% were on Saturdays and <1% on Sundays. These 200 hours fell into 36 separate episodes of complete occupancy, each lasting 122 hours. Each patient, on average, received 3.1 hours of NCCS during each episode of full occupancy (range, 111 hours). Within these 200 hours at capacity, we identified only 15 hours (8%) when all 24 beds were used for CCS. For 72% of the time, there was at least 1 bed with NCCS, and for 37% at least 2 beds. A small portion of the time (7%), the lack of beds was affected by occupancy by patients who should have been in the NICU, CICU, CCU, or PCU.
Data collected through direct observation can be used to understand aggregated and averaged experiences, but also more specific time periods. For example, we identified 1 week with the highest consistent level of occupancy and turnover: March 915 had empty unassigned beds for only 4% of the week. Of the 168 hours in the week, 68 (40%) had full capacity. However, for 90% of the time, at least 1 bed was used for a NCCS. Other analytic options included varying the assumptions around time needed for logistics. Overall, NCCS time on necessary logistics changes from 23% to 46% using 1 hour versus 2 hours as the cutoff. For floor bed assignments, assuming that the first hour of this activity is necessary logistics and any hour thereafter is not, 817 hours were wasted. Even after assuming 2 hours of necessary logistical time (which may also include steps such as nursing and physician sign‐out to the receiving team, often not recorded in the observations), this left 715 hours of NCCS time in which patients waited to be placed elsewhere in the hospital. For room cleaning, because recordings were hourly, but room cleaning could take less time, we performed a sensitivity analysis, converting all 1‐hour recordings to half‐hour recordings to half‐hour recordings (an exaggerated shortening since industry‐standard cleaning may take longer).
Of the 219 patients directly observed, 15 were noted to be waiting for a transfer out of the PICU but experienced a change in disposition before the transfer. On average, these patients waited 8 hours for a floor bed assignment (range, 221) before reverting to a CCS, which then lasted an average of 16.5 hours (range, 149). (Included in this group are 2 patients who experienced this change in disposition twice.) In post hoc review across the 5 weeks, no patients were transferred back to the PICU within 48 hours after being transferred out. During the study period, 19 patients were discharged directly from the PICU (8 to home, 7 by transport to another facility, and 4 to rehabilitation). One patient returned to the hospital (but not the PICU) within 48 hours of being discharged home from the PICU.
During the study period, using the highest census value for recorded for each 24‐hour period and the number of beds available that day, median hospital‐wide occupancy was 93% (interquartile range, 90%96%). During the 35 days of observation, 71% of the days had occupancy >90%, 29% of days had occupancy >95%, and 3% of days had occupancy >100%.
DISCUSSION
In this direct observation of a PICU, we found high usage of beds for delivery of CCS. We identified many episodes in which the half of the PICU we observed was fully occupied (200 of 824 hours), but not necessarily delivering PICU‐level care to all patients. In fact, 75% of the full‐capacity hours had at least 1 patient receiving NCCS and 37% had at least 2. Patients waiting for a floor bed assignment represented nearly 5% of bed‐hours observed (mean 9 hours per patient). That full occupancy was not random, but rather clustered on weekdays, is consistent with other work showing that hospitals are at greater risk for midweek crowding due to the way in which scheduled admissions enter and leave.1925
Our methods provide the basis for operational analysis and improvement to patient flow, such as value stream mapping.9, 26 Process improvement work could be directed to areas of delay uncovered through this analysis and inform clinical and nonclinical management. For example, one of the key problems faced by the PICU was finding floor bed assignments for patients leaving the unit. Simply building more beds in the PICU will not solve this problemand at an estimated cost of $2 million to add a bed, it is likely not an efficient means of responding to poor flow. In these cases, the problem seems to lie downstream, and could suggest shortage of regular floor beds or inefficient bed assignment procedures within the hospital. The output also suggests that variation in nonclinical processes should be a target for improvement, such as time to clean rooms, because variation is known to be a source of nonvalue‐added time in many operations.9, 26 High occupancy on weekdays but low occupancy on weekends also emphasizes the potential for smoothing occupancy to reduce the risk of midweek crowding and to better manage bed utilization and staffing.24, 25
When seeking to reduce nonvalue‐added time, one must weigh the risks of increased efficiency against clinical outcomes. For example, if patients could be transferred out of the PICU faster, would the risk of returns to the PICU be higher? In this study, 15 patients (7%) had a change in disposition from awaiting transfer back to a CCS. The fact that transfers did not happen instantaneously may serve as a safety check to reduce rapid returns, but it is not possible for us to evaluate the reasons why patients did not actually complete the pending transfers. Specifically, we cannot determine whether the patient's clinical status objectively deteriorated, the ICU team made a judgment call to hold the patient, or the floor team refused to accept the transfer. Given this fact, although it appears in this study (and in the health care system more broadly) that there are opportunities to increase efficiency and reduce nonvalue‐added time, it is not realistic (nor advisable) that such time be reduced to zero. Along this line, one must consider separately purely nonclinical functions such as room cleaning and those that include some clinical element, such as time waiting for a patient to be transferred.
Beyond the direct findings of this study, the method should be replicable in other settings and can reveal important information about health care efficiency, capacity, and flexibility. The bottlenecks identified would have been difficult to identify through administrative record review. The exact amount of time to spend on observation may vary from place to place and would depend on the expected variation over time and the level of detail sought. In general, the more common the event and the less variation, the less time needed to observe it.
This study has several limitations that should be considered in terms of interpreting the results and in seeking to reproduce the approach. First, hourly recordings may not be discrete enough for events that took less than 1 hour. To assess the degree to which this would affect our results, we reanalyzed all NCCS by subtracting 30 minutes (0.5 hour) from all recordings, which increased total CCS from 82% to 87% and decreased NCCS by the same 5 percentage points. In a related fashion, our recordings were truncated at the start and end of each 1‐week period, so we could only observe a maximum of 168 hours for any given activity and did not record how long an activity was happening before or after the recordings started or stopped, respectively. Second, each recording could only be for 1 activity per hour. Separate from the level of granularity already noted, this also limits interpretation of critical care activities that may have been simultaneous. However, because the goal of the study was not to describe the provision of critical care services, but rather the times when they were not being delivered, this does not influence our conclusions. For movement of patients, however, we missed instances of physician and nursing calling sign‐out on patients to receiving units, as these events last less than 1 hour (and in the case of surgical patients, generally do not occur as the team provides continuous coverage). The time for such events is then included in other activities. To the extent that this may influence the results, it would increase the perceived time for nonvalue‐added services, but to a limited degree, and never by more than 59 minutes. Third, the overnight hours (11:00 PM to 8:00 AM) were not directly observed, but retrospectively recorded each morning by reviewing the records and discussing the overnight events with the clinical staff. For example, if a patient was intubated at 11:00 PM and at 8:00 AM, the observer would confirm this and record that status for the intervening hours. This is unlikely to result in a substantial impact on the findings, because the overnight hours have a relative degree of stability even for unstable patients in terms of their status of needing or not needing a CCS. Fourth, we did not evaluate the appropriateness of CCS delivered (eg, how long a patient was ventilated). Our definitions for CCS and NCSS were based on Children's Hospital of Philadelphia practices, which may not be the same as those of other facilities. The categorization of CCS was objective for activities such as ventilation or continuous infusion, but was less clear for the not otherwise specified recordings, which represented patients with a complex illness or projected organ, respiratory, cardiac, or neurological failure. These patients were not receiving a specific critical care intervention, but were deemed to need to be in the PICU as opposed to a regular floor (eg, for frequent monitoring of potential respiratory failure). It would also include patients receiving combinations of therapies more efficiently delivered in the PICU. For that, the observers relied on the judgment of clinicians (primarily nurses) to determine whether the patient needed to be in the PICU or not; if no specific reason could be provided, not otherwise specified was applied. These 192 instances accounted for 2982 aggregate bed‐hours (15% of total). It is difficult to judge the direction of bias, because overestimation of need to be in the PICU may be as likely to occur as underestimation. Fifth, the very presence of the observers may have changed behavior. Knowing that they were being observed staff may have acted with greater efficiency than otherwise. We expect that such a finding would lead to less time appearing as necessary logistics or NCCS. Finally, results may not be generalizable to other hospitals or hospital settings. There are clearly important contextual factors, not only for the location but also for the duration. For example, staffing was never an issue during the 5 weeks of observation, but there are locations where an empty bed is not necessarily usable due to lack of staffing. Nonetheless, we believe the results provide a generalizable approach and methodology for other settings (and staffing could be a reason for an empty bed).
In terms of the setting, as noted, we observed one discrete 24‐bed unit, which comprises half of the total PICU. Thus, statements that the PICU was at full capacity must be interpreted in the context that additional rooms may have been available on the other side. Patients are generally admitted alternately to each unit, so the occupancies should parallel each other. We recorded the census every 4 hours for both sides from the electronic system (Sunrise Clinical Manager [SCM]). However, this only accounts for patients physically in beds, not beds held for patients in other locations. Thus, we would expect a discrepancy between direct observation and the SCM value. Through analysis of the entire pediatric intensive care unit,* that part which observed directly, and that which we did not observe directly using census data, we think it reasonable to assert that both units of the total PICU had constrained capacity during the times we directly observed and recorded such constraint on one side.
This study demonstrates the use of direct observation for inpatient settings to learn about resource utilization and identification of value‐added services. PubMed searches for the terms efficiency, flow, process redesign, and time management bring up many more references for operating rooms than for ICUs or inpatient beds. Some examples of ICU‐directed work include videography of an ICU in Australia27 and human factor analysis in ICU nursing.5 Time‐motion studies have also been conducted on clinical staff, such as physicians.28, 29
In conclusion, we found that direct observation provided important insights into the utilization of patient rooms in an important inpatient setting. Data such as these are valuable for clinical and process improvement work, as well as understanding how best to match capacity to patient need. Finally, the methodology is reproducible for other settings and would be an additional tool to measuring and improving the efficiency and value of the health system. When appropriate, this approach can also evaluate the effectiveness of process improvement, help identify and reduce waste,13 and contribute to the growing field that merges operations management with hospital administration and clinical care: in other words, evidence‐based management.30
Acknowledgements
The authors thank Paula Agosto, Patricia Hubbs, Heidi Martin, and Annette Bollig for contributions to the study design.
In comparing direct observation to the SCM count, we found perfect concordance for 110 hours (55%) during which 0 beds were available. For the other 90 hours, SCM reported 1 bed being available in 46 hours (23%), 2 beds being available in 24 hours (12%), 3 beds being available in 17 hours (9%), and 4 beds being available in 3 hours (2%)all while we directly observed 0 beds being available. Thus, cumulatively, 90% of the hours observed with no beds had an SCM report availability of 02 beds; 99% of the time that was 03 beds. Applying this rate of mismatch to the unit that we did not observe directly, SCM reported 0 beds for 46 (23%) of the 200 hours the observation unit was full; SCM reported 1 bed available in 70 hours (35%), 2 beds open in 42 hours (21%), 3 beds open in 26 hours (13%), and 4 beds open in 16 hours (8%). Cumulatively, that is 79% of the time with 02 beds and 92% at 03 beds. From this, we conclude that the combined PICU for both sides was likely functionally full at least 158 of the 200 hours that the side we observed was full (79% 200 hours) and likely had very constrained capacity during the other 42 hours.
- ,,,,.The effect of hospital occupancy on emergency department length of stay and patient disposition.Acad Emerg Med.2003;10:127–133.
- ,,,.The effect of hospital bed occupancy on throughput in the pediatric emergency department.Ann Emerg Med.2009;53:767–776.
- ,,,.A Comparison of in‐hospital mortality risk conferred by high hospital occupancy, differences in nurse staffing levels, weekend admission, and seasonal influenza.Medical Care.2010;48:224–232.
- .Intensive care unit occupancy: making room for more patients.Crit Care Med.2009;37:1794–1795.
- ,.A human factors engineering conceptual framework of nursing workload and patient safety in intensive care units.Intensive Crit Care Nurs.2005;21:284–301.
- ,,, et al.High level of burnout in intensivists: prevalence and associated factors.Am J Respir Crit Care Med.2007;175:686–692.
- ,,.Length of stay and efficiency in pediatric intensive care units.J Pediatr.1998;133:79–85.
- ,.Variability in duration of stay in pediatric intensive care units: a multiinstitutional study.J Pediatr.1996;128:35–44.
- ,.Matching Supply with Demand: An Introduction to Operations Management.New York, NY:McGraw‐Hill;2006.
- ,.Impact of workload on service time and patient safety: an econometric analysis of hospital operations.Management Science.2009;55:1486–1498.
- .OPIM 631: Operations Management.Philadelphia, PA:Wharton School, University of Pennsylvania;2008.
- ,,.From waste to value in health care.JAMA.2008;299:568–571.
- .Eliminating “waste” in health care.JAMA.2009;302:2481–2482.
- .Toyota Production System: Beyond Large‐scale Production.London, UK:Productivity Press;1995.
- ,.Interdisciplinary work flow assessment and redesign decreases operating room turnover time and allows for additional caseload.Arch Surg.2006;141:65–69.
- ,,,.Improving operating room efficiency through process redesign.Surgery.2006;140:509–514.
- ,,,.Successful strategies for improving operating room efficiency at academic institutions.Anesth Analg.1998year="1998"1998;86:896–906.
- ,,.Efficiency of the operating room suite.Am J Surg.2003;185:244–250.
- ,,, et al.Children's hospitals do not acutely respond to high occupancy.Pediatrics.2010;125:974–981.
- Institute for Healthcare Improvement. Smoothing elective surgical admissions. Available at: http://www.ihi.org/IHI/Topics/Flow/Patient Flow/EmergingContent/SmoothingElectiveSurgicalAdmissions.htm. Accessed October 24,2008.
- Boston Hospital Sees Big Impact from Smoothing Elective Schedule.OR Manager. Volume 20, no. 12,2004.
- ,.Rethinking rapid response teams.JAMA.2010;304:1375–1376.
- Litvak E, ed.Managing Patient Flow in Hospitals: Strategies and Solutions.2nd ed.Oak Brook, IL:Joint Commission Resources;2009.
- ,,,,,.Scheduled admissions and high occupancy at a children's hospital.J Hosp Med.2011;6:81–87.
- ,,, et al.Addressing inpatient crowding by smoothing occupancy at children's hospitals.J Hosp Med.2011;6:466–473.
- ,.Learning to See: Value Stream Mapping to Add Value and Eliminate MUDA.Cambridge, MA:Lean Enterprise Institute;1999.
- ,,.Reshaping ICU ward round practices using video‐reflexive ethnography.Qual Health Res.2008;18:380–390.
- ,,.How hospitalists spend their time: insights on efficiency and safety.J Hosp Med.2006;1:88–93.
- ,,, et al.Where did the day go? A time‐motion study of hospitalists.J Hosp Med2010;5:323–238.
- ,,.Improving patient care by linking evidence‐based medicine and evidence‐based management.JAMA.2007;298:673–676.
- ,,,,.The effect of hospital occupancy on emergency department length of stay and patient disposition.Acad Emerg Med.2003;10:127–133.
- ,,,.The effect of hospital bed occupancy on throughput in the pediatric emergency department.Ann Emerg Med.2009;53:767–776.
- ,,,.A Comparison of in‐hospital mortality risk conferred by high hospital occupancy, differences in nurse staffing levels, weekend admission, and seasonal influenza.Medical Care.2010;48:224–232.
- .Intensive care unit occupancy: making room for more patients.Crit Care Med.2009;37:1794–1795.
- ,.A human factors engineering conceptual framework of nursing workload and patient safety in intensive care units.Intensive Crit Care Nurs.2005;21:284–301.
- ,,, et al.High level of burnout in intensivists: prevalence and associated factors.Am J Respir Crit Care Med.2007;175:686–692.
- ,,.Length of stay and efficiency in pediatric intensive care units.J Pediatr.1998;133:79–85.
- ,.Variability in duration of stay in pediatric intensive care units: a multiinstitutional study.J Pediatr.1996;128:35–44.
- ,.Matching Supply with Demand: An Introduction to Operations Management.New York, NY:McGraw‐Hill;2006.
- ,.Impact of workload on service time and patient safety: an econometric analysis of hospital operations.Management Science.2009;55:1486–1498.
- .OPIM 631: Operations Management.Philadelphia, PA:Wharton School, University of Pennsylvania;2008.
- ,,.From waste to value in health care.JAMA.2008;299:568–571.
- .Eliminating “waste” in health care.JAMA.2009;302:2481–2482.
- .Toyota Production System: Beyond Large‐scale Production.London, UK:Productivity Press;1995.
- ,.Interdisciplinary work flow assessment and redesign decreases operating room turnover time and allows for additional caseload.Arch Surg.2006;141:65–69.
- ,,,.Improving operating room efficiency through process redesign.Surgery.2006;140:509–514.
- ,,,.Successful strategies for improving operating room efficiency at academic institutions.Anesth Analg.1998year="1998"1998;86:896–906.
- ,,.Efficiency of the operating room suite.Am J Surg.2003;185:244–250.
- ,,, et al.Children's hospitals do not acutely respond to high occupancy.Pediatrics.2010;125:974–981.
- Institute for Healthcare Improvement. Smoothing elective surgical admissions. Available at: http://www.ihi.org/IHI/Topics/Flow/Patient Flow/EmergingContent/SmoothingElectiveSurgicalAdmissions.htm. Accessed October 24,2008.
- Boston Hospital Sees Big Impact from Smoothing Elective Schedule.OR Manager. Volume 20, no. 12,2004.
- ,.Rethinking rapid response teams.JAMA.2010;304:1375–1376.
- Litvak E, ed.Managing Patient Flow in Hospitals: Strategies and Solutions.2nd ed.Oak Brook, IL:Joint Commission Resources;2009.
- ,,,,,.Scheduled admissions and high occupancy at a children's hospital.J Hosp Med.2011;6:81–87.
- ,,, et al.Addressing inpatient crowding by smoothing occupancy at children's hospitals.J Hosp Med.2011;6:466–473.
- ,.Learning to See: Value Stream Mapping to Add Value and Eliminate MUDA.Cambridge, MA:Lean Enterprise Institute;1999.
- ,,.Reshaping ICU ward round practices using video‐reflexive ethnography.Qual Health Res.2008;18:380–390.
- ,,.How hospitalists spend their time: insights on efficiency and safety.J Hosp Med.2006;1:88–93.
- ,,, et al.Where did the day go? A time‐motion study of hospitalists.J Hosp Med2010;5:323–238.
- ,,.Improving patient care by linking evidence‐based medicine and evidence‐based management.JAMA.2007;298:673–676.
Copyright © 2011 Society of Hospital Medicine
Addressing Inpatient Crowding
High levels of hospital occupancy are associated with compromises to quality of care and access (often referred to as crowding), 18 while low occupancy may be inefficient and also impact quality. 9, 10 Despite this, hospitals typically have uneven occupancy. Although some demand for services is driven by factors beyond the control of a hospital (eg, seasonal variation in viral illness), approximately 15%30% of admissions to children's hospitals are scheduled from days to months in advance, with usual arrivals on weekdays. 1114 For example, of the 3.4 million elective admissions in the 2006 Healthcare Cost and Utilization Project Kids Inpatient Database (HCUP KID), only 13% were admitted on weekends. 14 Combined with short length of stay (LOS) for such patients, this leads to higher midweek and lower weekend occupancy. 12
Hospitals respond to crowding in a number of ways, but often focus on reducing LOS to make room for new patients. 11, 15, 16 For hospitals that are relatively efficient in terms of LOS, efforts to reduce it may not increase functional capacity adequately. In children's hospitals, median lengths of stay are 2 to 3 days, and one‐third of hospitalizations are 1 day or less. 17 Thus, even 10%20% reductions in LOS trims hours, not days, from typical stays. Practical barriers (eg, reluctance to discharge in the middle of the night, or family preferences and work schedules) and undesired outcomes (eg, increased hospital re‐visits) are additional pitfalls encountered by relying on throughput enhancement alone.
