Eduard E. Vasilevskis, MD

How Does the Framework Address the Extent to Which Readmissions Reflect Hospital Quality?

One of the chief problems with readmissionrates as a hospital quality measure is that many of the factors that influence readmission may not currently be under the hospital's control. The healthcare environment to which a patient is being discharged (and was admitted from in the first place) is an important determinant of readmission.[42] In this context, it is noteworthy that successful interventions to reduce readmission are generally those that focus on outpatient follow‐up, while inpatient‐only interventions have had less success.[7] This is reflected in our framework above, informed by the literature, highlighting the importance of coordination between inpatient and outpatient providers and the importance of postdischarge care, including monitoring and managing symptoms after discharge, prompt follow‐up appointments, the continuation of patient self‐management activities, monitoring for drug‐related problems after discharge, and the effective utilization of community supports. Accountable care organizations, once established, would be responsible for several components of this environment, including the provision of prompt and effective follow‐up care.

The implication of the framework is that if a hospital does not have control over most of the factors that influence its readmission rate, it should see financial incentives to reduce readmission rates as an opportunity to invest in relationships with the outpatient environment from which their patients are admitted and to which they are discharged. One can envision hospitals growing ever‐closer relationships with their network of primary care physician groups, community agencies, and home health services, rehabilitation facilities, and nursing homes through coordinated discharge planning, medication management, patient education, shared electronic medical records, structured handoffs in care, and systems of intensive outpatient monitoring. Our proposed framework, in other words, emphasizes that hospitals cannot reduce their readmission rates by focusing on aspects of care within their walls. They must forge new and stronger relationships with their communities if they are to be successful.

How Does the Framework Help Us Understand Which Readmissions Are Preventable?

Public reporting and financial penalties are currently tied to all‐cause readmission, but preventable readmissions are a more appealing outcome to target. In one study, the ranking of hospitals by all‐cause readmission rate had very little correlation with the ranking by preventable readmission rate.[5] However, researchers have struggled to establish standardized, valid, and reliable measures for determining what proportion of readmissions are in fact preventable, with estimates ranging from 5% to 79% in the published literature.[43]

The difficulty of accurately determining preventability stems from an inadequate understanding of the roles that patient comorbidities, transitional processes of care, individual patient behaviors, and social and environmental determinants of health play in the complex process of hospital recidivism. Our proposed elements of an ideal transition in care provide a structure to frame this discussion and suggest future research opportunities to allow a more accurate and reliable understanding of the spectrum of preventability. Care system leadership can use the framework to rigorously evaluate their readmissions and determine the extent to which the transitions process approached the ideal. For example, if a readmission occurs despite care processes that addressed most of the domains with high fidelity, it becomes much less likely that the readmission was preventable. It should be noted that the converse is not always true: When a transition falls well short of the ideal, it does not always imply that provision of a more ideal transition would necessarily have prevented the readmission, but it does make it more likely.

For educators, the framework may provide insights for trainees into the complexity of the transitions process and vulnerability of patients during this time, highlighting preventable aspects of readmissions that are within the grasp of the discharging clinician or team. It highlights the importance of medication reconciliation, synchronous communication, and predischarge teaching, which are measurable and teachable skills for non‐physician providers, housestaff, and medical students. It also may allow for more structured feedback, for example, on the quality of discharge summaries produced by trainees.

How Could the Framework Improve Risk Adjustment for Between‐Hospital Comparisons?

Under the Patient Protection and Affordable Care Act (PPACA), hospitals will be compared to one another using risk‐standardized readmission rates as a way to penalize poorly performing hospitals. However, risk‐adjustment models have only modest ability to predict hospital readmission.[6] Moreover, current approaches predominantly adjust for patients' medical comorbidities (which are easily measurable), but they do not adequately take into account the growing literature on other factors that influence readmission rates, including a patient's health literacy, visual or cognitive impairment, functional status, language barriers, and community‐level factors such as social supports.[44, 45]

The Ideal Transition of Care provides a comprehensive framework of hospital discharge quality that provides additional process measures on which hospitals could be compared rather than focusing solely on (inadequately) risk‐adjusted readmission rates. Indeed, most other quality and safety measures (such as the National Quality Forum's Safe Practices[46] and The Joint Commission's National Patient Safety Goals),[47] emphasize process over outcome, in part because of issues of fairness. Process measures are less subject to differences in patient populations and also change the focus from simply reducing readmissions to improving transitional care more broadly. These process measures should be based on our framework and should attempt to capture as many dimensions of an optimal care transition as possible.

Possible examples of process measures include: the accuracy of medication reconciliation at admission and discharge; provision of prompt outpatient follow‐up; provision of adequate systems to monitor and manage symptoms after discharge; advanced care planning in appropriate patients; and the quality of discharge education, incorporating measurements of the patient's understanding and ability to self‐manage their illness. At least some of these could be used now as part of a performance measurement set that highlights opportunities for immediate system change and can serve as performance milestones.

The framework could also be used to validate risk‐adjustment techniques. After accounting for patient factors, the remaining variability in outcomes should be accounted for by processes of care that are in the transitions framework. Once these processes are accurately measured, one can determine if indeed the remaining variability is due to transitions processes, or rather unaccounted factors that are not being measured and that hospitals may have little control over. Such work can lead to iterative refinement of patient risk‐adjustment models.

What Does the Framework Imply About Best Practices for Reducing Readmission Rates?

Despite the limitations of readmission rates as a quality measure noted above, hospitals presently face potentially large financial penalties for readmissions and are allocating resources to readmission reduction efforts. However, hospitals currently may not have enough guidance to know what actions to take to reduce readmissions, and thus could be spending money inefficiently and reducing the value proposition of focusing on readmissions.

A recent systematic review of interventions hospitals could employ to reduce readmissions identified several positive studies, but also many negative studies, and there were significant barriers to understanding what works to reduce readmissions.[7] For example, most of the interventions described in both positive and negative studies were multifaceted, and the authors were unable to identify which components of the intervention were most effective. Also, while several studies have identified risk factors for readmission,[6, 48, 49] very few studies have identified which subgroups of patients benefit most from specific interventions. Few of the studies described key contextual factors that may have led to successful or failed implementation, or the fidelity with which the intervention was implemented.[50, 51, 52]

Few if any of the studies were guided by a concept of the ideal transition in care.[10] Such a framework will better guide development of multifaceted interventions and provide an improved means for interpreting the results. Clearly, rigorously conducted, multicenter studies of readmission prevention interventions are needed to move the field forward. These studies should: 1) correlate implementation of specific intervention components with reductions in readmission rates to better understand the most effective components; 2) be adequately powered to show effect modification, ie, which patients benefit most from these interventions; and 3) rigorously measure environmental context and intervention fidelity, and employ mixed methods to better understand predictors of implementation success and failure.

Our framework can be used in the design and evaluation of such interventions. For example, interventions could be designed that incorporate as many of the domains of an ideal transition as possible, in particular those that span the inpatient and outpatient settings. Processes of care metrics can be developed that measure the extent to which each domain is delivered, analogous to the way the Joint Commission might aggregate individual scores on the 10 items in Acute Myocardial Infarction Core Measure Set[53] to provide a composite of the quality of care provided to patients with this diagnosis. These can be used to correlate certain intervention components with success in reducing readmissions and also in measuring intervention fidelity.

How Does the Framework Address the Extent to Which Readmissions Reflect Hospital Quality?

One of the chief problems with readmissionrates as a hospital quality measure is that many of the factors that influence readmission may not currently be under the hospital's control. The healthcare environment to which a patient is being discharged (and was admitted from in the first place) is an important determinant of readmission.[42] In this context, it is noteworthy that successful interventions to reduce readmission are generally those that focus on outpatient follow‐up, while inpatient‐only interventions have had less success.[7] This is reflected in our framework above, informed by the literature, highlighting the importance of coordination between inpatient and outpatient providers and the importance of postdischarge care, including monitoring and managing symptoms after discharge, prompt follow‐up appointments, the continuation of patient self‐management activities, monitoring for drug‐related problems after discharge, and the effective utilization of community supports. Accountable care organizations, once established, would be responsible for several components of this environment, including the provision of prompt and effective follow‐up care.

The implication of the framework is that if a hospital does not have control over most of the factors that influence its readmission rate, it should see financial incentives to reduce readmission rates as an opportunity to invest in relationships with the outpatient environment from which their patients are admitted and to which they are discharged. One can envision hospitals growing ever‐closer relationships with their network of primary care physician groups, community agencies, and home health services, rehabilitation facilities, and nursing homes through coordinated discharge planning, medication management, patient education, shared electronic medical records, structured handoffs in care, and systems of intensive outpatient monitoring. Our proposed framework, in other words, emphasizes that hospitals cannot reduce their readmission rates by focusing on aspects of care within their walls. They must forge new and stronger relationships with their communities if they are to be successful.

How Does the Framework Help Us Understand Which Readmissions Are Preventable?

