Leora I. Horwitz, MD, MHS

Given the high cost of readmissions to the healthcare system, there has been a substantial push to reduce readmissions by policymakers.¹ Among these is the Hospital Readmissions Reduction Program (HRRP), in which hospitals with higher than expected readmission rates receive reduced payments from Medicare.² Recent evidence has suggested the success of such policy changes, with multiple reports demonstrating a decrease in 30-day readmission rates in the Medicare population starting in 2010.^3-8

Initiatives to reduce readmissions can also have an effect on total number of admissions.^9,10 Indeed, along with the recent reduction in readmission, there has been a reduction in all admissions among Medicare beneficiaries.^11,12 Some studies have found that as admissions have decreased, the burden of comorbidity has increased among hospitalized patients,^3,11 suggesting that hospitals may be increasingly filled with patients at high risk of readmission. However, whether readmission risk among hospitalized patients has changed remains unknown, and understanding changes in risk profile could help inform which patients to target with future interventions to reduce readmissions.

Hospital efforts to reduce readmissions may have differential effects on types of patients by risk. For instance, low-intensity, system-wide interventions such as standardized discharge instructions or medicine reconciliation may have a stronger effect on patients at relatively low risk of readmission who may have a few important drivers of readmission that are easily overcome. Alternatively, the impact of intensive care transitions management might be greatest for high-risk patients, who have the most need for postdischarge medications, follow-up, and self-care.

The purpose of this study was therefore twofold: (1) to observe changes in average monthly risk of readmission among hospitalized Medicare patients and (2) to examine changes in readmission rates for Medicare patients at various risk of readmission. We hypothesized that readmission risk in the Medicare population would increase in recent years, as overall number of admissions and readmissions have fallen.^7,11 Additionally, we hypothesized that standardized readmission rates would decline less in highest risk patients as compared with the lowest risk patients because transitional care interventions may not be able to mitigate the large burden of comorbidity and social issues present in many high-risk patients.^13,14

We performed a retrospective cohort study of hospitalizations to US nonfederal short-term acute care facilities by Medicare beneficiaries between January 2009 and June 2015. The design involved 4 steps. First, we estimated a predictive model for unplanned readmissions within 30 days of discharge. Second, we assigned each hospitalization a predicted risk of readmission based on the model. Third, we studied trends in mean predicted risk of readmission during the study period. Fourth, we examined trends in observed to expected (O/E) readmission for hospitalizations in the lowest, middle, and highest categories of predicted risk of readmission to determine whether reductions in readmissions were more substantial in certain risk groups than in others.

Data were obtained from the Centers for Medicare and Medicaid Services (CMS) Inpatient Standard Analytic File and the Medicare Enrollment Data Base. We included hospitalizations of fee-for-service Medicare beneficiaries age ≥65 with continuous enrollment in Part A Medicare fee-for-service for at least 1 year prior and 30 days after the hospitalization.¹⁵ Hospitalizations with a discharge disposition of death, transfer to another acute hospital, and left against medical advice (AMA) were excluded. We also excluded patients with enrollment in hospice care prior to hospitalization. We excluded hospitalizations in June 2012 because of an irregularity in data availability for that month.

Hospitalizations were categorized into 5 specialty cohorts according to service line. The 5 cohorts were those used for the CMS hospital-wide readmission measure and included surgery/gynecology, medicine, cardiovascular, cardiorespiratory, and neurology.¹⁵ Among the 3 clinical conditions tracked as part of HRRP, heart failure and pneumonia were a subset of the cardiorespiratory cohort, while acute myocardial infarction was a subset of the cardiovascular cohort. Our use of cohorts was threefold: first, the average risk of readmission differs substantially across these cohorts, so pooling them produces heterogeneous risk strata; second, risk variables perform differently in different cohorts, so one single model may not be as accurate for calculating risk; and, third, the use of disease cohorts makes our results comparable to the CMS model and similar to other readmission studies in Medicare.^7,8,15

For development of the risk model, the outcome was 30-day unplanned hospital readmission. Planned readmissions were excluded; these were defined by the CMS algorithm as readmissions in which a typically planned procedure occurred in a hospitalization with a nonacute principal diagnosis.¹⁶ Independent variables included age and comorbidities in the final hospital-wide readmission models for each of the 5 specialty cohorts.¹⁵ In order to produce the best possible individual risk prediction for each patient, we added additional independent variables that CMS avoids for hospital quality measurement purposes but that contribute to risk of readmission: sex, race, dual eligibility status, number of prior AMA discharges, intensive care unit stay during current hospitalization, coronary care unit stay during current hospitalization, and hospitalization in the prior 30, 90, and 180 days. We also included an indicator variable for hospitalizations with more than 9 discharge diagnosis codes on or after January 2011, the time at which Medicare allowed an increase of the number of International Classification of Diseases, 9th Revision-Clinical Modification diagnosis billing codes from 9 to 25.¹⁷ This indicator adjusts for the increased availability of comorbidity codes, which might otherwise inflate the predicted risk relative to hospitalizations prior to that date.

Based on the risk models, each hospitalization was assigned a predicted risk of readmission. For each specialty cohort, we pooled all hospitalizations across all study years and divided them into risk quintiles. We categorized hospitalizations as high risk if in the highest quintile, medium risk if in the middle 3 quintiles, and low risk if in the lowest quintile of predicted risk for all study hospitalizations in a given specialty cohort.

For our time trend analyses, we studied 2 outcomes: monthly mean predicted risk and monthly ratio of observed readmissions to expected readmissions for patients in the lowest, middle, and highest categories of predicted risk of readmission. We studied monthly predicted risk to determine whether the average readmission risk of patients was changing over time as admission and readmission rates were declining. We studied the ratio of O/E readmissions to determine whether the decline in overall readmissions was more substantial in particular risk strata; we used the ratio of O/E readmissions, which measures number of readmissions divided by number of readmissions predicted by the model, rather than crude observed readmissions, as O/E readmissions account for any changes in risk profiles over time within each risk stratum. Independent variables in our trend analyses were year—entered as a continuous variable—and indicators for postintroduction of the Affordable Care Act (ACA, March 2010) and for postintroduction of HRRP (October 2012); these time indicators were included because of prior studies demonstrating that the introduction of ACA was associated with a decrease from baseline in readmission rates, which leveled off after introduction of HRRP.⁷ We also included an indicator for calendar quarter to account for seasonal effects.

Statistical Analysis

We developed generalized estimating equation models to predict 30-day unplanned readmission for each of the 5 specialty cohorts. The 5 models were fit using all patients in each cohort for the included time period and were adjusted for clustering by hospital. We assessed discrimination by calculating area under the receiver operating characteristic curve (AUC) for the 5 models; the AUCs measured the models’ ability to distinguish patients who were readmitted versus those who were not.¹⁸ We also calculated AUCs for each year to examine model performance over time.

Using these models, we calculated predicted risk for each hospitalization and averaged these to obtain mean predicted risk for each specialty cohort for each month. To test for trends in mean risk, we estimated 5 time series models, one for each cohort, with the dependent variable of monthly mean predicted risk. For each cohort, we first estimated a series of 12 empty autoregressive models, each with a different autoregressive term (1, 2...12). For each model, we calculated χ² for the test that the autocorrelation was 0; based on a comparison of chi-squared values, we specified an autocorrelation of 1 month for all models. Accordingly, a 1-month lag was used to estimate one final model for each cohort. Independent variables included year and indicators for post-ACA and post-HRRP; these variables captured the effect of trends over time and the introduction of these policy changes, respectively.¹⁹

To determine whether changes in risk over time were a result of changes in particular risk groups, we categorized hospitalizations into risk strata based on quintiles of predicted risk for each specialty cohort for the entire study period. For each individual year, we calculated the proportion of hospitalizations in the highest, middle, and lowest readmission risk strata for each cohort.

We calculated the monthly ratio of O/E readmission for hospitalizations in the lowest 20%, middle 60%, and highest 20% of readmission risk by month; O/E reflects the excess or deficit observed events relative to the number predicted by the model. Using this monthly O/E as the dependent variable, we developed autoregressive time series models as above, again with a 1-month lag, for each of these 3 risk strata in each cohort. As before, independent variables were year as a continuous variable, indicator variables for post-ACA and post-HRRP, and a categorical variable for calendar quarter.