Managing scheduled admissions through smoothing is an alternative strategy to reduce variability and high occupancy. 6, 12, 1820 The concept is to proactively control the entry of patients, when possible, to achieve more even levels of occupancy, instead of the peaks and troughs commonly encountered. Nonetheless, it is not a widely used approach. 18, 20, 21 We hypothesized that children's hospitals had substantial unused capacity that could be used to smooth occupancy, which would reduce weekday crowding. While it is obvious that smoothing will reduce peaks to average levels (and also raise troughs), we sought to quantify just how large this difference wasand thereby quantify the potential of smoothing to reduce inpatient crowding (or, conversely, expose more patients to high levels of occupancy). Is there enough variation to justify smoothing, and, if a hospital does smooth, what is the expected result? If the number of patients removed from exposure to high occupancy is not substantial, other means to address inpatient crowding might be of more value. Our aims were to quantify the difference in weekday versus weekend occupancy, report on mathematical feasibility of such an approach, and determine the difference in number of patients exposed to various levels of high occupancy.
Methods
Data Source
This retrospective study was conducted with resource‐utilization data from 39 freestanding, tertiary‐care children's hospitals in the Pediatric Health Information System (PHIS). Participating hospitals are located in noncompeting markets of 23 states, plus the District of Columbia, and affiliated with the Child Health Corporation of America (CHCA, Shawnee Mission, KS). They account for 80% of freestanding, and 20% of all general, tertiary‐care children's hospitals. Data quality and reliability are assured through joint ongoing, systematic monitoring. The Children's Hospital of Philadelphia Committees for the Protection of Human Subjects approved the protocol with a waiver of informed consent.
Patients
Patients admitted January 1December 31, 2007 were eligible for inclusion. Due to variation in the presence of birthing, neonatal intensive care, and behavioral health units across hospitals, these beds and associated patients were excluded. Inpatients enter hospitals either as scheduled (often referred to as elective) or unscheduled (emergent or urgent) admissions. Because PHIS does not include these data, KID was used to standardize the PHIS data for proportion of scheduled admissions. 22 (KID is a healthcare database of 23 million pediatric inpatient discharges developed through federalstateindustry partnership, and sponsored by the Agency for Healthcare Research and Quality [AHRQ].) Each encounter in KID includes a principal International Classification of Diseases, 9th revision (ICD‐9) discharge diagnosis code, and is designated by the hospital as elective (ranging from chemotherapy to tonsillectomy) or not elective. Because admissions, rather than diagnoses, are scheduled, a proportion of patients with each primary diagnosis in KID are scheduled (eg, 28% of patients with a primary diagnosis of esophageal reflux). Proportions in KID were matched to principal diagnoses in PHIS.
Definitions
The census was the number of patients registered as inpatients (including those physically in the emergency department [ED] from time of ED arrival)whether observation or inpatient statusat midnight, the conclusion of the day. Hospital capacity was set using CHCA data (and confirmed by each hospital's administrative personnel) as the number of licensed in‐service beds available for patients in 2007; we assumed beds were staffed and capacity fixed for the year. Occupancy was calculated by dividing census by capacity. Maximum occupancy in a week referred to the highest occupancy level achieved in a seven‐day period (MondaySunday). We analyzed a set of thresholds for high‐occupancy (85%, 90%, 95%, and 100%), because there is no consistent definition for when hospitals are at high occupancy or when crowding occurs, though crowding has been described as starting at 85% occupancy. 2325
Analysis
The hospital was the unit of analysis. We report hospital characteristics, including capacity, number of discharges, and census region, and annual standardized length of stay ratio (SLOSR) as observed‐to‐expected LOS.
Smoothing Technique
A retrospective smoothing algorithm set each hospital's daily occupancy during a week to that hospital's mean occupancy for the week; effectively spreading the week's volume of patients evenly across the days of the week. While inter‐week and inter‐month smoothing were considered, intra‐week smoothing was deemed more practical for the largest number of patients, as it would not mean delaying care by more than one week. In the case of a planned treatment course (eg, chemotherapy), only intra‐week smoothing would maintain the necessary scheduled intervals of treatment.
Mathematical Feasibility
To approximate the number of patient admissions that would require different scheduling during a particular week to achieve smoothed weekly occupancy, we determined the total number of patient‐days in the week that required different scheduling and divided by the average LOS for the week. We then divided the number of admissions‐to‐move by total weekly admissions to compute the percentage at each hospital across 52 weeks of the year.
Measuring the Impact of Smoothing
We focused on the frequency and severity of high occupancy and the number of patients exposed to it. This framework led to 4 measures that assess the opportunity and effect of smoothing:
Difference in hospital weekdayweekend occupancy: Equal to 12‐month median of difference between mean weekday occupancy and mean weekend occupancy for each hospital‐week.
Difference in hospital maximummean occupancy: Equal to median of difference between maximum one‐day occupancy and weekly mean (smoothed) occupancy for each hospital‐week. A regression line was derived from the data for the 39 hospitals to report expected reduction in peak occupancy based on the magnitude of the difference between weekday and weekend occupancy.
Difference in number of hospitals exposed to above‐threshold occupancy: Equal to difference, pre‐ and post‐smoothing, in number of hospitals facing high‐occupancy conditions on an average of at least one weekday midnight per week during the year at different occupancy thresholds.
Difference in number of patients exposed to above‐threshold occupancy: Equal to difference, pre‐ and post‐smoothing, in number of patients exposed to hospital midnight occupancy at the thresholds. We utilized patient‐days for the calculation to avoid double‐counting, and divided this by average LOS, in order to determine the number of patients who would no longer be exposed to over‐threshold occupancy after smoothing, while also adjusting for patients newly exposed to over‐threshold occupancy levels.
All analyses were performed separately for each hospital for the entire year and then for winter (DecemberMarch), the period during which most crowding occurred. Analyses were performed using SAS (version 9.2, SAS Institute, Inc, Cary, NC); P values <0.05 were considered statistically significant.
Results
The characteristics of the 39 hospitals are provided in Table 1. Based on standardization with KID, 23.6% of PHIS admissions were scheduled (range: 18.1%35.8%) or a median of 81.5 scheduled admissions per week per hospital; 26.6% of weekday admissions were scheduled versus 16.1% for weekends. Overall, 12.4% of scheduled admissions entered on weekends. For all patients, median LOS was three days (interquartile range [IQR]: twofive days), but median LOS for scheduled admissions was two days (IQR: onefour days). The median LOS and IQR were the same by day of admission for all days of the week. Most hospitals had an overall SLOSR close to one (median: 0.9, IQR: 0.91.1). Overall, hospital mean midnight occupancy ranged from 70.9% to 108.1% on weekdays and 65.7% to 94.9% on weekends. Uniformly, weekday occupancy exceeded weekend occupancy, with a median difference of 8.2% points (IQR: 7.2%9.5% points). There was a wide range of median hospital weekdayweekend occupancy differences across hospitals (Figure 1). The overall difference was less in winter (median difference: 7.7% points; IQR: 6.3%8.8% points) than in summer (median difference: 8.6% points; IQR: 7.4%9.8% points (Wilcoxon Sign Rank test, P < 0.001). Thirty‐five hospitals (89.7%) exceeded the 85% occupancy threshold and 29 (74.4%) exceeded the 95% occupancy threshold on at least 20% of weekdays (Table 2). Across all the hospitals, the median difference in weekly maximum and weekly mean occupancy was 6.6% points (IQR: 6.2%7.4% points) (Figure 2).
| Characteristics | No. (%) |
|---|---|
| |
| Licensed in‐service beds | n = 39 hospitals |
| <200 beds | 6 (15.4) |
| 200249 beds | 10 (25.6) |
| 250300 beds | 14 (35.9) |
| >300 beds | 9 (23.1) |
| No. of discharges | |
| <10,000 | 5 (12.8) |
| 10,00013,999 | 14 (35.9) |
| 14,00017,999 | 11 (28.2) |
| >18,000 | 9 (23.1) |
| Census region | |
| West | 9 (23.1) |
| Midwest | 11 (28.2) |
| Northeast | 6 (15.4) |
| South | 13 (33.3) |
| Admissions | n = 590,352 admissions |
| Medical scheduled admissions* | 79,683 |
| Surgical scheduled admissions* | 59,640 |
| Total scheduled admissions* (% of all admissions) | 139,323 (23.6) |
| Weekend medical scheduled admissions* (% of all medical scheduled admissions) | 13,546 (17.0) |
| Weekend surgical scheduled admissions* (% of all surgical scheduled admissions) | 3,757 (6.3) |
| Weekend total scheduled admissions* (% of total scheduled admissions) | 17,276 (12.4) |
| Entire Year | >85% | Occupancy Threshold | >95% | >100% |
|---|---|---|---|---|
| >90% | ||||
| ||||
| No. of hospitals (n = 39) with mean weekday occupancy above threshold | ||||
| Before smoothing (current state) | 33 | 25 | 14 | 6 |
| After smoothing | 32 | 22 | 10 | 1 |
| No. of hospitals (n = 39) above threshold 20% of weekdays | ||||
| Before smoothing (current state) | 35 | 34 | 29 | 14 |
| After smoothing | 35 | 32 | 21 | 9 |
| Median (IQR) no. of patient‐days per hospital not exposed to occupancy above threshold by smoothing | 3,071 | 281 | 3236 | 3281 |
| (5,552, 919) | (5,288, 3,103) | (0, 7,083) | (962, 8,517) | |
| Median (IQR) no. of patients per hospital not exposed to occupancy above threshold by smoothing | 596 | 50 | 630 | 804 |
| (1,190, 226) | (916, 752) | (0, 1,492) | (231, 2,195) | |
Smoothing reduced the number of hospitals at each occupancy threshold, except 85% (Table 2). As a linear relationship, the reduction in weekday peak occupancy (y) based on a hospital's median difference in weekly maximum and weekly mean occupancy (x) was y = 2.69 + 0.48x. Thus, a hospital with a 10% point difference between weekday and weekend occupancy could reduce weekday peak by 7.5% points.
Smoothing increased the number of patients exposed to the lower thresholds (85% and 90%), but decreased the number of patients exposed to >95% occupancy (Table 2). For example, smoothing at the 95% threshold resulted in 630 fewer patients per hospital exposed to that threshold. If all 39 hospitals had within‐week smoothing, a net of 39,607 patients would have been protected from exposure to >95% occupancy and a net of 50,079 patients from 100% occupancy.
To demonstrate the varied effects of smoothing, Table 3 and Figure 3 present representative categories of response to smoothing depending on pre‐smoothing patterns. While not all hospitals decreased occupancy to below thresholds after smoothing (Types B and D), the overall occupancy was reduced and fewer patients were exposed to extreme levels of high occupancy (eg, >100%).
| Category | Before Smoothing Hospital Description | After Smoothing Hospital Description | No. of Hospitals at 85% Threshold (n = 39) | No. of Hospitals at 95% Threshold (n = 39) |
|---|---|---|---|---|
| ||||
| Type A | Weekdays above threshold | All days below threshold, resulting in net decrease in patients exposed to occupancies above threshold | 3 | 1 |
| Weekends below threshold | ||||
| Type B | Weekdays above threshold | All days above threshold, resulting in net increase in patients exposed to occupancies above threshold | 12 | 18 |
| Weekends below threshold | ||||
| Type C | All days of week below threshold | All days of week below threshold | 6 | 19 |
| Type D | All days of week above threshold | All days of week above threshold, resulting in net decrease in patients exposed to extreme high occupancy | 18 | 1 |
To achieve within‐week smoothing, a median of 7.4 patient‐admissions per week (range: 2.314.4) would have to be scheduled on a different day of the week. This equates to a median of 2.6% (IQR: 2.25%, 2.99%; range: 0.02%9.2%) of all admissionsor 9% of a typical hospital‐week's scheduled admissions.
Discussion
This analysis of 39 children's hospitals found high levels of occupancy and weekend occupancy lower than weekday occupancy (median difference: 8.2% points). Only 12.4% of scheduled admissions entered on weekends. Thus, weekend capacity is available to offset high weekday occupancy. Hospitals at the higher end of the occupancy thresholds (95%, 100%) would reduce the number of days operating at very high occupancy and the number of patients exposed to such levels by smoothing. This change is mathematically feasible, as a median of 7.4 patients would have to be proactively scheduled differently each week, just under one‐tenth of scheduled admissions. Since LOS by day of admission was the same (median: two days), the opportunity to affect occupancy by shifting patients should be relatively similar for all days of the week. In addition, these admissions were short, conferring greater flexibility. Implementing smoothing over the course of the week does not necessarily require admitting patients on weekends. For example, Monday admissions with an anticipated three‐day LOS could enter on Friday with anticipated discharge on Monday to alleviate midweek crowding and take advantage of unoccupied weekend beds. 26
At the highest levels of occupancy, smoothing reduces the frequency of reaching these maximum levels, but can have the effect of actually exposing more patient‐days to a higher occupancy. For example, for nine hospitals in our analysis with >20% of days over 100%, smoothing decreased days over 100%, but exposed weekend patients to higher levels of occupancy (Figure 3). Since most admissions are short and most scheduled admissions currently occur on weekdays, the number of individual patients (not patient‐days) newly exposed to such high occupancy may not increase much after smoothing at these facilities. Regardless, hospitals with such a pattern may not be able to rely solely on smoothing to avoid weekday crowding, and, if they are operating efficiently in terms of SLOSR, might be justified in building more capacity.
Consistent with our findings, the Institute for Healthcare Improvement, the Institute for Healthcare Optimization, and the American Hospital Association Quality Center stress that addressing artificial variability of scheduled admissions is a critical first step to improving patient flow and quality of care while reducing costs. 18, 21, 27 Our study suggests that small numbers of patients need to be proactively scheduled differently to decrease midweek peak occupancy, so only a small proportion of families would need to find this desirable to make it attractive for hospitals and patients. This type of proactive smoothing decreases peak occupancy on weekdays, reducing the safety risks associated with high occupancy, improving acute access for emergent patients, shortening wait‐times and loss of scheduled patients to another facility, and increasing procedure volume (3%74% in one study). 28 Smoothing may also increase quality and safety on weekends, as emergent patients admitted on weekends experience more delays in necessary treatment and have worse outcomes. 2932 In addition, increasing scheduled admissions to span weekends may appeal to some families wishing to avoid absence from work to be with their hospitalized child, to parents concerned about school performanceand may also appeal to staff members seeking flexible schedules. Increasing weekend hospital capacity is safe, feasible, and economical, even when considering the increased wages for weekend work. 33, 34 Finally, smoothing over the whole week allows fixed costs (eg, surgical suites, imaging equipment) to be allocated over 7 days rather than 5, and allows for better matching of revenue to the fixed expenses.
Rather than a prescriptive approach, our work suggests hospitals need to identify only a small number of patients to proactively shift, providing them opportunities to adapt the approach to local circumstances. The particular patients to move around may also depend on the costs and benefits of services (eg, radiologic, laboratory, operative) and the hospital's existing patterns of staffing. A number of hospitals that have engaged in similar work have achieved sustainable results, such as Seattle Children's Hospital, Boston Medical Center, St. John's Regional Health Center, and New York University Langone Medical Center. 19, 26, 3537 In these cases, proactive smoothing took advantage of unused capacity and decreased crowding on days that had been traditionally very full. Hospitals that rarely or never have high‐occupancy days, and that do not expect growth in volume, may not need to employ smoothing, whereas others that have crowding issues primarily in the winter may wish to implement smoothing techniques seasonally.
Aside from attempting to reduce high‐occupancy through modification of admission patterns, other proactive approaches include optimizing staffing and processes around care, improving efficiency of care, and building additional beds. 16, 25, 38, 39 However, the expense of construction and the scarcity of capital often preclude this last option. Among children's hospitals, with SLOSR close to one, implementing strategies to reduce the LOS during periods of high occupancy may not result in meaningful reductions in LOS, as such approaches would only decrease the typical child's hospitalization by hours, not days. In addition to proactive strategies, hospitals also rely on reactive approaches, such as ED boarding, placing patients in hallways on units, diverting ambulances or transfers, or canceling scheduled admissions at the last moment, to decrease crowding. 16, 39, 40
This study has several limitations. First, use of administrative data precluded modeling all responses. For example, some hospitals may be better able to accommodate fluctuations in census or high occupancy without compromising quality or access. Second, we only considered intra‐week smoothing, but hospitals may benefit from smoothing over longer periods of time, especially since children's hospitals are busier in winter months, but incoming scheduled volume is often not reduced. 11 Hospitals with large occupancy variations across months may want to consider broadening the time horizon for smoothing, and weigh the costs and benefits over that period of time, including parental and clinician concerns and preferences for not delaying treatment. At the individual hospital level, discrete‐event simulation would likely be useful to consider the trade‐offs of smoothing to different levels and over different periods of time. Third, we assumed a fixed number of beds for the year, an approach that may not accurately reflect actual available beds on specific days. This limitation was minimized by counting all beds for each hospital as available for all the days of the year, so that hospitals with a high census when all available beds are included would have an even higher percent occupancy if some of those beds were not actually open. In a related way, then, we also do not consider how staffing may need to be altered or augmented to care for additional patients on certain days. Fourth, midnight census, the only universally available measure, was used to determine occupancy rather than peak census. Midnight census provides a standard snapshot, but is lower than mid‐day peak census. 41 In order to account for these limitations, we considered several different thresholds of high occupancy. Fifth, we smoothed at the hospital level, but differential effects may exist at the unit level. Sixth, to determine proportion of scheduled admissions, we used HCUP KID proportions on PHIS admissions. Overall, this approach likely overestimated scheduled medical admissions on weekends, thus biasing our result towards the null hypothesis. Finally, only freestanding children's hospitals were included in this study. While this may limit generalizability, the general concept of smoothing occupancy should apply in any setting with substantial and consistent variation.
In summary, our study revealed that children's hospitals often face high midweek occupancy, but also have substantial unused weekend capacity. Hospitals facing challenges with high weekday occupancy could proactively use a smoothing approach to decrease the frequency and severity of high occupancy. Further qualitative evaluation is also warranted around child, family, and staff preferences concerning scheduled admissions, school, and work.
- , , , . A comparison of in‐hospital mortality risk conferred by high hospital occupancy, differences in nurse staffing levels, weekend admission, and seasonal influenza. Medical Care. 2010;48(3):224–232.
- , , , et al. Hospital workload and adverse events. Med Care. 2007;45(5):448–455.
- , , , , . Impact of admission‐day crowding on the length of stay of pediatric hospitalizations. Pediatrics. 2008;121(4):e718–e730.
- , , , . The effect of hospital bed occupancy on throughput in the pediatric emergency department. Ann Emerg Med. 2009;53(6):767–776.
- . The tipping point: the relationship between volume and patient harm. Am J Med Qual. 2008;23(5):336–341.
- , , , , , . Managing unnecessary variability in patient demand to reduce nursing stress and improve patient safety. Jt Comm J Qual Patient Saf. 2005;31(6):330–338.
- Hospital‐Based Emergency Care: At the Breaking Point. Washington, DC: Institute of Medicine Committee on the Future of Emergency Care in the United States Health System; 2006.
- , , , , . The effect of hospital occupancy on emergency department length of stay and patient disposition. Acad Emerg Med. 2003;10(2):127–133.
- . Interpreting the Volume‐Outcome Relationship in the Context of Health Care Quality: Workshop Summary. Washington, DC: National Academies Press; 2000.
- , , , , . Has recognition of the relationship between mortality rates and hospital volume for major cancer surgery in California made a difference? A follow‐up analysis of another decade. Ann Surg. 2009;250(3):472–483.
- , , , et al. Children's hospitals do not acutely respond to high occupancy. Pediatrics. 2010;125:974–981.
- , , , , , . Scheduled admissions and high occupancy at a children's hospital. J Hosp Med. 2011;6(2):81–87.
