Jeffrey J. Glasheen, MD

METHODS

RESULTS

DISCUSSION

METHODS

RESULTS

DISCUSSION

METHODS

RESULTS

DISCUSSION

On January 1, 2014, several major provisions of the Affordable Care Act (ACA) took effect, including introduction of the individual mandate for health insurance coverage, opening of the Health Insurance Marketplace, and expansion of Medicaid eligibility to Americans earning up to 133% of the federal poverty level.[1] Nearly 9 million US adults have enrolled in Medicaid since that time, primarily in the 31 states and Washington, DC that have opted into Medicaid expansion.[2, 3] ACA implementation has also had a significant impact on hospital payer mix, primarily by reducing the volume of uncompensated care in Medicaid‐expansion states.[4, 5]

The differential shift in payer mix in Medicaid‐expansion versus nonexpansion states may be relevant to hospitals beyond reimbursement. Medicaid insurance has historically been associated with longer hospitalizations and higher in‐hospital mortality in diverse patient populations, more so than commercial insurance and often even uninsured payer status.[6, 7, 8, 9, 10, 11, 12, 13, 14, 15] The disparity in outcomes between patients with Medicaid versus other insurance persists even after adjustment for disease severity and baseline comorbidities. Insurance type may influence the delivery of inpatient care through variation in access to invasive procedures and adherence to guideline‐concordant medical therapies.[9, 10, 11, 12] Medicaid patients may be more likely than uninsured patients to remain hospitalized pending postacute care placement rather than be discharged home with family support.[16] Medicaid patients are also less likely to leave against medical advice than uninsured patients.[17]

Currently, little is known about the impact of state Medicaid expansion status on length of stay (LOS) or mortality nationally. It is possible that hospitals in Medicaid‐expansion states have experienced relative worsening in LOS and mortality as their share of Medicaid patients has grown. Determining the impact of ACA implementation on payer mix and patient outcomes is particularly important for academic medical centers (AMCs), as they traditionally care for the greatest percentage of both Medicaid and uninsured patients.[18] We sought to characterize the impact of state Medicaid expansion status on payer mix, LOS, and in‐hospital mortality for general medicine patients at AMCs in the United States.

METHODS

The University HealthSystem Consortium (UHC) is an alliance of 117 AMCs and 310 affiliated hospitals, representing >90% of such institutions in the US. We queried the online UHC Clinical Data Base/Resource Manager (CDB/RM) to obtain hospital‐level insurance, LOS, and mortality data for inpatients discharged from a general medicine service between October 1, 2012 and September 30, 2015. We excluded hospitals that were missing data for any month within the study period. No patient‐level data were accessed.

Our outcomes of interest were the proportion of discharges by primary payer (Medicare, commercial, Medicaid, uninsured, or other [eg, Tri‐Care or Workers' Compensation]), as well as the LOS index and mortality index. Both indices were defined as the ratio of the observed to expected values. To determine the expected LOS and mortality, the UHC 2015 risk adjustment models were applied to all cases, adjusting for variables such as patient demographics, low socioeconomic status, admit source and status, severity of illness, and comorbid conditions, as described by International Classification of Diseases, Ninth Revision codes. These models have been validated and are used for research and quality benchmarking for member institutions.[19]

We next stratified hospitals according to state Medicaid expansion status. We defined Medicaid‐expansion states as those that had expanded Medicaid by the end of the study period: Arizona, Arkansas, California, Colorado, Connecticut, Illinois, Indiana, Iowa, Kentucky, Maryland, Massachusetts, Michigan, Minnesota, Nevada, New Hampshire, New Jersey, New Mexico, New York, Ohio, Oregon, Pennsylvania, Rhode Island, Washington, Washington DC, and West Virginia. Nonexpansion states included Alabama, Florida, Georgia, Kansas, Louisiana, Missouri, Nebraska, North Carolina, South Carolina, Tennessee, Texas, Utah, Virginia, and Wisconsin. We excluded 12 states due to incomplete data: Alaska, Delaware, Hawaii, Idaho, North Dakota, Maine, Mississippi, Montana, Oklahoma, South Dakota, Vermont, and Wyoming.

We then identified our pre‐ and post‐ACA implementation periods. Medicaid coverage expansion took effect in all expansion states on January 1, 2014, with the exception of Michigan (April 1, 2014), New Hampshire (August 15, 2014), Pennsylvania (January 1, 2015), and Indiana (February 1, 2015).[3] We therefore defined October 1, 2012 to December 31, 2013 as the pre‐ACA implementation period and January 1, 2014 to September 30, 2015 as the post‐ACA implementation period for all states except for Michigan, New Hampshire, Pennsylvania, and Indiana. For these 4 states, we customized the pre‐ and post‐ACA implementation periods to their respective dates of Medicaid expansion; for New Hampshire, we designated October 1, 2012 to July 31, 2014 as the pre‐ACA implementation period and September 1, 2014 to September 30, 2015 as the post‐ACA implementation period, as we were unable to distinguish before versus after data in August 2014 based on the midmonth expansion of Medicaid.

After stratifying hospitals into groups based on whether they were located in Medicaid‐expansion or nonexpansion states, the proportion of discharges by payer was compared between pre‐ and post‐ACA implementation periods both graphically by quarter and using linear regression models weighted for the number of cases from each hospital. Next, for both Medicaid‐expansion and nonexpansion hospitals, LOS index and mortality index were compared before and after ACA implementation using linear regression models weighted for the number of cases from each hospital, both overall and by payer. Difference‐in‐differences estimations were then completed to compare the proportion of discharges by payer, LOS index, and mortality index between Medicaid‐expansion and nonexpansion hospitals before and after ACA implementation. Post hoc linear regression analyses were completed to evaluate the effect of clustering by state level strata on payer mix and LOS and mortality indices. A 2‐sided P value of <0.05 was considered statistically significant. Data analyses were performed using Stata 12.0 (StataCorp, College Station, TX).

RESULTS

We identified 4,258,952 discharges among general medicine patients from 211 hospitals in 38 states and Washington, DC between October 1, 2012, and September 30, 2015. This included 3,144,488 discharges from 156 hospitals in 24 Medicaid‐expansion states and Washington, DC and 1,114,464 discharges from 55 hospitals in 14 nonexpansion states.

Figure 1 shows the trends in payer mix over time for hospitals in both Medicaid‐expansion and nonexpansion states. As summarized in Table 1, hospitals in Medicaid‐expansion states experienced a significant 3.7‐percentage point increase in Medicaid discharges (P = 0.013) and 2.9‐percentage point decrease in uninsured discharges (P < 0.001) after ACA implementation. This represented an approximately 19% jump and 60% drop in Medicaid and uninsured discharges, respectively. Hospitals in nonexpansion states saw no significant change in the proportion of discharges by payer after ACA implementation. In the difference‐in‐differences analysis, there was a trend toward a greater change in the proportion of Medicaid discharges pre‐ to post‐ACA implementation among hospitals in Medicaid‐expansion states compared to hospitals in nonexpansion states (mean difference‐in‐differences 4.1%, 95% confidence interval [CI]: 0.3%, 8.6%, P = 0.070).

Figure 1

Table 1 shows that the overall LOS index remained unchanged pre‐ to post‐ACA implementation for both Medicaid‐expansion (1.017 to 1.006, P = 0.488) and nonexpansion hospitals (1.008 to 0.995, P = 0.574). LOS indices for each payer type also remained unchanged. The overall mortality index significantly improved pre‐ to post‐ACA implementation for both Medicaid‐expansion (1.000 to 0.878, P < 0.001) and nonexpansion hospitals (0.997 to 0.850, P = 0.001). Among both Medicaid‐expansion and nonexpansion hospitals, the mortality index significantly improved for Medicare, commercial, and Medicaid discharges but not for uninsured or other discharges. In the difference‐in‐differences analysis, the changes in LOS indices and mortality indices pre‐ to post‐ACA implementation did not differ significantly between hospitals in Medicaid‐expansion versus nonexpansion states.

In post hoc linear regression analyses of payer mix and LOS and mortality indices clustered by state‐level strata, point estimates were minimally changed. Although 95% CIs were slightly wider, statistical significance was unchanged from our primary analyses (data not shown).

DISCUSSION

We found that ACA implementation had a significant impact on payer mix for general medicine patients at AMCs in the United States, primarily by increasing the number of Medicaid beneficiaries and by decreasing the number of uninsured patients in Medicaid‐expansion states. State Medicaid expansion status did not appear to influence either LOS or in‐hospital mortality.

Our study offers some of the longest‐term data currently available on the impact of ACA implementation on payer mix trends and encompasses more states than others have previously. Although we uniquely focused on general medicine patients at AMCs, our results are similar to those seen for US hospitals overall. Nikpay and colleagues evaluated payer mix trends for non‐Medicare adult inpatient stays in 16 states through the second quarter of 2014 using the Healthcare Cost and Utilization Project database through the Agency for Healthcare Research and Quality.[4] They found a relative 20% increase and 50% decrease in Medicaid and uninsured discharges in Medicaid‐expansion states, along with nonsignificant changes in nonexpansion states. Hempstead and Cantor assessed payer mix for non‐Medicare discharges using state hospital association data from 21 states through the fourth quarter of 2014 and found a significant increase in Medicaid patients as well as a nearly significant decrease in uninsured patients in expansion states relative to nonexpansion states.[5] The Department of Health and Human Services also reported that uninsured/self‐pay discharges fell substantially (65%73%) in Medicaid‐expansion states by the end of 2014, with slight decreases in nonexpansion states.[20]

In contrast to our hypothesis, the overall LOS and in‐hospital mortality indices were not influenced by state Medicaid expansion status. From a purely mathematical standpoint, the contribution of Medicaid patients to the overall LOS and mortality indices may have been eclipsed by Medicare and commercially insured patients, who represented a higher proportion of total discharges. The lack of impact of state Medicaid expansion status on overall LOS and mortality indices did not appear to occur as a result of indices for Medicaid patients trending toward the mean. As predicted based on observational studies, Medicaid patients in our study tended to have a higher LOS index than those with other insurance types. Medicaid patients actually tended to have a lower mortality index in our analysis; the reason for this latter finding is unclear and in contrast to other published studies.[6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 21]

To our knowledge, no other studies have evaluated the effect of payer mix changes under the ACA on inpatient outcomes. However, new evidence is emerging on outpatient outcomes. Low‐income adults in Medicaid‐expansion states have reported greater gains in access to primary care services and in the diagnosis of certain chronic health conditions than those in nonexpansion states as a result of ACA implementation.[22, 23] Such improvements in the outpatient setting might be expected to reduce patient acuity on admission. However, they would not necessarily translate to relative improvements in LOS or mortality indices for Medicaid‐expansion hospitals, as the UHC risk adjustment models controlled for disease severity on admission.

Similarly, few studies have assessed the impact of payer mix changes under previous state Medicaid expansions on inpatient outcomes. After Massachusetts expanded Medicaid and enacted near‐universal healthcare coverage in 2006, a minimal LOS reduction of just 0.05 days was observed.[24] New York expanded Medicaid eligibility to nondisabled childless adults with incomes below 100% of the federal poverty level in September 2001, whereas Arizona did so in November 2001 and Maine in October 2002. A study comparing outcomes in these states to 4 neighboring nonexpansion states found a relative reduction in annual all‐cause mortality of 6.1% population wide; however, it did not assess in‐hospital mortality.[25] The Oregon Health Insurance Experiment that randomized low‐income adults to expanded Medicaid coverage or not in 2008 has also reported on outpatient rather than inpatient outcomes.[26]

Our findings have potential implications for health policymakers. That Medicaid expansion status had a neutral effect on both LOS and mortality indices in our analysis should be reassuring for states contemplating Medicaid expansion in the future. Our results also highlight the need for further efforts to reduce disparities in inpatient care based on payer status. For example, although Medicare, commercially insured, and Medicaid patients witnessed significant improvements in mortality indices pre‐ to post‐ACA implementation in hospitals in both Medicaid‐expansion and nonexpansion states, uninsured patients did not.

This study has several limitations. First, our analysis of the impact of ACA implementation on payer mix did not account for concurrent socioeconomic trends that may have influenced insurance coverage across the United States. However, the main goal of this analysis was to demonstrate that changes in payer mix did in fact occur over time, to provide rationale for our subsequent LOS and mortality analyses. Second, we could not control for variation in the design and implementation of Medicaid expansions across states as permitted under the federal Section 1115 waiver process. Third, we only had access to hospital‐level data through the UHC CDB/RM, rather than individual patient data. We attempted to mitigate this limitation by weighting data according to the number of cases per hospital. Lastly, additional patient‐level factors that may influence LOS or mortality may not be included in the UHC risk adjustment models.

In summary, the differential shift in payer mix between Medicaid‐expansion and nonexpansion states did not influence overall LOS or in‐hospital mortality for general medicine patients at AMCs in the United States. Additional research could help to determine the impact of ACA implementation on other patient outcomes that may be dependent on insurance status, such as readmissions or hospital‐acquired complications.

Disclosures: M.E.A. conceived of the study concept and design, assisted with data acquisition, and drafted the manuscript. J.J.G. assisted with study design and made critical revisions to the manuscript. D.A. assisted with study design and made critical revisions to the manuscript. R.P. assisted with study design and made critical revisions to the manuscript. M.L. assisted with study design and data acquisition and made critical revisions to the manuscript. C.D.J. assisted with study design, performed data analyses, and made critical revisions to the manuscript. A modified abstract was presented in poster format at the University HealthSystem Consortium Annual Conference held September 30 to October 2, 2015 in Orlando, Florida, as well as at the Society of Hospital Medicine Research, Innovations, and Vignettes 2016 Annual Meeting held March 69, 2016, in San Diego, California. The authors report no conflicts of interest.

On January 1, 2014, several major provisions of the Affordable Care Act (ACA) took effect, including introduction of the individual mandate for health insurance coverage, opening of the Health Insurance Marketplace, and expansion of Medicaid eligibility to Americans earning up to 133% of the federal poverty level.[1] Nearly 9 million US adults have enrolled in Medicaid since that time, primarily in the 31 states and Washington, DC that have opted into Medicaid expansion.[2, 3] ACA implementation has also had a significant impact on hospital payer mix, primarily by reducing the volume of uncompensated care in Medicaid‐expansion states.[4, 5]

The differential shift in payer mix in Medicaid‐expansion versus nonexpansion states may be relevant to hospitals beyond reimbursement. Medicaid insurance has historically been associated with longer hospitalizations and higher in‐hospital mortality in diverse patient populations, more so than commercial insurance and often even uninsured payer status.[6, 7, 8, 9, 10, 11, 12, 13, 14, 15] The disparity in outcomes between patients with Medicaid versus other insurance persists even after adjustment for disease severity and baseline comorbidities. Insurance type may influence the delivery of inpatient care through variation in access to invasive procedures and adherence to guideline‐concordant medical therapies.[9, 10, 11, 12] Medicaid patients may be more likely than uninsured patients to remain hospitalized pending postacute care placement rather than be discharged home with family support.[16] Medicaid patients are also less likely to leave against medical advice than uninsured patients.[17]

Currently, little is known about the impact of state Medicaid expansion status on length of stay (LOS) or mortality nationally. It is possible that hospitals in Medicaid‐expansion states have experienced relative worsening in LOS and mortality as their share of Medicaid patients has grown. Determining the impact of ACA implementation on payer mix and patient outcomes is particularly important for academic medical centers (AMCs), as they traditionally care for the greatest percentage of both Medicaid and uninsured patients.[18] We sought to characterize the impact of state Medicaid expansion status on payer mix, LOS, and in‐hospital mortality for general medicine patients at AMCs in the United States.