All analyses were done in SAS version 9.3 (SAS Institute Inc., Cary, NC) and Stata version 14.2 (StataCorp LLC, College Station, TX).

We included 47,288,961 hospitalizations in the study, of which 11,231,242 (23.8%) were in the surgery/gynecology cohort, 19,548,711 (41.3%) were in the medicine cohort, 5,433,125 (11.5%) were in the cardiovascular cohort, 8,179,691 (17.3%) were in the cardiorespiratory cohort, and 2,896,192 (6.1%) were in the neurology cohort. The readmission rate was 16.2% (n = 7,642,161) overall, with the highest rates observed in the cardiorespiratory (20.5%) and medicine (17.6%) cohorts and the lowest rates observed in the surgery/gynecology (11.8%) and neurology (13.8%) cohorts.

The final predictive models for each cohort ranged in number of parameters from 56 for the cardiorespiratory cohort to 264 for the surgery/gynecology cohort. The models had AUCs of 0.70, 0.65, 0.67, 0.65, and 0.63 for the surgery/gynecology, medicine, cardiovascular, cardiorespiratory, and neurology cohorts, respectively; AUCs remained fairly stable over time for all disease cohorts (Appendix Table 1).

A number of mechanisms may account for the across-the-board improvements in readmission reduction. Many hospitals have instituted system-wide interventions, including patient education, medicine reconciliation, and early postdischarge follow-up,²⁰ which may have reduced readmissions across all patient risk strata. Alternatively, hospitals may have implemented interventions that disproportionally benefited low-risk patients while simultaneously utilizing interventions that only benefited high-risk patients. For instance, increasing threshold for admission⁷ may have the greatest effect on low-risk patients who could be most easily managed at home, while many intensive transitional care interventions have been developed to target only high-risk patients.^21,22

With the introduction of HRRP, there have been a number of concerns about the readmission measure used to penalize hospitals for high readmission rates. One major concern has been that the readmission metric may be flawed in its ability to capture continued improvement related to readmission.²³ Some have suggested that with better population health management, admissions will decrease, patient risk of the remaining patients will increase, and hospitals will be increasingly filled with patients who have high likelihood of readmission. This potential for increased risk with HRRP was suggested by a recent study that found that comorbidities increased in hospitalized Medicare beneficiaries between 2010 and 2013.¹¹ Our results were mixed in supporting this potential phenomenon because we examined global risk of readmission and found that some of the cohorts had increased risk over time while others did not. Others have expressed concern that readmission measure does not account for socioeconomic status, which has been associated with readmission rates.^24-27 Although we did not directly examine socioeconomic status in our study, we found that hospitals have been able to reduce readmission across all levels of risk, which includes markers of socioeconomic status, including race and Medicaid eligibility status.

Although we hypothesized that readmission risk would increase as number of hospitalizations decreased over time, we found no increase in readmission risk among the cohorts with HRRP diagnoses that had the largest decrease in readmission rates.^7,8 Conversely, readmission risk did increase—with a concurrent increase in the proportion of high-risk hospitalizations—in the surgery/gynecology and neurology cohorts that were not subject to HRRP penalties. Nonetheless, rehospitalizations were reduced for all risk categories in these 2 cohorts. Notably, surgery/gynecology and neurology had the lowest readmission rates overall. These findings suggest that initiatives to prevent initial hospitalizations, such as increasing the threshold for postoperative admission, may have had a greater effect on low- versus high-risk patients in low-risk hospitalizations. However, once a patient is hospitalized, multidisciplinary strategies appear to be effective at reducing readmissions for all risk classes in these cohorts.

For the 3 cohorts in which we observed an increase in readmission risk among hospitalized patients, the risk appeared to increase in early 2011. This time was about 10 months after passage of ACA, the timing of which was previously associated with a drop in readmission rates,^7,8 but well before HRRP went into effect in October 2012. The increase in readmission risk coincided with an increase in the number of diagnostic codes that could be included on a hospital claim to Medicare.¹⁷ This increase in allowable codes allowed us to capture more diagnoses for some patients, potentially resulting in an increase in apparent predicted risk of readmissions. While we adjusted for this in our predictive models, we may not have fully accounted for differences in risk related to coding change. As a result, some of the observed differences in risk in our study may be attributable to coding differences. More broadly, studies demonstrating the success of HRRP have typically examined risk-adjusted rates of readmission.^3,7 It is possible that a small portion of the observed reduction in risk-adjusted readmission rates may be related to the increase in predicted risk of readmission observed in our study. Future assessment of trends in readmission during this period should consider accounting for change in the number of allowed billing codes.

Other limitations should be considered in the interpretation of this study. First, like many predictive models for readmission,¹⁴ ours had imperfect discrimination, which could affect our results. Second, our study was based on older Medicare patients, so findings may not be applicable to younger patients. Third, while we accounted for surrogates for socioeconomic status, including dual eligibility and race, our models lacked other socioeconomic and community factors that can influence readmission.^24-26 Nonetheless, 1 study suggested that easily measured socioeconomic factors may not have a strong influence on the readmission metric used by Medicare.²⁸ Fourth, while our study included over 47 million hospitalizations, our time trend analyses used calendar month as the primary independent variable. As our study included 77 months, we may not have had sufficient power to detect small changes in risk over time.

Medicare readmissions have declined steadily in recent years, presumably at least partly in response to policy changes including HRRP. We found that hospitals have been effective at reducing readmissions across a range of patient risk strata and clinical conditions. As a result, the overall risk of readmission for hospitalized patients has remained constant for some but not all conditions. Whether institutions can continue to reduce readmission rates for most types of patients remains to be seen.

Acknowledgments

This study was supported by the Agency for Healthcare Research and Quality (AHRQ) grant R01HS022882. Dr. Blecker was supported by the AHRQ grant K08HS23683. The authors would like to thank Shawn Hoke and Jane Padikkala for administrative support.

Disclosure

This study was supported by the Agency for Healthcare Research and Quality (AHRQ) grants R01HS022882 and K08HS23683. The authors have no conflicts to report.

Given the high cost of readmissions to the healthcare system, there has been a substantial push to reduce readmissions by policymakers.¹ Among these is the Hospital Readmissions Reduction Program (HRRP), in which hospitals with higher than expected readmission rates receive reduced payments from Medicare.² Recent evidence has suggested the success of such policy changes, with multiple reports demonstrating a decrease in 30-day readmission rates in the Medicare population starting in 2010.^3-8

Initiatives to reduce readmissions can also have an effect on total number of admissions.^9,10 Indeed, along with the recent reduction in readmission, there has been a reduction in all admissions among Medicare beneficiaries.^11,12 Some studies have found that as admissions have decreased, the burden of comorbidity has increased among hospitalized patients,^3,11 suggesting that hospitals may be increasingly filled with patients at high risk of readmission. However, whether readmission risk among hospitalized patients has changed remains unknown, and understanding changes in risk profile could help inform which patients to target with future interventions to reduce readmissions.

Hospital efforts to reduce readmissions may have differential effects on types of patients by risk. For instance, low-intensity, system-wide interventions such as standardized discharge instructions or medicine reconciliation may have a stronger effect on patients at relatively low risk of readmission who may have a few important drivers of readmission that are easily overcome. Alternatively, the impact of intensive care transitions management might be greatest for high-risk patients, who have the most need for postdischarge medications, follow-up, and self-care.

The purpose of this study was therefore twofold: (1) to observe changes in average monthly risk of readmission among hospitalized Medicare patients and (2) to examine changes in readmission rates for Medicare patients at various risk of readmission. We hypothesized that readmission risk in the Medicare population would increase in recent years, as overall number of admissions and readmissions have fallen.^7,11 Additionally, we hypothesized that standardized readmission rates would decline less in highest risk patients as compared with the lowest risk patients because transitional care interventions may not be able to mitigate the large burden of comorbidity and social issues present in many high-risk patients.^13,14

We performed a retrospective cohort study of hospitalizations to US nonfederal short-term acute care facilities by Medicare beneficiaries between January 2009 and June 2015. The design involved 4 steps. First, we estimated a predictive model for unplanned readmissions within 30 days of discharge. Second, we assigned each hospitalization a predicted risk of readmission based on the model. Third, we studied trends in mean predicted risk of readmission during the study period. Fourth, we examined trends in observed to expected (O/E) readmission for hospitalizations in the lowest, middle, and highest categories of predicted risk of readmission to determine whether reductions in readmissions were more substantial in certain risk groups than in others.