- , , . Characteristics of weekday and weekend hospital admissions. HCUP Statistical Brief. 2010;87.
- Agency for Healthcare Research and Quality. HCUP databases, Healthcare Cost and Utilization Project (HCUP); 2008. Available at: http://www.hcup‐us.ahrq.gov/kidoverview.jsp. Accessed July 15, 2009.
- , et al. Managing capacity to reduce emergency department overcrowding and ambulance diversions. J Qual Patient Saf. 2006;32(5):239–245.
- Institute for Healthcare Improvement. Flow initiatives; 2008. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed February 20, 2008.
- , , , , , . Trends in high‐turnover stays among children hospitalized in the United States, 1993–2003. Pediatrics. 2009;123(3):996–1002.
- Institute for Healthcare Improvement. Smoothing elective surgical admissions. Available at: http://www.ihi.org/IHI/Topics/Flow/PatientFlow/EmergingContent/SmoothingElectiveSurgicalAdmissions.htm. Accessed October 24, 2008.
- Boston hospital sees big impact from smoothing elective schedule. OR Manager. 2004;20:12.
- . Managing Variability in Patient Flow Is the Key to Improving Access to Care, Nursing Staffing, Quality of Care, and Reducing Its Cost. Paper presented at Institute of Medicine, Washington, DC; June 24, 2004.
- American Hospital Association Quality Center. Available at: http://www.ahaqualitycenter.org/ahaqualitycenter/. Accessed October 14, 2008.
- Healthcare Cost and Utilization Project (HCUP). Kids' Inpatient Database (KID); July 2008. Available at: http://www.hcup‐us.ahrq.gov/kidoverview.jsp. Accessed September 10, 2008.
- , , . Using a queuing model to help plan bed allocation in a department of geriatric medicine. Health Care Manag Sci. 2002;5(4):307–313.
- . How many hospital beds? Inquiry. 2002;39(4):400–412.
- . Institute for Healthcare Improvement. Patient flow comments. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed September 10, 2008.
- . Factory efficiency comes to the hospital. New York Times. July 9, 2010.
- Institute for Healthcare Improvement. Re‐engineering the operating room. Available at: http://www.ihi.org/IHI/Programs/ConferencesAndSeminars/ReengineeringtheOperatingRoomSept08.htm. Accessed November 8, 2008.
- , . Enhanced weekend service: an affordable means to increased hospital procedure volume. CMAJ. 2005;172(4):503–504.
- , . Mortality among patients admitted to hospitals on weekends as compared with weekdays. N Engl J Med. 2001;345:663–668.
- , , , , , . Weekend versus weekday admission and mortality from myocardial infarction. N Engl J Med. 2007;356:1099–1109.
- , . Waiting for urgent procedures on the weekend among emergently hospitalized patients. Am J Med. 2004;117:175–181.
- . Do hospitals provide lower quality care on weekends? Health Serv Res. 2007;42:1589–1612.
- . Hospital saves by working weekends. Mod Healthc. 1996;26:82–99.
- , , , , . Weekend and holiday exercise testing in patients with chest pain. J Gen Intern Med. 1999;14:10–14.
- . Boston Medical Center Case Study: Institute of Healthcare Optimization; 2006. Available at: http://www.ihoptimize.org/8f16e142‐eeaa‐4898–9e62–660218f19ffb/download.htm. Accessed October 3, 2010.
- , , , . The impact of IMPACT on St John's Regional Health Center. Mo Med. 2003;100:590–592.
- NYU Langone Medical Center Extends Access to Non‐Emergent Care as Part of Commitment to Patient‐Centered Care (June 23, 2010). Available at: http://communications.med.nyu.edu/news/2010/nyu‐langone‐medical‐center‐extends‐access‐non‐emergent‐care‐part‐commitment‐patient‐center. Accessed October 3, 2010.
- Carondelet St. Mary's Hospital. A pragmatic approach to improving patient efficiency throughput. Improvement Report 2005. Available at: http://www.ihi.org/IHI/Topics/Flow/PatientFlow/ImprovementStories/APragmaticApproachtoImprovingPatientEfficiencyThroughput.htm. Accessed October 3, 2010.
- AHA Solutions. Patient Flow Challenges Assessment 2009. Chicago, IL; 2009.
- , , , , , . A conceptual model of emergency department crowding. Ann Emerg Med. 2003;42(2):173–180.
- . Annual bed statistics give a misleading picture of hospital surge capacity. Ann Emerg Med. 2006;48(4):384–388.
High levels of hospital occupancy are associated with compromises to quality of care and access (often referred to as crowding), 18 while low occupancy may be inefficient and also impact quality. 9, 10 Despite this, hospitals typically have uneven occupancy. Although some demand for services is driven by factors beyond the control of a hospital (eg, seasonal variation in viral illness), approximately 15%30% of admissions to children's hospitals are scheduled from days to months in advance, with usual arrivals on weekdays. 1114 For example, of the 3.4 million elective admissions in the 2006 Healthcare Cost and Utilization Project Kids Inpatient Database (HCUP KID), only 13% were admitted on weekends. 14 Combined with short length of stay (LOS) for such patients, this leads to higher midweek and lower weekend occupancy. 12
Hospitals respond to crowding in a number of ways, but often focus on reducing LOS to make room for new patients. 11, 15, 16 For hospitals that are relatively efficient in terms of LOS, efforts to reduce it may not increase functional capacity adequately. In children's hospitals, median lengths of stay are 2 to 3 days, and one‐third of hospitalizations are 1 day or less. 17 Thus, even 10%20% reductions in LOS trims hours, not days, from typical stays. Practical barriers (eg, reluctance to discharge in the middle of the night, or family preferences and work schedules) and undesired outcomes (eg, increased hospital re‐visits) are additional pitfalls encountered by relying on throughput enhancement alone.
Managing scheduled admissions through smoothing is an alternative strategy to reduce variability and high occupancy. 6, 12, 1820 The concept is to proactively control the entry of patients, when possible, to achieve more even levels of occupancy, instead of the peaks and troughs commonly encountered. Nonetheless, it is not a widely used approach. 18, 20, 21 We hypothesized that children's hospitals had substantial unused capacity that could be used to smooth occupancy, which would reduce weekday crowding. While it is obvious that smoothing will reduce peaks to average levels (and also raise troughs), we sought to quantify just how large this difference wasand thereby quantify the potential of smoothing to reduce inpatient crowding (or, conversely, expose more patients to high levels of occupancy). Is there enough variation to justify smoothing, and, if a hospital does smooth, what is the expected result? If the number of patients removed from exposure to high occupancy is not substantial, other means to address inpatient crowding might be of more value. Our aims were to quantify the difference in weekday versus weekend occupancy, report on mathematical feasibility of such an approach, and determine the difference in number of patients exposed to various levels of high occupancy.
Methods
Data Source
This retrospective study was conducted with resource‐utilization data from 39 freestanding, tertiary‐care children's hospitals in the Pediatric Health Information System (PHIS). Participating hospitals are located in noncompeting markets of 23 states, plus the District of Columbia, and affiliated with the Child Health Corporation of America (CHCA, Shawnee Mission, KS). They account for 80% of freestanding, and 20% of all general, tertiary‐care children's hospitals. Data quality and reliability are assured through joint ongoing, systematic monitoring. The Children's Hospital of Philadelphia Committees for the Protection of Human Subjects approved the protocol with a waiver of informed consent.
Patients
Patients admitted January 1December 31, 2007 were eligible for inclusion. Due to variation in the presence of birthing, neonatal intensive care, and behavioral health units across hospitals, these beds and associated patients were excluded. Inpatients enter hospitals either as scheduled (often referred to as elective) or unscheduled (emergent or urgent) admissions. Because PHIS does not include these data, KID was used to standardize the PHIS data for proportion of scheduled admissions. 22 (KID is a healthcare database of 23 million pediatric inpatient discharges developed through federalstateindustry partnership, and sponsored by the Agency for Healthcare Research and Quality [AHRQ].) Each encounter in KID includes a principal International Classification of Diseases, 9th revision (ICD‐9) discharge diagnosis code, and is designated by the hospital as elective (ranging from chemotherapy to tonsillectomy) or not elective. Because admissions, rather than diagnoses, are scheduled, a proportion of patients with each primary diagnosis in KID are scheduled (eg, 28% of patients with a primary diagnosis of esophageal reflux). Proportions in KID were matched to principal diagnoses in PHIS.
Definitions
The census was the number of patients registered as inpatients (including those physically in the emergency department [ED] from time of ED arrival)whether observation or inpatient statusat midnight, the conclusion of the day. Hospital capacity was set using CHCA data (and confirmed by each hospital's administrative personnel) as the number of licensed in‐service beds available for patients in 2007; we assumed beds were staffed and capacity fixed for the year. Occupancy was calculated by dividing census by capacity. Maximum occupancy in a week referred to the highest occupancy level achieved in a seven‐day period (MondaySunday). We analyzed a set of thresholds for high‐occupancy (85%, 90%, 95%, and 100%), because there is no consistent definition for when hospitals are at high occupancy or when crowding occurs, though crowding has been described as starting at 85% occupancy. 2325
Analysis
The hospital was the unit of analysis. We report hospital characteristics, including capacity, number of discharges, and census region, and annual standardized length of stay ratio (SLOSR) as observed‐to‐expected LOS.
Smoothing Technique
A retrospective smoothing algorithm set each hospital's daily occupancy during a week to that hospital's mean occupancy for the week; effectively spreading the week's volume of patients evenly across the days of the week. While inter‐week and inter‐month smoothing were considered, intra‐week smoothing was deemed more practical for the largest number of patients, as it would not mean delaying care by more than one week. In the case of a planned treatment course (eg, chemotherapy), only intra‐week smoothing would maintain the necessary scheduled intervals of treatment.
Mathematical Feasibility
To approximate the number of patient admissions that would require different scheduling during a particular week to achieve smoothed weekly occupancy, we determined the total number of patient‐days in the week that required different scheduling and divided by the average LOS for the week. We then divided the number of admissions‐to‐move by total weekly admissions to compute the percentage at each hospital across 52 weeks of the year.
Measuring the Impact of Smoothing
We focused on the frequency and severity of high occupancy and the number of patients exposed to it. This framework led to 4 measures that assess the opportunity and effect of smoothing:
Difference in hospital weekdayweekend occupancy: Equal to 12‐month median of difference between mean weekday occupancy and mean weekend occupancy for each hospital‐week.
Difference in hospital maximummean occupancy: Equal to median of difference between maximum one‐day occupancy and weekly mean (smoothed) occupancy for each hospital‐week. A regression line was derived from the data for the 39 hospitals to report expected reduction in peak occupancy based on the magnitude of the difference between weekday and weekend occupancy.
Difference in number of hospitals exposed to above‐threshold occupancy: Equal to difference, pre‐ and post‐smoothing, in number of hospitals facing high‐occupancy conditions on an average of at least one weekday midnight per week during the year at different occupancy thresholds.
Difference in number of patients exposed to above‐threshold occupancy: Equal to difference, pre‐ and post‐smoothing, in number of patients exposed to hospital midnight occupancy at the thresholds. We utilized patient‐days for the calculation to avoid double‐counting, and divided this by average LOS, in order to determine the number of patients who would no longer be exposed to over‐threshold occupancy after smoothing, while also adjusting for patients newly exposed to over‐threshold occupancy levels.
All analyses were performed separately for each hospital for the entire year and then for winter (DecemberMarch), the period during which most crowding occurred. Analyses were performed using SAS (version 9.2, SAS Institute, Inc, Cary, NC); P values <0.05 were considered statistically significant.
Results
The characteristics of the 39 hospitals are provided in Table 1. Based on standardization with KID, 23.6% of PHIS admissions were scheduled (range: 18.1%35.8%) or a median of 81.5 scheduled admissions per week per hospital; 26.6% of weekday admissions were scheduled versus 16.1% for weekends. Overall, 12.4% of scheduled admissions entered on weekends. For all patients, median LOS was three days (interquartile range [IQR]: twofive days), but median LOS for scheduled admissions was two days (IQR: onefour days). The median LOS and IQR were the same by day of admission for all days of the week. Most hospitals had an overall SLOSR close to one (median: 0.9, IQR: 0.91.1). Overall, hospital mean midnight occupancy ranged from 70.9% to 108.1% on weekdays and 65.7% to 94.9% on weekends. Uniformly, weekday occupancy exceeded weekend occupancy, with a median difference of 8.2% points (IQR: 7.2%9.5% points). There was a wide range of median hospital weekdayweekend occupancy differences across hospitals (Figure 1). The overall difference was less in winter (median difference: 7.7% points; IQR: 6.3%8.8% points) than in summer (median difference: 8.6% points; IQR: 7.4%9.8% points (Wilcoxon Sign Rank test, P < 0.001). Thirty‐five hospitals (89.7%) exceeded the 85% occupancy threshold and 29 (74.4%) exceeded the 95% occupancy threshold on at least 20% of weekdays (Table 2). Across all the hospitals, the median difference in weekly maximum and weekly mean occupancy was 6.6% points (IQR: 6.2%7.4% points) (Figure 2).
| Characteristics | No. (%) |
|---|---|
| |
| Licensed in‐service beds | n = 39 hospitals |
| <200 beds | 6 (15.4) |
| 200249 beds | 10 (25.6) |
| 250300 beds | 14 (35.9) |
| >300 beds | 9 (23.1) |
| No. of discharges | |
| <10,000 | 5 (12.8) |
| 10,00013,999 | 14 (35.9) |
| 14,00017,999 | 11 (28.2) |
| >18,000 | 9 (23.1) |
| Census region | |
| West | 9 (23.1) |
| Midwest | 11 (28.2) |
| Northeast | 6 (15.4) |
| South | 13 (33.3) |
| Admissions | n = 590,352 admissions |
| Medical scheduled admissions* | 79,683 |
| Surgical scheduled admissions* | 59,640 |
| Total scheduled admissions* (% of all admissions) | 139,323 (23.6) |
| Weekend medical scheduled admissions* (% of all medical scheduled admissions) | 13,546 (17.0) |
| Weekend surgical scheduled admissions* (% of all surgical scheduled admissions) | 3,757 (6.3) |
| Weekend total scheduled admissions* (% of total scheduled admissions) | 17,276 (12.4) |
| Entire Year | >85% | Occupancy Threshold | >95% | >100% |
|---|---|---|---|---|
| >90% | ||||
| ||||
| No. of hospitals (n = 39) with mean weekday occupancy above threshold | ||||
| Before smoothing (current state) | 33 | 25 | 14 | 6 |
| After smoothing | 32 | 22 | 10 | 1 |
| No. of hospitals (n = 39) above threshold 20% of weekdays | ||||
| Before smoothing (current state) | 35 | 34 | 29 | 14 |
| After smoothing | 35 | 32 | 21 | 9 |
| Median (IQR) no. of patient‐days per hospital not exposed to occupancy above threshold by smoothing | 3,071 | 281 | 3236 | 3281 |
| (5,552, 919) | (5,288, 3,103) | (0, 7,083) | (962, 8,517) | |
| Median (IQR) no. of patients per hospital not exposed to occupancy above threshold by smoothing | 596 | 50 | 630 | 804 |
| (1,190, 226) | (916, 752) | (0, 1,492) | (231, 2,195) | |
Smoothing reduced the number of hospitals at each occupancy threshold, except 85% (Table 2). As a linear relationship, the reduction in weekday peak occupancy (y) based on a hospital's median difference in weekly maximum and weekly mean occupancy (x) was y = 2.69 + 0.48x. Thus, a hospital with a 10% point difference between weekday and weekend occupancy could reduce weekday peak by 7.5% points.
Smoothing increased the number of patients exposed to the lower thresholds (85% and 90%), but decreased the number of patients exposed to >95% occupancy (Table 2). For example, smoothing at the 95% threshold resulted in 630 fewer patients per hospital exposed to that threshold. If all 39 hospitals had within‐week smoothing, a net of 39,607 patients would have been protected from exposure to >95% occupancy and a net of 50,079 patients from 100% occupancy.
To demonstrate the varied effects of smoothing, Table 3 and Figure 3 present representative categories of response to smoothing depending on pre‐smoothing patterns. While not all hospitals decreased occupancy to below thresholds after smoothing (Types B and D), the overall occupancy was reduced and fewer patients were exposed to extreme levels of high occupancy (eg, >100%).
| Category | Before Smoothing Hospital Description | After Smoothing Hospital Description | No. of Hospitals at 85% Threshold (n = 39) | No. of Hospitals at 95% Threshold (n = 39) |
|---|---|---|---|---|
| ||||
| Type A | Weekdays above threshold | All days below threshold, resulting in net decrease in patients exposed to occupancies above threshold | 3 | 1 |
| Weekends below threshold | ||||
| Type B | Weekdays above threshold | All days above threshold, resulting in net increase in patients exposed to occupancies above threshold | 12 | 18 |
| Weekends below threshold | ||||
| Type C | All days of week below threshold | All days of week below threshold | 6 | 19 |
| Type D | All days of week above threshold | All days of week above threshold, resulting in net decrease in patients exposed to extreme high occupancy | 18 | 1 |
To achieve within‐week smoothing, a median of 7.4 patient‐admissions per week (range: 2.314.4) would have to be scheduled on a different day of the week. This equates to a median of 2.6% (IQR: 2.25%, 2.99%; range: 0.02%9.2%) of all admissionsor 9% of a typical hospital‐week's scheduled admissions.
Discussion
This analysis of 39 children's hospitals found high levels of occupancy and weekend occupancy lower than weekday occupancy (median difference: 8.2% points). Only 12.4% of scheduled admissions entered on weekends. Thus, weekend capacity is available to offset high weekday occupancy. Hospitals at the higher end of the occupancy thresholds (95%, 100%) would reduce the number of days operating at very high occupancy and the number of patients exposed to such levels by smoothing. This change is mathematically feasible, as a median of 7.4 patients would have to be proactively scheduled differently each week, just under one‐tenth of scheduled admissions. Since LOS by day of admission was the same (median: two days), the opportunity to affect occupancy by shifting patients should be relatively similar for all days of the week. In addition, these admissions were short, conferring greater flexibility. Implementing smoothing over the course of the week does not necessarily require admitting patients on weekends. For example, Monday admissions with an anticipated three‐day LOS could enter on Friday with anticipated discharge on Monday to alleviate midweek crowding and take advantage of unoccupied weekend beds. 26
At the highest levels of occupancy, smoothing reduces the frequency of reaching these maximum levels, but can have the effect of actually exposing more patient‐days to a higher occupancy. For example, for nine hospitals in our analysis with >20% of days over 100%, smoothing decreased days over 100%, but exposed weekend patients to higher levels of occupancy (Figure 3). Since most admissions are short and most scheduled admissions currently occur on weekdays, the number of individual patients (not patient‐days) newly exposed to such high occupancy may not increase much after smoothing at these facilities. Regardless, hospitals with such a pattern may not be able to rely solely on smoothing to avoid weekday crowding, and, if they are operating efficiently in terms of SLOSR, might be justified in building more capacity.
Consistent with our findings, the Institute for Healthcare Improvement, the Institute for Healthcare Optimization, and the American Hospital Association Quality Center stress that addressing artificial variability of scheduled admissions is a critical first step to improving patient flow and quality of care while reducing costs. 18, 21, 27 Our study suggests that small numbers of patients need to be proactively scheduled differently to decrease midweek peak occupancy, so only a small proportion of families would need to find this desirable to make it attractive for hospitals and patients. This type of proactive smoothing decreases peak occupancy on weekdays, reducing the safety risks associated with high occupancy, improving acute access for emergent patients, shortening wait‐times and loss of scheduled patients to another facility, and increasing procedure volume (3%74% in one study). 28 Smoothing may also increase quality and safety on weekends, as emergent patients admitted on weekends experience more delays in necessary treatment and have worse outcomes. 2932 In addition, increasing scheduled admissions to span weekends may appeal to some families wishing to avoid absence from work to be with their hospitalized child, to parents concerned about school performanceand may also appeal to staff members seeking flexible schedules. Increasing weekend hospital capacity is safe, feasible, and economical, even when considering the increased wages for weekend work. 33, 34 Finally, smoothing over the whole week allows fixed costs (eg, surgical suites, imaging equipment) to be allocated over 7 days rather than 5, and allows for better matching of revenue to the fixed expenses.