METHODS

The University HealthSystem Consortium (UHC) is an alliance of 117 AMCs and 310 affiliated hospitals, representing >90% of such institutions in the US. We queried the online UHC Clinical Data Base/Resource Manager (CDB/RM) to obtain hospital‐level insurance, LOS, and mortality data for inpatients discharged from a general medicine service between October 1, 2012 and September 30, 2015. We excluded hospitals that were missing data for any month within the study period. No patient‐level data were accessed.

Our outcomes of interest were the proportion of discharges by primary payer (Medicare, commercial, Medicaid, uninsured, or other [eg, Tri‐Care or Workers' Compensation]), as well as the LOS index and mortality index. Both indices were defined as the ratio of the observed to expected values. To determine the expected LOS and mortality, the UHC 2015 risk adjustment models were applied to all cases, adjusting for variables such as patient demographics, low socioeconomic status, admit source and status, severity of illness, and comorbid conditions, as described by International Classification of Diseases, Ninth Revision codes. These models have been validated and are used for research and quality benchmarking for member institutions.[19]

We next stratified hospitals according to state Medicaid expansion status. We defined Medicaid‐expansion states as those that had expanded Medicaid by the end of the study period: Arizona, Arkansas, California, Colorado, Connecticut, Illinois, Indiana, Iowa, Kentucky, Maryland, Massachusetts, Michigan, Minnesota, Nevada, New Hampshire, New Jersey, New Mexico, New York, Ohio, Oregon, Pennsylvania, Rhode Island, Washington, Washington DC, and West Virginia. Nonexpansion states included Alabama, Florida, Georgia, Kansas, Louisiana, Missouri, Nebraska, North Carolina, South Carolina, Tennessee, Texas, Utah, Virginia, and Wisconsin. We excluded 12 states due to incomplete data: Alaska, Delaware, Hawaii, Idaho, North Dakota, Maine, Mississippi, Montana, Oklahoma, South Dakota, Vermont, and Wyoming.

We then identified our pre‐ and post‐ACA implementation periods. Medicaid coverage expansion took effect in all expansion states on January 1, 2014, with the exception of Michigan (April 1, 2014), New Hampshire (August 15, 2014), Pennsylvania (January 1, 2015), and Indiana (February 1, 2015).[3] We therefore defined October 1, 2012 to December 31, 2013 as the pre‐ACA implementation period and January 1, 2014 to September 30, 2015 as the post‐ACA implementation period for all states except for Michigan, New Hampshire, Pennsylvania, and Indiana. For these 4 states, we customized the pre‐ and post‐ACA implementation periods to their respective dates of Medicaid expansion; for New Hampshire, we designated October 1, 2012 to July 31, 2014 as the pre‐ACA implementation period and September 1, 2014 to September 30, 2015 as the post‐ACA implementation period, as we were unable to distinguish before versus after data in August 2014 based on the midmonth expansion of Medicaid.

After stratifying hospitals into groups based on whether they were located in Medicaid‐expansion or nonexpansion states, the proportion of discharges by payer was compared between pre‐ and post‐ACA implementation periods both graphically by quarter and using linear regression models weighted for the number of cases from each hospital. Next, for both Medicaid‐expansion and nonexpansion hospitals, LOS index and mortality index were compared before and after ACA implementation using linear regression models weighted for the number of cases from each hospital, both overall and by payer. Difference‐in‐differences estimations were then completed to compare the proportion of discharges by payer, LOS index, and mortality index between Medicaid‐expansion and nonexpansion hospitals before and after ACA implementation. Post hoc linear regression analyses were completed to evaluate the effect of clustering by state level strata on payer mix and LOS and mortality indices. A 2‐sided P value of <0.05 was considered statistically significant. Data analyses were performed using Stata 12.0 (StataCorp, College Station, TX).

RESULTS

We identified 4,258,952 discharges among general medicine patients from 211 hospitals in 38 states and Washington, DC between October 1, 2012, and September 30, 2015. This included 3,144,488 discharges from 156 hospitals in 24 Medicaid‐expansion states and Washington, DC and 1,114,464 discharges from 55 hospitals in 14 nonexpansion states.

Figure 1 shows the trends in payer mix over time for hospitals in both Medicaid‐expansion and nonexpansion states. As summarized in Table 1, hospitals in Medicaid‐expansion states experienced a significant 3.7‐percentage point increase in Medicaid discharges (P = 0.013) and 2.9‐percentage point decrease in uninsured discharges (P < 0.001) after ACA implementation. This represented an approximately 19% jump and 60% drop in Medicaid and uninsured discharges, respectively. Hospitals in nonexpansion states saw no significant change in the proportion of discharges by payer after ACA implementation. In the difference‐in‐differences analysis, there was a trend toward a greater change in the proportion of Medicaid discharges pre‐ to post‐ACA implementation among hospitals in Medicaid‐expansion states compared to hospitals in nonexpansion states (mean difference‐in‐differences 4.1%, 95% confidence interval [CI]: 0.3%, 8.6%, P = 0.070).

Figure 1

Table 1 shows that the overall LOS index remained unchanged pre‐ to post‐ACA implementation for both Medicaid‐expansion (1.017 to 1.006, P = 0.488) and nonexpansion hospitals (1.008 to 0.995, P = 0.574). LOS indices for each payer type also remained unchanged. The overall mortality index significantly improved pre‐ to post‐ACA implementation for both Medicaid‐expansion (1.000 to 0.878, P < 0.001) and nonexpansion hospitals (0.997 to 0.850, P = 0.001). Among both Medicaid‐expansion and nonexpansion hospitals, the mortality index significantly improved for Medicare, commercial, and Medicaid discharges but not for uninsured or other discharges. In the difference‐in‐differences analysis, the changes in LOS indices and mortality indices pre‐ to post‐ACA implementation did not differ significantly between hospitals in Medicaid‐expansion versus nonexpansion states.

In post hoc linear regression analyses of payer mix and LOS and mortality indices clustered by state‐level strata, point estimates were minimally changed. Although 95% CIs were slightly wider, statistical significance was unchanged from our primary analyses (data not shown).

DISCUSSION

We found that ACA implementation had a significant impact on payer mix for general medicine patients at AMCs in the United States, primarily by increasing the number of Medicaid beneficiaries and by decreasing the number of uninsured patients in Medicaid‐expansion states. State Medicaid expansion status did not appear to influence either LOS or in‐hospital mortality.

Our study offers some of the longest‐term data currently available on the impact of ACA implementation on payer mix trends and encompasses more states than others have previously. Although we uniquely focused on general medicine patients at AMCs, our results are similar to those seen for US hospitals overall. Nikpay and colleagues evaluated payer mix trends for non‐Medicare adult inpatient stays in 16 states through the second quarter of 2014 using the Healthcare Cost and Utilization Project database through the Agency for Healthcare Research and Quality.[4] They found a relative 20% increase and 50% decrease in Medicaid and uninsured discharges in Medicaid‐expansion states, along with nonsignificant changes in nonexpansion states. Hempstead and Cantor assessed payer mix for non‐Medicare discharges using state hospital association data from 21 states through the fourth quarter of 2014 and found a significant increase in Medicaid patients as well as a nearly significant decrease in uninsured patients in expansion states relative to nonexpansion states.[5] The Department of Health and Human Services also reported that uninsured/self‐pay discharges fell substantially (65%73%) in Medicaid‐expansion states by the end of 2014, with slight decreases in nonexpansion states.[20]

In contrast to our hypothesis, the overall LOS and in‐hospital mortality indices were not influenced by state Medicaid expansion status. From a purely mathematical standpoint, the contribution of Medicaid patients to the overall LOS and mortality indices may have been eclipsed by Medicare and commercially insured patients, who represented a higher proportion of total discharges. The lack of impact of state Medicaid expansion status on overall LOS and mortality indices did not appear to occur as a result of indices for Medicaid patients trending toward the mean. As predicted based on observational studies, Medicaid patients in our study tended to have a higher LOS index than those with other insurance types. Medicaid patients actually tended to have a lower mortality index in our analysis; the reason for this latter finding is unclear and in contrast to other published studies.[6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 21]

To our knowledge, no other studies have evaluated the effect of payer mix changes under the ACA on inpatient outcomes. However, new evidence is emerging on outpatient outcomes. Low‐income adults in Medicaid‐expansion states have reported greater gains in access to primary care services and in the diagnosis of certain chronic health conditions than those in nonexpansion states as a result of ACA implementation.[22, 23] Such improvements in the outpatient setting might be expected to reduce patient acuity on admission. However, they would not necessarily translate to relative improvements in LOS or mortality indices for Medicaid‐expansion hospitals, as the UHC risk adjustment models controlled for disease severity on admission.

Similarly, few studies have assessed the impact of payer mix changes under previous state Medicaid expansions on inpatient outcomes. After Massachusetts expanded Medicaid and enacted near‐universal healthcare coverage in 2006, a minimal LOS reduction of just 0.05 days was observed.[24] New York expanded Medicaid eligibility to nondisabled childless adults with incomes below 100% of the federal poverty level in September 2001, whereas Arizona did so in November 2001 and Maine in October 2002. A study comparing outcomes in these states to 4 neighboring nonexpansion states found a relative reduction in annual all‐cause mortality of 6.1% population wide; however, it did not assess in‐hospital mortality.[25] The Oregon Health Insurance Experiment that randomized low‐income adults to expanded Medicaid coverage or not in 2008 has also reported on outpatient rather than inpatient outcomes.[26]

Our findings have potential implications for health policymakers. That Medicaid expansion status had a neutral effect on both LOS and mortality indices in our analysis should be reassuring for states contemplating Medicaid expansion in the future. Our results also highlight the need for further efforts to reduce disparities in inpatient care based on payer status. For example, although Medicare, commercially insured, and Medicaid patients witnessed significant improvements in mortality indices pre‐ to post‐ACA implementation in hospitals in both Medicaid‐expansion and nonexpansion states, uninsured patients did not.

This study has several limitations. First, our analysis of the impact of ACA implementation on payer mix did not account for concurrent socioeconomic trends that may have influenced insurance coverage across the United States. However, the main goal of this analysis was to demonstrate that changes in payer mix did in fact occur over time, to provide rationale for our subsequent LOS and mortality analyses. Second, we could not control for variation in the design and implementation of Medicaid expansions across states as permitted under the federal Section 1115 waiver process. Third, we only had access to hospital‐level data through the UHC CDB/RM, rather than individual patient data. We attempted to mitigate this limitation by weighting data according to the number of cases per hospital. Lastly, additional patient‐level factors that may influence LOS or mortality may not be included in the UHC risk adjustment models.

In summary, the differential shift in payer mix between Medicaid‐expansion and nonexpansion states did not influence overall LOS or in‐hospital mortality for general medicine patients at AMCs in the United States. Additional research could help to determine the impact of ACA implementation on other patient outcomes that may be dependent on insurance status, such as readmissions or hospital‐acquired complications.

Disclosures: M.E.A. conceived of the study concept and design, assisted with data acquisition, and drafted the manuscript. J.J.G. assisted with study design and made critical revisions to the manuscript. D.A. assisted with study design and made critical revisions to the manuscript. R.P. assisted with study design and made critical revisions to the manuscript. M.L. assisted with study design and data acquisition and made critical revisions to the manuscript. C.D.J. assisted with study design, performed data analyses, and made critical revisions to the manuscript. A modified abstract was presented in poster format at the University HealthSystem Consortium Annual Conference held September 30 to October 2, 2015 in Orlando, Florida, as well as at the Society of Hospital Medicine Research, Innovations, and Vignettes 2016 Annual Meeting held March 69, 2016, in San Diego, California. The authors report no conflicts of interest.

On January 1, 2014, several major provisions of the Affordable Care Act (ACA) took effect, including introduction of the individual mandate for health insurance coverage, opening of the Health Insurance Marketplace, and expansion of Medicaid eligibility to Americans earning up to 133% of the federal poverty level.[1] Nearly 9 million US adults have enrolled in Medicaid since that time, primarily in the 31 states and Washington, DC that have opted into Medicaid expansion.[2, 3] ACA implementation has also had a significant impact on hospital payer mix, primarily by reducing the volume of uncompensated care in Medicaid‐expansion states.[4, 5]

The differential shift in payer mix in Medicaid‐expansion versus nonexpansion states may be relevant to hospitals beyond reimbursement. Medicaid insurance has historically been associated with longer hospitalizations and higher in‐hospital mortality in diverse patient populations, more so than commercial insurance and often even uninsured payer status.[6, 7, 8, 9, 10, 11, 12, 13, 14, 15] The disparity in outcomes between patients with Medicaid versus other insurance persists even after adjustment for disease severity and baseline comorbidities. Insurance type may influence the delivery of inpatient care through variation in access to invasive procedures and adherence to guideline‐concordant medical therapies.[9, 10, 11, 12] Medicaid patients may be more likely than uninsured patients to remain hospitalized pending postacute care placement rather than be discharged home with family support.[16] Medicaid patients are also less likely to leave against medical advice than uninsured patients.[17]

Currently, little is known about the impact of state Medicaid expansion status on length of stay (LOS) or mortality nationally. It is possible that hospitals in Medicaid‐expansion states have experienced relative worsening in LOS and mortality as their share of Medicaid patients has grown. Determining the impact of ACA implementation on payer mix and patient outcomes is particularly important for academic medical centers (AMCs), as they traditionally care for the greatest percentage of both Medicaid and uninsured patients.[18] We sought to characterize the impact of state Medicaid expansion status on payer mix, LOS, and in‐hospital mortality for general medicine patients at AMCs in the United States.

METHODS

The University HealthSystem Consortium (UHC) is an alliance of 117 AMCs and 310 affiliated hospitals, representing >90% of such institutions in the US. We queried the online UHC Clinical Data Base/Resource Manager (CDB/RM) to obtain hospital‐level insurance, LOS, and mortality data for inpatients discharged from a general medicine service between October 1, 2012 and September 30, 2015. We excluded hospitals that were missing data for any month within the study period. No patient‐level data were accessed.