Data were obtained from the Centers for Medicare and Medicaid Services (CMS) Inpatient Standard Analytic File and the Medicare Enrollment Data Base. We included hospitalizations of fee-for-service Medicare beneficiaries age ≥65 with continuous enrollment in Part A Medicare fee-for-service for at least 1 year prior and 30 days after the hospitalization.¹⁵ Hospitalizations with a discharge disposition of death, transfer to another acute hospital, and left against medical advice (AMA) were excluded. We also excluded patients with enrollment in hospice care prior to hospitalization. We excluded hospitalizations in June 2012 because of an irregularity in data availability for that month.

Hospitalizations were categorized into 5 specialty cohorts according to service line. The 5 cohorts were those used for the CMS hospital-wide readmission measure and included surgery/gynecology, medicine, cardiovascular, cardiorespiratory, and neurology.¹⁵ Among the 3 clinical conditions tracked as part of HRRP, heart failure and pneumonia were a subset of the cardiorespiratory cohort, while acute myocardial infarction was a subset of the cardiovascular cohort. Our use of cohorts was threefold: first, the average risk of readmission differs substantially across these cohorts, so pooling them produces heterogeneous risk strata; second, risk variables perform differently in different cohorts, so one single model may not be as accurate for calculating risk; and, third, the use of disease cohorts makes our results comparable to the CMS model and similar to other readmission studies in Medicare.^7,8,15

For development of the risk model, the outcome was 30-day unplanned hospital readmission. Planned readmissions were excluded; these were defined by the CMS algorithm as readmissions in which a typically planned procedure occurred in a hospitalization with a nonacute principal diagnosis.¹⁶ Independent variables included age and comorbidities in the final hospital-wide readmission models for each of the 5 specialty cohorts.¹⁵ In order to produce the best possible individual risk prediction for each patient, we added additional independent variables that CMS avoids for hospital quality measurement purposes but that contribute to risk of readmission: sex, race, dual eligibility status, number of prior AMA discharges, intensive care unit stay during current hospitalization, coronary care unit stay during current hospitalization, and hospitalization in the prior 30, 90, and 180 days. We also included an indicator variable for hospitalizations with more than 9 discharge diagnosis codes on or after January 2011, the time at which Medicare allowed an increase of the number of International Classification of Diseases, 9th Revision-Clinical Modification diagnosis billing codes from 9 to 25.¹⁷ This indicator adjusts for the increased availability of comorbidity codes, which might otherwise inflate the predicted risk relative to hospitalizations prior to that date.

Based on the risk models, each hospitalization was assigned a predicted risk of readmission. For each specialty cohort, we pooled all hospitalizations across all study years and divided them into risk quintiles. We categorized hospitalizations as high risk if in the highest quintile, medium risk if in the middle 3 quintiles, and low risk if in the lowest quintile of predicted risk for all study hospitalizations in a given specialty cohort.

For our time trend analyses, we studied 2 outcomes: monthly mean predicted risk and monthly ratio of observed readmissions to expected readmissions for patients in the lowest, middle, and highest categories of predicted risk of readmission. We studied monthly predicted risk to determine whether the average readmission risk of patients was changing over time as admission and readmission rates were declining. We studied the ratio of O/E readmissions to determine whether the decline in overall readmissions was more substantial in particular risk strata; we used the ratio of O/E readmissions, which measures number of readmissions divided by number of readmissions predicted by the model, rather than crude observed readmissions, as O/E readmissions account for any changes in risk profiles over time within each risk stratum. Independent variables in our trend analyses were year—entered as a continuous variable—and indicators for postintroduction of the Affordable Care Act (ACA, March 2010) and for postintroduction of HRRP (October 2012); these time indicators were included because of prior studies demonstrating that the introduction of ACA was associated with a decrease from baseline in readmission rates, which leveled off after introduction of HRRP.⁷ We also included an indicator for calendar quarter to account for seasonal effects.

Statistical Analysis

We developed generalized estimating equation models to predict 30-day unplanned readmission for each of the 5 specialty cohorts. The 5 models were fit using all patients in each cohort for the included time period and were adjusted for clustering by hospital. We assessed discrimination by calculating area under the receiver operating characteristic curve (AUC) for the 5 models; the AUCs measured the models’ ability to distinguish patients who were readmitted versus those who were not.¹⁸ We also calculated AUCs for each year to examine model performance over time.

Using these models, we calculated predicted risk for each hospitalization and averaged these to obtain mean predicted risk for each specialty cohort for each month. To test for trends in mean risk, we estimated 5 time series models, one for each cohort, with the dependent variable of monthly mean predicted risk. For each cohort, we first estimated a series of 12 empty autoregressive models, each with a different autoregressive term (1, 2...12). For each model, we calculated χ² for the test that the autocorrelation was 0; based on a comparison of chi-squared values, we specified an autocorrelation of 1 month for all models. Accordingly, a 1-month lag was used to estimate one final model for each cohort. Independent variables included year and indicators for post-ACA and post-HRRP; these variables captured the effect of trends over time and the introduction of these policy changes, respectively.¹⁹

To determine whether changes in risk over time were a result of changes in particular risk groups, we categorized hospitalizations into risk strata based on quintiles of predicted risk for each specialty cohort for the entire study period. For each individual year, we calculated the proportion of hospitalizations in the highest, middle, and lowest readmission risk strata for each cohort.

We calculated the monthly ratio of O/E readmission for hospitalizations in the lowest 20%, middle 60%, and highest 20% of readmission risk by month; O/E reflects the excess or deficit observed events relative to the number predicted by the model. Using this monthly O/E as the dependent variable, we developed autoregressive time series models as above, again with a 1-month lag, for each of these 3 risk strata in each cohort. As before, independent variables were year as a continuous variable, indicator variables for post-ACA and post-HRRP, and a categorical variable for calendar quarter.

All analyses were done in SAS version 9.3 (SAS Institute Inc., Cary, NC) and Stata version 14.2 (StataCorp LLC, College Station, TX).

We included 47,288,961 hospitalizations in the study, of which 11,231,242 (23.8%) were in the surgery/gynecology cohort, 19,548,711 (41.3%) were in the medicine cohort, 5,433,125 (11.5%) were in the cardiovascular cohort, 8,179,691 (17.3%) were in the cardiorespiratory cohort, and 2,896,192 (6.1%) were in the neurology cohort. The readmission rate was 16.2% (n = 7,642,161) overall, with the highest rates observed in the cardiorespiratory (20.5%) and medicine (17.6%) cohorts and the lowest rates observed in the surgery/gynecology (11.8%) and neurology (13.8%) cohorts.

The final predictive models for each cohort ranged in number of parameters from 56 for the cardiorespiratory cohort to 264 for the surgery/gynecology cohort. The models had AUCs of 0.70, 0.65, 0.67, 0.65, and 0.63 for the surgery/gynecology, medicine, cardiovascular, cardiorespiratory, and neurology cohorts, respectively; AUCs remained fairly stable over time for all disease cohorts (Appendix Table 1).

A number of mechanisms may account for the across-the-board improvements in readmission reduction. Many hospitals have instituted system-wide interventions, including patient education, medicine reconciliation, and early postdischarge follow-up,²⁰ which may have reduced readmissions across all patient risk strata. Alternatively, hospitals may have implemented interventions that disproportionally benefited low-risk patients while simultaneously utilizing interventions that only benefited high-risk patients. For instance, increasing threshold for admission⁷ may have the greatest effect on low-risk patients who could be most easily managed at home, while many intensive transitional care interventions have been developed to target only high-risk patients.^21,22

With the introduction of HRRP, there have been a number of concerns about the readmission measure used to penalize hospitals for high readmission rates. One major concern has been that the readmission metric may be flawed in its ability to capture continued improvement related to readmission.²³ Some have suggested that with better population health management, admissions will decrease, patient risk of the remaining patients will increase, and hospitals will be increasingly filled with patients who have high likelihood of readmission. This potential for increased risk with HRRP was suggested by a recent study that found that comorbidities increased in hospitalized Medicare beneficiaries between 2010 and 2013.¹¹ Our results were mixed in supporting this potential phenomenon because we examined global risk of readmission and found that some of the cohorts had increased risk over time while others did not. Others have expressed concern that readmission measure does not account for socioeconomic status, which has been associated with readmission rates.^24-27 Although we did not directly examine socioeconomic status in our study, we found that hospitals have been able to reduce readmission across all levels of risk, which includes markers of socioeconomic status, including race and Medicaid eligibility status.