Public reporting and financial penalties are currently tied to all‐cause readmission, but preventable readmissions are a more appealing outcome to target. In one study, the ranking of hospitals by all‐cause readmission rate had very little correlation with the ranking by preventable readmission rate.[5] However, researchers have struggled to establish standardized, valid, and reliable measures for determining what proportion of readmissions are in fact preventable, with estimates ranging from 5% to 79% in the published literature.[43]

The difficulty of accurately determining preventability stems from an inadequate understanding of the roles that patient comorbidities, transitional processes of care, individual patient behaviors, and social and environmental determinants of health play in the complex process of hospital recidivism. Our proposed elements of an ideal transition in care provide a structure to frame this discussion and suggest future research opportunities to allow a more accurate and reliable understanding of the spectrum of preventability. Care system leadership can use the framework to rigorously evaluate their readmissions and determine the extent to which the transitions process approached the ideal. For example, if a readmission occurs despite care processes that addressed most of the domains with high fidelity, it becomes much less likely that the readmission was preventable. It should be noted that the converse is not always true: When a transition falls well short of the ideal, it does not always imply that provision of a more ideal transition would necessarily have prevented the readmission, but it does make it more likely.

For educators, the framework may provide insights for trainees into the complexity of the transitions process and vulnerability of patients during this time, highlighting preventable aspects of readmissions that are within the grasp of the discharging clinician or team. It highlights the importance of medication reconciliation, synchronous communication, and predischarge teaching, which are measurable and teachable skills for non‐physician providers, housestaff, and medical students. It also may allow for more structured feedback, for example, on the quality of discharge summaries produced by trainees.

How Could the Framework Improve Risk Adjustment for Between‐Hospital Comparisons?

Under the Patient Protection and Affordable Care Act (PPACA), hospitals will be compared to one another using risk‐standardized readmission rates as a way to penalize poorly performing hospitals. However, risk‐adjustment models have only modest ability to predict hospital readmission.[6] Moreover, current approaches predominantly adjust for patients' medical comorbidities (which are easily measurable), but they do not adequately take into account the growing literature on other factors that influence readmission rates, including a patient's health literacy, visual or cognitive impairment, functional status, language barriers, and community‐level factors such as social supports.[44, 45]

The Ideal Transition of Care provides a comprehensive framework of hospital discharge quality that provides additional process measures on which hospitals could be compared rather than focusing solely on (inadequately) risk‐adjusted readmission rates. Indeed, most other quality and safety measures (such as the National Quality Forum's Safe Practices[46] and The Joint Commission's National Patient Safety Goals),[47] emphasize process over outcome, in part because of issues of fairness. Process measures are less subject to differences in patient populations and also change the focus from simply reducing readmissions to improving transitional care more broadly. These process measures should be based on our framework and should attempt to capture as many dimensions of an optimal care transition as possible.

Possible examples of process measures include: the accuracy of medication reconciliation at admission and discharge; provision of prompt outpatient follow‐up; provision of adequate systems to monitor and manage symptoms after discharge; advanced care planning in appropriate patients; and the quality of discharge education, incorporating measurements of the patient's understanding and ability to self‐manage their illness. At least some of these could be used now as part of a performance measurement set that highlights opportunities for immediate system change and can serve as performance milestones.

The framework could also be used to validate risk‐adjustment techniques. After accounting for patient factors, the remaining variability in outcomes should be accounted for by processes of care that are in the transitions framework. Once these processes are accurately measured, one can determine if indeed the remaining variability is due to transitions processes, or rather unaccounted factors that are not being measured and that hospitals may have little control over. Such work can lead to iterative refinement of patient risk‐adjustment models.

What Does the Framework Imply About Best Practices for Reducing Readmission Rates?

Despite the limitations of readmission rates as a quality measure noted above, hospitals presently face potentially large financial penalties for readmissions and are allocating resources to readmission reduction efforts. However, hospitals currently may not have enough guidance to know what actions to take to reduce readmissions, and thus could be spending money inefficiently and reducing the value proposition of focusing on readmissions.

A recent systematic review of interventions hospitals could employ to reduce readmissions identified several positive studies, but also many negative studies, and there were significant barriers to understanding what works to reduce readmissions.[7] For example, most of the interventions described in both positive and negative studies were multifaceted, and the authors were unable to identify which components of the intervention were most effective. Also, while several studies have identified risk factors for readmission,[6, 48, 49] very few studies have identified which subgroups of patients benefit most from specific interventions. Few of the studies described key contextual factors that may have led to successful or failed implementation, or the fidelity with which the intervention was implemented.[50, 51, 52]

Few if any of the studies were guided by a concept of the ideal transition in care.[10] Such a framework will better guide development of multifaceted interventions and provide an improved means for interpreting the results. Clearly, rigorously conducted, multicenter studies of readmission prevention interventions are needed to move the field forward. These studies should: 1) correlate implementation of specific intervention components with reductions in readmission rates to better understand the most effective components; 2) be adequately powered to show effect modification, ie, which patients benefit most from these interventions; and 3) rigorously measure environmental context and intervention fidelity, and employ mixed methods to better understand predictors of implementation success and failure.

Our framework can be used in the design and evaluation of such interventions. For example, interventions could be designed that incorporate as many of the domains of an ideal transition as possible, in particular those that span the inpatient and outpatient settings. Processes of care metrics can be developed that measure the extent to which each domain is delivered, analogous to the way the Joint Commission might aggregate individual scores on the 10 items in Acute Myocardial Infarction Core Measure Set[53] to provide a composite of the quality of care provided to patients with this diagnosis. These can be used to correlate certain intervention components with success in reducing readmissions and also in measuring intervention fidelity.

How Does the Framework Address the Extent to Which Readmissions Reflect Hospital Quality?

One of the chief problems with readmissionrates as a hospital quality measure is that many of the factors that influence readmission may not currently be under the hospital's control. The healthcare environment to which a patient is being discharged (and was admitted from in the first place) is an important determinant of readmission.[42] In this context, it is noteworthy that successful interventions to reduce readmission are generally those that focus on outpatient follow‐up, while inpatient‐only interventions have had less success.[7] This is reflected in our framework above, informed by the literature, highlighting the importance of coordination between inpatient and outpatient providers and the importance of postdischarge care, including monitoring and managing symptoms after discharge, prompt follow‐up appointments, the continuation of patient self‐management activities, monitoring for drug‐related problems after discharge, and the effective utilization of community supports. Accountable care organizations, once established, would be responsible for several components of this environment, including the provision of prompt and effective follow‐up care.

The implication of the framework is that if a hospital does not have control over most of the factors that influence its readmission rate, it should see financial incentives to reduce readmission rates as an opportunity to invest in relationships with the outpatient environment from which their patients are admitted and to which they are discharged. One can envision hospitals growing ever‐closer relationships with their network of primary care physician groups, community agencies, and home health services, rehabilitation facilities, and nursing homes through coordinated discharge planning, medication management, patient education, shared electronic medical records, structured handoffs in care, and systems of intensive outpatient monitoring. Our proposed framework, in other words, emphasizes that hospitals cannot reduce their readmission rates by focusing on aspects of care within their walls. They must forge new and stronger relationships with their communities if they are to be successful.

How Does the Framework Help Us Understand Which Readmissions Are Preventable?

Public reporting and financial penalties are currently tied to all‐cause readmission, but preventable readmissions are a more appealing outcome to target. In one study, the ranking of hospitals by all‐cause readmission rate had very little correlation with the ranking by preventable readmission rate.[5] However, researchers have struggled to establish standardized, valid, and reliable measures for determining what proportion of readmissions are in fact preventable, with estimates ranging from 5% to 79% in the published literature.[43]

The difficulty of accurately determining preventability stems from an inadequate understanding of the roles that patient comorbidities, transitional processes of care, individual patient behaviors, and social and environmental determinants of health play in the complex process of hospital recidivism. Our proposed elements of an ideal transition in care provide a structure to frame this discussion and suggest future research opportunities to allow a more accurate and reliable understanding of the spectrum of preventability. Care system leadership can use the framework to rigorously evaluate their readmissions and determine the extent to which the transitions process approached the ideal. For example, if a readmission occurs despite care processes that addressed most of the domains with high fidelity, it becomes much less likely that the readmission was preventable. It should be noted that the converse is not always true: When a transition falls well short of the ideal, it does not always imply that provision of a more ideal transition would necessarily have prevented the readmission, but it does make it more likely.

For educators, the framework may provide insights for trainees into the complexity of the transitions process and vulnerability of patients during this time, highlighting preventable aspects of readmissions that are within the grasp of the discharging clinician or team. It highlights the importance of medication reconciliation, synchronous communication, and predischarge teaching, which are measurable and teachable skills for non‐physician providers, housestaff, and medical students. It also may allow for more structured feedback, for example, on the quality of discharge summaries produced by trainees.

How Could the Framework Improve Risk Adjustment for Between‐Hospital Comparisons?

Under the Patient Protection and Affordable Care Act (PPACA), hospitals will be compared to one another using risk‐standardized readmission rates as a way to penalize poorly performing hospitals. However, risk‐adjustment models have only modest ability to predict hospital readmission.[6] Moreover, current approaches predominantly adjust for patients' medical comorbidities (which are easily measurable), but they do not adequately take into account the growing literature on other factors that influence readmission rates, including a patient's health literacy, visual or cognitive impairment, functional status, language barriers, and community‐level factors such as social supports.[44, 45]

The Ideal Transition of Care provides a comprehensive framework of hospital discharge quality that provides additional process measures on which hospitals could be compared rather than focusing solely on (inadequately) risk‐adjusted readmission rates. Indeed, most other quality and safety measures (such as the National Quality Forum's Safe Practices[46] and The Joint Commission's National Patient Safety Goals),[47] emphasize process over outcome, in part because of issues of fairness. Process measures are less subject to differences in patient populations and also change the focus from simply reducing readmissions to improving transitional care more broadly. These process measures should be based on our framework and should attempt to capture as many dimensions of an optimal care transition as possible.