Rather than a prescriptive approach, our work suggests hospitals need to identify only a small number of patients to proactively shift, providing them opportunities to adapt the approach to local circumstances. The particular patients to move around may also depend on the costs and benefits of services (eg, radiologic, laboratory, operative) and the hospital's existing patterns of staffing. A number of hospitals that have engaged in similar work have achieved sustainable results, such as Seattle Children's Hospital, Boston Medical Center, St. John's Regional Health Center, and New York University Langone Medical Center. 19, 26, 3537 In these cases, proactive smoothing took advantage of unused capacity and decreased crowding on days that had been traditionally very full. Hospitals that rarely or never have high‐occupancy days, and that do not expect growth in volume, may not need to employ smoothing, whereas others that have crowding issues primarily in the winter may wish to implement smoothing techniques seasonally.
Aside from attempting to reduce high‐occupancy through modification of admission patterns, other proactive approaches include optimizing staffing and processes around care, improving efficiency of care, and building additional beds. 16, 25, 38, 39 However, the expense of construction and the scarcity of capital often preclude this last option. Among children's hospitals, with SLOSR close to one, implementing strategies to reduce the LOS during periods of high occupancy may not result in meaningful reductions in LOS, as such approaches would only decrease the typical child's hospitalization by hours, not days. In addition to proactive strategies, hospitals also rely on reactive approaches, such as ED boarding, placing patients in hallways on units, diverting ambulances or transfers, or canceling scheduled admissions at the last moment, to decrease crowding. 16, 39, 40
This study has several limitations. First, use of administrative data precluded modeling all responses. For example, some hospitals may be better able to accommodate fluctuations in census or high occupancy without compromising quality or access. Second, we only considered intra‐week smoothing, but hospitals may benefit from smoothing over longer periods of time, especially since children's hospitals are busier in winter months, but incoming scheduled volume is often not reduced. 11 Hospitals with large occupancy variations across months may want to consider broadening the time horizon for smoothing, and weigh the costs and benefits over that period of time, including parental and clinician concerns and preferences for not delaying treatment. At the individual hospital level, discrete‐event simulation would likely be useful to consider the trade‐offs of smoothing to different levels and over different periods of time. Third, we assumed a fixed number of beds for the year, an approach that may not accurately reflect actual available beds on specific days. This limitation was minimized by counting all beds for each hospital as available for all the days of the year, so that hospitals with a high census when all available beds are included would have an even higher percent occupancy if some of those beds were not actually open. In a related way, then, we also do not consider how staffing may need to be altered or augmented to care for additional patients on certain days. Fourth, midnight census, the only universally available measure, was used to determine occupancy rather than peak census. Midnight census provides a standard snapshot, but is lower than mid‐day peak census. 41 In order to account for these limitations, we considered several different thresholds of high occupancy. Fifth, we smoothed at the hospital level, but differential effects may exist at the unit level. Sixth, to determine proportion of scheduled admissions, we used HCUP KID proportions on PHIS admissions. Overall, this approach likely overestimated scheduled medical admissions on weekends, thus biasing our result towards the null hypothesis. Finally, only freestanding children's hospitals were included in this study. While this may limit generalizability, the general concept of smoothing occupancy should apply in any setting with substantial and consistent variation.
In summary, our study revealed that children's hospitals often face high midweek occupancy, but also have substantial unused weekend capacity. Hospitals facing challenges with high weekday occupancy could proactively use a smoothing approach to decrease the frequency and severity of high occupancy. Further qualitative evaluation is also warranted around child, family, and staff preferences concerning scheduled admissions, school, and work.
High levels of hospital occupancy are associated with compromises to quality of care and access (often referred to as crowding), 18 while low occupancy may be inefficient and also impact quality. 9, 10 Despite this, hospitals typically have uneven occupancy. Although some demand for services is driven by factors beyond the control of a hospital (eg, seasonal variation in viral illness), approximately 15%30% of admissions to children's hospitals are scheduled from days to months in advance, with usual arrivals on weekdays. 1114 For example, of the 3.4 million elective admissions in the 2006 Healthcare Cost and Utilization Project Kids Inpatient Database (HCUP KID), only 13% were admitted on weekends. 14 Combined with short length of stay (LOS) for such patients, this leads to higher midweek and lower weekend occupancy. 12
Hospitals respond to crowding in a number of ways, but often focus on reducing LOS to make room for new patients. 11, 15, 16 For hospitals that are relatively efficient in terms of LOS, efforts to reduce it may not increase functional capacity adequately. In children's hospitals, median lengths of stay are 2 to 3 days, and one‐third of hospitalizations are 1 day or less. 17 Thus, even 10%20% reductions in LOS trims hours, not days, from typical stays. Practical barriers (eg, reluctance to discharge in the middle of the night, or family preferences and work schedules) and undesired outcomes (eg, increased hospital re‐visits) are additional pitfalls encountered by relying on throughput enhancement alone.
Managing scheduled admissions through smoothing is an alternative strategy to reduce variability and high occupancy. 6, 12, 1820 The concept is to proactively control the entry of patients, when possible, to achieve more even levels of occupancy, instead of the peaks and troughs commonly encountered. Nonetheless, it is not a widely used approach. 18, 20, 21 We hypothesized that children's hospitals had substantial unused capacity that could be used to smooth occupancy, which would reduce weekday crowding. While it is obvious that smoothing will reduce peaks to average levels (and also raise troughs), we sought to quantify just how large this difference wasand thereby quantify the potential of smoothing to reduce inpatient crowding (or, conversely, expose more patients to high levels of occupancy). Is there enough variation to justify smoothing, and, if a hospital does smooth, what is the expected result? If the number of patients removed from exposure to high occupancy is not substantial, other means to address inpatient crowding might be of more value. Our aims were to quantify the difference in weekday versus weekend occupancy, report on mathematical feasibility of such an approach, and determine the difference in number of patients exposed to various levels of high occupancy.
Methods
Data Source
This retrospective study was conducted with resource‐utilization data from 39 freestanding, tertiary‐care children's hospitals in the Pediatric Health Information System (PHIS). Participating hospitals are located in noncompeting markets of 23 states, plus the District of Columbia, and affiliated with the Child Health Corporation of America (CHCA, Shawnee Mission, KS). They account for 80% of freestanding, and 20% of all general, tertiary‐care children's hospitals. Data quality and reliability are assured through joint ongoing, systematic monitoring. The Children's Hospital of Philadelphia Committees for the Protection of Human Subjects approved the protocol with a waiver of informed consent.
Patients
Patients admitted January 1December 31, 2007 were eligible for inclusion. Due to variation in the presence of birthing, neonatal intensive care, and behavioral health units across hospitals, these beds and associated patients were excluded. Inpatients enter hospitals either as scheduled (often referred to as elective) or unscheduled (emergent or urgent) admissions. Because PHIS does not include these data, KID was used to standardize the PHIS data for proportion of scheduled admissions. 22 (KID is a healthcare database of 23 million pediatric inpatient discharges developed through federalstateindustry partnership, and sponsored by the Agency for Healthcare Research and Quality [AHRQ].) Each encounter in KID includes a principal International Classification of Diseases, 9th revision (ICD‐9) discharge diagnosis code, and is designated by the hospital as elective (ranging from chemotherapy to tonsillectomy) or not elective. Because admissions, rather than diagnoses, are scheduled, a proportion of patients with each primary diagnosis in KID are scheduled (eg, 28% of patients with a primary diagnosis of esophageal reflux). Proportions in KID were matched to principal diagnoses in PHIS.
Definitions
The census was the number of patients registered as inpatients (including those physically in the emergency department [ED] from time of ED arrival)whether observation or inpatient statusat midnight, the conclusion of the day. Hospital capacity was set using CHCA data (and confirmed by each hospital's administrative personnel) as the number of licensed in‐service beds available for patients in 2007; we assumed beds were staffed and capacity fixed for the year. Occupancy was calculated by dividing census by capacity. Maximum occupancy in a week referred to the highest occupancy level achieved in a seven‐day period (MondaySunday). We analyzed a set of thresholds for high‐occupancy (85%, 90%, 95%, and 100%), because there is no consistent definition for when hospitals are at high occupancy or when crowding occurs, though crowding has been described as starting at 85% occupancy. 2325
Analysis
The hospital was the unit of analysis. We report hospital characteristics, including capacity, number of discharges, and census region, and annual standardized length of stay ratio (SLOSR) as observed‐to‐expected LOS.
Smoothing Technique
A retrospective smoothing algorithm set each hospital's daily occupancy during a week to that hospital's mean occupancy for the week; effectively spreading the week's volume of patients evenly across the days of the week. While inter‐week and inter‐month smoothing were considered, intra‐week smoothing was deemed more practical for the largest number of patients, as it would not mean delaying care by more than one week. In the case of a planned treatment course (eg, chemotherapy), only intra‐week smoothing would maintain the necessary scheduled intervals of treatment.
Mathematical Feasibility
To approximate the number of patient admissions that would require different scheduling during a particular week to achieve smoothed weekly occupancy, we determined the total number of patient‐days in the week that required different scheduling and divided by the average LOS for the week. We then divided the number of admissions‐to‐move by total weekly admissions to compute the percentage at each hospital across 52 weeks of the year.
Measuring the Impact of Smoothing
We focused on the frequency and severity of high occupancy and the number of patients exposed to it. This framework led to 4 measures that assess the opportunity and effect of smoothing:
Difference in hospital weekdayweekend occupancy: Equal to 12‐month median of difference between mean weekday occupancy and mean weekend occupancy for each hospital‐week.
Difference in hospital maximummean occupancy: Equal to median of difference between maximum one‐day occupancy and weekly mean (smoothed) occupancy for each hospital‐week. A regression line was derived from the data for the 39 hospitals to report expected reduction in peak occupancy based on the magnitude of the difference between weekday and weekend occupancy.
Difference in number of hospitals exposed to above‐threshold occupancy: Equal to difference, pre‐ and post‐smoothing, in number of hospitals facing high‐occupancy conditions on an average of at least one weekday midnight per week during the year at different occupancy thresholds.
Difference in number of patients exposed to above‐threshold occupancy: Equal to difference, pre‐ and post‐smoothing, in number of patients exposed to hospital midnight occupancy at the thresholds. We utilized patient‐days for the calculation to avoid double‐counting, and divided this by average LOS, in order to determine the number of patients who would no longer be exposed to over‐threshold occupancy after smoothing, while also adjusting for patients newly exposed to over‐threshold occupancy levels.
All analyses were performed separately for each hospital for the entire year and then for winter (DecemberMarch), the period during which most crowding occurred. Analyses were performed using SAS (version 9.2, SAS Institute, Inc, Cary, NC); P values <0.05 were considered statistically significant.
Results
The characteristics of the 39 hospitals are provided in Table 1. Based on standardization with KID, 23.6% of PHIS admissions were scheduled (range: 18.1%35.8%) or a median of 81.5 scheduled admissions per week per hospital; 26.6% of weekday admissions were scheduled versus 16.1% for weekends. Overall, 12.4% of scheduled admissions entered on weekends. For all patients, median LOS was three days (interquartile range [IQR]: twofive days), but median LOS for scheduled admissions was two days (IQR: onefour days). The median LOS and IQR were the same by day of admission for all days of the week. Most hospitals had an overall SLOSR close to one (median: 0.9, IQR: 0.91.1). Overall, hospital mean midnight occupancy ranged from 70.9% to 108.1% on weekdays and 65.7% to 94.9% on weekends. Uniformly, weekday occupancy exceeded weekend occupancy, with a median difference of 8.2% points (IQR: 7.2%9.5% points). There was a wide range of median hospital weekdayweekend occupancy differences across hospitals (Figure 1). The overall difference was less in winter (median difference: 7.7% points; IQR: 6.3%8.8% points) than in summer (median difference: 8.6% points; IQR: 7.4%9.8% points (Wilcoxon Sign Rank test, P < 0.001). Thirty‐five hospitals (89.7%) exceeded the 85% occupancy threshold and 29 (74.4%) exceeded the 95% occupancy threshold on at least 20% of weekdays (Table 2). Across all the hospitals, the median difference in weekly maximum and weekly mean occupancy was 6.6% points (IQR: 6.2%7.4% points) (Figure 2).
| Characteristics | No. (%) |
|---|---|
| |
| Licensed in‐service beds | n = 39 hospitals |
| <200 beds | 6 (15.4) |
| 200249 beds | 10 (25.6) |
| 250300 beds | 14 (35.9) |
| >300 beds | 9 (23.1) |
| No. of discharges | |
| <10,000 | 5 (12.8) |
| 10,00013,999 | 14 (35.9) |
| 14,00017,999 | 11 (28.2) |
| >18,000 | 9 (23.1) |
| Census region | |
| West | 9 (23.1) |
| Midwest | 11 (28.2) |
| Northeast | 6 (15.4) |
| South | 13 (33.3) |
| Admissions | n = 590,352 admissions |
| Medical scheduled admissions* | 79,683 |
| Surgical scheduled admissions* | 59,640 |
| Total scheduled admissions* (% of all admissions) | 139,323 (23.6) |
| Weekend medical scheduled admissions* (% of all medical scheduled admissions) | 13,546 (17.0) |
| Weekend surgical scheduled admissions* (% of all surgical scheduled admissions) | 3,757 (6.3) |
| Weekend total scheduled admissions* (% of total scheduled admissions) | 17,276 (12.4) |
| Entire Year | >85% | Occupancy Threshold | >95% | >100% |
|---|---|---|---|---|
| >90% | ||||
| ||||
| No. of hospitals (n = 39) with mean weekday occupancy above threshold | ||||
| Before smoothing (current state) | 33 | 25 | 14 | 6 |
| After smoothing | 32 | 22 | 10 | 1 |
| No. of hospitals (n = 39) above threshold 20% of weekdays | ||||
| Before smoothing (current state) | 35 | 34 | 29 | 14 |
| After smoothing | 35 | 32 | 21 | 9 |
| Median (IQR) no. of patient‐days per hospital not exposed to occupancy above threshold by smoothing | 3,071 | 281 | 3236 | 3281 |
| (5,552, 919) | (5,288, 3,103) | (0, 7,083) | (962, 8,517) | |
| Median (IQR) no. of patients per hospital not exposed to occupancy above threshold by smoothing | 596 | 50 | 630 | 804 |
| (1,190, 226) | (916, 752) | (0, 1,492) | (231, 2,195) | |
Smoothing reduced the number of hospitals at each occupancy threshold, except 85% (Table 2). As a linear relationship, the reduction in weekday peak occupancy (y) based on a hospital's median difference in weekly maximum and weekly mean occupancy (x) was y = 2.69 + 0.48x. Thus, a hospital with a 10% point difference between weekday and weekend occupancy could reduce weekday peak by 7.5% points.
Smoothing increased the number of patients exposed to the lower thresholds (85% and 90%), but decreased the number of patients exposed to >95% occupancy (Table 2). For example, smoothing at the 95% threshold resulted in 630 fewer patients per hospital exposed to that threshold. If all 39 hospitals had within‐week smoothing, a net of 39,607 patients would have been protected from exposure to >95% occupancy and a net of 50,079 patients from 100% occupancy.
To demonstrate the varied effects of smoothing, Table 3 and Figure 3 present representative categories of response to smoothing depending on pre‐smoothing patterns. While not all hospitals decreased occupancy to below thresholds after smoothing (Types B and D), the overall occupancy was reduced and fewer patients were exposed to extreme levels of high occupancy (eg, >100%).
| Category | Before Smoothing Hospital Description | After Smoothing Hospital Description | No. of Hospitals at 85% Threshold (n = 39) | No. of Hospitals at 95% Threshold (n = 39) |
|---|---|---|---|---|
| ||||
| Type A | Weekdays above threshold | All days below threshold, resulting in net decrease in patients exposed to occupancies above threshold | 3 | 1 |
| Weekends below threshold | ||||
| Type B | Weekdays above threshold | All days above threshold, resulting in net increase in patients exposed to occupancies above threshold | 12 | 18 |
| Weekends below threshold | ||||
| Type C | All days of week below threshold | All days of week below threshold | 6 | 19 |
| Type D | All days of week above threshold | All days of week above threshold, resulting in net decrease in patients exposed to extreme high occupancy | 18 | 1 |
To achieve within‐week smoothing, a median of 7.4 patient‐admissions per week (range: 2.314.4) would have to be scheduled on a different day of the week. This equates to a median of 2.6% (IQR: 2.25%, 2.99%; range: 0.02%9.2%) of all admissionsor 9% of a typical hospital‐week's scheduled admissions.
Discussion
This analysis of 39 children's hospitals found high levels of occupancy and weekend occupancy lower than weekday occupancy (median difference: 8.2% points). Only 12.4% of scheduled admissions entered on weekends. Thus, weekend capacity is available to offset high weekday occupancy. Hospitals at the higher end of the occupancy thresholds (95%, 100%) would reduce the number of days operating at very high occupancy and the number of patients exposed to such levels by smoothing. This change is mathematically feasible, as a median of 7.4 patients would have to be proactively scheduled differently each week, just under one‐tenth of scheduled admissions. Since LOS by day of admission was the same (median: two days), the opportunity to affect occupancy by shifting patients should be relatively similar for all days of the week. In addition, these admissions were short, conferring greater flexibility. Implementing smoothing over the course of the week does not necessarily require admitting patients on weekends. For example, Monday admissions with an anticipated three‐day LOS could enter on Friday with anticipated discharge on Monday to alleviate midweek crowding and take advantage of unoccupied weekend beds. 26
At the highest levels of occupancy, smoothing reduces the frequency of reaching these maximum levels, but can have the effect of actually exposing more patient‐days to a higher occupancy. For example, for nine hospitals in our analysis with >20% of days over 100%, smoothing decreased days over 100%, but exposed weekend patients to higher levels of occupancy (Figure 3). Since most admissions are short and most scheduled admissions currently occur on weekdays, the number of individual patients (not patient‐days) newly exposed to such high occupancy may not increase much after smoothing at these facilities. Regardless, hospitals with such a pattern may not be able to rely solely on smoothing to avoid weekday crowding, and, if they are operating efficiently in terms of SLOSR, might be justified in building more capacity.
Consistent with our findings, the Institute for Healthcare Improvement, the Institute for Healthcare Optimization, and the American Hospital Association Quality Center stress that addressing artificial variability of scheduled admissions is a critical first step to improving patient flow and quality of care while reducing costs. 18, 21, 27 Our study suggests that small numbers of patients need to be proactively scheduled differently to decrease midweek peak occupancy, so only a small proportion of families would need to find this desirable to make it attractive for hospitals and patients. This type of proactive smoothing decreases peak occupancy on weekdays, reducing the safety risks associated with high occupancy, improving acute access for emergent patients, shortening wait‐times and loss of scheduled patients to another facility, and increasing procedure volume (3%74% in one study). 28 Smoothing may also increase quality and safety on weekends, as emergent patients admitted on weekends experience more delays in necessary treatment and have worse outcomes. 2932 In addition, increasing scheduled admissions to span weekends may appeal to some families wishing to avoid absence from work to be with their hospitalized child, to parents concerned about school performanceand may also appeal to staff members seeking flexible schedules. Increasing weekend hospital capacity is safe, feasible, and economical, even when considering the increased wages for weekend work. 33, 34 Finally, smoothing over the whole week allows fixed costs (eg, surgical suites, imaging equipment) to be allocated over 7 days rather than 5, and allows for better matching of revenue to the fixed expenses.