Although we hypothesized that readmission risk would increase as number of hospitalizations decreased over time, we found no increase in readmission risk among the cohorts with HRRP diagnoses that had the largest decrease in readmission rates.^7,8 Conversely, readmission risk did increase—with a concurrent increase in the proportion of high-risk hospitalizations—in the surgery/gynecology and neurology cohorts that were not subject to HRRP penalties. Nonetheless, rehospitalizations were reduced for all risk categories in these 2 cohorts. Notably, surgery/gynecology and neurology had the lowest readmission rates overall. These findings suggest that initiatives to prevent initial hospitalizations, such as increasing the threshold for postoperative admission, may have had a greater effect on low- versus high-risk patients in low-risk hospitalizations. However, once a patient is hospitalized, multidisciplinary strategies appear to be effective at reducing readmissions for all risk classes in these cohorts.

For the 3 cohorts in which we observed an increase in readmission risk among hospitalized patients, the risk appeared to increase in early 2011. This time was about 10 months after passage of ACA, the timing of which was previously associated with a drop in readmission rates,^7,8 but well before HRRP went into effect in October 2012. The increase in readmission risk coincided with an increase in the number of diagnostic codes that could be included on a hospital claim to Medicare.¹⁷ This increase in allowable codes allowed us to capture more diagnoses for some patients, potentially resulting in an increase in apparent predicted risk of readmissions. While we adjusted for this in our predictive models, we may not have fully accounted for differences in risk related to coding change. As a result, some of the observed differences in risk in our study may be attributable to coding differences. More broadly, studies demonstrating the success of HRRP have typically examined risk-adjusted rates of readmission.^3,7 It is possible that a small portion of the observed reduction in risk-adjusted readmission rates may be related to the increase in predicted risk of readmission observed in our study. Future assessment of trends in readmission during this period should consider accounting for change in the number of allowed billing codes.

Other limitations should be considered in the interpretation of this study. First, like many predictive models for readmission,¹⁴ ours had imperfect discrimination, which could affect our results. Second, our study was based on older Medicare patients, so findings may not be applicable to younger patients. Third, while we accounted for surrogates for socioeconomic status, including dual eligibility and race, our models lacked other socioeconomic and community factors that can influence readmission.^24-26 Nonetheless, 1 study suggested that easily measured socioeconomic factors may not have a strong influence on the readmission metric used by Medicare.²⁸ Fourth, while our study included over 47 million hospitalizations, our time trend analyses used calendar month as the primary independent variable. As our study included 77 months, we may not have had sufficient power to detect small changes in risk over time.

Medicare readmissions have declined steadily in recent years, presumably at least partly in response to policy changes including HRRP. We found that hospitals have been effective at reducing readmissions across a range of patient risk strata and clinical conditions. As a result, the overall risk of readmission for hospitalized patients has remained constant for some but not all conditions. Whether institutions can continue to reduce readmission rates for most types of patients remains to be seen.

Acknowledgments

This study was supported by the Agency for Healthcare Research and Quality (AHRQ) grant R01HS022882. Dr. Blecker was supported by the AHRQ grant K08HS23683. The authors would like to thank Shawn Hoke and Jane Padikkala for administrative support.

Disclosure

This study was supported by the Agency for Healthcare Research and Quality (AHRQ) grants R01HS022882 and K08HS23683. The authors have no conflicts to report.

Given the high cost of readmissions to the healthcare system, there has been a substantial push to reduce readmissions by policymakers.¹ Among these is the Hospital Readmissions Reduction Program (HRRP), in which hospitals with higher than expected readmission rates receive reduced payments from Medicare.² Recent evidence has suggested the success of such policy changes, with multiple reports demonstrating a decrease in 30-day readmission rates in the Medicare population starting in 2010.^3-8

Initiatives to reduce readmissions can also have an effect on total number of admissions.^9,10 Indeed, along with the recent reduction in readmission, there has been a reduction in all admissions among Medicare beneficiaries.^11,12 Some studies have found that as admissions have decreased, the burden of comorbidity has increased among hospitalized patients,^3,11 suggesting that hospitals may be increasingly filled with patients at high risk of readmission. However, whether readmission risk among hospitalized patients has changed remains unknown, and understanding changes in risk profile could help inform which patients to target with future interventions to reduce readmissions.

Hospital efforts to reduce readmissions may have differential effects on types of patients by risk. For instance, low-intensity, system-wide interventions such as standardized discharge instructions or medicine reconciliation may have a stronger effect on patients at relatively low risk of readmission who may have a few important drivers of readmission that are easily overcome. Alternatively, the impact of intensive care transitions management might be greatest for high-risk patients, who have the most need for postdischarge medications, follow-up, and self-care.

The purpose of this study was therefore twofold: (1) to observe changes in average monthly risk of readmission among hospitalized Medicare patients and (2) to examine changes in readmission rates for Medicare patients at various risk of readmission. We hypothesized that readmission risk in the Medicare population would increase in recent years, as overall number of admissions and readmissions have fallen.^7,11 Additionally, we hypothesized that standardized readmission rates would decline less in highest risk patients as compared with the lowest risk patients because transitional care interventions may not be able to mitigate the large burden of comorbidity and social issues present in many high-risk patients.^13,14

We performed a retrospective cohort study of hospitalizations to US nonfederal short-term acute care facilities by Medicare beneficiaries between January 2009 and June 2015. The design involved 4 steps. First, we estimated a predictive model for unplanned readmissions within 30 days of discharge. Second, we assigned each hospitalization a predicted risk of readmission based on the model. Third, we studied trends in mean predicted risk of readmission during the study period. Fourth, we examined trends in observed to expected (O/E) readmission for hospitalizations in the lowest, middle, and highest categories of predicted risk of readmission to determine whether reductions in readmissions were more substantial in certain risk groups than in others.

Data were obtained from the Centers for Medicare and Medicaid Services (CMS) Inpatient Standard Analytic File and the Medicare Enrollment Data Base. We included hospitalizations of fee-for-service Medicare beneficiaries age ≥65 with continuous enrollment in Part A Medicare fee-for-service for at least 1 year prior and 30 days after the hospitalization.¹⁵ Hospitalizations with a discharge disposition of death, transfer to another acute hospital, and left against medical advice (AMA) were excluded. We also excluded patients with enrollment in hospice care prior to hospitalization. We excluded hospitalizations in June 2012 because of an irregularity in data availability for that month.

Hospitalizations were categorized into 5 specialty cohorts according to service line. The 5 cohorts were those used for the CMS hospital-wide readmission measure and included surgery/gynecology, medicine, cardiovascular, cardiorespiratory, and neurology.¹⁵ Among the 3 clinical conditions tracked as part of HRRP, heart failure and pneumonia were a subset of the cardiorespiratory cohort, while acute myocardial infarction was a subset of the cardiovascular cohort. Our use of cohorts was threefold: first, the average risk of readmission differs substantially across these cohorts, so pooling them produces heterogeneous risk strata; second, risk variables perform differently in different cohorts, so one single model may not be as accurate for calculating risk; and, third, the use of disease cohorts makes our results comparable to the CMS model and similar to other readmission studies in Medicare.^7,8,15

For development of the risk model, the outcome was 30-day unplanned hospital readmission. Planned readmissions were excluded; these were defined by the CMS algorithm as readmissions in which a typically planned procedure occurred in a hospitalization with a nonacute principal diagnosis.¹⁶ Independent variables included age and comorbidities in the final hospital-wide readmission models for each of the 5 specialty cohorts.¹⁵ In order to produce the best possible individual risk prediction for each patient, we added additional independent variables that CMS avoids for hospital quality measurement purposes but that contribute to risk of readmission: sex, race, dual eligibility status, number of prior AMA discharges, intensive care unit stay during current hospitalization, coronary care unit stay during current hospitalization, and hospitalization in the prior 30, 90, and 180 days. We also included an indicator variable for hospitalizations with more than 9 discharge diagnosis codes on or after January 2011, the time at which Medicare allowed an increase of the number of International Classification of Diseases, 9th Revision-Clinical Modification diagnosis billing codes from 9 to 25.¹⁷ This indicator adjusts for the increased availability of comorbidity codes, which might otherwise inflate the predicted risk relative to hospitalizations prior to that date.