Possible examples of process measures include: the accuracy of medication reconciliation at admission and discharge; provision of prompt outpatient follow‐up; provision of adequate systems to monitor and manage symptoms after discharge; advanced care planning in appropriate patients; and the quality of discharge education, incorporating measurements of the patient's understanding and ability to self‐manage their illness. At least some of these could be used now as part of a performance measurement set that highlights opportunities for immediate system change and can serve as performance milestones.

The framework could also be used to validate risk‐adjustment techniques. After accounting for patient factors, the remaining variability in outcomes should be accounted for by processes of care that are in the transitions framework. Once these processes are accurately measured, one can determine if indeed the remaining variability is due to transitions processes, or rather unaccounted factors that are not being measured and that hospitals may have little control over. Such work can lead to iterative refinement of patient risk‐adjustment models.

What Does the Framework Imply About Best Practices for Reducing Readmission Rates?

Despite the limitations of readmission rates as a quality measure noted above, hospitals presently face potentially large financial penalties for readmissions and are allocating resources to readmission reduction efforts. However, hospitals currently may not have enough guidance to know what actions to take to reduce readmissions, and thus could be spending money inefficiently and reducing the value proposition of focusing on readmissions.

A recent systematic review of interventions hospitals could employ to reduce readmissions identified several positive studies, but also many negative studies, and there were significant barriers to understanding what works to reduce readmissions.[7] For example, most of the interventions described in both positive and negative studies were multifaceted, and the authors were unable to identify which components of the intervention were most effective. Also, while several studies have identified risk factors for readmission,[6, 48, 49] very few studies have identified which subgroups of patients benefit most from specific interventions. Few of the studies described key contextual factors that may have led to successful or failed implementation, or the fidelity with which the intervention was implemented.[50, 51, 52]

Few if any of the studies were guided by a concept of the ideal transition in care.[10] Such a framework will better guide development of multifaceted interventions and provide an improved means for interpreting the results. Clearly, rigorously conducted, multicenter studies of readmission prevention interventions are needed to move the field forward. These studies should: 1) correlate implementation of specific intervention components with reductions in readmission rates to better understand the most effective components; 2) be adequately powered to show effect modification, ie, which patients benefit most from these interventions; and 3) rigorously measure environmental context and intervention fidelity, and employ mixed methods to better understand predictors of implementation success and failure.

Our framework can be used in the design and evaluation of such interventions. For example, interventions could be designed that incorporate as many of the domains of an ideal transition as possible, in particular those that span the inpatient and outpatient settings. Processes of care metrics can be developed that measure the extent to which each domain is delivered, analogous to the way the Joint Commission might aggregate individual scores on the 10 items in Acute Myocardial Infarction Core Measure Set[53] to provide a composite of the quality of care provided to patients with this diagnosis. These can be used to correlate certain intervention components with success in reducing readmissions and also in measuring intervention fidelity.

Study Sites

We examined the performance of 209 hospitals (63% of all 334 non‐federal facilities in California) participating in the California Hospital Assessment and Reporting Taskforce (CHART) at the time of the survey. CHART is a voluntary quality reporting initiative that began publicly reporting hospital quality data in January 2006.

Hospital‐level Organizational, Case‐mix, and Quality Data

Hospital organizational characteristics (eg, bed size) were obtained from publicly available discharge and utilization data sets from the California Office of Statewide Health Planning and Development (OSHPD). We also linked hospital‐level patient‐mix data (eg, race) from these OSHPD files.

We obtained quality of care data from CHART for January 2006 through June 2007, the time period corresponding to the survey. Quality metrics included 16 measures collected by the Center for Medicare and Medicaid Services (www.cms.hhs.gov) and extensively used in quality research.1, 4, 13, 15‐17 Rather than define a single measure, we examined multiple process measures, anticipating differential impacts of hospitalists on various processes of care for AMI, CHF, and pneumonia. Measures were further divided among those that are usually measured upon initial presentation to the hospital and those that are measured throughout the entire hospitalization and discharge. This division reflects the division of care in the hospital, where emergency room physicians are likely to have a more critical role for admission processes.

Survey Process

We surveyed all nonfederal, acute care hospitals in California that participated in CHART.2 We first identified contacts at each site via professional society mailing lists. We then sent web‐based surveys to all with available email addresses and a fax/paper survey to the remainder. We surveyed individuals between October 2006 and April 2007 and repeated the process at intervals of 1 to 3 weeks. For remaining nonrespondents, we placed a direct call unless consent to survey had been specifically refused. We contacted the following persons in sequence: (1) hospital executives or administrative leaders; (2) hospital medicine department leaders; (3) admitting emergency room personnel or medical staff officers; and (4) hospital website information. In the case of multiple responses with disagreement, the hospital/hospitalist leader's response was treated as the primary source. At each step, respondents were asked to answer questions only if they had a direct working knowledge of their hospitalist services.

Survey Data

Our key survey question to all respondents included whether the respondents could confirm their hospitals had at least one hospitalist medicine group. Hospital leaders were also asked to participate in a more comprehensive survey of their organizational and clinical characteristics. Within the comprehensive survey, leaders also provided estimates of the percent of general medical patients admitted by hospitalists. This measure, used in prior surveys of hospital leaders,9 was intended to be an easily understood approximation of the intensity of hospitalist utilization in any given hospital. A more rigorous, direct measure was not feasible due to the complexity of obtaining admission data over such a large, diverse set of hospitals.

Process Performance Measures

AMI measures assessed at admission included aspirin and ‐blocker administration within 24 hours of arrival. AMI measures assessed at discharge included aspirin administration, ‐blocker administration, angiotensin converting enzyme inhibitor (ACE‐I) (or angiotensin receptor blocker [ARB]) administration for left ventricular (LV) dysfunction, and smoking cessation counseling. There were no CHF admission measures. CHF discharge measures included assessment of LV function, the use of an ACE‐I or ARB for LV dysfunction, and smoking cessation counseling. Pneumonia admission measures included the drawing of blood cultures prior to the receipt of antibiotics, timely administration of initial antibiotics (<8 hours), and antibiotics consistent with recommendations. Pneumonia discharge measures included pneumococcal vaccination, flu vaccination, and smoking cessation counseling.

For each performance measure, we quantified the percentage of missed quality opportunities, defined as the number of patients who did not receive a care process divided by the number of eligible patients, multiplied by 100. In addition, we calculated composite scores for admission and discharge measures across each condition. We summed the numerators and denominators of individual performance measures to generate a disease‐specific composite numerator and denominator. Both individual and composite scores were produced using methodology outlined by the Center for Medicare & Medicaid Services.18 In order to retain as representative a sample of hospitals as possible, we calculated composite scores for hospitals that had a minimum of 25 observations in at least 2 of the quality indicators that made up each composite score.

Statistical Analysis

We used chi‐square tests, Student t tests, and Mann‐Whitney tests, where appropriate, to compare hospital‐level characteristics of hospitals that utilized hospitalists vs. those that did not. Similar analyses were performed among the subset of hospitals that utilized hospitalists. Among this subgroup of hospitals, we compared hospital‐level characteristics between hospitals that provided information regarding the percent of patients admitted by hospitalists vs. those who did not provide this information.

We used multivariable, generalized linear regression models to assess the relationship between having at least 1 hospitalist group and the percentage of missed quality of care measures. Because percentages were not normally distributed (ie, a majority of hospitals had few missed opportunities, while a minority had many), multivariable models employed log‐link functions with a gamma distribution.19, 20 Coefficients for our key predictor (presence of hospitalists) were transformed back to the original units (percentage of missed quality opportunities) so that a positive coefficient represented a higher number of quality measures missed relative to hospitals without hospitalists. Models were adjusted for factors previously reported to be associated with care quality. Hospital organizational characteristics included the number of beds, teaching status, registered nursing (RN) hours per adjusted patient day, and hospital ownership (for‐profit vs. not‐for‐profit). Hospital patient mix factors included annual percentage of admissions by insurance status (Medicare, Medicaid, other), annual percentage of admissions by race (white vs. nonwhite), annual percentage of do‐not‐resuscitate status at admission, and mean diagnosis‐related group‐based case‐mix index.21 We additionally adjusted for the number of cardiac catheterizations, a measure that moderately correlates with the number of cardiologists and technology utilization.22‐24 In our subset analysis among those hospitals with hospitalists, our key predictor for regression analyses was the percentage of patients admitted by hospitalists. For ease of interpretation, the percentage of patients admitted by hospitalists was centered on the mean across all respondent hospitals, and we report the effect of increasing by 10% the percentage of patients admitted by hospitalists. Models were adjusted for the same hospital organizational characteristics listed above. For those models, a positive coefficient also meant a higher number of measures missed.