Rather than a prescriptive approach, our work suggests hospitals need to identify only a small number of patients to proactively shift, providing them opportunities to adapt the approach to local circumstances. The particular patients to move around may also depend on the costs and benefits of services (eg, radiologic, laboratory, operative) and the hospital's existing patterns of staffing. A number of hospitals that have engaged in similar work have achieved sustainable results, such as Seattle Children's Hospital, Boston Medical Center, St. John's Regional Health Center, and New York University Langone Medical Center. 19, 26, 3537 In these cases, proactive smoothing took advantage of unused capacity and decreased crowding on days that had been traditionally very full. Hospitals that rarely or never have high‐occupancy days, and that do not expect growth in volume, may not need to employ smoothing, whereas others that have crowding issues primarily in the winter may wish to implement smoothing techniques seasonally.
Aside from attempting to reduce high‐occupancy through modification of admission patterns, other proactive approaches include optimizing staffing and processes around care, improving efficiency of care, and building additional beds. 16, 25, 38, 39 However, the expense of construction and the scarcity of capital often preclude this last option. Among children's hospitals, with SLOSR close to one, implementing strategies to reduce the LOS during periods of high occupancy may not result in meaningful reductions in LOS, as such approaches would only decrease the typical child's hospitalization by hours, not days. In addition to proactive strategies, hospitals also rely on reactive approaches, such as ED boarding, placing patients in hallways on units, diverting ambulances or transfers, or canceling scheduled admissions at the last moment, to decrease crowding. 16, 39, 40
This study has several limitations. First, use of administrative data precluded modeling all responses. For example, some hospitals may be better able to accommodate fluctuations in census or high occupancy without compromising quality or access. Second, we only considered intra‐week smoothing, but hospitals may benefit from smoothing over longer periods of time, especially since children's hospitals are busier in winter months, but incoming scheduled volume is often not reduced. 11 Hospitals with large occupancy variations across months may want to consider broadening the time horizon for smoothing, and weigh the costs and benefits over that period of time, including parental and clinician concerns and preferences for not delaying treatment. At the individual hospital level, discrete‐event simulation would likely be useful to consider the trade‐offs of smoothing to different levels and over different periods of time. Third, we assumed a fixed number of beds for the year, an approach that may not accurately reflect actual available beds on specific days. This limitation was minimized by counting all beds for each hospital as available for all the days of the year, so that hospitals with a high census when all available beds are included would have an even higher percent occupancy if some of those beds were not actually open. In a related way, then, we also do not consider how staffing may need to be altered or augmented to care for additional patients on certain days. Fourth, midnight census, the only universally available measure, was used to determine occupancy rather than peak census. Midnight census provides a standard snapshot, but is lower than mid‐day peak census. 41 In order to account for these limitations, we considered several different thresholds of high occupancy. Fifth, we smoothed at the hospital level, but differential effects may exist at the unit level. Sixth, to determine proportion of scheduled admissions, we used HCUP KID proportions on PHIS admissions. Overall, this approach likely overestimated scheduled medical admissions on weekends, thus biasing our result towards the null hypothesis. Finally, only freestanding children's hospitals were included in this study. While this may limit generalizability, the general concept of smoothing occupancy should apply in any setting with substantial and consistent variation.
In summary, our study revealed that children's hospitals often face high midweek occupancy, but also have substantial unused weekend capacity. Hospitals facing challenges with high weekday occupancy could proactively use a smoothing approach to decrease the frequency and severity of high occupancy. Further qualitative evaluation is also warranted around child, family, and staff preferences concerning scheduled admissions, school, and work.
- , , , . A comparison of in‐hospital mortality risk conferred by high hospital occupancy, differences in nurse staffing levels, weekend admission, and seasonal influenza. Medical Care. 2010;48(3):224–232.
- , , , et al. Hospital workload and adverse events. Med Care. 2007;45(5):448–455.
- , , , , . Impact of admission‐day crowding on the length of stay of pediatric hospitalizations. Pediatrics. 2008;121(4):e718–e730.
- , , , . The effect of hospital bed occupancy on throughput in the pediatric emergency department. Ann Emerg Med. 2009;53(6):767–776.
- . The tipping point: the relationship between volume and patient harm. Am J Med Qual. 2008;23(5):336–341.
- , , , , , . Managing unnecessary variability in patient demand to reduce nursing stress and improve patient safety. Jt Comm J Qual Patient Saf. 2005;31(6):330–338.
- Hospital‐Based Emergency Care: At the Breaking Point. Washington, DC: Institute of Medicine Committee on the Future of Emergency Care in the United States Health System; 2006.
- , , , , . The effect of hospital occupancy on emergency department length of stay and patient disposition. Acad Emerg Med. 2003;10(2):127–133.
- . Interpreting the Volume‐Outcome Relationship in the Context of Health Care Quality: Workshop Summary. Washington, DC: National Academies Press; 2000.
- , , , , . Has recognition of the relationship between mortality rates and hospital volume for major cancer surgery in California made a difference? A follow‐up analysis of another decade. Ann Surg. 2009;250(3):472–483.
- , , , et al. Children's hospitals do not acutely respond to high occupancy. Pediatrics. 2010;125:974–981.
- , , , , , . Scheduled admissions and high occupancy at a children's hospital. J Hosp Med. 2011;6(2):81–87.
- , , . Characteristics of weekday and weekend hospital admissions. HCUP Statistical Brief. 2010;87.
- Agency for Healthcare Research and Quality. HCUP databases, Healthcare Cost and Utilization Project (HCUP); 2008. Available at: http://www.hcup‐us.ahrq.gov/kidoverview.jsp. Accessed July 15, 2009.
- , et al. Managing capacity to reduce emergency department overcrowding and ambulance diversions. J Qual Patient Saf. 2006;32(5):239–245.
- Institute for Healthcare Improvement. Flow initiatives; 2008. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed February 20, 2008.
- , , , , , . Trends in high‐turnover stays among children hospitalized in the United States, 1993–2003. Pediatrics. 2009;123(3):996–1002.
- Institute for Healthcare Improvement. Smoothing elective surgical admissions. Available at: http://www.ihi.org/IHI/Topics/Flow/PatientFlow/EmergingContent/SmoothingElectiveSurgicalAdmissions.htm. Accessed October 24, 2008.
- Boston hospital sees big impact from smoothing elective schedule. OR Manager. 2004;20:12.
- . Managing Variability in Patient Flow Is the Key to Improving Access to Care, Nursing Staffing, Quality of Care, and Reducing Its Cost. Paper presented at Institute of Medicine, Washington, DC; June 24, 2004.
- American Hospital Association Quality Center. Available at: http://www.ahaqualitycenter.org/ahaqualitycenter/. Accessed October 14, 2008.
- Healthcare Cost and Utilization Project (HCUP). Kids' Inpatient Database (KID); July 2008. Available at: http://www.hcup‐us.ahrq.gov/kidoverview.jsp. Accessed September 10, 2008.
- , , . Using a queuing model to help plan bed allocation in a department of geriatric medicine. Health Care Manag Sci. 2002;5(4):307–313.
- . How many hospital beds? Inquiry. 2002;39(4):400–412.
- . Institute for Healthcare Improvement. Patient flow comments. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed September 10, 2008.
- . Factory efficiency comes to the hospital. New York Times. July 9, 2010.
- Institute for Healthcare Improvement. Re‐engineering the operating room. Available at: http://www.ihi.org/IHI/Programs/ConferencesAndSeminars/ReengineeringtheOperatingRoomSept08.htm. Accessed November 8, 2008.
- , . Enhanced weekend service: an affordable means to increased hospital procedure volume. CMAJ. 2005;172(4):503–504.
- , . Mortality among patients admitted to hospitals on weekends as compared with weekdays. N Engl J Med. 2001;345:663–668.
- , , , , , . Weekend versus weekday admission and mortality from myocardial infarction. N Engl J Med. 2007;356:1099–1109.
- , . Waiting for urgent procedures on the weekend among emergently hospitalized patients. Am J Med. 2004;117:175–181.
- . Do hospitals provide lower quality care on weekends? Health Serv Res. 2007;42:1589–1612.
- . Hospital saves by working weekends. Mod Healthc. 1996;26:82–99.
- , , , , . Weekend and holiday exercise testing in patients with chest pain. J Gen Intern Med. 1999;14:10–14.
- . Boston Medical Center Case Study: Institute of Healthcare Optimization; 2006. Available at: http://www.ihoptimize.org/8f16e142‐eeaa‐4898–9e62–660218f19ffb/download.htm. Accessed October 3, 2010.
- , , , . The impact of IMPACT on St John's Regional Health Center. Mo Med. 2003;100:590–592.
- NYU Langone Medical Center Extends Access to Non‐Emergent Care as Part of Commitment to Patient‐Centered Care (June 23, 2010). Available at: http://communications.med.nyu.edu/news/2010/nyu‐langone‐medical‐center‐extends‐access‐non‐emergent‐care‐part‐commitment‐patient‐center. Accessed October 3, 2010.
- Carondelet St. Mary's Hospital. A pragmatic approach to improving patient efficiency throughput. Improvement Report 2005. Available at: http://www.ihi.org/IHI/Topics/Flow/PatientFlow/ImprovementStories/APragmaticApproachtoImprovingPatientEfficiencyThroughput.htm. Accessed October 3, 2010.
- AHA Solutions. Patient Flow Challenges Assessment 2009. Chicago, IL; 2009.
- , , , , , . A conceptual model of emergency department crowding. Ann Emerg Med. 2003;42(2):173–180.
- . Annual bed statistics give a misleading picture of hospital surge capacity. Ann Emerg Med. 2006;48(4):384–388.
- , , , . A comparison of in‐hospital mortality risk conferred by high hospital occupancy, differences in nurse staffing levels, weekend admission, and seasonal influenza. Medical Care. 2010;48(3):224–232.
- , , , et al. Hospital workload and adverse events. Med Care. 2007;45(5):448–455.
- , , , , . Impact of admission‐day crowding on the length of stay of pediatric hospitalizations. Pediatrics. 2008;121(4):e718–e730.
- , , , . The effect of hospital bed occupancy on throughput in the pediatric emergency department. Ann Emerg Med. 2009;53(6):767–776.
- . The tipping point: the relationship between volume and patient harm. Am J Med Qual. 2008;23(5):336–341.
- , , , , , . Managing unnecessary variability in patient demand to reduce nursing stress and improve patient safety. Jt Comm J Qual Patient Saf. 2005;31(6):330–338.
- Hospital‐Based Emergency Care: At the Breaking Point. Washington, DC: Institute of Medicine Committee on the Future of Emergency Care in the United States Health System; 2006.
- , , , , . The effect of hospital occupancy on emergency department length of stay and patient disposition. Acad Emerg Med. 2003;10(2):127–133.
- . Interpreting the Volume‐Outcome Relationship in the Context of Health Care Quality: Workshop Summary. Washington, DC: National Academies Press; 2000.
- , , , , . Has recognition of the relationship between mortality rates and hospital volume for major cancer surgery in California made a difference? A follow‐up analysis of another decade. Ann Surg. 2009;250(3):472–483.
- , , , et al. Children's hospitals do not acutely respond to high occupancy. Pediatrics. 2010;125:974–981.
- , , , , , . Scheduled admissions and high occupancy at a children's hospital. J Hosp Med. 2011;6(2):81–87.
- , , . Characteristics of weekday and weekend hospital admissions. HCUP Statistical Brief. 2010;87.
- Agency for Healthcare Research and Quality. HCUP databases, Healthcare Cost and Utilization Project (HCUP); 2008. Available at: http://www.hcup‐us.ahrq.gov/kidoverview.jsp. Accessed July 15, 2009.
- , et al. Managing capacity to reduce emergency department overcrowding and ambulance diversions. J Qual Patient Saf. 2006;32(5):239–245.
- Institute for Healthcare Improvement. Flow initiatives; 2008. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed February 20, 2008.
- , , , , , . Trends in high‐turnover stays among children hospitalized in the United States, 1993–2003. Pediatrics. 2009;123(3):996–1002.
- Institute for Healthcare Improvement. Smoothing elective surgical admissions. Available at: http://www.ihi.org/IHI/Topics/Flow/PatientFlow/EmergingContent/SmoothingElectiveSurgicalAdmissions.htm. Accessed October 24, 2008.
- Boston hospital sees big impact from smoothing elective schedule. OR Manager. 2004;20:12.
- . Managing Variability in Patient Flow Is the Key to Improving Access to Care, Nursing Staffing, Quality of Care, and Reducing Its Cost. Paper presented at Institute of Medicine, Washington, DC; June 24, 2004.
- American Hospital Association Quality Center. Available at: http://www.ahaqualitycenter.org/ahaqualitycenter/. Accessed October 14, 2008.
- Healthcare Cost and Utilization Project (HCUP). Kids' Inpatient Database (KID); July 2008. Available at: http://www.hcup‐us.ahrq.gov/kidoverview.jsp. Accessed September 10, 2008.
- , , . Using a queuing model to help plan bed allocation in a department of geriatric medicine. Health Care Manag Sci. 2002;5(4):307–313.
- . How many hospital beds? Inquiry. 2002;39(4):400–412.
- . Institute for Healthcare Improvement. Patient flow comments. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed September 10, 2008.
- . Factory efficiency comes to the hospital. New York Times. July 9, 2010.
- Institute for Healthcare Improvement. Re‐engineering the operating room. Available at: http://www.ihi.org/IHI/Programs/ConferencesAndSeminars/ReengineeringtheOperatingRoomSept08.htm. Accessed November 8, 2008.
- , . Enhanced weekend service: an affordable means to increased hospital procedure volume. CMAJ. 2005;172(4):503–504.
- , . Mortality among patients admitted to hospitals on weekends as compared with weekdays. N Engl J Med. 2001;345:663–668.
- , , , , , . Weekend versus weekday admission and mortality from myocardial infarction. N Engl J Med. 2007;356:1099–1109.
- , . Waiting for urgent procedures on the weekend among emergently hospitalized patients. Am J Med. 2004;117:175–181.
- . Do hospitals provide lower quality care on weekends? Health Serv Res. 2007;42:1589–1612.
- . Hospital saves by working weekends. Mod Healthc. 1996;26:82–99.
- , , , , . Weekend and holiday exercise testing in patients with chest pain. J Gen Intern Med. 1999;14:10–14.
- . Boston Medical Center Case Study: Institute of Healthcare Optimization; 2006. Available at: http://www.ihoptimize.org/8f16e142‐eeaa‐4898–9e62–660218f19ffb/download.htm. Accessed October 3, 2010.
- , , , . The impact of IMPACT on St John's Regional Health Center. Mo Med. 2003;100:590–592.
- NYU Langone Medical Center Extends Access to Non‐Emergent Care as Part of Commitment to Patient‐Centered Care (June 23, 2010). Available at: http://communications.med.nyu.edu/news/2010/nyu‐langone‐medical‐center‐extends‐access‐non‐emergent‐care‐part‐commitment‐patient‐center. Accessed October 3, 2010.
- Carondelet St. Mary's Hospital. A pragmatic approach to improving patient efficiency throughput. Improvement Report 2005. Available at: http://www.ihi.org/IHI/Topics/Flow/PatientFlow/ImprovementStories/APragmaticApproachtoImprovingPatientEfficiencyThroughput.htm. Accessed October 3, 2010.
- AHA Solutions. Patient Flow Challenges Assessment 2009. Chicago, IL; 2009.
- , , , , , . A conceptual model of emergency department crowding. Ann Emerg Med. 2003;42(2):173–180.
- . Annual bed statistics give a misleading picture of hospital surge capacity. Ann Emerg Med. 2006;48(4):384–388.
Copyright © 2011 Society of Hospital Medicine
Scheduled Admissions and Occupancy
Patient flow in a hospital refers to the management and movement of patients through the facility. Optimizing patient flow is considered of great importance to improvement of quality (including safety, efficiency, timeliness, equity, effectiveness, and patient‐centeredness), as well as finance, staff satisfaction, education and overall healthcare value.18 Central to concerns about patient flow at hospitals is occupancy, which is the census (number of patients at a point in time) divided by the bed capacity. Occupancy that is too high is associated with challenges to quality and access,913 while occupancy that is too low may underutilize resources and be costly.14, 15 Occupancy is determined by the pattern of admission and discharge, thus including length of stay (LOS) as a factor. While all related, admissions, census, occupancy, and LOS convey different aspects of hospital operations and may point to different opportunities to improve patient flow.
Variability in patient flow over time has been noted as a common occurrence in adult hospitals, due to uneven patterns of scheduled (elective) admissions, as well as uncontrollable variability of emergent admissions.2, 45, 16 Typically very few patients are scheduled to enter hospitals over weekends. In addition, when the admission is expected to be 5 days or less, clinical and operational staff may schedule those admissions early in the week to avoid patients staying the weekend. This artificial variability has been shown to lead to uneven levels of occupancy, with crowding on some days of the week more than others.2, 45, 16 As hospital crowding adversely affects access to emergent and elective care, quality and safety of care, and patient and staff satisfaction, many groups are focusing attention on patient flow and strategies to avoid high occupancy.19, 17 This is true for children's hospitals, as well, particularly as these scarce resources have ever increasing demand placed on them.1820
Patient flow improvements can be made by increasing efficiency of throughput, primarily measured by decreased LOS, or by addressing artificial variability in how hospital beds are used. As children's hospitals have short LOSs and are relatively efficient (as measured by standardized LOS ratios), we sought to evaluate how much artificial variability was active at 1 large children's hospital. We did this to both evaluate flow at 1 institution and to create methodology for other hospitals to use in order to better understand and improve their flow.
Our specific aims were to describe daily and monthly variability in admission, discharge, LOS, and occupancy patterns at a large children's hospital and assess the relationship between scheduled admissions and occupancy.
Methods
This retrospective administrative data analysis was performed with admission‐discharge‐transfer (ADT) data for inpatient admissions from one urban, tertiary‐care children's hospital for the period July 1, 2007 to June 30, 2008. The dataset included the date and time of all arrivals and departures from all inpatient units (including observation‐status patients), as entered by the unit clerks into the electronic ADT system. The dataset also included categorization of the admission as emergent, urgent, or elective (hereafter referred to as scheduled.) Registration staff entered these codes at or prior to admission. Using the timestamps, LOS was calculated by subtracting admission date and time from discharge date and time. An SAS macro was applied to the timestamps to calculate a hospital census for every hour of each calendar day. Peak census figures were extracted for each day. Occupancy was calculated as census over number of beds in use (monthly average). Data for the hospital's peak daily census and occupancy were utilized to analyze patterns of occupancy by day of week and month of year. To express variability, coefficient of variation (CV) (standard deviation [SD] divided by its mean) was used, as it is helpful when samples sizes are different.21
Analysis of number of admissions per day of week and month by type was performed with descriptive statistics and t‐tests for significant differences across seasons. We calculated a measure of patient hours generated by day of admission based on the LOS generated by each admission, in which the average number of admissions for each day of the week was multiplied by the average LOS (in hours) for those admissions. In order to remove outliers and focus on patients whose occupancy would affect weekly variation, we analyzed in detail the admissions with LOS 30 days and 7 days, respectively.
Statistical analyses were performed with SAS 9.2 (SAS Institute, Cary, NC), Stata 10.0 (StataCorp, College Station, TX) and Microsoft Excel (Microsoft, Redmond, WA). The study was approved by the Human Subjects Committee of the hospital's Institutional Review Board.