Based on the risk models, each hospitalization was assigned a predicted risk of readmission. For each specialty cohort, we pooled all hospitalizations across all study years and divided them into risk quintiles. We categorized hospitalizations as high risk if in the highest quintile, medium risk if in the middle 3 quintiles, and low risk if in the lowest quintile of predicted risk for all study hospitalizations in a given specialty cohort.

For our time trend analyses, we studied 2 outcomes: monthly mean predicted risk and monthly ratio of observed readmissions to expected readmissions for patients in the lowest, middle, and highest categories of predicted risk of readmission. We studied monthly predicted risk to determine whether the average readmission risk of patients was changing over time as admission and readmission rates were declining. We studied the ratio of O/E readmissions to determine whether the decline in overall readmissions was more substantial in particular risk strata; we used the ratio of O/E readmissions, which measures number of readmissions divided by number of readmissions predicted by the model, rather than crude observed readmissions, as O/E readmissions account for any changes in risk profiles over time within each risk stratum. Independent variables in our trend analyses were year—entered as a continuous variable—and indicators for postintroduction of the Affordable Care Act (ACA, March 2010) and for postintroduction of HRRP (October 2012); these time indicators were included because of prior studies demonstrating that the introduction of ACA was associated with a decrease from baseline in readmission rates, which leveled off after introduction of HRRP.⁷ We also included an indicator for calendar quarter to account for seasonal effects.

Statistical Analysis

We developed generalized estimating equation models to predict 30-day unplanned readmission for each of the 5 specialty cohorts. The 5 models were fit using all patients in each cohort for the included time period and were adjusted for clustering by hospital. We assessed discrimination by calculating area under the receiver operating characteristic curve (AUC) for the 5 models; the AUCs measured the models’ ability to distinguish patients who were readmitted versus those who were not.¹⁸ We also calculated AUCs for each year to examine model performance over time.

Using these models, we calculated predicted risk for each hospitalization and averaged these to obtain mean predicted risk for each specialty cohort for each month. To test for trends in mean risk, we estimated 5 time series models, one for each cohort, with the dependent variable of monthly mean predicted risk. For each cohort, we first estimated a series of 12 empty autoregressive models, each with a different autoregressive term (1, 2...12). For each model, we calculated χ² for the test that the autocorrelation was 0; based on a comparison of chi-squared values, we specified an autocorrelation of 1 month for all models. Accordingly, a 1-month lag was used to estimate one final model for each cohort. Independent variables included year and indicators for post-ACA and post-HRRP; these variables captured the effect of trends over time and the introduction of these policy changes, respectively.¹⁹

To determine whether changes in risk over time were a result of changes in particular risk groups, we categorized hospitalizations into risk strata based on quintiles of predicted risk for each specialty cohort for the entire study period. For each individual year, we calculated the proportion of hospitalizations in the highest, middle, and lowest readmission risk strata for each cohort.

We calculated the monthly ratio of O/E readmission for hospitalizations in the lowest 20%, middle 60%, and highest 20% of readmission risk by month; O/E reflects the excess or deficit observed events relative to the number predicted by the model. Using this monthly O/E as the dependent variable, we developed autoregressive time series models as above, again with a 1-month lag, for each of these 3 risk strata in each cohort. As before, independent variables were year as a continuous variable, indicator variables for post-ACA and post-HRRP, and a categorical variable for calendar quarter.

All analyses were done in SAS version 9.3 (SAS Institute Inc., Cary, NC) and Stata version 14.2 (StataCorp LLC, College Station, TX).

We included 47,288,961 hospitalizations in the study, of which 11,231,242 (23.8%) were in the surgery/gynecology cohort, 19,548,711 (41.3%) were in the medicine cohort, 5,433,125 (11.5%) were in the cardiovascular cohort, 8,179,691 (17.3%) were in the cardiorespiratory cohort, and 2,896,192 (6.1%) were in the neurology cohort. The readmission rate was 16.2% (n = 7,642,161) overall, with the highest rates observed in the cardiorespiratory (20.5%) and medicine (17.6%) cohorts and the lowest rates observed in the surgery/gynecology (11.8%) and neurology (13.8%) cohorts.

The final predictive models for each cohort ranged in number of parameters from 56 for the cardiorespiratory cohort to 264 for the surgery/gynecology cohort. The models had AUCs of 0.70, 0.65, 0.67, 0.65, and 0.63 for the surgery/gynecology, medicine, cardiovascular, cardiorespiratory, and neurology cohorts, respectively; AUCs remained fairly stable over time for all disease cohorts (Appendix Table 1).

A number of mechanisms may account for the across-the-board improvements in readmission reduction. Many hospitals have instituted system-wide interventions, including patient education, medicine reconciliation, and early postdischarge follow-up,²⁰ which may have reduced readmissions across all patient risk strata. Alternatively, hospitals may have implemented interventions that disproportionally benefited low-risk patients while simultaneously utilizing interventions that only benefited high-risk patients. For instance, increasing threshold for admission⁷ may have the greatest effect on low-risk patients who could be most easily managed at home, while many intensive transitional care interventions have been developed to target only high-risk patients.^21,22

With the introduction of HRRP, there have been a number of concerns about the readmission measure used to penalize hospitals for high readmission rates. One major concern has been that the readmission metric may be flawed in its ability to capture continued improvement related to readmission.²³ Some have suggested that with better population health management, admissions will decrease, patient risk of the remaining patients will increase, and hospitals will be increasingly filled with patients who have high likelihood of readmission. This potential for increased risk with HRRP was suggested by a recent study that found that comorbidities increased in hospitalized Medicare beneficiaries between 2010 and 2013.¹¹ Our results were mixed in supporting this potential phenomenon because we examined global risk of readmission and found that some of the cohorts had increased risk over time while others did not. Others have expressed concern that readmission measure does not account for socioeconomic status, which has been associated with readmission rates.^24-27 Although we did not directly examine socioeconomic status in our study, we found that hospitals have been able to reduce readmission across all levels of risk, which includes markers of socioeconomic status, including race and Medicaid eligibility status.

Although we hypothesized that readmission risk would increase as number of hospitalizations decreased over time, we found no increase in readmission risk among the cohorts with HRRP diagnoses that had the largest decrease in readmission rates.^7,8 Conversely, readmission risk did increase—with a concurrent increase in the proportion of high-risk hospitalizations—in the surgery/gynecology and neurology cohorts that were not subject to HRRP penalties. Nonetheless, rehospitalizations were reduced for all risk categories in these 2 cohorts. Notably, surgery/gynecology and neurology had the lowest readmission rates overall. These findings suggest that initiatives to prevent initial hospitalizations, such as increasing the threshold for postoperative admission, may have had a greater effect on low- versus high-risk patients in low-risk hospitalizations. However, once a patient is hospitalized, multidisciplinary strategies appear to be effective at reducing readmissions for all risk classes in these cohorts.

For the 3 cohorts in which we observed an increase in readmission risk among hospitalized patients, the risk appeared to increase in early 2011. This time was about 10 months after passage of ACA, the timing of which was previously associated with a drop in readmission rates,^7,8 but well before HRRP went into effect in October 2012. The increase in readmission risk coincided with an increase in the number of diagnostic codes that could be included on a hospital claim to Medicare.¹⁷ This increase in allowable codes allowed us to capture more diagnoses for some patients, potentially resulting in an increase in apparent predicted risk of readmissions. While we adjusted for this in our predictive models, we may not have fully accounted for differences in risk related to coding change. As a result, some of the observed differences in risk in our study may be attributable to coding differences. More broadly, studies demonstrating the success of HRRP have typically examined risk-adjusted rates of readmission.^3,7 It is possible that a small portion of the observed reduction in risk-adjusted readmission rates may be related to the increase in predicted risk of readmission observed in our study. Future assessment of trends in readmission during this period should consider accounting for change in the number of allowed billing codes.