Our outcomes of interest were the proportion of discharges by primary payer (Medicare, commercial, Medicaid, uninsured, or other [eg, Tri‐Care or Workers' Compensation]), as well as the LOS index and mortality index. Both indices were defined as the ratio of the observed to expected values. To determine the expected LOS and mortality, the UHC 2015 risk adjustment models were applied to all cases, adjusting for variables such as patient demographics, low socioeconomic status, admit source and status, severity of illness, and comorbid conditions, as described by International Classification of Diseases, Ninth Revision codes. These models have been validated and are used for research and quality benchmarking for member institutions.[19]

We next stratified hospitals according to state Medicaid expansion status. We defined Medicaid‐expansion states as those that had expanded Medicaid by the end of the study period: Arizona, Arkansas, California, Colorado, Connecticut, Illinois, Indiana, Iowa, Kentucky, Maryland, Massachusetts, Michigan, Minnesota, Nevada, New Hampshire, New Jersey, New Mexico, New York, Ohio, Oregon, Pennsylvania, Rhode Island, Washington, Washington DC, and West Virginia. Nonexpansion states included Alabama, Florida, Georgia, Kansas, Louisiana, Missouri, Nebraska, North Carolina, South Carolina, Tennessee, Texas, Utah, Virginia, and Wisconsin. We excluded 12 states due to incomplete data: Alaska, Delaware, Hawaii, Idaho, North Dakota, Maine, Mississippi, Montana, Oklahoma, South Dakota, Vermont, and Wyoming.

We then identified our pre‐ and post‐ACA implementation periods. Medicaid coverage expansion took effect in all expansion states on January 1, 2014, with the exception of Michigan (April 1, 2014), New Hampshire (August 15, 2014), Pennsylvania (January 1, 2015), and Indiana (February 1, 2015).[3] We therefore defined October 1, 2012 to December 31, 2013 as the pre‐ACA implementation period and January 1, 2014 to September 30, 2015 as the post‐ACA implementation period for all states except for Michigan, New Hampshire, Pennsylvania, and Indiana. For these 4 states, we customized the pre‐ and post‐ACA implementation periods to their respective dates of Medicaid expansion; for New Hampshire, we designated October 1, 2012 to July 31, 2014 as the pre‐ACA implementation period and September 1, 2014 to September 30, 2015 as the post‐ACA implementation period, as we were unable to distinguish before versus after data in August 2014 based on the midmonth expansion of Medicaid.

After stratifying hospitals into groups based on whether they were located in Medicaid‐expansion or nonexpansion states, the proportion of discharges by payer was compared between pre‐ and post‐ACA implementation periods both graphically by quarter and using linear regression models weighted for the number of cases from each hospital. Next, for both Medicaid‐expansion and nonexpansion hospitals, LOS index and mortality index were compared before and after ACA implementation using linear regression models weighted for the number of cases from each hospital, both overall and by payer. Difference‐in‐differences estimations were then completed to compare the proportion of discharges by payer, LOS index, and mortality index between Medicaid‐expansion and nonexpansion hospitals before and after ACA implementation. Post hoc linear regression analyses were completed to evaluate the effect of clustering by state level strata on payer mix and LOS and mortality indices. A 2‐sided P value of <0.05 was considered statistically significant. Data analyses were performed using Stata 12.0 (StataCorp, College Station, TX).

RESULTS

We identified 4,258,952 discharges among general medicine patients from 211 hospitals in 38 states and Washington, DC between October 1, 2012, and September 30, 2015. This included 3,144,488 discharges from 156 hospitals in 24 Medicaid‐expansion states and Washington, DC and 1,114,464 discharges from 55 hospitals in 14 nonexpansion states.

Figure 1 shows the trends in payer mix over time for hospitals in both Medicaid‐expansion and nonexpansion states. As summarized in Table 1, hospitals in Medicaid‐expansion states experienced a significant 3.7‐percentage point increase in Medicaid discharges (P = 0.013) and 2.9‐percentage point decrease in uninsured discharges (P < 0.001) after ACA implementation. This represented an approximately 19% jump and 60% drop in Medicaid and uninsured discharges, respectively. Hospitals in nonexpansion states saw no significant change in the proportion of discharges by payer after ACA implementation. In the difference‐in‐differences analysis, there was a trend toward a greater change in the proportion of Medicaid discharges pre‐ to post‐ACA implementation among hospitals in Medicaid‐expansion states compared to hospitals in nonexpansion states (mean difference‐in‐differences 4.1%, 95% confidence interval [CI]: 0.3%, 8.6%, P = 0.070).

Figure 1

Table 1 shows that the overall LOS index remained unchanged pre‐ to post‐ACA implementation for both Medicaid‐expansion (1.017 to 1.006, P = 0.488) and nonexpansion hospitals (1.008 to 0.995, P = 0.574). LOS indices for each payer type also remained unchanged. The overall mortality index significantly improved pre‐ to post‐ACA implementation for both Medicaid‐expansion (1.000 to 0.878, P < 0.001) and nonexpansion hospitals (0.997 to 0.850, P = 0.001). Among both Medicaid‐expansion and nonexpansion hospitals, the mortality index significantly improved for Medicare, commercial, and Medicaid discharges but not for uninsured or other discharges. In the difference‐in‐differences analysis, the changes in LOS indices and mortality indices pre‐ to post‐ACA implementation did not differ significantly between hospitals in Medicaid‐expansion versus nonexpansion states.

In post hoc linear regression analyses of payer mix and LOS and mortality indices clustered by state‐level strata, point estimates were minimally changed. Although 95% CIs were slightly wider, statistical significance was unchanged from our primary analyses (data not shown).

DISCUSSION

We found that ACA implementation had a significant impact on payer mix for general medicine patients at AMCs in the United States, primarily by increasing the number of Medicaid beneficiaries and by decreasing the number of uninsured patients in Medicaid‐expansion states. State Medicaid expansion status did not appear to influence either LOS or in‐hospital mortality.

Our study offers some of the longest‐term data currently available on the impact of ACA implementation on payer mix trends and encompasses more states than others have previously. Although we uniquely focused on general medicine patients at AMCs, our results are similar to those seen for US hospitals overall. Nikpay and colleagues evaluated payer mix trends for non‐Medicare adult inpatient stays in 16 states through the second quarter of 2014 using the Healthcare Cost and Utilization Project database through the Agency for Healthcare Research and Quality.[4] They found a relative 20% increase and 50% decrease in Medicaid and uninsured discharges in Medicaid‐expansion states, along with nonsignificant changes in nonexpansion states. Hempstead and Cantor assessed payer mix for non‐Medicare discharges using state hospital association data from 21 states through the fourth quarter of 2014 and found a significant increase in Medicaid patients as well as a nearly significant decrease in uninsured patients in expansion states relative to nonexpansion states.[5] The Department of Health and Human Services also reported that uninsured/self‐pay discharges fell substantially (65%73%) in Medicaid‐expansion states by the end of 2014, with slight decreases in nonexpansion states.[20]

In contrast to our hypothesis, the overall LOS and in‐hospital mortality indices were not influenced by state Medicaid expansion status. From a purely mathematical standpoint, the contribution of Medicaid patients to the overall LOS and mortality indices may have been eclipsed by Medicare and commercially insured patients, who represented a higher proportion of total discharges. The lack of impact of state Medicaid expansion status on overall LOS and mortality indices did not appear to occur as a result of indices for Medicaid patients trending toward the mean. As predicted based on observational studies, Medicaid patients in our study tended to have a higher LOS index than those with other insurance types. Medicaid patients actually tended to have a lower mortality index in our analysis; the reason for this latter finding is unclear and in contrast to other published studies.[6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 21]

To our knowledge, no other studies have evaluated the effect of payer mix changes under the ACA on inpatient outcomes. However, new evidence is emerging on outpatient outcomes. Low‐income adults in Medicaid‐expansion states have reported greater gains in access to primary care services and in the diagnosis of certain chronic health conditions than those in nonexpansion states as a result of ACA implementation.[22, 23] Such improvements in the outpatient setting might be expected to reduce patient acuity on admission. However, they would not necessarily translate to relative improvements in LOS or mortality indices for Medicaid‐expansion hospitals, as the UHC risk adjustment models controlled for disease severity on admission.

Similarly, few studies have assessed the impact of payer mix changes under previous state Medicaid expansions on inpatient outcomes. After Massachusetts expanded Medicaid and enacted near‐universal healthcare coverage in 2006, a minimal LOS reduction of just 0.05 days was observed.[24] New York expanded Medicaid eligibility to nondisabled childless adults with incomes below 100% of the federal poverty level in September 2001, whereas Arizona did so in November 2001 and Maine in October 2002. A study comparing outcomes in these states to 4 neighboring nonexpansion states found a relative reduction in annual all‐cause mortality of 6.1% population wide; however, it did not assess in‐hospital mortality.[25] The Oregon Health Insurance Experiment that randomized low‐income adults to expanded Medicaid coverage or not in 2008 has also reported on outpatient rather than inpatient outcomes.[26]

Our findings have potential implications for health policymakers. That Medicaid expansion status had a neutral effect on both LOS and mortality indices in our analysis should be reassuring for states contemplating Medicaid expansion in the future. Our results also highlight the need for further efforts to reduce disparities in inpatient care based on payer status. For example, although Medicare, commercially insured, and Medicaid patients witnessed significant improvements in mortality indices pre‐ to post‐ACA implementation in hospitals in both Medicaid‐expansion and nonexpansion states, uninsured patients did not.

This study has several limitations. First, our analysis of the impact of ACA implementation on payer mix did not account for concurrent socioeconomic trends that may have influenced insurance coverage across the United States. However, the main goal of this analysis was to demonstrate that changes in payer mix did in fact occur over time, to provide rationale for our subsequent LOS and mortality analyses. Second, we could not control for variation in the design and implementation of Medicaid expansions across states as permitted under the federal Section 1115 waiver process. Third, we only had access to hospital‐level data through the UHC CDB/RM, rather than individual patient data. We attempted to mitigate this limitation by weighting data according to the number of cases per hospital. Lastly, additional patient‐level factors that may influence LOS or mortality may not be included in the UHC risk adjustment models.

In summary, the differential shift in payer mix between Medicaid‐expansion and nonexpansion states did not influence overall LOS or in‐hospital mortality for general medicine patients at AMCs in the United States. Additional research could help to determine the impact of ACA implementation on other patient outcomes that may be dependent on insurance status, such as readmissions or hospital‐acquired complications.

Disclosures: M.E.A. conceived of the study concept and design, assisted with data acquisition, and drafted the manuscript. J.J.G. assisted with study design and made critical revisions to the manuscript. D.A. assisted with study design and made critical revisions to the manuscript. R.P. assisted with study design and made critical revisions to the manuscript. M.L. assisted with study design and data acquisition and made critical revisions to the manuscript. C.D.J. assisted with study design, performed data analyses, and made critical revisions to the manuscript. A modified abstract was presented in poster format at the University HealthSystem Consortium Annual Conference held September 30 to October 2, 2015 in Orlando, Florida, as well as at the Society of Hospital Medicine Research, Innovations, and Vignettes 2016 Annual Meeting held March 69, 2016, in San Diego, California. The authors report no conflicts of interest.

Hospital medicine is an emerging specialty comprised predominantly of early‐career faculty, often less than 5 years postresidency and predominately at instructor or assistant professor level.[1] Effective mentoring has been identified as a critical component of academic success.[2, 3] Published data suggest that most academic hospitalists do not have a mentor, and when they do, the majority of them spend less than 4 hours per year with their mentor.[2] The reasons for this are multifactorial but largely result from the lack of structure, opportunities, and local senior academic hospitalists.[1, 4] Early‐career faculty have difficulty establishing external mentoring relationships, and new models beyond the traditional intrainstitutional dyad are needed.[3, 4] The need for mentors and structured mentorship networks may be particularly high in hospital medicine.[5]

The Visiting Professorship in Hospital Medicine Program was designed to promote cross‐institutional mentorship, share hospitalist innovations, and facilitate academic collaboration between hospitalist groups. We describe the design and early experience with this program across 5 academic hospital medicine programs.

PROGRAM DESIGN

RESULTS

To date, 5 academic medical centers have participated in the visiting professorship program, with 7 visiting professors interacting with 29 early‐career faculty. Of the 29 early‐career faculty, 72% (21/29) were at the rank of assistant professor, 17% (5/29) instructor, 7% (2/29) residents with plans to hire, and 3% (1/29) associate professor. The median was 2 years in academic medicine and 1 year at current academic rank. Forty‐one percent (12/29) were women and 7% (2/29) identified as ethnic minority. Of the 7 visiting professors, 57% (4/7) were assistant professor and 43% (3/7) were associate professors. The median was 5 years in academic medicine, 29% (2/7) were women, and none identified as ethnic minority.

Immediate postvisit survey response was obtained for all participating faculty. In the immediate postvisit survey, on a 5‐point Likert scale, the 29 early‐career faculty rated the visit: 4.4 for promoting cross‐institutional dissemination of innovations, 4.2 for advancing my academic career, 4.2 for fostering cross‐institutional mentor‐mentee relationships, and 4.4 for promoting cross‐institutional collaborations. Ninety‐three percent (26/28 accounting for 1 nonresponse to this question) reported the visiting professorship had high potential to disseminate innovation (rated greater than 3 on the 5‐point Likert score). Eighty‐three percent (24/29) of the early‐career faculty rated the visit highly useful in advancing their career, 76% (22/29) responded that the visit was highly likely to foster external mentorship relationships, and 90% (26/29) reported the visit highly effective in promoting cross‐institutional collaborations. In the immediate postvisit survey, the 7 visiting professors rated the visit 4.9 for promoting cross‐institutional dissemination of innovations, 4.3 for advancing my academic career, 4.0 for fostering cross‐institutional mentor‐mentee relationships, and 4.3 for promoting cross‐institutional collaborations.

Free‐text comments from both visiting professors and early‐career faculty were generally favorable (Table 1). Some comments offered constructive input on appropriate matching of faculty, previsit preparation, or desire for more time in sessions (Table 1).