Results
A total of 22,310 patients were admitted over the period July 1, 2007 to June 30, 2008, including 4957 (22%) coded as scheduled and 17,353 (78%) coded as emergent. (Only 200 patients were registered as urgent and these were recoded as emergent for this analysis). Details on admission types and discharging departments are provided in Table 1. Overall, mean LOS was 5.6 days (median 2.29 days). For patients with LOS 30 days, mean LOS was 3.88 days (median 2.22 days). For patients staying 7 days, mean LOS was 2.4 days (median 1.98 days). Among patients with LOS 7 days, mean LOS for scheduled patients was longer for those admitted on Monday than on any other weekday (2.49 vs. 2.08 days, P < 0.0001). In contrast, mean LOS for emergent patients was longer for patients admitted on Friday and Saturday than the rest of the week (2.57 vs. 2.44 days, P < 0.0001).
| All | Scheduled | Emergent | |
|---|---|---|---|
| |||
| Total Admissions, n (%)* | 22,310 | 4957 (22) | 17,353 (78) |
| Median LOS (days) | 2.29 | 1.93 | 2.50 |
| Mean LOS (days) (95% CI) | 5.60 (5.41, 5.79) | 4.20 (3.95, 4.45) | 5.78 (5.596.0) |
| % Patients with LOS 30 days (%) | 97 | 98 | 96 |
| % Patients with LOS 7 days (%) | 84 | 89 | 83 |
| Medical patients at discharge, n (%) | 16,586 (74) | 2363 (48) | 14,403 (83) |
| Surgical patients at discharge, n (%) | 4276 (19) | 2450 (49) | 1826 (10.5) |
| Critical care patients at discharge (NICU, PICU, CICU), n (%) | 1433 (6) | 140 (3) | 1293 (7.5) |
Total admissions per month (Figure 1) averaged 1937 in October through April and 1751 in May through September (P = 0.03). Variation in the number of emergent and scheduled patients over months of the year were similar (CV 10% for each), but emergent admissions did decrease in summer (mean 1299 for June‐September vs. 1520 for the rest of the year, P = 0.003). Conversely, scheduled admissions remained relatively stable all year‐long: mean 423 per month for May through September vs. mean 413 per month for October through April (P = 0.48). Even just the summer months of June‐August, when school‐age children are on vacation, were not significantly different from other months (440 vs. 404, P = 0.2).
Variation in volume of admissions was large over days of the week, driven primarily by the pattern of scheduled admissions (CV 65.3%), which dropped off completely on weekends (Table 2, Figure 2). In contrast, there was much less variation in the number of emergent admissions across days of the week (CV 12%). For both emergent and scheduled admissions, more patients came in on Mondays than any other day of the week, but even more so for scheduled patients. While emergent admissions did decline on weekends, it was driven primarily by a decrease in physician referrals (ie, direct admission) from clinics (mean 7.48 per weekday vs. 0.73 per weekend day for the entire year, P < 0.001), while emergency department (ED) admissions remained relatively stable (mean 35.8 per weekday vs. 32.7 per weekend day, P = 0.08). Emergency transports were also stable (mean 7.15 per weekday vs. 6.49 per weekend day, P = 0.10).
| All (%) | Scheduled (%) | Emergent (%) | |
|---|---|---|---|
| |||
| CV on admissions by month | 8 | 10 | 10 |
| CV on admissions over days of week (including weekends) | 24 | 65 | 12 |
| CV on admissions over days of week (excluding weekends) | 6 | 10 | 5 |
| CV on monthly occupancy over 12 months | 4 | 14 | 2 |
Although scheduled patients contributed less to the hospital's overall occupancy, they conferred most of the variability by day of week. Over the days of the week, variation for scheduled occupancy was nearly twice that for emergent occupancy (CV 19% vs. 10%). Within the higher‐volume period of October to April, the differential was even more evident (CV 19% for scheduled occupancy versus 6% for emergent).
For scheduled patients with LOS 30 days (98% of scheduled patients), Mondays and Tuesdays together accounted for 42.5% of admission volume and 44.7% of the patient‐hours generated. For scheduled patients with LOS 7 days (89% of scheduled patients), Mondays and Tuesdays together accounted for 42% of admission volume and 45.2% of the patient‐hours generated. This combined impact of volume and LOS from admissions earlier in the week (restricted to patients with LOS 7days) is displayed graphically in Figure 3, which depicts the unevenness of scheduled admissions and their time in the hospital, with many patients overlapping in the middle of the week. Together with the more steady flow of emergent patients, this variability in scheduled occupancy contributed to mid‐week crowding, with higher risk of the hospital being >90% and >95% occupied on Wednesday through Friday (Figure 4). Detailed hourly analysis (not displayed) showed this risk to be highest from Wednesday afternoon to Friday afternoon. Due to higher emergent census, certain months also had a higher risk of high occupancy at daily peak. For example, while the entire year had 50% to 60% of Wednesdays and Thursdays with occupancy >90%, during the months of November through February, 70% to 85% of those days had occupancy at that level or higher (all these patterns were seen for both stays with LOS 30 days and 7days).
Discussion
In this study, we found that a large children's hospital was frequently at high occupancy in certain months and on certain days more than others, driven largely by predictable seasonal increases in emergent admissions and variability in scheduled admissions by day of week, respectively. Patient‐hours generated by day of admission varied as a result of both volume and LOS, both of which were larger in the early part of the week and diminished as the week progressed for scheduled admissions. But, the cumulative effect of many admissions with relatively‐longer LOS on Monday through Wednesday contributed to high occupancy on Wednesday afternoon to Friday morning, underscoring the importance of admission patterns on census later in the week. Our finding that the occupancy of scheduled patientsthe result of both the admission pattern and their LOSis also highly variable suggests that managing the inflow of scheduled patients could decrease crowding on weekdays, assure a consistent supply of capacity for regular admissions and surges, and improve the value of the delivery system.18 This inflow management would ideally consider both admissions and associated LOS, since rescheduling patients with a longer LOS (eg, 34 days) would have a greater impact on occupancy than rescheduling patients with a shorter LOS (eg, 12 days).
Not surprisingly, total admissions decreased in summer months, especially in July and August, due primarily to fewer emergent admissions. In fact, scheduled admissions per month remained relatively stable over the entire year. The decrease in summer emergent admissions may present an opportunity to stepwise shift a proportion of scheduled admissions from the spring and fall into the summer months, and winter into spring and fall, to alleviate crowding in the winter (Figure 1). Assuming clinical conditions, families and staff members were amenable to this change, hospitals with similar patterns could use this method to reduce the crowding (eg, days over 90% or 95% occupancy) that occurs in the winter.
Using patient‐hours (or days) generated by day of admission, it is evident that admission of more and longer‐stay patients at the start of the week contributes to higher occupancy later in the week (Figure 4). Mid‐week crowding could potentially contribute to a number of operational issues, including delays of new admissions, decreases in physician referrals and patient satisfaction, and an increased use of nontraditional beds (eg, treatment rooms, playrooms, doubling up single rooms) that lead to excessive patient to staff ratios and burnout for clinical staff.
The reasons for these patterns of admissions may include clinician or patient preference to avoid weekend admissions, lack of availability of particular services or resources on weekends, or concerns about safety and efficiency (due to relatively lower staffing on weekends in many hospitals).2230 While clinicians may prefer to avoid additional work on weekends, there are benefits to smoothing occupancy, including less risk of excessive work mid‐week and potential revenue opportunities. In addition, when contrasted with the estimated $1 million to $2 million cost per bed of construction, the marginal cost of staffing to provide safe, high‐quality care on weekends is dramatically lower than that of adding more beds (when faced with mid‐week crowding and unused weekend capacity). In addition, empty beds also do not generate revenue to cover fixed or variable costs, meaning that hospitals are not matching revenue to cost when there is unused capacity due to artificial variability.15, 31 Hospitals looking to make greater use of weekends, however, must be sensitive to staff concerns and the organizational dynamics of 7‐day operations, including the risk for burn‐out and attrition. Such factors should not be perceived as insurmountable barriers, particularly in light of opportunities for flexible scheduling and gain‐sharing.
Patients' and parents' preferences may partially drive admitting patterns, but a reasonable proportion of them may prefer to minimize the number of work and school days missed by being admitted near or on weekends. For example, an expected 3‐day admission could start on Friday and end on Sunday or Monday, rather than the current practice which appears to be to admit on Monday and discharge before the weekend. This may not only meet preferences among some parents to avoid missing work or school, but also by consideration of educational outcomes for hospitalized children.32
In addition, higher mean LOS for emergent patients on the weekends suggests that some services are currently unavailable on weekends to treat patients who are admitted on Fridays through Sundays.2, 25, 29, 33 More even staffing and provision of diagnostic and therapeutic services on weekends (eg, advanced radiology, consult, and laboratory services) would not only remove the barrier to weekend occupancy, it would also improve efficiency, timeliness, patient‐centeredness, and potentially effectiveness and safety for emergent patients. Running hospitals at full functionality on only 5 days of the week means that 2 out of 7 days puts patients at risk for disparate care, which may be appearing in this analysis as prolonged LOS due to lack of services over the weekenda pattern reported in the literature for adult hospitals.
Operations management and queuing theory suggest that systems function well up to 85% to 90% of capacity.34 Hospitals that plan ahead and ensure a buffer for unscheduled admissions during months or days when that demand is known to rise are less likely to cross into high occupancy. On the other hand, hospitals that do not anticipate increases in unscheduled admissions are more likely to encounter excess capacity problems.35 Aligning incentives with all staff can assist in this planning and maximize control of capacity.
Adopting the use of CV in health care operations would also be of value as a way to better express and track variation in admissions, occupancy, and discharges. Since different patient populations, different units, different hospitals, and different months have different scales, SD is not easily comparable across these settings. CV allows for comparison of variation by normalizing on the mean. In this study, it clearly differentiated the variation in elective admissions (CV 65%) over days of the week from the relative stability of emergent admissions (CV 12%). As variability and its management are important to operations, quality control, and quality improvement, use of CV can play an important role in hospital management and health services research. As days with high levels of activity may put more stress on the system, tracking this variation could lead to improvements in quality and value.
This study has several limitations. Data were analyzed for 1 children's hospital, so the analysis may or may not generally apply to other hospitals. However, in a separate study, we analyzed data from the Pediatric Health Information System database, and observed similar patterns.18 In addition, the proportion of elective patients shown in this study was similar to the national data in Kids Inpatient Database (KID, about 15% of all admissions elective).36 Moreover, the methods are reproducible for other settings, which would be useful to clinical and hospital leadership. Second, the trends depicted in the data only reflected data for one year. Third, coding of the admission as emergent or elective was done by registrars at or before arrival and may not reflect actual clinical need. In addition, those admissions coded as elective included a heterogeneous population (eg, chemotherapy to research studies).
Further studies should analyze trends for other hospitals and evaluate the effect of high peak census and high levels of variation with quality, safety, efficiency, patient satisfaction, financial, and educational outcomes for those receiving care, working, or learning at hospitals. In addition, a qualitative study that develops insights into clinician and patient/parent preferences would help answer questions in regard to usage of weekends for scheduled patients.
Conclusions
Scheduled admissions drive most variability in day‐to‐day occupancy despite the fact that they are a smaller proportion of the inpatient population. Variation in scheduled admissions by day of week provides hospitals with an opportunity to address crowding without adding beds or delaying admissions. Rather, fully utilizing capacity by smoothing occupancy over all days of the week can reduce the risk of high occupancy and thereby improve accessibility and patient/parent satisfaction. While family and staff preferences need to be considered, some combination of within‐week smoothing and shifting admissions towards summer are likely to achieve dramatic improvements in patient flow without large expenditures of capital. The key, then, is to ensure that organizational dynamic factors support these changes, so that staff members do not become stressed working at a 7‐day facility. Taken together, these strategies would better match revenue to capacity, and ultimately increase the quality and value of healthcare operations.
Acknowledgements
Authors' contributions: Study concept and design: Fieldston, Ragavan. Analysis and interpretation of data: Ragavan, Fieldston, Jayaraman, Pati. Drafting of the manuscript: Ragavan, Fieldston. Critical Revision of the manuscript for important intellectual content: Fieldston, Ragavan, Pati, Metlay. Statistical analysis: Fieldston, Jayaraman, Ragavan, Allebach. Study supervision: Fieldston, Pati, Metlay.
Additional contributions: The authors the fellows and faculty of the Robert Wood Johnson Foundation Clinical Scholars Program at the University of Pennsylvania and members of its Community Advisory Board for their suggestions to this work. They also wish to thank Tracy Kish, Jennifer Massenburg, and Brian Smith for assistance with access to and interpretation of hospital census and bed capacity data.
- ,,,,,.Relationship between resident workload and self‐perceived learning on inpatient medicine wards: a longitudinal study.BMC Med Educ.2006;6(1):35.
- AHA Solutions, Patient Flow Challenges Assessment 2009. Chicago, IL.2009.
- ,.Patient flow in hospitals: understanding and controlling it better.Front Health Serv Manage.2004;20:3–15.
- . Managing Variability in Patient Flow is the Key to Improving Access to Care, Nursing Staffing, Quality of Care, and Reducing Its Cost. Paper presented at: Institute of Medicine; June 24,2004.
- ,,,,,.Managing unnecessary variability in patient demand to reduce nursing stress and improve patient safety.Jt Comm J Qual Patient Saf.2005;31(6):330–338.
- ,,.Developing models for patient flow and daily surge capacity research.Acad Emerg Med.2006;13(11):1109–1113.
- Institute for Healthcare Improvement, Flow initiatives. 2008. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed June2010.
- ,,, et al.Hospital workload and adverse events.Med Care.2007;45(5):448–455.
- ,,,,.Impact of admission‐day crowding on the length of stay of pediatric hospitalizations.Pediatrics.2008;121(4):e718–730.
- ,,,,.Emergency department crowding, Part 1: concept, causes, and moral consequences.Ann Emerg Med.2009;53(5):605–611.
- ,.Emergency department overcrowding and ambulance diversion: the impact and potential solutions of extended boarding of admitted patients in the emergency department.J Emerg Med.2006;30(3):351–356.
- ,,,.A comparison of in‐hospital mortality risk conferred by high hospital occupancy, differences in nurse staffing levels, weekend admission, and seasonal influenza.Medical Care.2010;48(3):224–232.
- .,,.The effect of hospital bed occupancy on throughput in the pediatric emergency department.Ann Emerg Med.2009;53(6):767–776.e763.
- .Interpreting the Volume‐Outcome Relationship in the Context of Health Care Quality: Workshop Summary.Washington, DC:National Academies Press;2000.
- ,.Uncertain demand, the structure of hospital costs, and the cost of empty hospital beds.J Health Econ.1995;14(3):291–317.
- ,,, et al.Variability in surgical caseload and access to intensive care services.Anesthesiology.2003;98(6):1491–1496.
- ,,,,.Effects of hospital care environment on patient mortality and nurse outcomes.J Nurs Adm.2009;39(7/8):S45–S51.
- ,,, et al.Children's hospitals do not acutely respond to high occupancy.Pediatrics.2010;125:974–981.
- ,.Easing the strain on a pediatric tertiary care center: use of a redistribution system.Arch Pediatr Adolesc Med.2007;161(9):870–876.
- ,,.Lengths of stay and costs associated with Children's Hospitals.Pediatrics.2005;115:839–844.
- ,.Matching Supply with Demand: An introduction to operations management.New York:McGraw‐Hill;2006.
- .Do hospitals provide lower quality care on weekends?Health Serv Res.2007;42:1589–1612.
- ,.Waiting for urgent procedures on the weekend among emergently hospitalized patients.Am J Med.2004;117:175–181.
- ,.Mortality among patients admitted to hospitals on weekends as compared with weekdays.N Engl J Med.2001;345:663–668.
- ,.Enhanced weekend service: an affordable means to increased hospital procedure volume.CMAJ.2005;172(4):503–504.
- .Hospital deaths and weekend admissions‐how do we leap across a chasm?Clin Nurse Spec.2002;16:91–92.
- ,.Weekend birth and higher neonatal mortality: a problem of patient acuity or quality of care?J Obstet Gynecol Neonatal Nurs.2003;32(6):724–733.
- ,,, et al.Weekend versus weekday admission and mortality from myocardial infarction.N Engl J Med.2007;356:1099–1109.
- .Weekend admission and treatment of patients with renal colic: a case of avoidable variation?Urology.2009;73(4):720–724.
- ,,.Characteristics of weekday and weekend hospital admissions.HCUP Statistical Brief.2010;87. http://www.hcup‐us.ahrq.gov/reports/statbriefs/sb87.pdf.
- ,. Hospital Costs and the Cost of Empty Hospital Beds (NBER Working Paper No. W3872).1991.
- ,,.The effects of timing of pediatric knee ligament surgery on short‐term academic performance in school‐aged athletes.Am J Sports Med.2009;37(9):1684–1691.
- Institute for Healthcare Improvement, Smoothing Elective Surgical Admissions. Available at: http://www.ihi.org/IHI/Topics/Flow/PatientFlow/EmergingContent/SmoothingElectiveSurgicalAdmissions.htm. Accessed June2010.
- . Institute for Healthcare Improvement, Patient Flow Comments. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed June2010.
- .OPIM 631: Operations Management.Wharton School, University of Pennsylvania.Philadelphia, PA.2008.
- Agency for Healthcare Research and Quality. HCUP Databases, Healthcare Cost and Utilization Project (HCUP). 2008. Available at: www.hcup‐us.ahrq.gov/kidoverview.jsp. Accessed June2010.
Patient flow in a hospital refers to the management and movement of patients through the facility. Optimizing patient flow is considered of great importance to improvement of quality (including safety, efficiency, timeliness, equity, effectiveness, and patient‐centeredness), as well as finance, staff satisfaction, education and overall healthcare value.18 Central to concerns about patient flow at hospitals is occupancy, which is the census (number of patients at a point in time) divided by the bed capacity. Occupancy that is too high is associated with challenges to quality and access,913 while occupancy that is too low may underutilize resources and be costly.14, 15 Occupancy is determined by the pattern of admission and discharge, thus including length of stay (LOS) as a factor. While all related, admissions, census, occupancy, and LOS convey different aspects of hospital operations and may point to different opportunities to improve patient flow.
Variability in patient flow over time has been noted as a common occurrence in adult hospitals, due to uneven patterns of scheduled (elective) admissions, as well as uncontrollable variability of emergent admissions.2, 45, 16 Typically very few patients are scheduled to enter hospitals over weekends. In addition, when the admission is expected to be 5 days or less, clinical and operational staff may schedule those admissions early in the week to avoid patients staying the weekend. This artificial variability has been shown to lead to uneven levels of occupancy, with crowding on some days of the week more than others.2, 45, 16 As hospital crowding adversely affects access to emergent and elective care, quality and safety of care, and patient and staff satisfaction, many groups are focusing attention on patient flow and strategies to avoid high occupancy.19, 17 This is true for children's hospitals, as well, particularly as these scarce resources have ever increasing demand placed on them.1820
Patient flow improvements can be made by increasing efficiency of throughput, primarily measured by decreased LOS, or by addressing artificial variability in how hospital beds are used. As children's hospitals have short LOSs and are relatively efficient (as measured by standardized LOS ratios), we sought to evaluate how much artificial variability was active at 1 large children's hospital. We did this to both evaluate flow at 1 institution and to create methodology for other hospitals to use in order to better understand and improve their flow.
Our specific aims were to describe daily and monthly variability in admission, discharge, LOS, and occupancy patterns at a large children's hospital and assess the relationship between scheduled admissions and occupancy.
Methods
This retrospective administrative data analysis was performed with admission‐discharge‐transfer (ADT) data for inpatient admissions from one urban, tertiary‐care children's hospital for the period July 1, 2007 to June 30, 2008. The dataset included the date and time of all arrivals and departures from all inpatient units (including observation‐status patients), as entered by the unit clerks into the electronic ADT system. The dataset also included categorization of the admission as emergent, urgent, or elective (hereafter referred to as scheduled.) Registration staff entered these codes at or prior to admission. Using the timestamps, LOS was calculated by subtracting admission date and time from discharge date and time. An SAS macro was applied to the timestamps to calculate a hospital census for every hour of each calendar day. Peak census figures were extracted for each day. Occupancy was calculated as census over number of beds in use (monthly average). Data for the hospital's peak daily census and occupancy were utilized to analyze patterns of occupancy by day of week and month of year. To express variability, coefficient of variation (CV) (standard deviation [SD] divided by its mean) was used, as it is helpful when samples sizes are different.21
Analysis of number of admissions per day of week and month by type was performed with descriptive statistics and t‐tests for significant differences across seasons. We calculated a measure of patient hours generated by day of admission based on the LOS generated by each admission, in which the average number of admissions for each day of the week was multiplied by the average LOS (in hours) for those admissions. In order to remove outliers and focus on patients whose occupancy would affect weekly variation, we analyzed in detail the admissions with LOS 30 days and 7 days, respectively.