Other limitations should be considered in the interpretation of this study. First, like many predictive models for readmission,¹⁴ ours had imperfect discrimination, which could affect our results. Second, our study was based on older Medicare patients, so findings may not be applicable to younger patients. Third, while we accounted for surrogates for socioeconomic status, including dual eligibility and race, our models lacked other socioeconomic and community factors that can influence readmission.^24-26 Nonetheless, 1 study suggested that easily measured socioeconomic factors may not have a strong influence on the readmission metric used by Medicare.²⁸ Fourth, while our study included over 47 million hospitalizations, our time trend analyses used calendar month as the primary independent variable. As our study included 77 months, we may not have had sufficient power to detect small changes in risk over time.

Medicare readmissions have declined steadily in recent years, presumably at least partly in response to policy changes including HRRP. We found that hospitals have been effective at reducing readmissions across a range of patient risk strata and clinical conditions. As a result, the overall risk of readmission for hospitalized patients has remained constant for some but not all conditions. Whether institutions can continue to reduce readmission rates for most types of patients remains to be seen.

Acknowledgments

This study was supported by the Agency for Healthcare Research and Quality (AHRQ) grant R01HS022882. Dr. Blecker was supported by the AHRQ grant K08HS23683. The authors would like to thank Shawn Hoke and Jane Padikkala for administrative support.

Disclosure

This study was supported by the Agency for Healthcare Research and Quality (AHRQ) grants R01HS022882 and K08HS23683. The authors have no conflicts to report.

In this issue of the Journal of Hospital Medicine, Ellimoottil and colleagues examine characteristics of readmissions identified as planned by the planned readmission algorithm developed for the Center for Medicare & Medicaid Services (CMS) by using Medicare claims data from 131 hospitals in Michigan.¹ They found that a substantial portion of readmissions currently classified as planned by the algorithm appear to be nonelective, as defined by the presence of a charge by an emergency medicine physician or an admission type of emergent or urgent, making those hospitalizations unlikely to be planned. They suggest that the algorithm could be modified to exclude such cases from the planned designation.

To determine whether modifying the algorithm as recommended is a good idea, it is helpful to examine the origins of the existing planned readmission algorithm. The algorithm originated as a consequence of hospital accountability measures for readmissions and was developed by this author in collaboration with colleagues at Yale University and elsewhere.² Readmission measures have been controversial in part because clearly some (undetermined) fraction of readmissions is unavoidable. Many commentators have asked that readmission measures therefore capture only avoidable or related readmissions. Avoidable readmissions are those that could have been prevented by members of the healthcare system through actions taken during or after hospitalization, such as patient counseling, communication among team members, and guideline-concordant medical care. Related readmissions are those directly stemming from the index admission. However, reliably and accurately defining such events has proven elusive. One study, for instance, found the rate of physician-assessed preventability in published studies ranged from 9% to 48%.³ The challenge is even greater in trying to determine preventability using just claims data, without physician review of charts. Imagine, for instance, a patient with heart failure who is readmitted with heart failure exacerbation. The readmission preceded by a large fast-food meal is likely preventable; although even in this case, some would argue the healthcare system should not be held accountable for a readmission if the patient had been properly counseled about avoiding salty food. The one preceded by progressively worsening systolic function in a patient who reliably takes medications, weighs herself daily, and watches her diet is likely not. But both appear identical in claims. Related is also a difficult concept to operationalize. A recently hospitalized patient readmitted with pneumonia might have acquired it in the hospital (related) or from her grandchild 2 weeks later (unrelated). Again, both appear identical in claims.

In the ideal world, clinicians would be held accountable only for preventable readmissions. In practice, that has not proven to be possible.

Instead, the CMS readmission measures omit readmissions that are thought to be planned in advance: necessary and intentional readmissions. Defining a planned readmission is conceptually easier than defining a preventable readmission, yet even this is not always straightforward. The clearest case might be a person with a longstanding plan to have an elective surgery (say, a hip replacement) who is briefly admitted with something minor enough not to delay a subsequent admission for the scheduled surgery. Other patients are admitted with acute problems that require follow-up hospitalization (for instance, an acute myocardial infarction that requires a coronary artery bypass graft 2 weeks later).⁴ More ambiguous are patients who are sent home on a course of treatment with a plan for rehospitalization if it fails; for instance, a patient with gangrene is sent home on intravenous antibiotics but fails to improve and is rehospitalized for an amputation. Is that readmission planned or unplanned? Reasonable people might disagree.

Nonetheless, assuming it is desirable to at least try to identify and remove planned readmissions from measures, there are a number of ways in which one might do so. Perhaps the simplest would be to classify each hospitalization as planned or not on the UB-04 claim form. Such a process would be very feasible but also subject to gaming or coding variability. Given that there is some ambiguity and no standard about what types of readmissions are planned and that current policy provides incentives to reduce unplanned readmission rates, hospitals might vary in the cases to which they would apply such a code. This approach, therefore, has not been favored by payers to date. An alternative is to prospectively flag admissions that are expected to result in planned readmissions. In fiscal year 2014, the CMS implemented this option for newborns and patients with acute myocardial infarction by creating new discharge status codes of “discharged to [location] with a planned acute care hospital inpatient readmission.” Institutions can flag discharges that they know at the time of discharge will be followed by a readmission, such as a newborn who requires a repeat hospitalization for repair of a congenital anomaly.⁵ There is no time span required for the planned readmission to qualify. However, the difficulty in broadening the applicability of this option to all discharges lies in identification and matching; there also remains a possibility for gaming. The code does not specify when the readmission is expected nor for what diagnosis or procedure. How, then, do we know if the subsequent readmission is the one anticipated? Unexpected readmissions may still occur in the interim. Conversely, what if the discharging clinicians don’t know about an anticipated planned procedure? What would stop hospitals from labeling every discharge as expected to be followed by a planned readmission? These considerations have largely prevented the CMS from asking hospitals to apply the new code widely or from applying the code to identify planned readmissions.

Instead, the existing algorithm attempts to identify procedures that might be done on an elective basis and assumes readmissions with these procedures are planned if paired with a nonurgent diagnosis. Ellimoottil and colleagues attempt to verify whether this is accurate using a creative approach of seeking emergency department (ED) charges and admission type of emergent or urgent, and they found that roughly half of planned readmissions are, in fact, likely unplanned. This figure agrees closely with the original chart review validation of the algorithm. In particular, they found that some procedures, such as percutaneous cardiac interventions, appear to be paired regularly with a nonurgent principal diagnosis, such as coronary artery disease, even when done on an urgent basis.

This validation was performed prior to the availability of version 4.0 of the planned readmission algorithm, which removes several high-frequency procedures from the potentially planned readmission list (including cardiac devices and diagnostic cardiac catheterizations) that were very frequently mischaracterized as planned in the original chart validation.⁶ At least 8 such cases were also identified in this validation according to the table. Therefore, the misclassification rate of the current algorithm version is probably less than that reported in this article. Nonetheless, percutaneous transluminal coronary angioplasty remains on the planned procedure list in version 4.0 and appears to account for a substantial error rate, and it is likely that the authors’ approach would improve the accuracy even of the newer version of the algorithm.

The advantages of the suggested modifications are that they do not require chart review and could be readily adopted by the CMS. Although seeking ED charges for Medicare is somewhat cumbersome in that they are recorded in a different data set than the inpatient hospitalizations, there is no absolute barrier to adding this step to the algorithm, and doing so has substantial face validity. That said, identifying ED visits is not straightforward because nonemergency services can be provided in the ED (ie, critical care or observation care) and because facilities and providers have different billing requirements, producing different estimates depending on the data set used.⁷ Including admission type would be easier, but it would be less conservative and likely less accurate, as this field has not been validated and is not typically audited. Nonetheless, adding the presence of ED charges seems likely to improve the accuracy of the algorithm. As the CMS continues to refine the planned readmission algorithm, these proposed changes would be very reasonable to study with chart validation and, if valid, to consider adopting.

Disclosure 

Dr. Horwitz reports grants from Center for Medicare & Medicaid Services, grants from Agency for Healthcare Research and Quality, during the conduct of the study.