One‐year follow‐up was obtained for all but 1 early‐career faculty member receiving the follow‐up survey, and all 3 visiting professors. Of the 3 visiting professorships that occurred more than 1 year ago, 16 mentorship contacts occurred in total (phone, e‐mail, or in person) between 13 early‐career faculty and 3 visiting professors in the year after the initial visits (range, 04 contacts). Follow‐up contact occurred for 3 of 4 early‐career faculty from the first visiting professorship, 3 of 5 from the second visiting professorship, and 2 of 4 from the third visiting professorship. One early‐career faculty member from each host academic medical center had 3 or more additional contacts with the visiting professor in the year following the initial visit. Overall, 8/13 (62%) of early‐career faculty had at least 1 follow‐up mentoring discussion. On 1‐year follow‐up, overall utility for professional development was rated an average of 3.5 by early‐career faculty (with a trend of higher ratings of efficacy with increasing number of follow‐up contacts) and 4.7 by visiting professors. Half (8/16) of the involved faculty report having seen evidence of cross‐institutional dissemination of innovation. Ninety‐four percent (15/16) of participants at 1‐year follow‐up felt there was benefit to their institution in continuing the program for the next year.

Objective evidence of cross‐institutional scholarship, assessed by email query of both visiting professors and senior hospitalists coordinating the visits, includes 2 collaborative peer reviewed publications including mentors and mentees participating in the visiting professorship.[6, 7] Joint educational curriculum development on high‐value care between sites is planned. The Visiting Professorship in Hospital Medicine Program has resulted in 1 external letter to support a visiting professor's promotion to date.

Figure 1

DISCUSSION

Hospital Medicine is a young, rapidly growing field, hence the number of experienced academic hospitalist mentors with expertise in successfully navigating an academic career is limited. A national study of hospitalist leaders found that 75% of clinician‐educators and 58% of research faculty feel that lack of mentorship is a major issue.[1] Mentorship for hospitalist clinician‐investigators is often delivered by nonhospitalists.[2, 8] There is little evidence of external mentorship for academic clinician‐educators in hospital medicine.[1] Without explicit programmatic support, many faculty may find this to be a barrier to career advancement. A study of successfully promoted hospitalists identified difficulty identifying external senior hospitalists to write letters in support of promotion as an obstacle.[9] Our study of the Visiting Professorship in Hospital Medicine Program found that early‐career faculty rated the visit as useful in advancing their career and fostered external mentorship relationships. Subsequent experience suggests more than half of the early‐career faculty will maintain contact with the visiting professor over the year following the visit. Visiting professors rate the experience particularly highly in their own career advancement.

The hospitalist movement is built on a foundation of innovation. The focus of each presentation was on an innovation developed by the visiting professor, and each visit showcased an innovation of the visited institution. This is distinct from traditional Hospital Grand Rounds, which more often focus on basic science research or clinical pathophysiology/disease management based on subspecialty topics.[10] The Visiting Professorship in Hospital Medicine Program was judged by participants to be an effective means of spreading innovation.

Insights from experience with the Visiting Professorship in Hospital Medicine Program include the importance of preliminary work prior to each visit. Program directors need to attend closely to the fit between the interests and career path of the visiting professor and those of the early‐career faculty. The innovations being shared should be aligned with organizational interests to maximize the chance of subsequent spread of the innovation and future collaboration. Providing faculty information about the objectives of the program in advance of the visit and arranging an exchange of curricula vitae between the early‐career faculty and the visiting professor allows participants to prepare for the in‐person coaching. Based on comments from participants, prompting contact from the visiting professor after the visit may be helpful to initiate the longitudinal relationship. We also found that early‐career faculty may not be aware of how to effectively use a mentoring relationship with an external faculty member. Training sessions for both mentors and mentees on effective mentorship relationships before visiting professorships might improve early‐career faculty confidence in initiating relationships and maximize value from mentor coaching.

A key issue is finding the right level of career maturity for the visiting professor. Our approach in selecting visiting professors was congruent with utilization of midcareer peer coaches employed by intrainstitutional hospital medicine mentoring programs.[11] The visiting professor should have sufficient experience and accomplishments to be able to effectively counsel junior faculty. However, it is important that the visiting professor also has sufficient time and interest to take on additional mentees and to be a full participant in shared scholarship projects emerging from the experience.

This study represents the experience of 5 mature academic hospitalist groups, and results may not be generalizable to dissimilar institutions or if only the most senior faculty are selected to perform visits. There is an inherent selection bias in the choice of both visiting professor and early‐career faculty. The small sample size of the faculty exposed to this program is a limitation to generalizability of the results of this evaluation. Whether this program will result in greater success in promotion of academic hospitalists cannot be assessed based on the follow‐up available. The Visiting Professorship in Hospital Medicine Program has continued to be sustained with an additional academic medical center enrolled and 2 additional site visits planned. The costs of the program are low, largely air travel and a night of lodging, as well as nominal administrative logistical support. Perceived benefits by participants and academic medical centers make this modest investment worth considering for academic hospitalist groups.

CONCLUSIONS

The Visiting Professorship in Hospital Medicine Program offers structure, opportunities, and access to senior mentors to advance the development of early‐career hospitalists while spreading innovation to distant sites. It is assessed by participants to facilitate external mentoring relationships and has the potential to advance the careers of both early‐career faculty as well as the visiting professors.

Hospital medicine is an emerging specialty comprised predominantly of early‐career faculty, often less than 5 years postresidency and predominately at instructor or assistant professor level.[1] Effective mentoring has been identified as a critical component of academic success.[2, 3] Published data suggest that most academic hospitalists do not have a mentor, and when they do, the majority of them spend less than 4 hours per year with their mentor.[2] The reasons for this are multifactorial but largely result from the lack of structure, opportunities, and local senior academic hospitalists.[1, 4] Early‐career faculty have difficulty establishing external mentoring relationships, and new models beyond the traditional intrainstitutional dyad are needed.[3, 4] The need for mentors and structured mentorship networks may be particularly high in hospital medicine.[5]

The Visiting Professorship in Hospital Medicine Program was designed to promote cross‐institutional mentorship, share hospitalist innovations, and facilitate academic collaboration between hospitalist groups. We describe the design and early experience with this program across 5 academic hospital medicine programs.

PROGRAM DESIGN

RESULTS

To date, 5 academic medical centers have participated in the visiting professorship program, with 7 visiting professors interacting with 29 early‐career faculty. Of the 29 early‐career faculty, 72% (21/29) were at the rank of assistant professor, 17% (5/29) instructor, 7% (2/29) residents with plans to hire, and 3% (1/29) associate professor. The median was 2 years in academic medicine and 1 year at current academic rank. Forty‐one percent (12/29) were women and 7% (2/29) identified as ethnic minority. Of the 7 visiting professors, 57% (4/7) were assistant professor and 43% (3/7) were associate professors. The median was 5 years in academic medicine, 29% (2/7) were women, and none identified as ethnic minority.

Immediate postvisit survey response was obtained for all participating faculty. In the immediate postvisit survey, on a 5‐point Likert scale, the 29 early‐career faculty rated the visit: 4.4 for promoting cross‐institutional dissemination of innovations, 4.2 for advancing my academic career, 4.2 for fostering cross‐institutional mentor‐mentee relationships, and 4.4 for promoting cross‐institutional collaborations. Ninety‐three percent (26/28 accounting for 1 nonresponse to this question) reported the visiting professorship had high potential to disseminate innovation (rated greater than 3 on the 5‐point Likert score). Eighty‐three percent (24/29) of the early‐career faculty rated the visit highly useful in advancing their career, 76% (22/29) responded that the visit was highly likely to foster external mentorship relationships, and 90% (26/29) reported the visit highly effective in promoting cross‐institutional collaborations. In the immediate postvisit survey, the 7 visiting professors rated the visit 4.9 for promoting cross‐institutional dissemination of innovations, 4.3 for advancing my academic career, 4.0 for fostering cross‐institutional mentor‐mentee relationships, and 4.3 for promoting cross‐institutional collaborations.

Free‐text comments from both visiting professors and early‐career faculty were generally favorable (Table 1). Some comments offered constructive input on appropriate matching of faculty, previsit preparation, or desire for more time in sessions (Table 1).

One‐year follow‐up was obtained for all but 1 early‐career faculty member receiving the follow‐up survey, and all 3 visiting professors. Of the 3 visiting professorships that occurred more than 1 year ago, 16 mentorship contacts occurred in total (phone, e‐mail, or in person) between 13 early‐career faculty and 3 visiting professors in the year after the initial visits (range, 04 contacts). Follow‐up contact occurred for 3 of 4 early‐career faculty from the first visiting professorship, 3 of 5 from the second visiting professorship, and 2 of 4 from the third visiting professorship. One early‐career faculty member from each host academic medical center had 3 or more additional contacts with the visiting professor in the year following the initial visit. Overall, 8/13 (62%) of early‐career faculty had at least 1 follow‐up mentoring discussion. On 1‐year follow‐up, overall utility for professional development was rated an average of 3.5 by early‐career faculty (with a trend of higher ratings of efficacy with increasing number of follow‐up contacts) and 4.7 by visiting professors. Half (8/16) of the involved faculty report having seen evidence of cross‐institutional dissemination of innovation. Ninety‐four percent (15/16) of participants at 1‐year follow‐up felt there was benefit to their institution in continuing the program for the next year.

Objective evidence of cross‐institutional scholarship, assessed by email query of both visiting professors and senior hospitalists coordinating the visits, includes 2 collaborative peer reviewed publications including mentors and mentees participating in the visiting professorship.[6, 7] Joint educational curriculum development on high‐value care between sites is planned. The Visiting Professorship in Hospital Medicine Program has resulted in 1 external letter to support a visiting professor's promotion to date.

Figure 1

DISCUSSION

Hospital Medicine is a young, rapidly growing field, hence the number of experienced academic hospitalist mentors with expertise in successfully navigating an academic career is limited. A national study of hospitalist leaders found that 75% of clinician‐educators and 58% of research faculty feel that lack of mentorship is a major issue.[1] Mentorship for hospitalist clinician‐investigators is often delivered by nonhospitalists.[2, 8] There is little evidence of external mentorship for academic clinician‐educators in hospital medicine.[1] Without explicit programmatic support, many faculty may find this to be a barrier to career advancement. A study of successfully promoted hospitalists identified difficulty identifying external senior hospitalists to write letters in support of promotion as an obstacle.[9] Our study of the Visiting Professorship in Hospital Medicine Program found that early‐career faculty rated the visit as useful in advancing their career and fostered external mentorship relationships. Subsequent experience suggests more than half of the early‐career faculty will maintain contact with the visiting professor over the year following the visit. Visiting professors rate the experience particularly highly in their own career advancement.

The hospitalist movement is built on a foundation of innovation. The focus of each presentation was on an innovation developed by the visiting professor, and each visit showcased an innovation of the visited institution. This is distinct from traditional Hospital Grand Rounds, which more often focus on basic science research or clinical pathophysiology/disease management based on subspecialty topics.[10] The Visiting Professorship in Hospital Medicine Program was judged by participants to be an effective means of spreading innovation.

Insights from experience with the Visiting Professorship in Hospital Medicine Program include the importance of preliminary work prior to each visit. Program directors need to attend closely to the fit between the interests and career path of the visiting professor and those of the early‐career faculty. The innovations being shared should be aligned with organizational interests to maximize the chance of subsequent spread of the innovation and future collaboration. Providing faculty information about the objectives of the program in advance of the visit and arranging an exchange of curricula vitae between the early‐career faculty and the visiting professor allows participants to prepare for the in‐person coaching. Based on comments from participants, prompting contact from the visiting professor after the visit may be helpful to initiate the longitudinal relationship. We also found that early‐career faculty may not be aware of how to effectively use a mentoring relationship with an external faculty member. Training sessions for both mentors and mentees on effective mentorship relationships before visiting professorships might improve early‐career faculty confidence in initiating relationships and maximize value from mentor coaching.

A key issue is finding the right level of career maturity for the visiting professor. Our approach in selecting visiting professors was congruent with utilization of midcareer peer coaches employed by intrainstitutional hospital medicine mentoring programs.[11] The visiting professor should have sufficient experience and accomplishments to be able to effectively counsel junior faculty. However, it is important that the visiting professor also has sufficient time and interest to take on additional mentees and to be a full participant in shared scholarship projects emerging from the experience.

This study represents the experience of 5 mature academic hospitalist groups, and results may not be generalizable to dissimilar institutions or if only the most senior faculty are selected to perform visits. There is an inherent selection bias in the choice of both visiting professor and early‐career faculty. The small sample size of the faculty exposed to this program is a limitation to generalizability of the results of this evaluation. Whether this program will result in greater success in promotion of academic hospitalists cannot be assessed based on the follow‐up available. The Visiting Professorship in Hospital Medicine Program has continued to be sustained with an additional academic medical center enrolled and 2 additional site visits planned. The costs of the program are low, largely air travel and a night of lodging, as well as nominal administrative logistical support. Perceived benefits by participants and academic medical centers make this modest investment worth considering for academic hospitalist groups.

CONCLUSIONS

The Visiting Professorship in Hospital Medicine Program offers structure, opportunities, and access to senior mentors to advance the development of early‐career hospitalists while spreading innovation to distant sites. It is assessed by participants to facilitate external mentoring relationships and has the potential to advance the careers of both early‐career faculty as well as the visiting professors.

Hospital medicine is an emerging specialty comprised predominantly of early‐career faculty, often less than 5 years postresidency and predominately at instructor or assistant professor level.[1] Effective mentoring has been identified as a critical component of academic success.[2, 3] Published data suggest that most academic hospitalists do not have a mentor, and when they do, the majority of them spend less than 4 hours per year with their mentor.[2] The reasons for this are multifactorial but largely result from the lack of structure, opportunities, and local senior academic hospitalists.[1, 4] Early‐career faculty have difficulty establishing external mentoring relationships, and new models beyond the traditional intrainstitutional dyad are needed.[3, 4] The need for mentors and structured mentorship networks may be particularly high in hospital medicine.[5]

The Visiting Professorship in Hospital Medicine Program was designed to promote cross‐institutional mentorship, share hospitalist innovations, and facilitate academic collaboration between hospitalist groups. We describe the design and early experience with this program across 5 academic hospital medicine programs.

PROGRAM DESIGN

RESULTS

To date, 5 academic medical centers have participated in the visiting professorship program, with 7 visiting professors interacting with 29 early‐career faculty. Of the 29 early‐career faculty, 72% (21/29) were at the rank of assistant professor, 17% (5/29) instructor, 7% (2/29) residents with plans to hire, and 3% (1/29) associate professor. The median was 2 years in academic medicine and 1 year at current academic rank. Forty‐one percent (12/29) were women and 7% (2/29) identified as ethnic minority. Of the 7 visiting professors, 57% (4/7) were assistant professor and 43% (3/7) were associate professors. The median was 5 years in academic medicine, 29% (2/7) were women, and none identified as ethnic minority.