For both sets of predictors, we additionally tested for the presence of interactions between the predictors and hospital bed size (both continuous as well as dichotomized at 150 beds) in composite measure performance, given the possibility that any hospitalist effect may be greater among smaller, resource‐limited hospitals. Tests for interaction were performed with the likelihood ratio test. In addition, to minimize any potential bias or loss of power that might result from limiting the analysis to hospitals with complete data, we used the multivariate imputation by chained equations method, as implemented in STATA 9.2 (StataCorp, College Station, TX), to create 10 imputed datasets.25 Imputation of missing values was restricted to confounding variables. Standard methods were then used to combine results over the 10 imputed datasets. We also applied Bonferroni corrections to composite measure tests based on the number of composites generated (n = 5). Thus, for the 5 inpatient composites created, standard definitions of significance (P 0.05) were corrected by dividing composite P values by 5, requiring P 0.01 for significance. The institutional review board of the University of California, San Francisco, approved the study. All analyses were performed using STATA 9.2.

Characteristics of Participating Sites

There were 209 eligible hospitals. All 209 (100%) hospitals provided data about the presence or absence of hospitalists via at least 1 of our survey strategies. The majority of identification of hospitalist utilization was via contact with either hospital or hospitalist leaders, n = 147 (70.3%). Web‐sites informed hospitalist prevalence in only 3 (1.4%) hospitals. There were 8 (3.8%) occurrences of disagreement between sources, all of which had available hospital/hospitalist leader responses. Only 1 (0.5%) hospital did not have the minimum 25 patients eligible for any disease‐specific quality measures during the data reporting period. Collectively, the remaining 208 hospitals accounted for 81% of California's acute care hospital population.

Comparisons of Sites With Hospitalists and Those Without

A total of 170 hospitals (82%) participating in CHART used hospitalists. Hospitals with and without hospitalists differed by a variety of characteristics (Table 1). Sites with hospitalists were larger, less likely to be for‐profit, had more registered nursing hours per day, and performed more cardiac catheterizations.

Relationship Between Hospitalist Group Utilization and the Percentage of Missed Quality Opportunities

Table 2 shows the frequency of missed quality opportunities in sites with hospitalists compared to those without. In general, for both individual and composite measures of quality, multivariable adjustment modestly attenuated the observed differences between the 2 groups of hospitals. We present only the more conservative adjusted estimates.

Compared to hospitals without hospitalists, those with hospitalists did not have any statistically significant differences in the individual and composite admission measures for each of the disease processes. In contrast, there were statistically significant differences between hospitalist and nonhospitalist sites for many individual cardiac processes of care that typically occur after admission from the emergency room (ie, LV function assessment for CHF) or those that occurred at discharge (ie, aspirin and ACE‐I/ARB at discharge for AMI). Similarly, the composite discharge scores for AMI and CHF revealed better overall process measure performance at sites with hospitalists, although the AMI composite did not meet statistical significance. There were no statistically significant differences between groups for the pneumonia process measures assessed at discharge. In addition, for composite measures there were no statistically significant interactions between hospitalist prevalence and bed size, although there was a trend (P = 0.06) for the CHF discharge composite, with a larger effect of hospitalists among smaller hospitals.

Percent of Patients Admitted by Hospitalists

Of the 171 hospitals with hospitalists, 71 (42%) estimated the percent of patients admitted by their hospitalist physicians. Among the respondents, the mean and median percentages of medical patients admitted by hospitalists were 51% (SD = 25%) and 49% (IQR = 30‐70%), respectively. Thirty hospitals were above the sample mean. Compared to nonrespondent sites, respondent hospitals took care of more white patients; otherwise, respondent and nonrespondent hospitals were similar in terms of bed size, location, performance across each measure, and other observable characteristics (Supporting Information, Appendix 1).

Relationship Between the Estimated Percentages of Medical Patients Admitted by Hospitalists and Missed Quality Opportunities

Table 3 displays the change in missed quality measures associated with each additional 10% of patients estimated to be admitted by hospitalists. A higher estimated percentage of patients admitted by hospitalists was associated with statistically significant improvements in quality of care across a majority of individual measures and for all composite discharge measures regardless of condition. For example, every 10% increase in the mean estimated number of patients admitted by hospitalists was associated with a mean of 0.6% (P < 0.001), 0.5% (P = 0.004), and 1.5% (P = 0.006) fewer missed quality opportunities for AMI, CHF, and pneumonia discharge process measures composites, respectively. In addition, for these composite measures, there were no statistically significant interactions between the estimated percentage of patients admitted by hospitalists and bed size (dichotomized at 150 beds), although there was a trend (P = 0.09) for the AMI discharge composite, with a larger effect of hospitalists among smaller hospitals.

In order to test the robustness of our results, we carried out 2 secondary analyses. First, we used multivariable models to generate a propensity score representing the predicted probability of being assigned to a hospital with hospitalists. We then used the propensity score as an additional covariate in subsequent multivariable models. In addition, we performed a complete‐case analysis (including only hospitals with complete data, n = 204) as a check on the sensitivity of our results to missing data. Neither analysis produced results substantially different from those presented.

Study Sites

We examined the performance of 209 hospitals (63% of all 334 non‐federal facilities in California) participating in the California Hospital Assessment and Reporting Taskforce (CHART) at the time of the survey. CHART is a voluntary quality reporting initiative that began publicly reporting hospital quality data in January 2006.

Hospital‐level Organizational, Case‐mix, and Quality Data

Hospital organizational characteristics (eg, bed size) were obtained from publicly available discharge and utilization data sets from the California Office of Statewide Health Planning and Development (OSHPD). We also linked hospital‐level patient‐mix data (eg, race) from these OSHPD files.

We obtained quality of care data from CHART for January 2006 through June 2007, the time period corresponding to the survey. Quality metrics included 16 measures collected by the Center for Medicare and Medicaid Services (www.cms.hhs.gov) and extensively used in quality research.1, 4, 13, 15‐17 Rather than define a single measure, we examined multiple process measures, anticipating differential impacts of hospitalists on various processes of care for AMI, CHF, and pneumonia. Measures were further divided among those that are usually measured upon initial presentation to the hospital and those that are measured throughout the entire hospitalization and discharge. This division reflects the division of care in the hospital, where emergency room physicians are likely to have a more critical role for admission processes.

Survey Process

We surveyed all nonfederal, acute care hospitals in California that participated in CHART.2 We first identified contacts at each site via professional society mailing lists. We then sent web‐based surveys to all with available email addresses and a fax/paper survey to the remainder. We surveyed individuals between October 2006 and April 2007 and repeated the process at intervals of 1 to 3 weeks. For remaining nonrespondents, we placed a direct call unless consent to survey had been specifically refused. We contacted the following persons in sequence: (1) hospital executives or administrative leaders; (2) hospital medicine department leaders; (3) admitting emergency room personnel or medical staff officers; and (4) hospital website information. In the case of multiple responses with disagreement, the hospital/hospitalist leader's response was treated as the primary source. At each step, respondents were asked to answer questions only if they had a direct working knowledge of their hospitalist services.

Survey Data

Our key survey question to all respondents included whether the respondents could confirm their hospitals had at least one hospitalist medicine group. Hospital leaders were also asked to participate in a more comprehensive survey of their organizational and clinical characteristics. Within the comprehensive survey, leaders also provided estimates of the percent of general medical patients admitted by hospitalists. This measure, used in prior surveys of hospital leaders,9 was intended to be an easily understood approximation of the intensity of hospitalist utilization in any given hospital. A more rigorous, direct measure was not feasible due to the complexity of obtaining admission data over such a large, diverse set of hospitals.

Process Performance Measures

AMI measures assessed at admission included aspirin and ‐blocker administration within 24 hours of arrival. AMI measures assessed at discharge included aspirin administration, ‐blocker administration, angiotensin converting enzyme inhibitor (ACE‐I) (or angiotensin receptor blocker [ARB]) administration for left ventricular (LV) dysfunction, and smoking cessation counseling. There were no CHF admission measures. CHF discharge measures included assessment of LV function, the use of an ACE‐I or ARB for LV dysfunction, and smoking cessation counseling. Pneumonia admission measures included the drawing of blood cultures prior to the receipt of antibiotics, timely administration of initial antibiotics (<8 hours), and antibiotics consistent with recommendations. Pneumonia discharge measures included pneumococcal vaccination, flu vaccination, and smoking cessation counseling.

For each performance measure, we quantified the percentage of missed quality opportunities, defined as the number of patients who did not receive a care process divided by the number of eligible patients, multiplied by 100. In addition, we calculated composite scores for admission and discharge measures across each condition. We summed the numerators and denominators of individual performance measures to generate a disease‐specific composite numerator and denominator. Both individual and composite scores were produced using methodology outlined by the Center for Medicare & Medicaid Services.18 In order to retain as representative a sample of hospitals as possible, we calculated composite scores for hospitals that had a minimum of 25 observations in at least 2 of the quality indicators that made up each composite score.

Statistical Analysis

We used chi‐square tests, Student t tests, and Mann‐Whitney tests, where appropriate, to compare hospital‐level characteristics of hospitals that utilized hospitalists vs. those that did not. Similar analyses were performed among the subset of hospitals that utilized hospitalists. Among this subgroup of hospitals, we compared hospital‐level characteristics between hospitals that provided information regarding the percent of patients admitted by hospitalists vs. those who did not provide this information.