In this issue of the Journal of Hospital Medicine, Ellimoottil and colleagues examine characteristics of readmissions identified as planned by the planned readmission algorithm developed for the Center for Medicare & Medicaid Services (CMS) by using Medicare claims data from 131 hospitals in Michigan.¹ They found that a substantial portion of readmissions currently classified as planned by the algorithm appear to be nonelective, as defined by the presence of a charge by an emergency medicine physician or an admission type of emergent or urgent, making those hospitalizations unlikely to be planned. They suggest that the algorithm could be modified to exclude such cases from the planned designation.

To determine whether modifying the algorithm as recommended is a good idea, it is helpful to examine the origins of the existing planned readmission algorithm. The algorithm originated as a consequence of hospital accountability measures for readmissions and was developed by this author in collaboration with colleagues at Yale University and elsewhere.² Readmission measures have been controversial in part because clearly some (undetermined) fraction of readmissions is unavoidable. Many commentators have asked that readmission measures therefore capture only avoidable or related readmissions. Avoidable readmissions are those that could have been prevented by members of the healthcare system through actions taken during or after hospitalization, such as patient counseling, communication among team members, and guideline-concordant medical care. Related readmissions are those directly stemming from the index admission. However, reliably and accurately defining such events has proven elusive. One study, for instance, found the rate of physician-assessed preventability in published studies ranged from 9% to 48%.³ The challenge is even greater in trying to determine preventability using just claims data, without physician review of charts. Imagine, for instance, a patient with heart failure who is readmitted with heart failure exacerbation. The readmission preceded by a large fast-food meal is likely preventable; although even in this case, some would argue the healthcare system should not be held accountable for a readmission if the patient had been properly counseled about avoiding salty food. The one preceded by progressively worsening systolic function in a patient who reliably takes medications, weighs herself daily, and watches her diet is likely not. But both appear identical in claims. Related is also a difficult concept to operationalize. A recently hospitalized patient readmitted with pneumonia might have acquired it in the hospital (related) or from her grandchild 2 weeks later (unrelated). Again, both appear identical in claims.

In the ideal world, clinicians would be held accountable only for preventable readmissions. In practice, that has not proven to be possible.

Instead, the CMS readmission measures omit readmissions that are thought to be planned in advance: necessary and intentional readmissions. Defining a planned readmission is conceptually easier than defining a preventable readmission, yet even this is not always straightforward. The clearest case might be a person with a longstanding plan to have an elective surgery (say, a hip replacement) who is briefly admitted with something minor enough not to delay a subsequent admission for the scheduled surgery. Other patients are admitted with acute problems that require follow-up hospitalization (for instance, an acute myocardial infarction that requires a coronary artery bypass graft 2 weeks later).⁴ More ambiguous are patients who are sent home on a course of treatment with a plan for rehospitalization if it fails; for instance, a patient with gangrene is sent home on intravenous antibiotics but fails to improve and is rehospitalized for an amputation. Is that readmission planned or unplanned? Reasonable people might disagree.

Nonetheless, assuming it is desirable to at least try to identify and remove planned readmissions from measures, there are a number of ways in which one might do so. Perhaps the simplest would be to classify each hospitalization as planned or not on the UB-04 claim form. Such a process would be very feasible but also subject to gaming or coding variability. Given that there is some ambiguity and no standard about what types of readmissions are planned and that current policy provides incentives to reduce unplanned readmission rates, hospitals might vary in the cases to which they would apply such a code. This approach, therefore, has not been favored by payers to date. An alternative is to prospectively flag admissions that are expected to result in planned readmissions. In fiscal year 2014, the CMS implemented this option for newborns and patients with acute myocardial infarction by creating new discharge status codes of “discharged to [location] with a planned acute care hospital inpatient readmission.” Institutions can flag discharges that they know at the time of discharge will be followed by a readmission, such as a newborn who requires a repeat hospitalization for repair of a congenital anomaly.⁵ There is no time span required for the planned readmission to qualify. However, the difficulty in broadening the applicability of this option to all discharges lies in identification and matching; there also remains a possibility for gaming. The code does not specify when the readmission is expected nor for what diagnosis or procedure. How, then, do we know if the subsequent readmission is the one anticipated? Unexpected readmissions may still occur in the interim. Conversely, what if the discharging clinicians don’t know about an anticipated planned procedure? What would stop hospitals from labeling every discharge as expected to be followed by a planned readmission? These considerations have largely prevented the CMS from asking hospitals to apply the new code widely or from applying the code to identify planned readmissions.

Instead, the existing algorithm attempts to identify procedures that might be done on an elective basis and assumes readmissions with these procedures are planned if paired with a nonurgent diagnosis. Ellimoottil and colleagues attempt to verify whether this is accurate using a creative approach of seeking emergency department (ED) charges and admission type of emergent or urgent, and they found that roughly half of planned readmissions are, in fact, likely unplanned. This figure agrees closely with the original chart review validation of the algorithm. In particular, they found that some procedures, such as percutaneous cardiac interventions, appear to be paired regularly with a nonurgent principal diagnosis, such as coronary artery disease, even when done on an urgent basis.

This validation was performed prior to the availability of version 4.0 of the planned readmission algorithm, which removes several high-frequency procedures from the potentially planned readmission list (including cardiac devices and diagnostic cardiac catheterizations) that were very frequently mischaracterized as planned in the original chart validation.⁶ At least 8 such cases were also identified in this validation according to the table. Therefore, the misclassification rate of the current algorithm version is probably less than that reported in this article. Nonetheless, percutaneous transluminal coronary angioplasty remains on the planned procedure list in version 4.0 and appears to account for a substantial error rate, and it is likely that the authors’ approach would improve the accuracy even of the newer version of the algorithm.

The advantages of the suggested modifications are that they do not require chart review and could be readily adopted by the CMS. Although seeking ED charges for Medicare is somewhat cumbersome in that they are recorded in a different data set than the inpatient hospitalizations, there is no absolute barrier to adding this step to the algorithm, and doing so has substantial face validity. That said, identifying ED visits is not straightforward because nonemergency services can be provided in the ED (ie, critical care or observation care) and because facilities and providers have different billing requirements, producing different estimates depending on the data set used.⁷ Including admission type would be easier, but it would be less conservative and likely less accurate, as this field has not been validated and is not typically audited. Nonetheless, adding the presence of ED charges seems likely to improve the accuracy of the algorithm. As the CMS continues to refine the planned readmission algorithm, these proposed changes would be very reasonable to study with chart validation and, if valid, to consider adopting.

Disclosure 

Dr. Horwitz reports grants from Center for Medicare & Medicaid Services, grants from Agency for Healthcare Research and Quality, during the conduct of the study.

In this issue of the Journal of Hospital Medicine, Ellimoottil and colleagues examine characteristics of readmissions identified as planned by the planned readmission algorithm developed for the Center for Medicare & Medicaid Services (CMS) by using Medicare claims data from 131 hospitals in Michigan.¹ They found that a substantial portion of readmissions currently classified as planned by the algorithm appear to be nonelective, as defined by the presence of a charge by an emergency medicine physician or an admission type of emergent or urgent, making those hospitalizations unlikely to be planned. They suggest that the algorithm could be modified to exclude such cases from the planned designation.

To determine whether modifying the algorithm as recommended is a good idea, it is helpful to examine the origins of the existing planned readmission algorithm. The algorithm originated as a consequence of hospital accountability measures for readmissions and was developed by this author in collaboration with colleagues at Yale University and elsewhere.² Readmission measures have been controversial in part because clearly some (undetermined) fraction of readmissions is unavoidable. Many commentators have asked that readmission measures therefore capture only avoidable or related readmissions. Avoidable readmissions are those that could have been prevented by members of the healthcare system through actions taken during or after hospitalization, such as patient counseling, communication among team members, and guideline-concordant medical care. Related readmissions are those directly stemming from the index admission. However, reliably and accurately defining such events has proven elusive. One study, for instance, found the rate of physician-assessed preventability in published studies ranged from 9% to 48%.³ The challenge is even greater in trying to determine preventability using just claims data, without physician review of charts. Imagine, for instance, a patient with heart failure who is readmitted with heart failure exacerbation. The readmission preceded by a large fast-food meal is likely preventable; although even in this case, some would argue the healthcare system should not be held accountable for a readmission if the patient had been properly counseled about avoiding salty food. The one preceded by progressively worsening systolic function in a patient who reliably takes medications, weighs herself daily, and watches her diet is likely not. But both appear identical in claims. Related is also a difficult concept to operationalize. A recently hospitalized patient readmitted with pneumonia might have acquired it in the hospital (related) or from her grandchild 2 weeks later (unrelated). Again, both appear identical in claims.