Immediate postvisit survey response was obtained for all participating faculty. In the immediate postvisit survey, on a 5‐point Likert scale, the 29 early‐career faculty rated the visit: 4.4 for promoting cross‐institutional dissemination of innovations, 4.2 for advancing my academic career, 4.2 for fostering cross‐institutional mentor‐mentee relationships, and 4.4 for promoting cross‐institutional collaborations. Ninety‐three percent (26/28 accounting for 1 nonresponse to this question) reported the visiting professorship had high potential to disseminate innovation (rated greater than 3 on the 5‐point Likert score). Eighty‐three percent (24/29) of the early‐career faculty rated the visit highly useful in advancing their career, 76% (22/29) responded that the visit was highly likely to foster external mentorship relationships, and 90% (26/29) reported the visit highly effective in promoting cross‐institutional collaborations. In the immediate postvisit survey, the 7 visiting professors rated the visit 4.9 for promoting cross‐institutional dissemination of innovations, 4.3 for advancing my academic career, 4.0 for fostering cross‐institutional mentor‐mentee relationships, and 4.3 for promoting cross‐institutional collaborations.

Free‐text comments from both visiting professors and early‐career faculty were generally favorable (Table 1). Some comments offered constructive input on appropriate matching of faculty, previsit preparation, or desire for more time in sessions (Table 1).

One‐year follow‐up was obtained for all but 1 early‐career faculty member receiving the follow‐up survey, and all 3 visiting professors. Of the 3 visiting professorships that occurred more than 1 year ago, 16 mentorship contacts occurred in total (phone, e‐mail, or in person) between 13 early‐career faculty and 3 visiting professors in the year after the initial visits (range, 04 contacts). Follow‐up contact occurred for 3 of 4 early‐career faculty from the first visiting professorship, 3 of 5 from the second visiting professorship, and 2 of 4 from the third visiting professorship. One early‐career faculty member from each host academic medical center had 3 or more additional contacts with the visiting professor in the year following the initial visit. Overall, 8/13 (62%) of early‐career faculty had at least 1 follow‐up mentoring discussion. On 1‐year follow‐up, overall utility for professional development was rated an average of 3.5 by early‐career faculty (with a trend of higher ratings of efficacy with increasing number of follow‐up contacts) and 4.7 by visiting professors. Half (8/16) of the involved faculty report having seen evidence of cross‐institutional dissemination of innovation. Ninety‐four percent (15/16) of participants at 1‐year follow‐up felt there was benefit to their institution in continuing the program for the next year.

Objective evidence of cross‐institutional scholarship, assessed by email query of both visiting professors and senior hospitalists coordinating the visits, includes 2 collaborative peer reviewed publications including mentors and mentees participating in the visiting professorship.[6, 7] Joint educational curriculum development on high‐value care between sites is planned. The Visiting Professorship in Hospital Medicine Program has resulted in 1 external letter to support a visiting professor's promotion to date.

Figure 1

DISCUSSION

Hospital Medicine is a young, rapidly growing field, hence the number of experienced academic hospitalist mentors with expertise in successfully navigating an academic career is limited. A national study of hospitalist leaders found that 75% of clinician‐educators and 58% of research faculty feel that lack of mentorship is a major issue.[1] Mentorship for hospitalist clinician‐investigators is often delivered by nonhospitalists.[2, 8] There is little evidence of external mentorship for academic clinician‐educators in hospital medicine.[1] Without explicit programmatic support, many faculty may find this to be a barrier to career advancement. A study of successfully promoted hospitalists identified difficulty identifying external senior hospitalists to write letters in support of promotion as an obstacle.[9] Our study of the Visiting Professorship in Hospital Medicine Program found that early‐career faculty rated the visit as useful in advancing their career and fostered external mentorship relationships. Subsequent experience suggests more than half of the early‐career faculty will maintain contact with the visiting professor over the year following the visit. Visiting professors rate the experience particularly highly in their own career advancement.

The hospitalist movement is built on a foundation of innovation. The focus of each presentation was on an innovation developed by the visiting professor, and each visit showcased an innovation of the visited institution. This is distinct from traditional Hospital Grand Rounds, which more often focus on basic science research or clinical pathophysiology/disease management based on subspecialty topics.[10] The Visiting Professorship in Hospital Medicine Program was judged by participants to be an effective means of spreading innovation.

Insights from experience with the Visiting Professorship in Hospital Medicine Program include the importance of preliminary work prior to each visit. Program directors need to attend closely to the fit between the interests and career path of the visiting professor and those of the early‐career faculty. The innovations being shared should be aligned with organizational interests to maximize the chance of subsequent spread of the innovation and future collaboration. Providing faculty information about the objectives of the program in advance of the visit and arranging an exchange of curricula vitae between the early‐career faculty and the visiting professor allows participants to prepare for the in‐person coaching. Based on comments from participants, prompting contact from the visiting professor after the visit may be helpful to initiate the longitudinal relationship. We also found that early‐career faculty may not be aware of how to effectively use a mentoring relationship with an external faculty member. Training sessions for both mentors and mentees on effective mentorship relationships before visiting professorships might improve early‐career faculty confidence in initiating relationships and maximize value from mentor coaching.

A key issue is finding the right level of career maturity for the visiting professor. Our approach in selecting visiting professors was congruent with utilization of midcareer peer coaches employed by intrainstitutional hospital medicine mentoring programs.[11] The visiting professor should have sufficient experience and accomplishments to be able to effectively counsel junior faculty. However, it is important that the visiting professor also has sufficient time and interest to take on additional mentees and to be a full participant in shared scholarship projects emerging from the experience.

This study represents the experience of 5 mature academic hospitalist groups, and results may not be generalizable to dissimilar institutions or if only the most senior faculty are selected to perform visits. There is an inherent selection bias in the choice of both visiting professor and early‐career faculty. The small sample size of the faculty exposed to this program is a limitation to generalizability of the results of this evaluation. Whether this program will result in greater success in promotion of academic hospitalists cannot be assessed based on the follow‐up available. The Visiting Professorship in Hospital Medicine Program has continued to be sustained with an additional academic medical center enrolled and 2 additional site visits planned. The costs of the program are low, largely air travel and a night of lodging, as well as nominal administrative logistical support. Perceived benefits by participants and academic medical centers make this modest investment worth considering for academic hospitalist groups.

CONCLUSIONS

The Visiting Professorship in Hospital Medicine Program offers structure, opportunities, and access to senior mentors to advance the development of early‐career hospitalists while spreading innovation to distant sites. It is assessed by participants to facilitate external mentoring relationships and has the potential to advance the careers of both early‐career faculty as well as the visiting professors.

Hospitalizations frequently last longer than warranted by medical necessity alone, due to inefficiencies within the US healthcare system.[1, 2] Discharge delays place patients at risk for hospital‐acquired complications and increase costs. With the growing emphasis on high‐value care, hospital length of stay (LOS) has emerged as a key metric for inpatient care and will remain a central focus of hospital‐based improvement initiatives for the foreseeable future.

Hospitals may find it difficult to identify the primary drivers of inpatient LOS in a dynamic and increasingly complex healthcare system. Multiple recent policy changes have affected inpatient care. The Health Information Technology for Economic and Clinical Health Act of 2009 has led to widespread adoption of electronic health records (EHRs) that have markedly impacted provider workflows.[3] In October 2013, the Centers for Medicare & Medicaid Services implemented the 2‐midnight rule, which reclassified lower acuity inpatients with an expected stay <48 hours to observation status.[4] In January 2014, expansion of insurance coverage under the Affordable Care Act (ACA) altered payer mix for hospitals nationwide.[5] At a local level, hospitals that are rapidly adjusting resource allocation, capital investments, and marketing efforts and making complex operational decisions (eg, to open new units or change admission or referral algorithms) may simultaneously experience shifts in patient volumes, case‐mix index, and staffing ratios with downstream effects on LOS.

Given the myriad factors influencing inpatient LOS, hospital leaders may encounter real challenges in designing effective LOS reduction strategies. For example, they may expend significant resources on real‐time demand‐capacity management systems to improve hospital‐wide patient flow, but the resultant emphasis on bed placement and early discharges may shave only hours off average LOS.[6, 7] An alternative approach may be to target the small percentage of patients with prolonged hospitalizations who contribute disproportionately to the average LOS, as other initiatives focused on high utilizers have done.[8, 9, 10]

Our institution noted an increase in the average inpatient LOS for general medicine patients from 2012 to 2014, prompting a call to action by hospital leaders. We sought to characterize the predictors of prolonged hospitalizations among medicine patients to guide future efforts aimed at mitigating the contribution of prolonged LOS to overall LOS.

METHODS

RESULTS

We identified 18,363 inpatient discharges among 12,511 medicine patients between January 1, 2012 and December 31, 2014. Of these discharges, 416 (2.3%) demonstrated prolonged LOS. Prolonged hospitalizations accounted for 18.6% of total inpatient days. The average LOS during the study period was 4.8 days including patients with prolonged LOS and 4.0 days excluding patients with prolonged LOS, a contribution of 0.8 days.

Table 1 compares the characteristics of patients with and without prolonged LOS. Age, insurance, discharge disposition, palliative care consults, ICU stays, and surgeries were among the variables that differed significantly between the 2 groups. Among patients undergoing surgery, those with prolonged LOS were more likely to have surgery >24 hours after admission than those without prolonged LOS (85.7% vs 51.4%, P<0.001).

Unspecified sepsis was the most frequent primary diagnosis, regardless of LOS category (Table 2). However, the second through fifth most frequent diagnoses differed for patients with and without prolonged LOS.

In an adjusted logistic regression model (Table 3), we found lower odds of prolonged LOS for each 10‐year increase in age and higher odds of prolonged LOS for Medicaid insurance, discharge to home with home health, discharge to a postacute‐care or long‐term acute‐care facility, and in‐hospital death. Methicillin‐resistant Staphylococcus aureus (MRSA) septicemia, palliative care consults, ICU stays, and surgeries were all also associated with increased odds of prolonged LOS. We identified a statistically significant interaction between ICU stay and surgical procedure (odds ratio: 2.53, 95% confidence interval: 1.51‐4.26, P<0.001).

Sensitivity analyses using LOS >14 and >30 days yielded similar results to LOS >21 days (see Supporting Appendix Table 1 in the online version of this article).

DISCUSSION

We found that a small proportion of medicine patients with prolonged hospitalizations contributed substantially to both total inpatient days and average inpatient LOS. Such disproportionate healthcare utilization is concerning in light of the Institute of Medicine's charge for health systems to deliver timely, efficient, and equitable care.[11]

Few studies in the United States have analyzed patient characteristics that predict prolonged LOS, and to our knowledge, none have evaluated prolonged LOS specifically in general medicine patients.[12, 13, 14] Among selected surgical populations, prolonged hospitalizations are most often related to placement difficulties, operational delays, and payer‐related issues, rather than severity of illness, baseline comorbidities, or in‐hospital complications.[12, 13] In our study, we found that patients with prolonged LOS were more likely to require a palliative care consult, ICU stay, or surgery, all proxies for disease severity. Patients with prolonged LOS were also more likely to undergo surgery >24 hours after admission than those without prolonged LOS, suggesting that the former were either too unstable to proceed directly to surgery or developed complications later during their hospitalization. Even after controlling for palliative care consults, ICU stays, and surgeries, placement at a postacute‐care facility was strongly associated with prolonged LOS. Patients with prolonged LOS were also more likely to have Medicaid compared to other insurance types.

Our findings have several potential implications for efforts aimed at decreasing the number of and length of prolonged hospitalizations. Although demographic and clinical factors such as Medicaid insurance, ICU stays, and surgeries are generally not modifiable, they could, particularly in combination, be used to trigger earlier and more intensive case management involvement. A streamlined insurance approval process for Medicaid pending inpatients could be beneficial, given the recent expansion of Medicaid eligibility under the ACA. Hospital partnerships with postacute‐care facilities could also relieve bottlenecks in placement.[8] Chart review of the patients with MRSA septicemia and prolonged LOS indicated that development of an intensive outpatient parenteral antimicrobial therapy pathway with substance abuse counseling could provide an alternative to extended inpatient treatment for intravenous drug users with complicated infections.[15]

This study has several limitations. First, given the lack of consensus in the literature regarding the definition of a prolonged hospitalization, it is difficult to directly compare our results with existing studies.[8, 12, 13, 14] However, we believe LOS >21 days to be a meaningful cutoff for our cohort. Most demographic and clinical variables that were predictive at >21 days were also predictive at >14 and >30 days, which reassured us that the relationship between variables and prolonged LOS was stable at different thresholds. Second, our database did not allow us to fully adjust for baseline comorbidities or categorize the reasons for discharge delays. Finally, this was a single‐center program evaluation. Although this limits generalization to other institutions, we believe our approach may serve as a guide for others interested in reducing prolonged hospitalizations.

In summary, prolonged hospitalizations represent a potentially high‐yield target for LOS reduction efforts. Prolonged hospitalizations among medicine patients at our institution particularly affected Medicaid enrollees with complex hospital stays who were not discharged home. Further studies are needed to determine the specific reasons for unnecessary hospital days in this population.

Hospitalizations frequently last longer than warranted by medical necessity alone, due to inefficiencies within the US healthcare system.[1, 2] Discharge delays place patients at risk for hospital‐acquired complications and increase costs. With the growing emphasis on high‐value care, hospital length of stay (LOS) has emerged as a key metric for inpatient care and will remain a central focus of hospital‐based improvement initiatives for the foreseeable future.

Hospitals may find it difficult to identify the primary drivers of inpatient LOS in a dynamic and increasingly complex healthcare system. Multiple recent policy changes have affected inpatient care. The Health Information Technology for Economic and Clinical Health Act of 2009 has led to widespread adoption of electronic health records (EHRs) that have markedly impacted provider workflows.[3] In October 2013, the Centers for Medicare & Medicaid Services implemented the 2‐midnight rule, which reclassified lower acuity inpatients with an expected stay <48 hours to observation status.[4] In January 2014, expansion of insurance coverage under the Affordable Care Act (ACA) altered payer mix for hospitals nationwide.[5] At a local level, hospitals that are rapidly adjusting resource allocation, capital investments, and marketing efforts and making complex operational decisions (eg, to open new units or change admission or referral algorithms) may simultaneously experience shifts in patient volumes, case‐mix index, and staffing ratios with downstream effects on LOS.

Given the myriad factors influencing inpatient LOS, hospital leaders may encounter real challenges in designing effective LOS reduction strategies. For example, they may expend significant resources on real‐time demand‐capacity management systems to improve hospital‐wide patient flow, but the resultant emphasis on bed placement and early discharges may shave only hours off average LOS.[6, 7] An alternative approach may be to target the small percentage of patients with prolonged hospitalizations who contribute disproportionately to the average LOS, as other initiatives focused on high utilizers have done.[8, 9, 10]

Our institution noted an increase in the average inpatient LOS for general medicine patients from 2012 to 2014, prompting a call to action by hospital leaders. We sought to characterize the predictors of prolonged hospitalizations among medicine patients to guide future efforts aimed at mitigating the contribution of prolonged LOS to overall LOS.