We used multivariable, generalized linear regression models to assess the relationship between having at least 1 hospitalist group and the percentage of missed quality of care measures. Because percentages were not normally distributed (ie, a majority of hospitals had few missed opportunities, while a minority had many), multivariable models employed log‐link functions with a gamma distribution.19, 20 Coefficients for our key predictor (presence of hospitalists) were transformed back to the original units (percentage of missed quality opportunities) so that a positive coefficient represented a higher number of quality measures missed relative to hospitals without hospitalists. Models were adjusted for factors previously reported to be associated with care quality. Hospital organizational characteristics included the number of beds, teaching status, registered nursing (RN) hours per adjusted patient day, and hospital ownership (for‐profit vs. not‐for‐profit). Hospital patient mix factors included annual percentage of admissions by insurance status (Medicare, Medicaid, other), annual percentage of admissions by race (white vs. nonwhite), annual percentage of do‐not‐resuscitate status at admission, and mean diagnosis‐related group‐based case‐mix index.21 We additionally adjusted for the number of cardiac catheterizations, a measure that moderately correlates with the number of cardiologists and technology utilization.22‐24 In our subset analysis among those hospitals with hospitalists, our key predictor for regression analyses was the percentage of patients admitted by hospitalists. For ease of interpretation, the percentage of patients admitted by hospitalists was centered on the mean across all respondent hospitals, and we report the effect of increasing by 10% the percentage of patients admitted by hospitalists. Models were adjusted for the same hospital organizational characteristics listed above. For those models, a positive coefficient also meant a higher number of measures missed.

For both sets of predictors, we additionally tested for the presence of interactions between the predictors and hospital bed size (both continuous as well as dichotomized at 150 beds) in composite measure performance, given the possibility that any hospitalist effect may be greater among smaller, resource‐limited hospitals. Tests for interaction were performed with the likelihood ratio test. In addition, to minimize any potential bias or loss of power that might result from limiting the analysis to hospitals with complete data, we used the multivariate imputation by chained equations method, as implemented in STATA 9.2 (StataCorp, College Station, TX), to create 10 imputed datasets.25 Imputation of missing values was restricted to confounding variables. Standard methods were then used to combine results over the 10 imputed datasets. We also applied Bonferroni corrections to composite measure tests based on the number of composites generated (n = 5). Thus, for the 5 inpatient composites created, standard definitions of significance (P 0.05) were corrected by dividing composite P values by 5, requiring P 0.01 for significance. The institutional review board of the University of California, San Francisco, approved the study. All analyses were performed using STATA 9.2.

Characteristics of Participating Sites

There were 209 eligible hospitals. All 209 (100%) hospitals provided data about the presence or absence of hospitalists via at least 1 of our survey strategies. The majority of identification of hospitalist utilization was via contact with either hospital or hospitalist leaders, n = 147 (70.3%). Web‐sites informed hospitalist prevalence in only 3 (1.4%) hospitals. There were 8 (3.8%) occurrences of disagreement between sources, all of which had available hospital/hospitalist leader responses. Only 1 (0.5%) hospital did not have the minimum 25 patients eligible for any disease‐specific quality measures during the data reporting period. Collectively, the remaining 208 hospitals accounted for 81% of California's acute care hospital population.

Comparisons of Sites With Hospitalists and Those Without

A total of 170 hospitals (82%) participating in CHART used hospitalists. Hospitals with and without hospitalists differed by a variety of characteristics (Table 1). Sites with hospitalists were larger, less likely to be for‐profit, had more registered nursing hours per day, and performed more cardiac catheterizations.

Relationship Between Hospitalist Group Utilization and the Percentage of Missed Quality Opportunities

Table 2 shows the frequency of missed quality opportunities in sites with hospitalists compared to those without. In general, for both individual and composite measures of quality, multivariable adjustment modestly attenuated the observed differences between the 2 groups of hospitals. We present only the more conservative adjusted estimates.

Compared to hospitals without hospitalists, those with hospitalists did not have any statistically significant differences in the individual and composite admission measures for each of the disease processes. In contrast, there were statistically significant differences between hospitalist and nonhospitalist sites for many individual cardiac processes of care that typically occur after admission from the emergency room (ie, LV function assessment for CHF) or those that occurred at discharge (ie, aspirin and ACE‐I/ARB at discharge for AMI). Similarly, the composite discharge scores for AMI and CHF revealed better overall process measure performance at sites with hospitalists, although the AMI composite did not meet statistical significance. There were no statistically significant differences between groups for the pneumonia process measures assessed at discharge. In addition, for composite measures there were no statistically significant interactions between hospitalist prevalence and bed size, although there was a trend (P = 0.06) for the CHF discharge composite, with a larger effect of hospitalists among smaller hospitals.

Percent of Patients Admitted by Hospitalists

Of the 171 hospitals with hospitalists, 71 (42%) estimated the percent of patients admitted by their hospitalist physicians. Among the respondents, the mean and median percentages of medical patients admitted by hospitalists were 51% (SD = 25%) and 49% (IQR = 30‐70%), respectively. Thirty hospitals were above the sample mean. Compared to nonrespondent sites, respondent hospitals took care of more white patients; otherwise, respondent and nonrespondent hospitals were similar in terms of bed size, location, performance across each measure, and other observable characteristics (Supporting Information, Appendix 1).

Relationship Between the Estimated Percentages of Medical Patients Admitted by Hospitalists and Missed Quality Opportunities

Table 3 displays the change in missed quality measures associated with each additional 10% of patients estimated to be admitted by hospitalists. A higher estimated percentage of patients admitted by hospitalists was associated with statistically significant improvements in quality of care across a majority of individual measures and for all composite discharge measures regardless of condition. For example, every 10% increase in the mean estimated number of patients admitted by hospitalists was associated with a mean of 0.6% (P < 0.001), 0.5% (P = 0.004), and 1.5% (P = 0.006) fewer missed quality opportunities for AMI, CHF, and pneumonia discharge process measures composites, respectively. In addition, for these composite measures, there were no statistically significant interactions between the estimated percentage of patients admitted by hospitalists and bed size (dichotomized at 150 beds), although there was a trend (P = 0.09) for the AMI discharge composite, with a larger effect of hospitalists among smaller hospitals.

In order to test the robustness of our results, we carried out 2 secondary analyses. First, we used multivariable models to generate a propensity score representing the predicted probability of being assigned to a hospital with hospitalists. We then used the propensity score as an additional covariate in subsequent multivariable models. In addition, we performed a complete‐case analysis (including only hospitals with complete data, n = 204) as a check on the sensitivity of our results to missing data. Neither analysis produced results substantially different from those presented.

Study Sites

We examined the performance of 209 hospitals (63% of all 334 non‐federal facilities in California) participating in the California Hospital Assessment and Reporting Taskforce (CHART) at the time of the survey. CHART is a voluntary quality reporting initiative that began publicly reporting hospital quality data in January 2006.

Hospital‐level Organizational, Case‐mix, and Quality Data

Hospital organizational characteristics (eg, bed size) were obtained from publicly available discharge and utilization data sets from the California Office of Statewide Health Planning and Development (OSHPD). We also linked hospital‐level patient‐mix data (eg, race) from these OSHPD files.

We obtained quality of care data from CHART for January 2006 through June 2007, the time period corresponding to the survey. Quality metrics included 16 measures collected by the Center for Medicare and Medicaid Services (www.cms.hhs.gov) and extensively used in quality research.1, 4, 13, 15‐17 Rather than define a single measure, we examined multiple process measures, anticipating differential impacts of hospitalists on various processes of care for AMI, CHF, and pneumonia. Measures were further divided among those that are usually measured upon initial presentation to the hospital and those that are measured throughout the entire hospitalization and discharge. This division reflects the division of care in the hospital, where emergency room physicians are likely to have a more critical role for admission processes.

Survey Process

We surveyed all nonfederal, acute care hospitals in California that participated in CHART.2 We first identified contacts at each site via professional society mailing lists. We then sent web‐based surveys to all with available email addresses and a fax/paper survey to the remainder. We surveyed individuals between October 2006 and April 2007 and repeated the process at intervals of 1 to 3 weeks. For remaining nonrespondents, we placed a direct call unless consent to survey had been specifically refused. We contacted the following persons in sequence: (1) hospital executives or administrative leaders; (2) hospital medicine department leaders; (3) admitting emergency room personnel or medical staff officers; and (4) hospital website information. In the case of multiple responses with disagreement, the hospital/hospitalist leader's response was treated as the primary source. At each step, respondents were asked to answer questions only if they had a direct working knowledge of their hospitalist services.

Survey Data

Our key survey question to all respondents included whether the respondents could confirm their hospitals had at least one hospitalist medicine group. Hospital leaders were also asked to participate in a more comprehensive survey of their organizational and clinical characteristics. Within the comprehensive survey, leaders also provided estimates of the percent of general medical patients admitted by hospitalists. This measure, used in prior surveys of hospital leaders,9 was intended to be an easily understood approximation of the intensity of hospitalist utilization in any given hospital. A more rigorous, direct measure was not feasible due to the complexity of obtaining admission data over such a large, diverse set of hospitals.