Statistical analyses were performed with SAS 9.2 (SAS Institute, Cary, NC), Stata 10.0 (StataCorp, College Station, TX) and Microsoft Excel (Microsoft, Redmond, WA). The study was approved by the Human Subjects Committee of the hospital's Institutional Review Board.
Results
A total of 22,310 patients were admitted over the period July 1, 2007 to June 30, 2008, including 4957 (22%) coded as scheduled and 17,353 (78%) coded as emergent. (Only 200 patients were registered as urgent and these were recoded as emergent for this analysis). Details on admission types and discharging departments are provided in Table 1. Overall, mean LOS was 5.6 days (median 2.29 days). For patients with LOS 30 days, mean LOS was 3.88 days (median 2.22 days). For patients staying 7 days, mean LOS was 2.4 days (median 1.98 days). Among patients with LOS 7 days, mean LOS for scheduled patients was longer for those admitted on Monday than on any other weekday (2.49 vs. 2.08 days, P < 0.0001). In contrast, mean LOS for emergent patients was longer for patients admitted on Friday and Saturday than the rest of the week (2.57 vs. 2.44 days, P < 0.0001).
| All | Scheduled | Emergent | |
|---|---|---|---|
| |||
| Total Admissions, n (%)* | 22,310 | 4957 (22) | 17,353 (78) |
| Median LOS (days) | 2.29 | 1.93 | 2.50 |
| Mean LOS (days) (95% CI) | 5.60 (5.41, 5.79) | 4.20 (3.95, 4.45) | 5.78 (5.596.0) |
| % Patients with LOS 30 days (%) | 97 | 98 | 96 |
| % Patients with LOS 7 days (%) | 84 | 89 | 83 |
| Medical patients at discharge, n (%) | 16,586 (74) | 2363 (48) | 14,403 (83) |
| Surgical patients at discharge, n (%) | 4276 (19) | 2450 (49) | 1826 (10.5) |
| Critical care patients at discharge (NICU, PICU, CICU), n (%) | 1433 (6) | 140 (3) | 1293 (7.5) |
Total admissions per month (Figure 1) averaged 1937 in October through April and 1751 in May through September (P = 0.03). Variation in the number of emergent and scheduled patients over months of the year were similar (CV 10% for each), but emergent admissions did decrease in summer (mean 1299 for June‐September vs. 1520 for the rest of the year, P = 0.003). Conversely, scheduled admissions remained relatively stable all year‐long: mean 423 per month for May through September vs. mean 413 per month for October through April (P = 0.48). Even just the summer months of June‐August, when school‐age children are on vacation, were not significantly different from other months (440 vs. 404, P = 0.2).
Variation in volume of admissions was large over days of the week, driven primarily by the pattern of scheduled admissions (CV 65.3%), which dropped off completely on weekends (Table 2, Figure 2). In contrast, there was much less variation in the number of emergent admissions across days of the week (CV 12%). For both emergent and scheduled admissions, more patients came in on Mondays than any other day of the week, but even more so for scheduled patients. While emergent admissions did decline on weekends, it was driven primarily by a decrease in physician referrals (ie, direct admission) from clinics (mean 7.48 per weekday vs. 0.73 per weekend day for the entire year, P < 0.001), while emergency department (ED) admissions remained relatively stable (mean 35.8 per weekday vs. 32.7 per weekend day, P = 0.08). Emergency transports were also stable (mean 7.15 per weekday vs. 6.49 per weekend day, P = 0.10).
| All (%) | Scheduled (%) | Emergent (%) | |
|---|---|---|---|
| |||
| CV on admissions by month | 8 | 10 | 10 |
| CV on admissions over days of week (including weekends) | 24 | 65 | 12 |
| CV on admissions over days of week (excluding weekends) | 6 | 10 | 5 |
| CV on monthly occupancy over 12 months | 4 | 14 | 2 |
Although scheduled patients contributed less to the hospital's overall occupancy, they conferred most of the variability by day of week. Over the days of the week, variation for scheduled occupancy was nearly twice that for emergent occupancy (CV 19% vs. 10%). Within the higher‐volume period of October to April, the differential was even more evident (CV 19% for scheduled occupancy versus 6% for emergent).
For scheduled patients with LOS 30 days (98% of scheduled patients), Mondays and Tuesdays together accounted for 42.5% of admission volume and 44.7% of the patient‐hours generated. For scheduled patients with LOS 7 days (89% of scheduled patients), Mondays and Tuesdays together accounted for 42% of admission volume and 45.2% of the patient‐hours generated. This combined impact of volume and LOS from admissions earlier in the week (restricted to patients with LOS 7days) is displayed graphically in Figure 3, which depicts the unevenness of scheduled admissions and their time in the hospital, with many patients overlapping in the middle of the week. Together with the more steady flow of emergent patients, this variability in scheduled occupancy contributed to mid‐week crowding, with higher risk of the hospital being >90% and >95% occupied on Wednesday through Friday (Figure 4). Detailed hourly analysis (not displayed) showed this risk to be highest from Wednesday afternoon to Friday afternoon. Due to higher emergent census, certain months also had a higher risk of high occupancy at daily peak. For example, while the entire year had 50% to 60% of Wednesdays and Thursdays with occupancy >90%, during the months of November through February, 70% to 85% of those days had occupancy at that level or higher (all these patterns were seen for both stays with LOS 30 days and 7days).
Discussion
In this study, we found that a large children's hospital was frequently at high occupancy in certain months and on certain days more than others, driven largely by predictable seasonal increases in emergent admissions and variability in scheduled admissions by day of week, respectively. Patient‐hours generated by day of admission varied as a result of both volume and LOS, both of which were larger in the early part of the week and diminished as the week progressed for scheduled admissions. But, the cumulative effect of many admissions with relatively‐longer LOS on Monday through Wednesday contributed to high occupancy on Wednesday afternoon to Friday morning, underscoring the importance of admission patterns on census later in the week. Our finding that the occupancy of scheduled patientsthe result of both the admission pattern and their LOSis also highly variable suggests that managing the inflow of scheduled patients could decrease crowding on weekdays, assure a consistent supply of capacity for regular admissions and surges, and improve the value of the delivery system.18 This inflow management would ideally consider both admissions and associated LOS, since rescheduling patients with a longer LOS (eg, 34 days) would have a greater impact on occupancy than rescheduling patients with a shorter LOS (eg, 12 days).
Not surprisingly, total admissions decreased in summer months, especially in July and August, due primarily to fewer emergent admissions. In fact, scheduled admissions per month remained relatively stable over the entire year. The decrease in summer emergent admissions may present an opportunity to stepwise shift a proportion of scheduled admissions from the spring and fall into the summer months, and winter into spring and fall, to alleviate crowding in the winter (Figure 1). Assuming clinical conditions, families and staff members were amenable to this change, hospitals with similar patterns could use this method to reduce the crowding (eg, days over 90% or 95% occupancy) that occurs in the winter.
Using patient‐hours (or days) generated by day of admission, it is evident that admission of more and longer‐stay patients at the start of the week contributes to higher occupancy later in the week (Figure 4). Mid‐week crowding could potentially contribute to a number of operational issues, including delays of new admissions, decreases in physician referrals and patient satisfaction, and an increased use of nontraditional beds (eg, treatment rooms, playrooms, doubling up single rooms) that lead to excessive patient to staff ratios and burnout for clinical staff.
The reasons for these patterns of admissions may include clinician or patient preference to avoid weekend admissions, lack of availability of particular services or resources on weekends, or concerns about safety and efficiency (due to relatively lower staffing on weekends in many hospitals).2230 While clinicians may prefer to avoid additional work on weekends, there are benefits to smoothing occupancy, including less risk of excessive work mid‐week and potential revenue opportunities. In addition, when contrasted with the estimated $1 million to $2 million cost per bed of construction, the marginal cost of staffing to provide safe, high‐quality care on weekends is dramatically lower than that of adding more beds (when faced with mid‐week crowding and unused weekend capacity). In addition, empty beds also do not generate revenue to cover fixed or variable costs, meaning that hospitals are not matching revenue to cost when there is unused capacity due to artificial variability.15, 31 Hospitals looking to make greater use of weekends, however, must be sensitive to staff concerns and the organizational dynamics of 7‐day operations, including the risk for burn‐out and attrition. Such factors should not be perceived as insurmountable barriers, particularly in light of opportunities for flexible scheduling and gain‐sharing.
Patients' and parents' preferences may partially drive admitting patterns, but a reasonable proportion of them may prefer to minimize the number of work and school days missed by being admitted near or on weekends. For example, an expected 3‐day admission could start on Friday and end on Sunday or Monday, rather than the current practice which appears to be to admit on Monday and discharge before the weekend. This may not only meet preferences among some parents to avoid missing work or school, but also by consideration of educational outcomes for hospitalized children.32
In addition, higher mean LOS for emergent patients on the weekends suggests that some services are currently unavailable on weekends to treat patients who are admitted on Fridays through Sundays.2, 25, 29, 33 More even staffing and provision of diagnostic and therapeutic services on weekends (eg, advanced radiology, consult, and laboratory services) would not only remove the barrier to weekend occupancy, it would also improve efficiency, timeliness, patient‐centeredness, and potentially effectiveness and safety for emergent patients. Running hospitals at full functionality on only 5 days of the week means that 2 out of 7 days puts patients at risk for disparate care, which may be appearing in this analysis as prolonged LOS due to lack of services over the weekenda pattern reported in the literature for adult hospitals.
Operations management and queuing theory suggest that systems function well up to 85% to 90% of capacity.34 Hospitals that plan ahead and ensure a buffer for unscheduled admissions during months or days when that demand is known to rise are less likely to cross into high occupancy. On the other hand, hospitals that do not anticipate increases in unscheduled admissions are more likely to encounter excess capacity problems.35 Aligning incentives with all staff can assist in this planning and maximize control of capacity.
Adopting the use of CV in health care operations would also be of value as a way to better express and track variation in admissions, occupancy, and discharges. Since different patient populations, different units, different hospitals, and different months have different scales, SD is not easily comparable across these settings. CV allows for comparison of variation by normalizing on the mean. In this study, it clearly differentiated the variation in elective admissions (CV 65%) over days of the week from the relative stability of emergent admissions (CV 12%). As variability and its management are important to operations, quality control, and quality improvement, use of CV can play an important role in hospital management and health services research. As days with high levels of activity may put more stress on the system, tracking this variation could lead to improvements in quality and value.
This study has several limitations. Data were analyzed for 1 children's hospital, so the analysis may or may not generally apply to other hospitals. However, in a separate study, we analyzed data from the Pediatric Health Information System database, and observed similar patterns.18 In addition, the proportion of elective patients shown in this study was similar to the national data in Kids Inpatient Database (KID, about 15% of all admissions elective).36 Moreover, the methods are reproducible for other settings, which would be useful to clinical and hospital leadership. Second, the trends depicted in the data only reflected data for one year. Third, coding of the admission as emergent or elective was done by registrars at or before arrival and may not reflect actual clinical need. In addition, those admissions coded as elective included a heterogeneous population (eg, chemotherapy to research studies).
Further studies should analyze trends for other hospitals and evaluate the effect of high peak census and high levels of variation with quality, safety, efficiency, patient satisfaction, financial, and educational outcomes for those receiving care, working, or learning at hospitals. In addition, a qualitative study that develops insights into clinician and patient/parent preferences would help answer questions in regard to usage of weekends for scheduled patients.
Conclusions
Scheduled admissions drive most variability in day‐to‐day occupancy despite the fact that they are a smaller proportion of the inpatient population. Variation in scheduled admissions by day of week provides hospitals with an opportunity to address crowding without adding beds or delaying admissions. Rather, fully utilizing capacity by smoothing occupancy over all days of the week can reduce the risk of high occupancy and thereby improve accessibility and patient/parent satisfaction. While family and staff preferences need to be considered, some combination of within‐week smoothing and shifting admissions towards summer are likely to achieve dramatic improvements in patient flow without large expenditures of capital. The key, then, is to ensure that organizational dynamic factors support these changes, so that staff members do not become stressed working at a 7‐day facility. Taken together, these strategies would better match revenue to capacity, and ultimately increase the quality and value of healthcare operations.
Acknowledgements
Authors' contributions: Study concept and design: Fieldston, Ragavan. Analysis and interpretation of data: Ragavan, Fieldston, Jayaraman, Pati. Drafting of the manuscript: Ragavan, Fieldston. Critical Revision of the manuscript for important intellectual content: Fieldston, Ragavan, Pati, Metlay. Statistical analysis: Fieldston, Jayaraman, Ragavan, Allebach. Study supervision: Fieldston, Pati, Metlay.
Additional contributions: The authors the fellows and faculty of the Robert Wood Johnson Foundation Clinical Scholars Program at the University of Pennsylvania and members of its Community Advisory Board for their suggestions to this work. They also wish to thank Tracy Kish, Jennifer Massenburg, and Brian Smith for assistance with access to and interpretation of hospital census and bed capacity data.
Patient flow in a hospital refers to the management and movement of patients through the facility. Optimizing patient flow is considered of great importance to improvement of quality (including safety, efficiency, timeliness, equity, effectiveness, and patient‐centeredness), as well as finance, staff satisfaction, education and overall healthcare value.18 Central to concerns about patient flow at hospitals is occupancy, which is the census (number of patients at a point in time) divided by the bed capacity. Occupancy that is too high is associated with challenges to quality and access,913 while occupancy that is too low may underutilize resources and be costly.14, 15 Occupancy is determined by the pattern of admission and discharge, thus including length of stay (LOS) as a factor. While all related, admissions, census, occupancy, and LOS convey different aspects of hospital operations and may point to different opportunities to improve patient flow.
Variability in patient flow over time has been noted as a common occurrence in adult hospitals, due to uneven patterns of scheduled (elective) admissions, as well as uncontrollable variability of emergent admissions.2, 45, 16 Typically very few patients are scheduled to enter hospitals over weekends. In addition, when the admission is expected to be 5 days or less, clinical and operational staff may schedule those admissions early in the week to avoid patients staying the weekend. This artificial variability has been shown to lead to uneven levels of occupancy, with crowding on some days of the week more than others.2, 45, 16 As hospital crowding adversely affects access to emergent and elective care, quality and safety of care, and patient and staff satisfaction, many groups are focusing attention on patient flow and strategies to avoid high occupancy.19, 17 This is true for children's hospitals, as well, particularly as these scarce resources have ever increasing demand placed on them.1820
Patient flow improvements can be made by increasing efficiency of throughput, primarily measured by decreased LOS, or by addressing artificial variability in how hospital beds are used. As children's hospitals have short LOSs and are relatively efficient (as measured by standardized LOS ratios), we sought to evaluate how much artificial variability was active at 1 large children's hospital. We did this to both evaluate flow at 1 institution and to create methodology for other hospitals to use in order to better understand and improve their flow.
Our specific aims were to describe daily and monthly variability in admission, discharge, LOS, and occupancy patterns at a large children's hospital and assess the relationship between scheduled admissions and occupancy.
Methods
This retrospective administrative data analysis was performed with admission‐discharge‐transfer (ADT) data for inpatient admissions from one urban, tertiary‐care children's hospital for the period July 1, 2007 to June 30, 2008. The dataset included the date and time of all arrivals and departures from all inpatient units (including observation‐status patients), as entered by the unit clerks into the electronic ADT system. The dataset also included categorization of the admission as emergent, urgent, or elective (hereafter referred to as scheduled.) Registration staff entered these codes at or prior to admission. Using the timestamps, LOS was calculated by subtracting admission date and time from discharge date and time. An SAS macro was applied to the timestamps to calculate a hospital census for every hour of each calendar day. Peak census figures were extracted for each day. Occupancy was calculated as census over number of beds in use (monthly average). Data for the hospital's peak daily census and occupancy were utilized to analyze patterns of occupancy by day of week and month of year. To express variability, coefficient of variation (CV) (standard deviation [SD] divided by its mean) was used, as it is helpful when samples sizes are different.21
Analysis of number of admissions per day of week and month by type was performed with descriptive statistics and t‐tests for significant differences across seasons. We calculated a measure of patient hours generated by day of admission based on the LOS generated by each admission, in which the average number of admissions for each day of the week was multiplied by the average LOS (in hours) for those admissions. In order to remove outliers and focus on patients whose occupancy would affect weekly variation, we analyzed in detail the admissions with LOS 30 days and 7 days, respectively.
Statistical analyses were performed with SAS 9.2 (SAS Institute, Cary, NC), Stata 10.0 (StataCorp, College Station, TX) and Microsoft Excel (Microsoft, Redmond, WA). The study was approved by the Human Subjects Committee of the hospital's Institutional Review Board.
Results
A total of 22,310 patients were admitted over the period July 1, 2007 to June 30, 2008, including 4957 (22%) coded as scheduled and 17,353 (78%) coded as emergent. (Only 200 patients were registered as urgent and these were recoded as emergent for this analysis). Details on admission types and discharging departments are provided in Table 1. Overall, mean LOS was 5.6 days (median 2.29 days). For patients with LOS 30 days, mean LOS was 3.88 days (median 2.22 days). For patients staying 7 days, mean LOS was 2.4 days (median 1.98 days). Among patients with LOS 7 days, mean LOS for scheduled patients was longer for those admitted on Monday than on any other weekday (2.49 vs. 2.08 days, P < 0.0001). In contrast, mean LOS for emergent patients was longer for patients admitted on Friday and Saturday than the rest of the week (2.57 vs. 2.44 days, P < 0.0001).
| All | Scheduled | Emergent | |
|---|---|---|---|
| |||
| Total Admissions, n (%)* | 22,310 | 4957 (22) | 17,353 (78) |
| Median LOS (days) | 2.29 | 1.93 | 2.50 |
| Mean LOS (days) (95% CI) | 5.60 (5.41, 5.79) | 4.20 (3.95, 4.45) | 5.78 (5.596.0) |
| % Patients with LOS 30 days (%) | 97 | 98 | 96 |
| % Patients with LOS 7 days (%) | 84 | 89 | 83 |
| Medical patients at discharge, n (%) | 16,586 (74) | 2363 (48) | 14,403 (83) |
| Surgical patients at discharge, n (%) | 4276 (19) | 2450 (49) | 1826 (10.5) |
| Critical care patients at discharge (NICU, PICU, CICU), n (%) | 1433 (6) | 140 (3) | 1293 (7.5) |
Total admissions per month (Figure 1) averaged 1937 in October through April and 1751 in May through September (P = 0.03). Variation in the number of emergent and scheduled patients over months of the year were similar (CV 10% for each), but emergent admissions did decrease in summer (mean 1299 for June‐September vs. 1520 for the rest of the year, P = 0.003). Conversely, scheduled admissions remained relatively stable all year‐long: mean 423 per month for May through September vs. mean 413 per month for October through April (P = 0.48). Even just the summer months of June‐August, when school‐age children are on vacation, were not significantly different from other months (440 vs. 404, P = 0.2).