METHODS

RESULTS

We identified 18,363 inpatient discharges among 12,511 medicine patients between January 1, 2012 and December 31, 2014. Of these discharges, 416 (2.3%) demonstrated prolonged LOS. Prolonged hospitalizations accounted for 18.6% of total inpatient days. The average LOS during the study period was 4.8 days including patients with prolonged LOS and 4.0 days excluding patients with prolonged LOS, a contribution of 0.8 days.

Table 1 compares the characteristics of patients with and without prolonged LOS. Age, insurance, discharge disposition, palliative care consults, ICU stays, and surgeries were among the variables that differed significantly between the 2 groups. Among patients undergoing surgery, those with prolonged LOS were more likely to have surgery >24 hours after admission than those without prolonged LOS (85.7% vs 51.4%, P<0.001).

Unspecified sepsis was the most frequent primary diagnosis, regardless of LOS category (Table 2). However, the second through fifth most frequent diagnoses differed for patients with and without prolonged LOS.

In an adjusted logistic regression model (Table 3), we found lower odds of prolonged LOS for each 10‐year increase in age and higher odds of prolonged LOS for Medicaid insurance, discharge to home with home health, discharge to a postacute‐care or long‐term acute‐care facility, and in‐hospital death. Methicillin‐resistant Staphylococcus aureus (MRSA) septicemia, palliative care consults, ICU stays, and surgeries were all also associated with increased odds of prolonged LOS. We identified a statistically significant interaction between ICU stay and surgical procedure (odds ratio: 2.53, 95% confidence interval: 1.51‐4.26, P<0.001).

Sensitivity analyses using LOS >14 and >30 days yielded similar results to LOS >21 days (see Supporting Appendix Table 1 in the online version of this article).

DISCUSSION

We found that a small proportion of medicine patients with prolonged hospitalizations contributed substantially to both total inpatient days and average inpatient LOS. Such disproportionate healthcare utilization is concerning in light of the Institute of Medicine's charge for health systems to deliver timely, efficient, and equitable care.[11]

Few studies in the United States have analyzed patient characteristics that predict prolonged LOS, and to our knowledge, none have evaluated prolonged LOS specifically in general medicine patients.[12, 13, 14] Among selected surgical populations, prolonged hospitalizations are most often related to placement difficulties, operational delays, and payer‐related issues, rather than severity of illness, baseline comorbidities, or in‐hospital complications.[12, 13] In our study, we found that patients with prolonged LOS were more likely to require a palliative care consult, ICU stay, or surgery, all proxies for disease severity. Patients with prolonged LOS were also more likely to undergo surgery >24 hours after admission than those without prolonged LOS, suggesting that the former were either too unstable to proceed directly to surgery or developed complications later during their hospitalization. Even after controlling for palliative care consults, ICU stays, and surgeries, placement at a postacute‐care facility was strongly associated with prolonged LOS. Patients with prolonged LOS were also more likely to have Medicaid compared to other insurance types.

Our findings have several potential implications for efforts aimed at decreasing the number of and length of prolonged hospitalizations. Although demographic and clinical factors such as Medicaid insurance, ICU stays, and surgeries are generally not modifiable, they could, particularly in combination, be used to trigger earlier and more intensive case management involvement. A streamlined insurance approval process for Medicaid pending inpatients could be beneficial, given the recent expansion of Medicaid eligibility under the ACA. Hospital partnerships with postacute‐care facilities could also relieve bottlenecks in placement.[8] Chart review of the patients with MRSA septicemia and prolonged LOS indicated that development of an intensive outpatient parenteral antimicrobial therapy pathway with substance abuse counseling could provide an alternative to extended inpatient treatment for intravenous drug users with complicated infections.[15]

This study has several limitations. First, given the lack of consensus in the literature regarding the definition of a prolonged hospitalization, it is difficult to directly compare our results with existing studies.[8, 12, 13, 14] However, we believe LOS >21 days to be a meaningful cutoff for our cohort. Most demographic and clinical variables that were predictive at >21 days were also predictive at >14 and >30 days, which reassured us that the relationship between variables and prolonged LOS was stable at different thresholds. Second, our database did not allow us to fully adjust for baseline comorbidities or categorize the reasons for discharge delays. Finally, this was a single‐center program evaluation. Although this limits generalization to other institutions, we believe our approach may serve as a guide for others interested in reducing prolonged hospitalizations.

In summary, prolonged hospitalizations represent a potentially high‐yield target for LOS reduction efforts. Prolonged hospitalizations among medicine patients at our institution particularly affected Medicaid enrollees with complex hospital stays who were not discharged home. Further studies are needed to determine the specific reasons for unnecessary hospital days in this population.

Hospitalizations frequently last longer than warranted by medical necessity alone, due to inefficiencies within the US healthcare system.[1, 2] Discharge delays place patients at risk for hospital‐acquired complications and increase costs. With the growing emphasis on high‐value care, hospital length of stay (LOS) has emerged as a key metric for inpatient care and will remain a central focus of hospital‐based improvement initiatives for the foreseeable future.

Hospitals may find it difficult to identify the primary drivers of inpatient LOS in a dynamic and increasingly complex healthcare system. Multiple recent policy changes have affected inpatient care. The Health Information Technology for Economic and Clinical Health Act of 2009 has led to widespread adoption of electronic health records (EHRs) that have markedly impacted provider workflows.[3] In October 2013, the Centers for Medicare & Medicaid Services implemented the 2‐midnight rule, which reclassified lower acuity inpatients with an expected stay <48 hours to observation status.[4] In January 2014, expansion of insurance coverage under the Affordable Care Act (ACA) altered payer mix for hospitals nationwide.[5] At a local level, hospitals that are rapidly adjusting resource allocation, capital investments, and marketing efforts and making complex operational decisions (eg, to open new units or change admission or referral algorithms) may simultaneously experience shifts in patient volumes, case‐mix index, and staffing ratios with downstream effects on LOS.

Given the myriad factors influencing inpatient LOS, hospital leaders may encounter real challenges in designing effective LOS reduction strategies. For example, they may expend significant resources on real‐time demand‐capacity management systems to improve hospital‐wide patient flow, but the resultant emphasis on bed placement and early discharges may shave only hours off average LOS.[6, 7] An alternative approach may be to target the small percentage of patients with prolonged hospitalizations who contribute disproportionately to the average LOS, as other initiatives focused on high utilizers have done.[8, 9, 10]

Our institution noted an increase in the average inpatient LOS for general medicine patients from 2012 to 2014, prompting a call to action by hospital leaders. We sought to characterize the predictors of prolonged hospitalizations among medicine patients to guide future efforts aimed at mitigating the contribution of prolonged LOS to overall LOS.

METHODS

RESULTS

We identified 18,363 inpatient discharges among 12,511 medicine patients between January 1, 2012 and December 31, 2014. Of these discharges, 416 (2.3%) demonstrated prolonged LOS. Prolonged hospitalizations accounted for 18.6% of total inpatient days. The average LOS during the study period was 4.8 days including patients with prolonged LOS and 4.0 days excluding patients with prolonged LOS, a contribution of 0.8 days.

Table 1 compares the characteristics of patients with and without prolonged LOS. Age, insurance, discharge disposition, palliative care consults, ICU stays, and surgeries were among the variables that differed significantly between the 2 groups. Among patients undergoing surgery, those with prolonged LOS were more likely to have surgery >24 hours after admission than those without prolonged LOS (85.7% vs 51.4%, P<0.001).

Unspecified sepsis was the most frequent primary diagnosis, regardless of LOS category (Table 2). However, the second through fifth most frequent diagnoses differed for patients with and without prolonged LOS.

In an adjusted logistic regression model (Table 3), we found lower odds of prolonged LOS for each 10‐year increase in age and higher odds of prolonged LOS for Medicaid insurance, discharge to home with home health, discharge to a postacute‐care or long‐term acute‐care facility, and in‐hospital death. Methicillin‐resistant Staphylococcus aureus (MRSA) septicemia, palliative care consults, ICU stays, and surgeries were all also associated with increased odds of prolonged LOS. We identified a statistically significant interaction between ICU stay and surgical procedure (odds ratio: 2.53, 95% confidence interval: 1.51‐4.26, P<0.001).

Sensitivity analyses using LOS >14 and >30 days yielded similar results to LOS >21 days (see Supporting Appendix Table 1 in the online version of this article).

DISCUSSION

We found that a small proportion of medicine patients with prolonged hospitalizations contributed substantially to both total inpatient days and average inpatient LOS. Such disproportionate healthcare utilization is concerning in light of the Institute of Medicine's charge for health systems to deliver timely, efficient, and equitable care.[11]

Few studies in the United States have analyzed patient characteristics that predict prolonged LOS, and to our knowledge, none have evaluated prolonged LOS specifically in general medicine patients.[12, 13, 14] Among selected surgical populations, prolonged hospitalizations are most often related to placement difficulties, operational delays, and payer‐related issues, rather than severity of illness, baseline comorbidities, or in‐hospital complications.[12, 13] In our study, we found that patients with prolonged LOS were more likely to require a palliative care consult, ICU stay, or surgery, all proxies for disease severity. Patients with prolonged LOS were also more likely to undergo surgery >24 hours after admission than those without prolonged LOS, suggesting that the former were either too unstable to proceed directly to surgery or developed complications later during their hospitalization. Even after controlling for palliative care consults, ICU stays, and surgeries, placement at a postacute‐care facility was strongly associated with prolonged LOS. Patients with prolonged LOS were also more likely to have Medicaid compared to other insurance types.

Our findings have several potential implications for efforts aimed at decreasing the number of and length of prolonged hospitalizations. Although demographic and clinical factors such as Medicaid insurance, ICU stays, and surgeries are generally not modifiable, they could, particularly in combination, be used to trigger earlier and more intensive case management involvement. A streamlined insurance approval process for Medicaid pending inpatients could be beneficial, given the recent expansion of Medicaid eligibility under the ACA. Hospital partnerships with postacute‐care facilities could also relieve bottlenecks in placement.[8] Chart review of the patients with MRSA septicemia and prolonged LOS indicated that development of an intensive outpatient parenteral antimicrobial therapy pathway with substance abuse counseling could provide an alternative to extended inpatient treatment for intravenous drug users with complicated infections.[15]

This study has several limitations. First, given the lack of consensus in the literature regarding the definition of a prolonged hospitalization, it is difficult to directly compare our results with existing studies.[8, 12, 13, 14] However, we believe LOS >21 days to be a meaningful cutoff for our cohort. Most demographic and clinical variables that were predictive at >21 days were also predictive at >14 and >30 days, which reassured us that the relationship between variables and prolonged LOS was stable at different thresholds. Second, our database did not allow us to fully adjust for baseline comorbidities or categorize the reasons for discharge delays. Finally, this was a single‐center program evaluation. Although this limits generalization to other institutions, we believe our approach may serve as a guide for others interested in reducing prolonged hospitalizations.

In summary, prolonged hospitalizations represent a potentially high‐yield target for LOS reduction efforts. Prolonged hospitalizations among medicine patients at our institution particularly affected Medicaid enrollees with complex hospital stays who were not discharged home. Further studies are needed to determine the specific reasons for unnecessary hospital days in this population.

In this issue of the Journal of Hospital Medicine, the results of 2 inpatient handoff studies further shape our evolving understanding of in‐hospital care transitions. Schouten and colleagues,[1] report no difference in adverse outcomes when admissions were handed off to the primary team using face‐to‐face compared to nonface‐to‐face interactions. Meanwhile, Hanson and colleagues[2] report that a written handoff tool is used infrequently by covering interns.

Schouten et al.'s study attempted to isolate the impact of the verbal portion of the handoff between admitting and accepting team by evaluating whether early adverse outcomes differed between patients whose teams performed a face‐to‐face handoffs compared to those who did not. Their study was a retrospective chart review, and no additional process changes, training, or instruction regarding handoffs were implemented or measured. Handoffs occurred primarily between advanced practice providers, hospitalists, and a small number of resident physicians, so generalizability of this study to other institutions may be limited. No difference in adverse events was noted between admissions with face‐to‐face compared to those without face‐to‐face handoffs (2.6% vs 3.2%). Unfortunately, this study was likely underpowered to detect significant changes in adverse events, with a sample size of 805 total patients with a 3% baseline rate of adverse events (by our estimate, over 5000 patients would be needed in each group10,000 overallto detect a 30% relative difference in event rates). Further, this study did not examine other outcomes that could be impacted by the handoff process such as provider efficiency or patient experience.

Face‐to‐face handoffs, the gold standard for handoffs between providers, was 1 of the sign‐out approaches examined in a study by Graham and colleagues.[3] This study, in contrast to the Schouten et al. study, prospectively evaluated adverse events before and after implementation of face‐to‐face handoffs, with structured written sign‐out from the primary team to nighttime covering physicians. Prior to implementation, handoffs consisted of a double handoff involving an intermediary physician and unstructured written sign‐out. Although no statistically significant reduction in adverse events was found in the Graham et al. study, significant improvements were noted in physician satisfaction, documentation of key elements in handoffs, and reduced data omissions; importantly, a trend of fewer near misses was noted comparing the pre‐ and postintervention periods. Although the Schouten et al. and Graham et al. studies suggest questionable benefit of face‐to‐face handoffs, we would caution that limitations in sample size and methodological sensitivity to detect adverse events in both studies could explain the lack of association between face‐to‐face handoffs and reduced adverse events. Furthermore, the promising findings of fewer data omissions and near misses in the intervention group in the Graham et al. study suggest benefit from a multipronged approach to improving handoffs including both face‐to‐face interactions and a structured written component.

In this issue, Hanson and colleagues also evaluated the use of a handoff tool by cross‐covering interns in a convenience sample of overnight clinical interactions. Despite finding that standard written documentation was considered beneficial by nearly all respondents (94.3%), the interns reported that the handoff tool was used in only 27.7% of encounters. This pales in comparison to the use of the nurse or chart in 94.4% of cross‐coverage encounters. The authors speculate that a handoff tool, for many years the only timely source of information, may not be as useful when information can be easily accessed in an electronic health record. Yet, in a prior systematic review that included 6 studies of computerized handoff tools, Li and colleagues found that computerized handoff tools may improve physician efficiency, enhance the completeness of handoff information, and even potentially reduce adverse events.[4]

The Schouten et al. and Hanson et al. studies raise important questions for the fields of hospital medicine and patient safety. Is it time to do away with the written and verbal portions of the handoff process? Should the handoff of patients simply consist of transferring a list of patients to covering providers? We do not believe this is the correct course of action. Rather, we recommend a more evolutionary, not revolutionary, interpretation of these results, especially when considered as part of a broader story of in‐hospital transitions of care.