Process Performance Measures

AMI measures assessed at admission included aspirin and ‐blocker administration within 24 hours of arrival. AMI measures assessed at discharge included aspirin administration, ‐blocker administration, angiotensin converting enzyme inhibitor (ACE‐I) (or angiotensin receptor blocker [ARB]) administration for left ventricular (LV) dysfunction, and smoking cessation counseling. There were no CHF admission measures. CHF discharge measures included assessment of LV function, the use of an ACE‐I or ARB for LV dysfunction, and smoking cessation counseling. Pneumonia admission measures included the drawing of blood cultures prior to the receipt of antibiotics, timely administration of initial antibiotics (<8 hours), and antibiotics consistent with recommendations. Pneumonia discharge measures included pneumococcal vaccination, flu vaccination, and smoking cessation counseling.

For each performance measure, we quantified the percentage of missed quality opportunities, defined as the number of patients who did not receive a care process divided by the number of eligible patients, multiplied by 100. In addition, we calculated composite scores for admission and discharge measures across each condition. We summed the numerators and denominators of individual performance measures to generate a disease‐specific composite numerator and denominator. Both individual and composite scores were produced using methodology outlined by the Center for Medicare & Medicaid Services.18 In order to retain as representative a sample of hospitals as possible, we calculated composite scores for hospitals that had a minimum of 25 observations in at least 2 of the quality indicators that made up each composite score.

Statistical Analysis

We used chi‐square tests, Student t tests, and Mann‐Whitney tests, where appropriate, to compare hospital‐level characteristics of hospitals that utilized hospitalists vs. those that did not. Similar analyses were performed among the subset of hospitals that utilized hospitalists. Among this subgroup of hospitals, we compared hospital‐level characteristics between hospitals that provided information regarding the percent of patients admitted by hospitalists vs. those who did not provide this information.

We used multivariable, generalized linear regression models to assess the relationship between having at least 1 hospitalist group and the percentage of missed quality of care measures. Because percentages were not normally distributed (ie, a majority of hospitals had few missed opportunities, while a minority had many), multivariable models employed log‐link functions with a gamma distribution.19, 20 Coefficients for our key predictor (presence of hospitalists) were transformed back to the original units (percentage of missed quality opportunities) so that a positive coefficient represented a higher number of quality measures missed relative to hospitals without hospitalists. Models were adjusted for factors previously reported to be associated with care quality. Hospital organizational characteristics included the number of beds, teaching status, registered nursing (RN) hours per adjusted patient day, and hospital ownership (for‐profit vs. not‐for‐profit). Hospital patient mix factors included annual percentage of admissions by insurance status (Medicare, Medicaid, other), annual percentage of admissions by race (white vs. nonwhite), annual percentage of do‐not‐resuscitate status at admission, and mean diagnosis‐related group‐based case‐mix index.21 We additionally adjusted for the number of cardiac catheterizations, a measure that moderately correlates with the number of cardiologists and technology utilization.22‐24 In our subset analysis among those hospitals with hospitalists, our key predictor for regression analyses was the percentage of patients admitted by hospitalists. For ease of interpretation, the percentage of patients admitted by hospitalists was centered on the mean across all respondent hospitals, and we report the effect of increasing by 10% the percentage of patients admitted by hospitalists. Models were adjusted for the same hospital organizational characteristics listed above. For those models, a positive coefficient also meant a higher number of measures missed.

For both sets of predictors, we additionally tested for the presence of interactions between the predictors and hospital bed size (both continuous as well as dichotomized at 150 beds) in composite measure performance, given the possibility that any hospitalist effect may be greater among smaller, resource‐limited hospitals. Tests for interaction were performed with the likelihood ratio test. In addition, to minimize any potential bias or loss of power that might result from limiting the analysis to hospitals with complete data, we used the multivariate imputation by chained equations method, as implemented in STATA 9.2 (StataCorp, College Station, TX), to create 10 imputed datasets.25 Imputation of missing values was restricted to confounding variables. Standard methods were then used to combine results over the 10 imputed datasets. We also applied Bonferroni corrections to composite measure tests based on the number of composites generated (n = 5). Thus, for the 5 inpatient composites created, standard definitions of significance (P 0.05) were corrected by dividing composite P values by 5, requiring P 0.01 for significance. The institutional review board of the University of California, San Francisco, approved the study. All analyses were performed using STATA 9.2.

Characteristics of Participating Sites

There were 209 eligible hospitals. All 209 (100%) hospitals provided data about the presence or absence of hospitalists via at least 1 of our survey strategies. The majority of identification of hospitalist utilization was via contact with either hospital or hospitalist leaders, n = 147 (70.3%). Web‐sites informed hospitalist prevalence in only 3 (1.4%) hospitals. There were 8 (3.8%) occurrences of disagreement between sources, all of which had available hospital/hospitalist leader responses. Only 1 (0.5%) hospital did not have the minimum 25 patients eligible for any disease‐specific quality measures during the data reporting period. Collectively, the remaining 208 hospitals accounted for 81% of California's acute care hospital population.

Comparisons of Sites With Hospitalists and Those Without

A total of 170 hospitals (82%) participating in CHART used hospitalists. Hospitals with and without hospitalists differed by a variety of characteristics (Table 1). Sites with hospitalists were larger, less likely to be for‐profit, had more registered nursing hours per day, and performed more cardiac catheterizations.

Relationship Between Hospitalist Group Utilization and the Percentage of Missed Quality Opportunities

Table 2 shows the frequency of missed quality opportunities in sites with hospitalists compared to those without. In general, for both individual and composite measures of quality, multivariable adjustment modestly attenuated the observed differences between the 2 groups of hospitals. We present only the more conservative adjusted estimates.

Compared to hospitals without hospitalists, those with hospitalists did not have any statistically significant differences in the individual and composite admission measures for each of the disease processes. In contrast, there were statistically significant differences between hospitalist and nonhospitalist sites for many individual cardiac processes of care that typically occur after admission from the emergency room (ie, LV function assessment for CHF) or those that occurred at discharge (ie, aspirin and ACE‐I/ARB at discharge for AMI). Similarly, the composite discharge scores for AMI and CHF revealed better overall process measure performance at sites with hospitalists, although the AMI composite did not meet statistical significance. There were no statistically significant differences between groups for the pneumonia process measures assessed at discharge. In addition, for composite measures there were no statistically significant interactions between hospitalist prevalence and bed size, although there was a trend (P = 0.06) for the CHF discharge composite, with a larger effect of hospitalists among smaller hospitals.

Percent of Patients Admitted by Hospitalists

Of the 171 hospitals with hospitalists, 71 (42%) estimated the percent of patients admitted by their hospitalist physicians. Among the respondents, the mean and median percentages of medical patients admitted by hospitalists were 51% (SD = 25%) and 49% (IQR = 30‐70%), respectively. Thirty hospitals were above the sample mean. Compared to nonrespondent sites, respondent hospitals took care of more white patients; otherwise, respondent and nonrespondent hospitals were similar in terms of bed size, location, performance across each measure, and other observable characteristics (Supporting Information, Appendix 1).

Relationship Between the Estimated Percentages of Medical Patients Admitted by Hospitalists and Missed Quality Opportunities

Table 3 displays the change in missed quality measures associated with each additional 10% of patients estimated to be admitted by hospitalists. A higher estimated percentage of patients admitted by hospitalists was associated with statistically significant improvements in quality of care across a majority of individual measures and for all composite discharge measures regardless of condition. For example, every 10% increase in the mean estimated number of patients admitted by hospitalists was associated with a mean of 0.6% (P < 0.001), 0.5% (P = 0.004), and 1.5% (P = 0.006) fewer missed quality opportunities for AMI, CHF, and pneumonia discharge process measures composites, respectively. In addition, for these composite measures, there were no statistically significant interactions between the estimated percentage of patients admitted by hospitalists and bed size (dichotomized at 150 beds), although there was a trend (P = 0.09) for the AMI discharge composite, with a larger effect of hospitalists among smaller hospitals.

In order to test the robustness of our results, we carried out 2 secondary analyses. First, we used multivariable models to generate a propensity score representing the predicted probability of being assigned to a hospital with hospitalists. We then used the propensity score as an additional covariate in subsequent multivariable models. In addition, we performed a complete‐case analysis (including only hospitals with complete data, n = 204) as a check on the sensitivity of our results to missing data. Neither analysis produced results substantially different from those presented.

Sites and Subjects

We targeted all nonfederal, nonspecialty, acute care hospitals in California (n = 334) for this survey. We limited our survey to California in order to maximize our local resources and to improve implementation of and response to the survey. Additionally, California's size and diversity gives it disproportionate impact and potential generalizability. At each site, we focused our efforts on identifying and surveying executives or administrative leaders involved in organizational and staff decisions, specifically the decision whether or not to hire and/or fund a hospitalist program and potentially direct its activities (described in more detail below). The University of California, San Francisco, Committee on Human Research approved the research protocol.

We identified hospital leaders at each site by merging information from multiple sources. These included the American Hospital Association database, the California Hospital Association, the Hospital Association of Southern California (HASC), the California Health Care Safety Net Institute, and individual hospital websites.

Survey Development

Our survey was based upon instruments used in previous research examining hospital medicine group organizational structure15, 22 and enhanced with questions developed by the research team (A.D.A., E.E.V., R.M.W.). The survey was pretested in an advisory group of 5 hospital Chief Executive Officers (CEOs), Chief Medical Officers (CMOs), and Vice Presidents for Medical Affairs (VPMAs) from sites across California. Based on their input, we removed, edited, or added questions to our survey. This advisory group also helped the research team design our survey process.