Variation in volume of admissions was large over days of the week, driven primarily by the pattern of scheduled admissions (CV 65.3%), which dropped off completely on weekends (Table 2, Figure 2). In contrast, there was much less variation in the number of emergent admissions across days of the week (CV 12%). For both emergent and scheduled admissions, more patients came in on Mondays than any other day of the week, but even more so for scheduled patients. While emergent admissions did decline on weekends, it was driven primarily by a decrease in physician referrals (ie, direct admission) from clinics (mean 7.48 per weekday vs. 0.73 per weekend day for the entire year, P < 0.001), while emergency department (ED) admissions remained relatively stable (mean 35.8 per weekday vs. 32.7 per weekend day, P = 0.08). Emergency transports were also stable (mean 7.15 per weekday vs. 6.49 per weekend day, P = 0.10).
| All (%) | Scheduled (%) | Emergent (%) | |
|---|---|---|---|
| |||
| CV on admissions by month | 8 | 10 | 10 |
| CV on admissions over days of week (including weekends) | 24 | 65 | 12 |
| CV on admissions over days of week (excluding weekends) | 6 | 10 | 5 |
| CV on monthly occupancy over 12 months | 4 | 14 | 2 |
Although scheduled patients contributed less to the hospital's overall occupancy, they conferred most of the variability by day of week. Over the days of the week, variation for scheduled occupancy was nearly twice that for emergent occupancy (CV 19% vs. 10%). Within the higher‐volume period of October to April, the differential was even more evident (CV 19% for scheduled occupancy versus 6% for emergent).
For scheduled patients with LOS 30 days (98% of scheduled patients), Mondays and Tuesdays together accounted for 42.5% of admission volume and 44.7% of the patient‐hours generated. For scheduled patients with LOS 7 days (89% of scheduled patients), Mondays and Tuesdays together accounted for 42% of admission volume and 45.2% of the patient‐hours generated. This combined impact of volume and LOS from admissions earlier in the week (restricted to patients with LOS 7days) is displayed graphically in Figure 3, which depicts the unevenness of scheduled admissions and their time in the hospital, with many patients overlapping in the middle of the week. Together with the more steady flow of emergent patients, this variability in scheduled occupancy contributed to mid‐week crowding, with higher risk of the hospital being >90% and >95% occupied on Wednesday through Friday (Figure 4). Detailed hourly analysis (not displayed) showed this risk to be highest from Wednesday afternoon to Friday afternoon. Due to higher emergent census, certain months also had a higher risk of high occupancy at daily peak. For example, while the entire year had 50% to 60% of Wednesdays and Thursdays with occupancy >90%, during the months of November through February, 70% to 85% of those days had occupancy at that level or higher (all these patterns were seen for both stays with LOS 30 days and 7days).
Discussion
In this study, we found that a large children's hospital was frequently at high occupancy in certain months and on certain days more than others, driven largely by predictable seasonal increases in emergent admissions and variability in scheduled admissions by day of week, respectively. Patient‐hours generated by day of admission varied as a result of both volume and LOS, both of which were larger in the early part of the week and diminished as the week progressed for scheduled admissions. But, the cumulative effect of many admissions with relatively‐longer LOS on Monday through Wednesday contributed to high occupancy on Wednesday afternoon to Friday morning, underscoring the importance of admission patterns on census later in the week. Our finding that the occupancy of scheduled patientsthe result of both the admission pattern and their LOSis also highly variable suggests that managing the inflow of scheduled patients could decrease crowding on weekdays, assure a consistent supply of capacity for regular admissions and surges, and improve the value of the delivery system.18 This inflow management would ideally consider both admissions and associated LOS, since rescheduling patients with a longer LOS (eg, 34 days) would have a greater impact on occupancy than rescheduling patients with a shorter LOS (eg, 12 days).
Not surprisingly, total admissions decreased in summer months, especially in July and August, due primarily to fewer emergent admissions. In fact, scheduled admissions per month remained relatively stable over the entire year. The decrease in summer emergent admissions may present an opportunity to stepwise shift a proportion of scheduled admissions from the spring and fall into the summer months, and winter into spring and fall, to alleviate crowding in the winter (Figure 1). Assuming clinical conditions, families and staff members were amenable to this change, hospitals with similar patterns could use this method to reduce the crowding (eg, days over 90% or 95% occupancy) that occurs in the winter.
Using patient‐hours (or days) generated by day of admission, it is evident that admission of more and longer‐stay patients at the start of the week contributes to higher occupancy later in the week (Figure 4). Mid‐week crowding could potentially contribute to a number of operational issues, including delays of new admissions, decreases in physician referrals and patient satisfaction, and an increased use of nontraditional beds (eg, treatment rooms, playrooms, doubling up single rooms) that lead to excessive patient to staff ratios and burnout for clinical staff.
The reasons for these patterns of admissions may include clinician or patient preference to avoid weekend admissions, lack of availability of particular services or resources on weekends, or concerns about safety and efficiency (due to relatively lower staffing on weekends in many hospitals).2230 While clinicians may prefer to avoid additional work on weekends, there are benefits to smoothing occupancy, including less risk of excessive work mid‐week and potential revenue opportunities. In addition, when contrasted with the estimated $1 million to $2 million cost per bed of construction, the marginal cost of staffing to provide safe, high‐quality care on weekends is dramatically lower than that of adding more beds (when faced with mid‐week crowding and unused weekend capacity). In addition, empty beds also do not generate revenue to cover fixed or variable costs, meaning that hospitals are not matching revenue to cost when there is unused capacity due to artificial variability.15, 31 Hospitals looking to make greater use of weekends, however, must be sensitive to staff concerns and the organizational dynamics of 7‐day operations, including the risk for burn‐out and attrition. Such factors should not be perceived as insurmountable barriers, particularly in light of opportunities for flexible scheduling and gain‐sharing.
Patients' and parents' preferences may partially drive admitting patterns, but a reasonable proportion of them may prefer to minimize the number of work and school days missed by being admitted near or on weekends. For example, an expected 3‐day admission could start on Friday and end on Sunday or Monday, rather than the current practice which appears to be to admit on Monday and discharge before the weekend. This may not only meet preferences among some parents to avoid missing work or school, but also by consideration of educational outcomes for hospitalized children.32
In addition, higher mean LOS for emergent patients on the weekends suggests that some services are currently unavailable on weekends to treat patients who are admitted on Fridays through Sundays.2, 25, 29, 33 More even staffing and provision of diagnostic and therapeutic services on weekends (eg, advanced radiology, consult, and laboratory services) would not only remove the barrier to weekend occupancy, it would also improve efficiency, timeliness, patient‐centeredness, and potentially effectiveness and safety for emergent patients. Running hospitals at full functionality on only 5 days of the week means that 2 out of 7 days puts patients at risk for disparate care, which may be appearing in this analysis as prolonged LOS due to lack of services over the weekenda pattern reported in the literature for adult hospitals.
Operations management and queuing theory suggest that systems function well up to 85% to 90% of capacity.34 Hospitals that plan ahead and ensure a buffer for unscheduled admissions during months or days when that demand is known to rise are less likely to cross into high occupancy. On the other hand, hospitals that do not anticipate increases in unscheduled admissions are more likely to encounter excess capacity problems.35 Aligning incentives with all staff can assist in this planning and maximize control of capacity.
Adopting the use of CV in health care operations would also be of value as a way to better express and track variation in admissions, occupancy, and discharges. Since different patient populations, different units, different hospitals, and different months have different scales, SD is not easily comparable across these settings. CV allows for comparison of variation by normalizing on the mean. In this study, it clearly differentiated the variation in elective admissions (CV 65%) over days of the week from the relative stability of emergent admissions (CV 12%). As variability and its management are important to operations, quality control, and quality improvement, use of CV can play an important role in hospital management and health services research. As days with high levels of activity may put more stress on the system, tracking this variation could lead to improvements in quality and value.
This study has several limitations. Data were analyzed for 1 children's hospital, so the analysis may or may not generally apply to other hospitals. However, in a separate study, we analyzed data from the Pediatric Health Information System database, and observed similar patterns.18 In addition, the proportion of elective patients shown in this study was similar to the national data in Kids Inpatient Database (KID, about 15% of all admissions elective).36 Moreover, the methods are reproducible for other settings, which would be useful to clinical and hospital leadership. Second, the trends depicted in the data only reflected data for one year. Third, coding of the admission as emergent or elective was done by registrars at or before arrival and may not reflect actual clinical need. In addition, those admissions coded as elective included a heterogeneous population (eg, chemotherapy to research studies).
Further studies should analyze trends for other hospitals and evaluate the effect of high peak census and high levels of variation with quality, safety, efficiency, patient satisfaction, financial, and educational outcomes for those receiving care, working, or learning at hospitals. In addition, a qualitative study that develops insights into clinician and patient/parent preferences would help answer questions in regard to usage of weekends for scheduled patients.
Conclusions
Scheduled admissions drive most variability in day‐to‐day occupancy despite the fact that they are a smaller proportion of the inpatient population. Variation in scheduled admissions by day of week provides hospitals with an opportunity to address crowding without adding beds or delaying admissions. Rather, fully utilizing capacity by smoothing occupancy over all days of the week can reduce the risk of high occupancy and thereby improve accessibility and patient/parent satisfaction. While family and staff preferences need to be considered, some combination of within‐week smoothing and shifting admissions towards summer are likely to achieve dramatic improvements in patient flow without large expenditures of capital. The key, then, is to ensure that organizational dynamic factors support these changes, so that staff members do not become stressed working at a 7‐day facility. Taken together, these strategies would better match revenue to capacity, and ultimately increase the quality and value of healthcare operations.
Acknowledgements
Authors' contributions: Study concept and design: Fieldston, Ragavan. Analysis and interpretation of data: Ragavan, Fieldston, Jayaraman, Pati. Drafting of the manuscript: Ragavan, Fieldston. Critical Revision of the manuscript for important intellectual content: Fieldston, Ragavan, Pati, Metlay. Statistical analysis: Fieldston, Jayaraman, Ragavan, Allebach. Study supervision: Fieldston, Pati, Metlay.
Additional contributions: The authors the fellows and faculty of the Robert Wood Johnson Foundation Clinical Scholars Program at the University of Pennsylvania and members of its Community Advisory Board for their suggestions to this work. They also wish to thank Tracy Kish, Jennifer Massenburg, and Brian Smith for assistance with access to and interpretation of hospital census and bed capacity data.
- ,,,,,.Relationship between resident workload and self‐perceived learning on inpatient medicine wards: a longitudinal study.BMC Med Educ.2006;6(1):35.
- AHA Solutions, Patient Flow Challenges Assessment 2009. Chicago, IL.2009.
- ,.Patient flow in hospitals: understanding and controlling it better.Front Health Serv Manage.2004;20:3–15.
- . Managing Variability in Patient Flow is the Key to Improving Access to Care, Nursing Staffing, Quality of Care, and Reducing Its Cost. Paper presented at: Institute of Medicine; June 24,2004.
- ,,,,,.Managing unnecessary variability in patient demand to reduce nursing stress and improve patient safety.Jt Comm J Qual Patient Saf.2005;31(6):330–338.
- ,,.Developing models for patient flow and daily surge capacity research.Acad Emerg Med.2006;13(11):1109–1113.
- Institute for Healthcare Improvement, Flow initiatives. 2008. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed June2010.
- ,,, et al.Hospital workload and adverse events.Med Care.2007;45(5):448–455.
- ,,,,.Impact of admission‐day crowding on the length of stay of pediatric hospitalizations.Pediatrics.2008;121(4):e718–730.
- ,,,,.Emergency department crowding, Part 1: concept, causes, and moral consequences.Ann Emerg Med.2009;53(5):605–611.
- ,.Emergency department overcrowding and ambulance diversion: the impact and potential solutions of extended boarding of admitted patients in the emergency department.J Emerg Med.2006;30(3):351–356.
- ,,,.A comparison of in‐hospital mortality risk conferred by high hospital occupancy, differences in nurse staffing levels, weekend admission, and seasonal influenza.Medical Care.2010;48(3):224–232.
- .,,.The effect of hospital bed occupancy on throughput in the pediatric emergency department.Ann Emerg Med.2009;53(6):767–776.e763.
- .Interpreting the Volume‐Outcome Relationship in the Context of Health Care Quality: Workshop Summary.Washington, DC:National Academies Press;2000.
- ,.Uncertain demand, the structure of hospital costs, and the cost of empty hospital beds.J Health Econ.1995;14(3):291–317.
- ,,, et al.Variability in surgical caseload and access to intensive care services.Anesthesiology.2003;98(6):1491–1496.
- ,,,,.Effects of hospital care environment on patient mortality and nurse outcomes.J Nurs Adm.2009;39(7/8):S45–S51.
- ,,, et al.Children's hospitals do not acutely respond to high occupancy.Pediatrics.2010;125:974–981.
- ,.Easing the strain on a pediatric tertiary care center: use of a redistribution system.Arch Pediatr Adolesc Med.2007;161(9):870–876.
- ,,.Lengths of stay and costs associated with Children's Hospitals.Pediatrics.2005;115:839–844.
- ,.Matching Supply with Demand: An introduction to operations management.New York:McGraw‐Hill;2006.
- .Do hospitals provide lower quality care on weekends?Health Serv Res.2007;42:1589–1612.
- ,.Waiting for urgent procedures on the weekend among emergently hospitalized patients.Am J Med.2004;117:175–181.
- ,.Mortality among patients admitted to hospitals on weekends as compared with weekdays.N Engl J Med.2001;345:663–668.
- ,.Enhanced weekend service: an affordable means to increased hospital procedure volume.CMAJ.2005;172(4):503–504.
- .Hospital deaths and weekend admissions‐how do we leap across a chasm?Clin Nurse Spec.2002;16:91–92.
- ,.Weekend birth and higher neonatal mortality: a problem of patient acuity or quality of care?J Obstet Gynecol Neonatal Nurs.2003;32(6):724–733.
- ,,, et al.Weekend versus weekday admission and mortality from myocardial infarction.N Engl J Med.2007;356:1099–1109.
- .Weekend admission and treatment of patients with renal colic: a case of avoidable variation?Urology.2009;73(4):720–724.
- ,,.Characteristics of weekday and weekend hospital admissions.HCUP Statistical Brief.2010;87. http://www.hcup‐us.ahrq.gov/reports/statbriefs/sb87.pdf.
- ,. Hospital Costs and the Cost of Empty Hospital Beds (NBER Working Paper No. W3872).1991.
- ,,.The effects of timing of pediatric knee ligament surgery on short‐term academic performance in school‐aged athletes.Am J Sports Med.2009;37(9):1684–1691.
- Institute for Healthcare Improvement, Smoothing Elective Surgical Admissions. Available at: http://www.ihi.org/IHI/Topics/Flow/PatientFlow/EmergingContent/SmoothingElectiveSurgicalAdmissions.htm. Accessed June2010.
- . Institute for Healthcare Improvement, Patient Flow Comments. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed June2010.
- .OPIM 631: Operations Management.Wharton School, University of Pennsylvania.Philadelphia, PA.2008.
- Agency for Healthcare Research and Quality. HCUP Databases, Healthcare Cost and Utilization Project (HCUP). 2008. Available at: www.hcup‐us.ahrq.gov/kidoverview.jsp. Accessed June2010.
- ,,,,,.Relationship between resident workload and self‐perceived learning on inpatient medicine wards: a longitudinal study.BMC Med Educ.2006;6(1):35.
- AHA Solutions, Patient Flow Challenges Assessment 2009. Chicago, IL.2009.
- ,.Patient flow in hospitals: understanding and controlling it better.Front Health Serv Manage.2004;20:3–15.
- . Managing Variability in Patient Flow is the Key to Improving Access to Care, Nursing Staffing, Quality of Care, and Reducing Its Cost. Paper presented at: Institute of Medicine; June 24,2004.
- ,,,,,.Managing unnecessary variability in patient demand to reduce nursing stress and improve patient safety.Jt Comm J Qual Patient Saf.2005;31(6):330–338.
- ,,.Developing models for patient flow and daily surge capacity research.Acad Emerg Med.2006;13(11):1109–1113.
- Institute for Healthcare Improvement, Flow initiatives. 2008. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed June2010.
- ,,, et al.Hospital workload and adverse events.Med Care.2007;45(5):448–455.
- ,,,,.Impact of admission‐day crowding on the length of stay of pediatric hospitalizations.Pediatrics.2008;121(4):e718–730.
- ,,,,.Emergency department crowding, Part 1: concept, causes, and moral consequences.Ann Emerg Med.2009;53(5):605–611.
- ,.Emergency department overcrowding and ambulance diversion: the impact and potential solutions of extended boarding of admitted patients in the emergency department.J Emerg Med.2006;30(3):351–356.
- ,,,.A comparison of in‐hospital mortality risk conferred by high hospital occupancy, differences in nurse staffing levels, weekend admission, and seasonal influenza.Medical Care.2010;48(3):224–232.
- .,,.The effect of hospital bed occupancy on throughput in the pediatric emergency department.Ann Emerg Med.2009;53(6):767–776.e763.
- .Interpreting the Volume‐Outcome Relationship in the Context of Health Care Quality: Workshop Summary.Washington, DC:National Academies Press;2000.
- ,.Uncertain demand, the structure of hospital costs, and the cost of empty hospital beds.J Health Econ.1995;14(3):291–317.
- ,,, et al.Variability in surgical caseload and access to intensive care services.Anesthesiology.2003;98(6):1491–1496.
- ,,,,.Effects of hospital care environment on patient mortality and nurse outcomes.J Nurs Adm.2009;39(7/8):S45–S51.
- ,,, et al.Children's hospitals do not acutely respond to high occupancy.Pediatrics.2010;125:974–981.
- ,.Easing the strain on a pediatric tertiary care center: use of a redistribution system.Arch Pediatr Adolesc Med.2007;161(9):870–876.
- ,,.Lengths of stay and costs associated with Children's Hospitals.Pediatrics.2005;115:839–844.
- ,.Matching Supply with Demand: An introduction to operations management.New York:McGraw‐Hill;2006.
- .Do hospitals provide lower quality care on weekends?Health Serv Res.2007;42:1589–1612.
- ,.Waiting for urgent procedures on the weekend among emergently hospitalized patients.Am J Med.2004;117:175–181.
- ,.Mortality among patients admitted to hospitals on weekends as compared with weekdays.N Engl J Med.2001;345:663–668.
- ,.Enhanced weekend service: an affordable means to increased hospital procedure volume.CMAJ.2005;172(4):503–504.
- .Hospital deaths and weekend admissions‐how do we leap across a chasm?Clin Nurse Spec.2002;16:91–92.
- ,.Weekend birth and higher neonatal mortality: a problem of patient acuity or quality of care?J Obstet Gynecol Neonatal Nurs.2003;32(6):724–733.
- ,,, et al.Weekend versus weekday admission and mortality from myocardial infarction.N Engl J Med.2007;356:1099–1109.
- .Weekend admission and treatment of patients with renal colic: a case of avoidable variation?Urology.2009;73(4):720–724.
- ,,.Characteristics of weekday and weekend hospital admissions.HCUP Statistical Brief.2010;87. http://www.hcup‐us.ahrq.gov/reports/statbriefs/sb87.pdf.
- ,. Hospital Costs and the Cost of Empty Hospital Beds (NBER Working Paper No. W3872).1991.
- ,,.The effects of timing of pediatric knee ligament surgery on short‐term academic performance in school‐aged athletes.Am J Sports Med.2009;37(9):1684–1691.
- Institute for Healthcare Improvement, Smoothing Elective Surgical Admissions. Available at: http://www.ihi.org/IHI/Topics/Flow/PatientFlow/EmergingContent/SmoothingElectiveSurgicalAdmissions.htm. Accessed June2010.
- . Institute for Healthcare Improvement, Patient Flow Comments. Available at: http://www.ihi.org/IHI/Topics/Flow. Accessed June2010.
- .OPIM 631: Operations Management.Wharton School, University of Pennsylvania.Philadelphia, PA.2008.
- Agency for Healthcare Research and Quality. HCUP Databases, Healthcare Cost and Utilization Project (HCUP). 2008. Available at: www.hcup‐us.ahrq.gov/kidoverview.jsp. Accessed June2010.
Copyright © 2010 Society of Hospital Medicine