For example, a recently published evaluation of a resident handoff‐improvement program in 9 hospitals and 10,740 patient admissions by Starmer and collegues[5] focused on a handoff bundle, I‐PASS, which is a pneumonic for Illness severity, Patient summary, Action items, Situation awareness and contingency planning, Synthesis by receiver. The authors report a reduction in medical errors and preventable adverse events without significant increases in the duration of oral handoff per patient. The handoff in this study included both oral and written elements in the I‐PASS format. Implementation was multipronged, and the I‐PASS bundle included (1) use of the I‐PASS pneumonic to standardize handoffs; (2) resident physician training in handoffs and communication through a 2‐hour workshop, followed by a 1‐hour role‐playing and simulation session, and a computer module for practice; (3) faculty development and observation with use of direct‐observation tools to provide structured feedback to residents; (4) active surveillance for errors (rather than relying on self‐report); and (5) a sustainability campaign to promote continuation of culture change. The complexity and robust nature of the I‐PASS handoff bundle suggests that having multiple structured components included in a handoff program with active, rather than retrospective, evaluation might increase the likelihood of improved, sustained outcomes. In addition, one might also conclude from the Starmer et al. study that it takes commitment from all levels, including residents, faculty, and administration, to improve handoffs between teams for inpatient care.

We commend Schouten et al. and Hanson et al. on their contributions to the literature, but believe that the story of the in‐hospital handoff has yet to be fully written. Although results from these 2 articles may cause speculation about the value of oral and written handoffs, we believe that the balance of evidence favors the use of a multipronged approach that involves both structured oral and written handoffs to improve the value and efficiency of handoffs. In addition, findings from the I‐PASS study support dedicated handoff training for providers, evaluation of handoffs using structured tools, and active surveillance for medical errors. Future areas of work should include a systematic review of the inpatient handoff literature and further evaluation of precisely which specific intervention components (eg, structured content of handoffs, sensemaking content) or modes of delivery (eg, face‐to‐face vs other) are most likely to reduce medical errors and improve patient outcomes. As the hospital medicine movement continues to grow, handoffs will continue to be paramount. Establishing the safest method to complete handoffs to promote patient safety should be a common goal for hospitalists.

The handoff story is still in evolution; as hospitalists, we are poised to be its author.

In this issue of the Journal of Hospital Medicine, the results of 2 inpatient handoff studies further shape our evolving understanding of in‐hospital care transitions. Schouten and colleagues,[1] report no difference in adverse outcomes when admissions were handed off to the primary team using face‐to‐face compared to nonface‐to‐face interactions. Meanwhile, Hanson and colleagues[2] report that a written handoff tool is used infrequently by covering interns.

Schouten et al.'s study attempted to isolate the impact of the verbal portion of the handoff between admitting and accepting team by evaluating whether early adverse outcomes differed between patients whose teams performed a face‐to‐face handoffs compared to those who did not. Their study was a retrospective chart review, and no additional process changes, training, or instruction regarding handoffs were implemented or measured. Handoffs occurred primarily between advanced practice providers, hospitalists, and a small number of resident physicians, so generalizability of this study to other institutions may be limited. No difference in adverse events was noted between admissions with face‐to‐face compared to those without face‐to‐face handoffs (2.6% vs 3.2%). Unfortunately, this study was likely underpowered to detect significant changes in adverse events, with a sample size of 805 total patients with a 3% baseline rate of adverse events (by our estimate, over 5000 patients would be needed in each group10,000 overallto detect a 30% relative difference in event rates). Further, this study did not examine other outcomes that could be impacted by the handoff process such as provider efficiency or patient experience.

Face‐to‐face handoffs, the gold standard for handoffs between providers, was 1 of the sign‐out approaches examined in a study by Graham and colleagues.[3] This study, in contrast to the Schouten et al. study, prospectively evaluated adverse events before and after implementation of face‐to‐face handoffs, with structured written sign‐out from the primary team to nighttime covering physicians. Prior to implementation, handoffs consisted of a double handoff involving an intermediary physician and unstructured written sign‐out. Although no statistically significant reduction in adverse events was found in the Graham et al. study, significant improvements were noted in physician satisfaction, documentation of key elements in handoffs, and reduced data omissions; importantly, a trend of fewer near misses was noted comparing the pre‐ and postintervention periods. Although the Schouten et al. and Graham et al. studies suggest questionable benefit of face‐to‐face handoffs, we would caution that limitations in sample size and methodological sensitivity to detect adverse events in both studies could explain the lack of association between face‐to‐face handoffs and reduced adverse events. Furthermore, the promising findings of fewer data omissions and near misses in the intervention group in the Graham et al. study suggest benefit from a multipronged approach to improving handoffs including both face‐to‐face interactions and a structured written component.

In this issue, Hanson and colleagues also evaluated the use of a handoff tool by cross‐covering interns in a convenience sample of overnight clinical interactions. Despite finding that standard written documentation was considered beneficial by nearly all respondents (94.3%), the interns reported that the handoff tool was used in only 27.7% of encounters. This pales in comparison to the use of the nurse or chart in 94.4% of cross‐coverage encounters. The authors speculate that a handoff tool, for many years the only timely source of information, may not be as useful when information can be easily accessed in an electronic health record. Yet, in a prior systematic review that included 6 studies of computerized handoff tools, Li and colleagues found that computerized handoff tools may improve physician efficiency, enhance the completeness of handoff information, and even potentially reduce adverse events.[4]

The Schouten et al. and Hanson et al. studies raise important questions for the fields of hospital medicine and patient safety. Is it time to do away with the written and verbal portions of the handoff process? Should the handoff of patients simply consist of transferring a list of patients to covering providers? We do not believe this is the correct course of action. Rather, we recommend a more evolutionary, not revolutionary, interpretation of these results, especially when considered as part of a broader story of in‐hospital transitions of care.

For example, a recently published evaluation of a resident handoff‐improvement program in 9 hospitals and 10,740 patient admissions by Starmer and collegues[5] focused on a handoff bundle, I‐PASS, which is a pneumonic for Illness severity, Patient summary, Action items, Situation awareness and contingency planning, Synthesis by receiver. The authors report a reduction in medical errors and preventable adverse events without significant increases in the duration of oral handoff per patient. The handoff in this study included both oral and written elements in the I‐PASS format. Implementation was multipronged, and the I‐PASS bundle included (1) use of the I‐PASS pneumonic to standardize handoffs; (2) resident physician training in handoffs and communication through a 2‐hour workshop, followed by a 1‐hour role‐playing and simulation session, and a computer module for practice; (3) faculty development and observation with use of direct‐observation tools to provide structured feedback to residents; (4) active surveillance for errors (rather than relying on self‐report); and (5) a sustainability campaign to promote continuation of culture change. The complexity and robust nature of the I‐PASS handoff bundle suggests that having multiple structured components included in a handoff program with active, rather than retrospective, evaluation might increase the likelihood of improved, sustained outcomes. In addition, one might also conclude from the Starmer et al. study that it takes commitment from all levels, including residents, faculty, and administration, to improve handoffs between teams for inpatient care.

We commend Schouten et al. and Hanson et al. on their contributions to the literature, but believe that the story of the in‐hospital handoff has yet to be fully written. Although results from these 2 articles may cause speculation about the value of oral and written handoffs, we believe that the balance of evidence favors the use of a multipronged approach that involves both structured oral and written handoffs to improve the value and efficiency of handoffs. In addition, findings from the I‐PASS study support dedicated handoff training for providers, evaluation of handoffs using structured tools, and active surveillance for medical errors. Future areas of work should include a systematic review of the inpatient handoff literature and further evaluation of precisely which specific intervention components (eg, structured content of handoffs, sensemaking content) or modes of delivery (eg, face‐to‐face vs other) are most likely to reduce medical errors and improve patient outcomes. As the hospital medicine movement continues to grow, handoffs will continue to be paramount. Establishing the safest method to complete handoffs to promote patient safety should be a common goal for hospitalists.

The handoff story is still in evolution; as hospitalists, we are poised to be its author.

In this issue of the Journal of Hospital Medicine, the results of 2 inpatient handoff studies further shape our evolving understanding of in‐hospital care transitions. Schouten and colleagues,[1] report no difference in adverse outcomes when admissions were handed off to the primary team using face‐to‐face compared to nonface‐to‐face interactions. Meanwhile, Hanson and colleagues[2] report that a written handoff tool is used infrequently by covering interns.

Schouten et al.'s study attempted to isolate the impact of the verbal portion of the handoff between admitting and accepting team by evaluating whether early adverse outcomes differed between patients whose teams performed a face‐to‐face handoffs compared to those who did not. Their study was a retrospective chart review, and no additional process changes, training, or instruction regarding handoffs were implemented or measured. Handoffs occurred primarily between advanced practice providers, hospitalists, and a small number of resident physicians, so generalizability of this study to other institutions may be limited. No difference in adverse events was noted between admissions with face‐to‐face compared to those without face‐to‐face handoffs (2.6% vs 3.2%). Unfortunately, this study was likely underpowered to detect significant changes in adverse events, with a sample size of 805 total patients with a 3% baseline rate of adverse events (by our estimate, over 5000 patients would be needed in each group10,000 overallto detect a 30% relative difference in event rates). Further, this study did not examine other outcomes that could be impacted by the handoff process such as provider efficiency or patient experience.

Face‐to‐face handoffs, the gold standard for handoffs between providers, was 1 of the sign‐out approaches examined in a study by Graham and colleagues.[3] This study, in contrast to the Schouten et al. study, prospectively evaluated adverse events before and after implementation of face‐to‐face handoffs, with structured written sign‐out from the primary team to nighttime covering physicians. Prior to implementation, handoffs consisted of a double handoff involving an intermediary physician and unstructured written sign‐out. Although no statistically significant reduction in adverse events was found in the Graham et al. study, significant improvements were noted in physician satisfaction, documentation of key elements in handoffs, and reduced data omissions; importantly, a trend of fewer near misses was noted comparing the pre‐ and postintervention periods. Although the Schouten et al. and Graham et al. studies suggest questionable benefit of face‐to‐face handoffs, we would caution that limitations in sample size and methodological sensitivity to detect adverse events in both studies could explain the lack of association between face‐to‐face handoffs and reduced adverse events. Furthermore, the promising findings of fewer data omissions and near misses in the intervention group in the Graham et al. study suggest benefit from a multipronged approach to improving handoffs including both face‐to‐face interactions and a structured written component.

In this issue, Hanson and colleagues also evaluated the use of a handoff tool by cross‐covering interns in a convenience sample of overnight clinical interactions. Despite finding that standard written documentation was considered beneficial by nearly all respondents (94.3%), the interns reported that the handoff tool was used in only 27.7% of encounters. This pales in comparison to the use of the nurse or chart in 94.4% of cross‐coverage encounters. The authors speculate that a handoff tool, for many years the only timely source of information, may not be as useful when information can be easily accessed in an electronic health record. Yet, in a prior systematic review that included 6 studies of computerized handoff tools, Li and colleagues found that computerized handoff tools may improve physician efficiency, enhance the completeness of handoff information, and even potentially reduce adverse events.[4]

The Schouten et al. and Hanson et al. studies raise important questions for the fields of hospital medicine and patient safety. Is it time to do away with the written and verbal portions of the handoff process? Should the handoff of patients simply consist of transferring a list of patients to covering providers? We do not believe this is the correct course of action. Rather, we recommend a more evolutionary, not revolutionary, interpretation of these results, especially when considered as part of a broader story of in‐hospital transitions of care.

For example, a recently published evaluation of a resident handoff‐improvement program in 9 hospitals and 10,740 patient admissions by Starmer and collegues[5] focused on a handoff bundle, I‐PASS, which is a pneumonic for Illness severity, Patient summary, Action items, Situation awareness and contingency planning, Synthesis by receiver. The authors report a reduction in medical errors and preventable adverse events without significant increases in the duration of oral handoff per patient. The handoff in this study included both oral and written elements in the I‐PASS format. Implementation was multipronged, and the I‐PASS bundle included (1) use of the I‐PASS pneumonic to standardize handoffs; (2) resident physician training in handoffs and communication through a 2‐hour workshop, followed by a 1‐hour role‐playing and simulation session, and a computer module for practice; (3) faculty development and observation with use of direct‐observation tools to provide structured feedback to residents; (4) active surveillance for errors (rather than relying on self‐report); and (5) a sustainability campaign to promote continuation of culture change. The complexity and robust nature of the I‐PASS handoff bundle suggests that having multiple structured components included in a handoff program with active, rather than retrospective, evaluation might increase the likelihood of improved, sustained outcomes. In addition, one might also conclude from the Starmer et al. study that it takes commitment from all levels, including residents, faculty, and administration, to improve handoffs between teams for inpatient care.

We commend Schouten et al. and Hanson et al. on their contributions to the literature, but believe that the story of the in‐hospital handoff has yet to be fully written. Although results from these 2 articles may cause speculation about the value of oral and written handoffs, we believe that the balance of evidence favors the use of a multipronged approach that involves both structured oral and written handoffs to improve the value and efficiency of handoffs. In addition, findings from the I‐PASS study support dedicated handoff training for providers, evaluation of handoffs using structured tools, and active surveillance for medical errors. Future areas of work should include a systematic review of the inpatient handoff literature and further evaluation of precisely which specific intervention components (eg, structured content of handoffs, sensemaking content) or modes of delivery (eg, face‐to‐face vs other) are most likely to reduce medical errors and improve patient outcomes. As the hospital medicine movement continues to grow, handoffs will continue to be paramount. Establishing the safest method to complete handoffs to promote patient safety should be a common goal for hospitalists.

The handoff story is still in evolution; as hospitalists, we are poised to be its author.