Our final survey defined a hospitalist as a physician who spends all or the majority of his or her clinical, administrative, educational, or research activities in the care of hospitalized patients.4 We collected data in 4 areas: (1) We asked hospital leaders to confirm the presence or absence of at least 1 hospitalist group practicing within the surveyed hospital. We also asked for the year the first hospitalist group began practicing within the specified hospital. (2) We asked hospital leaders to indicate, among a prespecified list of 11 choices, the reason(s) they implemented a hospitalist group at the surveyed hospital. Surveyed categories included: (a) care for uncovered patients (patients without an identified doctor and/or uninsured), (b) improve costs, (c) improve length of stay, (d) improve emergency department throughput, (e) primary care provider demand, (f) improve patient satisfaction, (g) improve emergency room staffing, (h) quality improvement needs, (i) specialist physician demand, (j) overnight coverage, and (k) surgical comanagement. Due to the close relationship between cost and length of stay, we combined these 2 categories into a single category for reporting and analysis. This resulted in 10 final categories. We asked leaders who did not identify a practicing hospitalist group about the likelihood of hospitalists practicing at their hospital within the next 5 years and the reason(s) for future implementation. (3) We asked leaders to describe the services currently provided among a prespecified list of clinical care duties that go beyond the scope of inpatient general internal medicine (eg, surgical comanagement, rapid response team leadership) as well as nonclinical duties (eg, quality improvement activities, systems project implementation). If hospitalists did not currently provide the identified service, we asked leaders to indicate if they would be inclined to involve hospitalists in the specified service in the future. (4) Finally, we asked hospital leaders their opinion regarding the need for further training or certification for hospitalists.

Survey Protocol

We administered surveys between October 2006 and April 2007. We initially emailed the survey. We repeated this process for nonrespondents at intervals of 1 to 3 weeks after the initial emailing. Next, we sent nonrespondents a physical mailing with a reminder letter. Finally, we made phone calls to those who had not responded within 4 weeks of the last mailed letter. We asked survey recipients to respond only if they felt they had an adequate working knowledge of the hospitalist service at their hospital. If they did not feel they could adequately answer all questions, we allowed them to forward the instrument to others with a better working knowledge of the service.

Because we allowed recipients to forward the survey, we occasionally received 2 surveys from 1 site. In this case, we selected the survey according to the following prioritization order: (1) CEOs/COOs, (2) CMOs, (3) VPMAs, and (4) other vice presidents (VPs) or executive/administrative leaders with staff organization knowledge and responsibilities.

Hospital Descriptive Data

We obtained hospital organizational data from the California Office of Statewide Health Planning and Development's (OSHPD) publicly available Case Mix Index Data, hospital Annual Financial Data, aggregated Patient Discharge Data, and Utilization Data from 2006.23 Organizational characteristics included hospital size, location, profit status, payor mix, and diagnosis‐related groupbased case‐mix. Teaching status was determined from the 2005 American Hospital Association database. Membership status in California's voluntary quality reporting initiative, California Hospital Assessment and Reporting Taskforce (CHART), was publicly available at http://www.calhospitalcompare.org.

Statistical Analyses

We performed univariable analyses to characterize survey respondents, followed by bivariable analyses to compare hospital characteristics and patient mix of responding and nonresponding hospitals. We used similar methods to characterize respondent hospitals with and without at least 1 hospitalist group. We compared continuous data with the Students t tests or Mann‐Whitney tests as appropriate and categorical data with chi‐square tests.

We then summarized the number of times a specific rationale was cited by hospital leaders for implementing a hospitalist group. Among hospitals that did not have a hospitalist system in place at the time of the survey, we asked if they were planning on starting one within the next 5 years. For these hospitals, we used content analysis to summarize open‐ended responses in order to understand factors that are currently influencing these hospital leaders to consider implementing a hospitalist group.

Next, we aimed to understand what clinical and nonclinical roles hospitalists were performing in hospitals with established hospitalist programs. Clinical activities were divided into general clinical areas, triage/emergency‐related, or administrative activities. First, we summarized the number and percent of programs performing each clinical and nonclinical activity. This was followed by logistic regression analyses to assess whether the time period that hospitalist groups began practicing or additional hospital characteristics predicted the performance of individual hospitalist activities. To guard against overfitting of models, analyses were limited to rationales that were cited a minimum of 50 times.24 Hospital factors were selected on the basis of face validity and advisory group input and included hospital bed size, ownership status (public vs. private), teaching status, and membership status in CHART. We divided the year of hospitalist program implementation into 3 time periods: (1) before 2002, (2) between 2002 and 2004, and (3) 2005 or later.

Finally, we described the percentage of hospitals that favored having their hospitalist group(s) perform each of the identified clinical or nonclinical activities, if they were not already performing them. We performed analyses with statistical software (Stata Version 9.2, College Station, TX).

Respondent Characteristics

We received 200 survey responses. Of those, we excluded 15 duplicates (eg, a survey from both the CEO and VPMA) and 6 responses identified as coming from hospitalists who did not have a leadership position in the hospital. Thus, the final hospital leader survey response rate was 54% (n = 179). Forty‐six percent of the final responses were from CEOs or COOs; 37% of responses were from CMOs, VPMAs, and medical directors; and the remaining 17% of responses were from other VPs or administrative directors.

Respondent and nonrespondent hospitals were statistically similar in terms of teaching status and participation in CHART. Hospital patient census, intensive care unit census, payer mix, and diagnosis‐related groupbased case‐mix revealed no statistically significant differences between groups (P > 0.05). Respondent hospitals tended to have fewer beds and were more often for‐profit compared to nonrespondents (P = 0.05 and P < 0.01, respectively).

Descriptive Characteristics of Hospitals with Hospitalists

Sixty‐four percent (n = 115) of hospital leaders stated that they utilized hospitalists for at least some patients. Hospitals with hospitalists were statistically more likely (P < 0.05) to be larger, a major teaching hospital, or a member of a voluntary quality reporting initiative (Table 1).

Among all hospitals with hospitalists, 39% estimated that hospitalists cared for at least one‐half of admitted medical patients, and 7% stated that hospitalists cared for all patients. Twenty‐four percent of respondents were unable to provide a quantitative estimate of the percent of patients cared for by hospitalists. When asked about expectations of growth in the coming year, 57% of respondents with hospitalists expected to see increases in the number of hospitalists at their hospital, and none expected a decrease. Among the 64 respondent hospitals that currently did not have a hospitalist program, 44% (n = 28) of the hospital leaders felt hospitalists would be managing patients in the future. Of those, 93% felt this would occur within the next 2 years.

Reasons for Implementing Hospitalists

Hospital leaders reported that the most important reasons for implementing a hospitalist model included caring for uncovered patients (68%), decreasing hospital costs and length of stay (63%), and improving throughput in the emergency room (62%). We provide additional reasons in Figure 1. In addition, leaders often identified multiple factors in the decision to utilize hospitalists, including demand from primary care doctors, patient satisfaction, and quality improvement. Among the 28 hospitals that currently did not have hospitalists but anticipated that they would soon (data not shown), the need to improve quality was the most commonly cited reason (54% of respondents) for expecting to start a program within 2 years, followed by demand from primary care doctors (46% of respondents).

Figure 1

Clinical Practice of Hospitalists and Expectations for Future Growth

Hospitalists perform a wide array of clinical and nonclinical duties (Figure 2). In addition to general medical care, the most common clinical activities of hospitalists included screening medical admissions from the emergency room for appropriateness of admission and triaging to appropriate level of care (67%), triaging patients transferred from an outside hospital (72%), and comanaging surgical patients (66%). The most common nonclinical activity was participation in quality improvement activities (72%). Multivariable analyses demonstrated that the performance of the most prevalent activities was not usually associated with the year of hospitalist implementation or hospital characteristics. An exception was that newly initiated programs had a statistically significant decreased odds of involvement in clinical guideline development (odds ratio [OR], 0.3; 95% confidence interval [CI], 0.1‐0.9) and a trend toward decreased leadership in quality improvement (OR, 0.3; 95% CI, 0.1‐1.1). Hospitalists at teaching hospitals had increased odds of managing patient transfers (OR, 4.7; 95% CI, 1.0‐21.2), whereas for‐profit hospitals had lower odds of screening patients in the emergency room (OR, 0.1; 95% CI, 0.0‐0.7).

Figure 2

Among those hospitals with hospitalists who were not presently involved in any of the above activities, there was a widespread interest among hospital leaders to have their hospitalist group(s) lead or participate in them (Figure 3). The most commonly cited activities included participation in inpatient clinical guideline development (85%), implementation of system‐wide projects (81%) (eg, computerized physician order entry system), participation on a rapid response team (80%), and caring for patients in an observation unit (80%).

Figure 3

Training and Certification for Hospitalists

About two‐thirds (64%) of hospital leaders with a hospitalist group(s) agreed or strongly agreed that hospitalists should have additional training and/or certification. Seventeen percent were undecided, whereas 11% either disagreed or strongly disagreed, and the remaining 8% did not provide an opinion.

Sites and Subjects

We targeted all nonfederal, nonspecialty, acute care hospitals in California (n = 334) for this survey. We limited our survey to California in order to maximize our local resources and to improve implementation of and response to the survey. Additionally, California's size and diversity gives it disproportionate impact and potential generalizability. At each site, we focused our efforts on identifying and surveying executives or administrative leaders involved in organizational and staff decisions, specifically the decision whether or not to hire and/or fund a hospitalist program and potentially direct its activities (described in more detail below). The University of California, San Francisco, Committee on Human Research approved the research protocol.

We identified hospital leaders at each site by merging information from multiple sources. These included the American Hospital Association database, the California Hospital Association, the Hospital Association of Southern California (HASC), the California Health Care Safety Net Institute, and individual hospital websites.

Survey Development

Our survey was based upon instruments used in previous research examining hospital medicine group organizational structure15, 22 and enhanced with questions developed by the research team (A.D.A., E.E.V., R.M.W.). The survey was pretested in an advisory group of 5 hospital Chief Executive Officers (CEOs), Chief Medical Officers (CMOs), and Vice Presidents for Medical Affairs (VPMAs) from sites across California. Based on their input, we removed, edited, or added questions to our survey. This advisory group also helped the research team design our survey process.

Our final survey defined a hospitalist as a physician who spends all or the majority of his or her clinical, administrative, educational, or research activities in the care of hospitalized patients.4 We collected data in 4 areas: (1) We asked hospital leaders to confirm the presence or absence of at least 1 hospitalist group practicing within the surveyed hospital. We also asked for the year the first hospitalist group began practicing within the specified hospital. (2) We asked hospital leaders to indicate, among a prespecified list of 11 choices, the reason(s) they implemented a hospitalist group at the surveyed hospital. Surveyed categories included: (a) care for uncovered patients (patients without an identified doctor and/or uninsured), (b) improve costs, (c) improve length of stay, (d) improve emergency department throughput, (e) primary care provider demand, (f) improve patient satisfaction, (g) improve emergency room staffing, (h) quality improvement needs, (i) specialist physician demand, (j) overnight coverage, and (k) surgical comanagement. Due to the close relationship between cost and length of stay, we combined these 2 categories into a single category for reporting and analysis. This resulted in 10 final categories. We asked leaders who did not identify a practicing hospitalist group about the likelihood of hospitalists practicing at their hospital within the next 5 years and the reason(s) for future implementation. (3) We asked leaders to describe the services currently provided among a prespecified list of clinical care duties that go beyond the scope of inpatient general internal medicine (eg, surgical comanagement, rapid response team leadership) as well as nonclinical duties (eg, quality improvement activities, systems project implementation). If hospitalists did not currently provide the identified service, we asked leaders to indicate if they would be inclined to involve hospitalists in the specified service in the future. (4) Finally, we asked hospital leaders their opinion regarding the need for further training or certification for hospitalists.

Survey Protocol

We administered surveys between October 2006 and April 2007. We initially emailed the survey. We repeated this process for nonrespondents at intervals of 1 to 3 weeks after the initial emailing. Next, we sent nonrespondents a physical mailing with a reminder letter. Finally, we made phone calls to those who had not responded within 4 weeks of the last mailed letter. We asked survey recipients to respond only if they felt they had an adequate working knowledge of the hospitalist service at their hospital. If they did not feel they could adequately answer all questions, we allowed them to forward the instrument to others with a better working knowledge of the service.

Because we allowed recipients to forward the survey, we occasionally received 2 surveys from 1 site. In this case, we selected the survey according to the following prioritization order: (1) CEOs/COOs, (2) CMOs, (3) VPMAs, and (4) other vice presidents (VPs) or executive/administrative leaders with staff organization knowledge and responsibilities.

Hospital Descriptive Data

We obtained hospital organizational data from the California Office of Statewide Health Planning and Development's (OSHPD) publicly available Case Mix Index Data, hospital Annual Financial Data, aggregated Patient Discharge Data, and Utilization Data from 2006.23 Organizational characteristics included hospital size, location, profit status, payor mix, and diagnosis‐related groupbased case‐mix. Teaching status was determined from the 2005 American Hospital Association database. Membership status in California's voluntary quality reporting initiative, California Hospital Assessment and Reporting Taskforce (CHART), was publicly available at http://www.calhospitalcompare.org.

Statistical Analyses

We performed univariable analyses to characterize survey respondents, followed by bivariable analyses to compare hospital characteristics and patient mix of responding and nonresponding hospitals. We used similar methods to characterize respondent hospitals with and without at least 1 hospitalist group. We compared continuous data with the Students t tests or Mann‐Whitney tests as appropriate and categorical data with chi‐square tests.

We then summarized the number of times a specific rationale was cited by hospital leaders for implementing a hospitalist group. Among hospitals that did not have a hospitalist system in place at the time of the survey, we asked if they were planning on starting one within the next 5 years. For these hospitals, we used content analysis to summarize open‐ended responses in order to understand factors that are currently influencing these hospital leaders to consider implementing a hospitalist group.

Next, we aimed to understand what clinical and nonclinical roles hospitalists were performing in hospitals with established hospitalist programs. Clinical activities were divided into general clinical areas, triage/emergency‐related, or administrative activities. First, we summarized the number and percent of programs performing each clinical and nonclinical activity. This was followed by logistic regression analyses to assess whether the time period that hospitalist groups began practicing or additional hospital characteristics predicted the performance of individual hospitalist activities. To guard against overfitting of models, analyses were limited to rationales that were cited a minimum of 50 times.24 Hospital factors were selected on the basis of face validity and advisory group input and included hospital bed size, ownership status (public vs. private), teaching status, and membership status in CHART. We divided the year of hospitalist program implementation into 3 time periods: (1) before 2002, (2) between 2002 and 2004, and (3) 2005 or later.

Finally, we described the percentage of hospitals that favored having their hospitalist group(s) perform each of the identified clinical or nonclinical activities, if they were not already performing them. We performed analyses with statistical software (Stata Version 9.2, College Station, TX).

Respondent Characteristics

We received 200 survey responses. Of those, we excluded 15 duplicates (eg, a survey from both the CEO and VPMA) and 6 responses identified as coming from hospitalists who did not have a leadership position in the hospital. Thus, the final hospital leader survey response rate was 54% (n = 179). Forty‐six percent of the final responses were from CEOs or COOs; 37% of responses were from CMOs, VPMAs, and medical directors; and the remaining 17% of responses were from other VPs or administrative directors.

Respondent and nonrespondent hospitals were statistically similar in terms of teaching status and participation in CHART. Hospital patient census, intensive care unit census, payer mix, and diagnosis‐related groupbased case‐mix revealed no statistically significant differences between groups (P > 0.05). Respondent hospitals tended to have fewer beds and were more often for‐profit compared to nonrespondents (P = 0.05 and P < 0.01, respectively).

Descriptive Characteristics of Hospitals with Hospitalists

Sixty‐four percent (n = 115) of hospital leaders stated that they utilized hospitalists for at least some patients. Hospitals with hospitalists were statistically more likely (P < 0.05) to be larger, a major teaching hospital, or a member of a voluntary quality reporting initiative (Table 1).

Among all hospitals with hospitalists, 39% estimated that hospitalists cared for at least one‐half of admitted medical patients, and 7% stated that hospitalists cared for all patients. Twenty‐four percent of respondents were unable to provide a quantitative estimate of the percent of patients cared for by hospitalists. When asked about expectations of growth in the coming year, 57% of respondents with hospitalists expected to see increases in the number of hospitalists at their hospital, and none expected a decrease. Among the 64 respondent hospitals that currently did not have a hospitalist program, 44% (n = 28) of the hospital leaders felt hospitalists would be managing patients in the future. Of those, 93% felt this would occur within the next 2 years.

Reasons for Implementing Hospitalists

Hospital leaders reported that the most important reasons for implementing a hospitalist model included caring for uncovered patients (68%), decreasing hospital costs and length of stay (63%), and improving throughput in the emergency room (62%). We provide additional reasons in Figure 1. In addition, leaders often identified multiple factors in the decision to utilize hospitalists, including demand from primary care doctors, patient satisfaction, and quality improvement. Among the 28 hospitals that currently did not have hospitalists but anticipated that they would soon (data not shown), the need to improve quality was the most commonly cited reason (54% of respondents) for expecting to start a program within 2 years, followed by demand from primary care doctors (46% of respondents).

Figure 1

Clinical Practice of Hospitalists and Expectations for Future Growth

Hospitalists perform a wide array of clinical and nonclinical duties (Figure 2). In addition to general medical care, the most common clinical activities of hospitalists included screening medical admissions from the emergency room for appropriateness of admission and triaging to appropriate level of care (67%), triaging patients transferred from an outside hospital (72%), and comanaging surgical patients (66%). The most common nonclinical activity was participation in quality improvement activities (72%). Multivariable analyses demonstrated that the performance of the most prevalent activities was not usually associated with the year of hospitalist implementation or hospital characteristics. An exception was that newly initiated programs had a statistically significant decreased odds of involvement in clinical guideline development (odds ratio [OR], 0.3; 95% confidence interval [CI], 0.1‐0.9) and a trend toward decreased leadership in quality improvement (OR, 0.3; 95% CI, 0.1‐1.1). Hospitalists at teaching hospitals had increased odds of managing patient transfers (OR, 4.7; 95% CI, 1.0‐21.2), whereas for‐profit hospitals had lower odds of screening patients in the emergency room (OR, 0.1; 95% CI, 0.0‐0.7).