For the frail older patient, hospitalization marks a period of high risk of poor outcomes and adverse events including functional decline, delirium, pressure ulcers, adverse drug events, nosocomial infections, and falls.1, 2 Physician recognition of elderly patients at risk for adverse outcomes is poor, making it difficult to intervene to prevent them.3, 4 Among frail, elderly inpatients at an urban academic medical center, doctors documented cognitive assessments in only 5% of patients. Functional assessments are appropriately documented in 40%80% of inpatients.3, 5

The Acute Care for Elders (ACE) unit is one of several models of comprehensive inpatient geriatric care that have been developed by geriatrician researchers to address the adverse events and functional decline that often accompany hospitalization.6 The ACE unit model generally incorporates: 1) a modified hospital environment, 2) early assessment and intensive management to minimize the adverse effects of hospital care, 3) early discharge planning, 4) patient centered care protocols, and 5) a consistent nursing staff.7 Two randomized, controlled trials have shown the ACE unit model to be successful in reducing functional decline among frail older inpatients during and after hospitalization.7, 8 While meta‐analyses data also suggests the ACE unit model reduces functional decline and future institutionalization, significant impact on other outcomes is not proven.9, 10

Several barriers have prevented the successful dissemination of the ACE unit model. The chief limitations are the upfront resources required to create and maintain a modified, dedicated unit, as well as the lack of a geriatrics trained workforce.7, 1113 The rapid growth of hospital medicine presents opportunities for innovation in the care of older patients. Still, a 2006 census demonstrated that few hospitalist groups had identified geriatric care as a priority.14

In response to these challenges, the University of Colorado Hospital Medicine Group created a hospitalist‐run inpatient medical service designed for the care of the frail older patient. This Hospitalist‐Acute Care for the Elderly (Hospitalist‐ACE) unit is a hybrid of a general medical service and an inpatient geriatrics unit.7 The goals of the Hospitalist‐ACE service are to provide high quality care tailored to older inpatients, thus minimizing the risks of functional decline and adverse events associate with hospitalization, and to provide a clinical geriatrics teaching experience for Hospitalist Training Track Residents within the Internal Medicine Residency Training Program and medical students at the University of Colorado Denver School of Medicine. The Hospitalist‐ACE unit is staffed with a core group of hospitalist attendings who have, at a minimum, attended an intensive mini‐course in inpatient geriatrics. The service employs interdisciplinary rounds; a brief, standardized geriatric assessment including screens of function, cognition, and mood; a clinical focus on mitigating the hazards of hospitalization, early discharge planning; and a novel geriatric educational curriculum for medicine residents and medical students.

This article will: 1) describe the creation of the Hospitalist‐ACE service at the University of Colorado Hospital; and 2) summarize the evaluation of the Hospitalist‐ACE service in a quasi‐randomized, controlled manner during its first year. We hypothesized that, when compared to patients receiving usual care, patients cared for on the Hospitalist‐ACE service would have increased recognition of abnormal functional status; recognition of abnormal cognitive status and delirium; equivalent lengths of stay and hospital charges; and decreased falls, 30‐day readmissions, and restraint use.

METHODS

Intervention

The Hospitalist‐ACE unit implemented an interdisciplinary team approach to identify and address geriatric syndromes in patients aged 70 and over. The Hospitalist‐ACE model of care consisted of clinical care provided by a hospitalist attending with additional training in geriatric medicine, administration of standardized geriatric screens assessing function, cognition, and mood, 15 minute daily (MondayFriday) interdisciplinary rounds focusing on recognition and management of geriatric syndromes and early discharge planning, and a standardized educational curriculum for medical residents and medical students addressing hazards of hospitalization.

The Hospitalist‐ACE service was a unique rotation within the Hospitalist Training Track of the Internal Medicine Residency that was developed with the support of the University of Colorado Hospital and the Internal Medicine Residency Training Program, and input from the Geriatrics Division at the University of Colorado Denver. The director received additional training from the Donald W. Reynolds FoundationUCLA Faculty Development to Advance Geriatric Education Mini‐Fellowship for hospitalist faculty. The mission of the service was to excel at educating the next generation of hospitalists while providing a model for excellence of care for hospitalized elderly patients. Important stakeholders were identified, and a leadership teamincluding representatives from nursing, physical and occupational therapy, pharmacy, social work, case management, and later, volunteer servicescreated the model daily interdisciplinary rounds. As geographic concentration was essential for the viability of interdisciplinary rounds, one unit (12 West) within the hospital was designated as the preferred location for patients admitted to the Hospitalist‐ACE service.

The Hospitalist‐ACE unit team consisted of one attending hospitalist, one resident, one intern, and medical students. The attending was one of five hospitalists, with additional training in geriatric medicine, who rotated attending responsibilities on the service. One of the hospitalists was board certified in geriatric medicine. Each of the other four hospitalists attended the Reynolds FoundationUCLA mini‐fellowship in geriatric medicine. Hospitalist‐ACE attendings rotated on a variety of other hospitalist services throughout the academic year, including the usual care services.

The brief standardized geriatric assessment consisted of six validated instruments, and was completed by house staff or medical students on admission, following instruction by the attending physician. The complete assessment tool is shown in Figure 1. The cognitive items included the Mini‐Cog,15 a two‐item depression screen,16 and the Confusion Assessment Method.17 The functional items included the Vulnerable Elders Survey (VES‐13),18 the Timed Get Up and Go test,19 and a two‐question falls screen.20 The elements of the assessment tool were selected by the Hospitalist‐ACE attendings for brevity and the potential to inform clinical management. To standardize the clinical and educational approach, the Hospitalist‐ACE attendings regularly discussed appropriate orders recommended in response to each positive screen, but no templated order sets were used during the study period.

Figure 1

Interdisciplinary rounds were attended by Hospitalist‐ACE physicians, nurses, case managers, social workers, physical or occupational therapists, pharmacists, and volunteers. Rounds were led by the attending or medical resident.

The educational curriculum encompassed 13 modules created by the attending faculty that cover delirium, falls, dementia, pressure ulcers, physiology of aging, movement disorders, medication safety, end of life care, advance directives, care transitions, financing of health care for the elderly, and ethical conundrums in the care of the elderly. A full table of contents appears in online Appendix 1. Additionally, portions of the curriculum have been published online.21, 22 Topic selection was guided by the Accreditation Council for Graduate Medical Education (ACGME) core geriatrics topics determined most relevant for the inpatient setting. Formal instruction of 3045 minutes duration occurred three to four days a week and was presented in addition to routine internal medicine educational conferences. Attendings coordinated teaching to ensure that each trainee was exposed to all of the content during the course of their four‐week rotation.

In contrast to the Hospitalist‐ACE service, usual care on the control general medical services consisted of either a hospitalist, a general internist, or an internal medicine subspecialist attending physician, with one medical resident, one intern, and medical students admitting every fourth day. The general medical teams attended daily discharge planning rounds with a discharge planner and social worker focused exclusively on discharge planning. The content of teaching rounds on the general medical services was largely left to the discretion of the attending physician.

This program evaluation of the Hospitalist‐ACE service was granted a waiver of consent and Health Insurance Portability and Accountability Act (HIPAA) by the Colorado Multiple Institutional Review Board.

RESULTS

Two hundred seventeen patients met the study entry criteria (Table 2): 122 were admitted to the Hospitalist‐ACE service, and 95 were admitted to usual care on the general medical services. The average age of the study patients was 80.5 years, 55.3% were female. Twenty‐eight percent of subjects were admitted for pulmonary diagnoses. The two groups of patients were similar with respect to age, gender, and distribution of primary diagnoses. The Hospitalist‐ACE patients had a mean MS‐DRG weight of 1.15, which was slightly higher than that of usual care patients at 1.05 (P = 0.06). Typically, 70% of Hospitalist‐ACE patients are admitted to the designated ACE medical unit (12 West).

DISCUSSION

We report the implementation and evaluation of a medical service tailored to the care of the acutely ill older patient that draws from elements of the hospitalist model and the ACE unit model.7, 14, 24 For this Hospitalist‐ACE service, we developed a specialized hospitalist workforce, assembled a brief geriatric assessment tailored to the inpatient setting, instituted an interdisciplinary rounding model, and created a novel inpatient geriatrics curriculum.

During the study period, we improved performance of important processes of care for hospitalized elders, including recognition of abnormal cognitive and functional status; maintained comparable resource use; and implemented a novel, inpatient‐focused geriatric medicine educational experience. We were unable to demonstrate an impact on key clinical outcomes such as falls, physical restraint use, and readmissions. Nonetheless, there is evidence that the performance of selected processes of care is associated with improved three‐year survival status in the community‐dwelling vulnerable older patient, and may also be associated with a mortality benefit in the hospitalized vulnerable older patient.25, 26 Therefore, methods to improve the performance of these processes of care may be of clinical importance.

The finding of increased use of DNAR orders in the face of equivalent documentation of code status is of interest and generates hypotheses for further study. It is possible that the educational experience and use of geriatric assessment provides a more complete context for the code status discussion (one that incorporates the patient's social, physical, and cognitive function). However, we do not know if the patients on the ACE service had improved concordance between their code status and their goals of care.

We believe that there was no difference in key clinical outcomes between Hospitalist‐ACE and control patients because the population in this study was relatively low acuity and, therefore, the occurrence of falls and the use of physical restraints were quite low in the study population. In particular, the readmission rate was much lower than is typical for the Medicare population at our hospital, making it challenging to draw conclusions about the impact of the intervention on readmissions, however, we cannot rule out the possibility that our early discharge planning did not address the determinants of readmission for this population.

The ACE unit paradigmcharacterized by 1) closed, modified hospital units; 2) staffing by geriatricians and nurses with geriatrics training; 3) employing geriatric nursing care protocolsrequires significant resources and is not feasible for all settings.6 There is a need for alternative models of comprehensive care for hospitalized elders that require fewer resources in the form of dedicated units and specialist personnel, and can be more responsive to institutional needs. For example, in a 2005 report, one institution reported the creation of a geriatric medicine service that utilized a geriatrician and hospitalist co‐attending model.14 More recently, a large geriatrics program replaced its inpatient geriatrics unit with a mobile inpatient geriatrics service staffed by an attending geriatricianhospitalist, a geriatrics fellow, and a nurse practitioner.27 While these innovative models have eliminated the dedicated unit, they rely on board certified geriatricians, a group in short supply nationally.28 Hospitalists are a rapidly growing provider group that, with appropriate training and building on the work of geriatricians, is poised to provide leadership in acute geriatric care.29, 30

In contrast to the comprehensive inpatient geriatric care models described above, the Hospitalist‐ACE service uses a specialized hospitalist workforce and is not dependent on continuous staffing by geriatricians. Although geographic concentration is important for the success of interdisciplinary rounds, the Hospitalist‐ACE service does not require a closed or modified unit. The nursing staff caring for Hospitalist‐ACE patients have generalist nursing training and, at the time of the study, did not utilize geriatric‐care protocols. Our results need to be interpreted in the light of these differences from the ACE unit model which is a significantly more intensive intervention than the Hospitalist‐ACE service. In addition, the current practice environment is quite different from the mid‐1990s when ACE units were developed and studied. Development and maintenance of models of comprehensive inpatient geriatric care require demonstration of both value as well as return on investment. The alignment of financial and regulatory incentives for programs that provide comprehensive care to complex patients, such as those anticipated by the Affordable Care Act, may encourage the growth of such models.

These data represent findings from a six‐month evaluation of a novel inpatient service in the middle of its first year. There are several limitations related to our study design. First, the results of this small study at a single academic medical center may be of limited generalizability to other settings. Second, the program was evaluated only three months after its inception; we did not capture further improvements in methods, training, and outcomes expected as the program matured. Third, most of the Hospitalist‐ACE service attendings and residents rotate on the UCH general medical services throughout the year. Consequently, we were unable to eliminate the possibility of contamination of the control group, and we were unable to blind the physicians to the study. Fourth, the study population had a relatively low severity of illnessthe average MS‐DRG weight was near 1and low rates of important adverse events such falls and restraint use. This may have occurred because we excluded patients transferred from the ICUs and other services. It is possible that the Hospitalist‐ACE intervention might have demonstrated a larger benefit in a sicker population that would have presented greater opportunities for reductions in length of stay, costs, and adverse events. Fifth, given the retrospective nature of the data collection, we were not able to prospectively assess the incidence of important geriatric outcomes such as delirium and functional decline, nor can we make conclusions about changes in function during the hospitalization.

While the outcome measures we used are conceptually similar to several measures developed by RAND's Assessing Care of Vulnerable Elders (ACOVE) project, this study did not explicitly rely on those constructs.31 To do so would have required prospective screening by clinical staff independent from the care team for vulnerability that was beyond the scope of this project. In addition, the ACOVE measures of interest for functional and cognitive decline are limited to documentation of cognitive or functional assessments in the medical record. The ACE service's adoption of a brief standardized geriatric assessment was almost certain to meet that documentation requirement. While documentation is important, it is not clear that documentation, in and of itself, improves outcomes. Therefore, we expanded upon the ACOVE constructs to include the need for the additional evidence of a treatment plan when abnormal physical or cognitive function was documented. These constructs are important process of care for vulnerable elders. While we demonstrated improvements in several of these important processes of care for elderly patients, we are unable to draw conclusions about the impact of these differences in care on important clinical outcomes such as development of delirium, long‐term institutionalization, or mortality.

CONCLUSIONS

The risks of hospitalization for older persons are numerous, and present challenges and opportunities for inpatient physicians. As the hospitalized population agesmirroring national demographic trends and trends in use of acute care hospitalsthe challenge of avoiding harm in the older hospitalized patient will intensify. Innovations in care to improve the experience and outcomes of hospitalization for older patients are needed in the face of limited geriatrics‐trained workforce and few discretionary funds for unit redesign. The Hospitalist‐ACE service is a promising strategy for hospitalist programs with sufficient numbers of older patients and hospitalists with interest in improving clinical care for older adults. It provides a model for hospitalists to employ geriatrics principles targeted at reducing harm to their most vulnerable patients. Hospitalist‐run geriatric care models offer great promise for improving the care of acutely ill elderly patients. Future investigation should focus on demonstrating the impact of such care on important clinical outcomes between admission and discharge; on model refinement and adaptation, such as determining what components of comprehensive geriatric care are essential to success; and on how complementary interventions, such as the use of templated orders for the hospitalized elderly, impact outcomes. Additional research is needed, with a focus on demonstrating value with regard to an array of outcomes including cost, readmissions, and preventable harms of care.

For the frail older patient, hospitalization marks a period of high risk of poor outcomes and adverse events including functional decline, delirium, pressure ulcers, adverse drug events, nosocomial infections, and falls.1, 2 Physician recognition of elderly patients at risk for adverse outcomes is poor, making it difficult to intervene to prevent them.3, 4 Among frail, elderly inpatients at an urban academic medical center, doctors documented cognitive assessments in only 5% of patients. Functional assessments are appropriately documented in 40%80% of inpatients.3, 5

The Acute Care for Elders (ACE) unit is one of several models of comprehensive inpatient geriatric care that have been developed by geriatrician researchers to address the adverse events and functional decline that often accompany hospitalization.6 The ACE unit model generally incorporates: 1) a modified hospital environment, 2) early assessment and intensive management to minimize the adverse effects of hospital care, 3) early discharge planning, 4) patient centered care protocols, and 5) a consistent nursing staff.7 Two randomized, controlled trials have shown the ACE unit model to be successful in reducing functional decline among frail older inpatients during and after hospitalization.7, 8 While meta‐analyses data also suggests the ACE unit model reduces functional decline and future institutionalization, significant impact on other outcomes is not proven.9, 10