In the ideal world, clinicians would be held accountable only for preventable readmissions. In practice, that has not proven to be possible.

Instead, the CMS readmission measures omit readmissions that are thought to be planned in advance: necessary and intentional readmissions. Defining a planned readmission is conceptually easier than defining a preventable readmission, yet even this is not always straightforward. The clearest case might be a person with a longstanding plan to have an elective surgery (say, a hip replacement) who is briefly admitted with something minor enough not to delay a subsequent admission for the scheduled surgery. Other patients are admitted with acute problems that require follow-up hospitalization (for instance, an acute myocardial infarction that requires a coronary artery bypass graft 2 weeks later).⁴ More ambiguous are patients who are sent home on a course of treatment with a plan for rehospitalization if it fails; for instance, a patient with gangrene is sent home on intravenous antibiotics but fails to improve and is rehospitalized for an amputation. Is that readmission planned or unplanned? Reasonable people might disagree.

Nonetheless, assuming it is desirable to at least try to identify and remove planned readmissions from measures, there are a number of ways in which one might do so. Perhaps the simplest would be to classify each hospitalization as planned or not on the UB-04 claim form. Such a process would be very feasible but also subject to gaming or coding variability. Given that there is some ambiguity and no standard about what types of readmissions are planned and that current policy provides incentives to reduce unplanned readmission rates, hospitals might vary in the cases to which they would apply such a code. This approach, therefore, has not been favored by payers to date. An alternative is to prospectively flag admissions that are expected to result in planned readmissions. In fiscal year 2014, the CMS implemented this option for newborns and patients with acute myocardial infarction by creating new discharge status codes of “discharged to [location] with a planned acute care hospital inpatient readmission.” Institutions can flag discharges that they know at the time of discharge will be followed by a readmission, such as a newborn who requires a repeat hospitalization for repair of a congenital anomaly.⁵ There is no time span required for the planned readmission to qualify. However, the difficulty in broadening the applicability of this option to all discharges lies in identification and matching; there also remains a possibility for gaming. The code does not specify when the readmission is expected nor for what diagnosis or procedure. How, then, do we know if the subsequent readmission is the one anticipated? Unexpected readmissions may still occur in the interim. Conversely, what if the discharging clinicians don’t know about an anticipated planned procedure? What would stop hospitals from labeling every discharge as expected to be followed by a planned readmission? These considerations have largely prevented the CMS from asking hospitals to apply the new code widely or from applying the code to identify planned readmissions.

Instead, the existing algorithm attempts to identify procedures that might be done on an elective basis and assumes readmissions with these procedures are planned if paired with a nonurgent diagnosis. Ellimoottil and colleagues attempt to verify whether this is accurate using a creative approach of seeking emergency department (ED) charges and admission type of emergent or urgent, and they found that roughly half of planned readmissions are, in fact, likely unplanned. This figure agrees closely with the original chart review validation of the algorithm. In particular, they found that some procedures, such as percutaneous cardiac interventions, appear to be paired regularly with a nonurgent principal diagnosis, such as coronary artery disease, even when done on an urgent basis.

This validation was performed prior to the availability of version 4.0 of the planned readmission algorithm, which removes several high-frequency procedures from the potentially planned readmission list (including cardiac devices and diagnostic cardiac catheterizations) that were very frequently mischaracterized as planned in the original chart validation.⁶ At least 8 such cases were also identified in this validation according to the table. Therefore, the misclassification rate of the current algorithm version is probably less than that reported in this article. Nonetheless, percutaneous transluminal coronary angioplasty remains on the planned procedure list in version 4.0 and appears to account for a substantial error rate, and it is likely that the authors’ approach would improve the accuracy even of the newer version of the algorithm.

The advantages of the suggested modifications are that they do not require chart review and could be readily adopted by the CMS. Although seeking ED charges for Medicare is somewhat cumbersome in that they are recorded in a different data set than the inpatient hospitalizations, there is no absolute barrier to adding this step to the algorithm, and doing so has substantial face validity. That said, identifying ED visits is not straightforward because nonemergency services can be provided in the ED (ie, critical care or observation care) and because facilities and providers have different billing requirements, producing different estimates depending on the data set used.⁷ Including admission type would be easier, but it would be less conservative and likely less accurate, as this field has not been validated and is not typically audited. Nonetheless, adding the presence of ED charges seems likely to improve the accuracy of the algorithm. As the CMS continues to refine the planned readmission algorithm, these proposed changes would be very reasonable to study with chart validation and, if valid, to consider adopting.

Disclosure 

Dr. Horwitz reports grants from Center for Medicare & Medicaid Services, grants from Agency for Healthcare Research and Quality, during the conduct of the study.

Millions of individuals with chronic diseases are hospitalized annually in the United States. More than 90% of hospitalized adults have at least 1 chronic disease,¹ and almost half of Medicare beneficiaries in the hospital have 4 or more chronic conditions.² While many patients are admitted for worsening of a single chronic disease, patients are hospitalized more commonly for other causes. For instance, although acute heart failure is among the most frequent causes of hospitalizations among older adults, three-fourths of hospitalizations of patients with heart failure are for reasons other than acute heart failure.³

When a patient with a chronic disease is hospitalized, the inpatient provider must consider whether to actively or passively manage the chronic disease. Studies have suggested that intervening in chronic diseases during hospitalizations can lead to long-term improvement in treatment;^4-6 for instance, stroke patients who were started on antihypertensive therapy at discharge were more likely to have their blood pressure controlled in the next year.⁵ However, some authors have argued that aggressive hypertension management by inpatient providers may result in patient harm.⁷ One case-based survey suggested that hospitalists were mixed in their interest in participating in chronic disease management in the hospital.⁸ This study found that providers were less likely to participate in chronic disease management if it was unrelated to the reason for hospitalization.⁸ However, to our knowledge, no studies have broadly evaluated inpatient provider attitudes, motivating factors, or barriers to participation in chronic disease management.

The purpose of this study was to understand provider attitudes towards chronic disease management for patients who are hospitalized for other causes. We were particularly interested in perceptions of barriers and facilitators to delivery of inpatient chronic disease management. Ultimately, such findings can inform future interventions to improve inpatient care of chronic disease.

In this qualitative study, we conducted in-depth interviews with providers to understand attitudes, barriers, and facilitators towards inpatient management of chronic disease; this study was part of a larger study to implement an electronic health record-based clinical decision-support system intervention to improve quality of care for hospitalized patients with heart failure.

We included providers who care for and can write medication orders for hospitalized adult patients at New York University (NYU) Langone Medical Center, an urban academic medical center. As patients with chronic conditions are commonly hospitalized for many reasons, we sought to interview providers from a range of clinical services without consideration of factors such as frequency of caring for patients with heart failure. We used a purposive sampling framework: we invited participants to ensure a range of services, including medicine, surgery, and neurology, and provider types, including attending physicians, resident physicians, nurse practitioners, and physician assistants. Potential participants, therefore, included all providers for adult hospitalized patients.

We identified potential participants through study team members, referrals from department heads and prior interviewees, and e-mails to department list serves. We did not formally track declinations to being interviewed, although we estimate them as fewer than 20% of providers directly approached. While we focused on inpatient providers at New York University Langone Medical Center, many of the attending physicians and residents spend a portion of their time at the Manhattan Veterans Affairs Hospital and Bellevue Hospital, a safety-net city hospital; providers could have outpatient responsibilities as well.

All participants provided verbal consent to participate. The study was approved by the New York University Institutional Review Board, which granted a waiver of documentation of consent. Participants received a $25 gift card following the interview.

We used a semi-structured interview guide (Appendix) to elicit in-depth accounts of provider attitudes, experiences with, and barriers and facilitators towards chronic disease management in the hospital. The interview began by asking about chronic disease in general and then asked more specific questions about heart failure; we included responses to both groups of questions in the current study. The interview also included questions related to the clinical decision-support system being developed as part of the larger implementation study, although we do not report on these results in the current study. The semi-structured interview guide was informed by the consolidated framework for advancing implementation science (CFIR), which offers an overarching typology for delineating factors that influence guideline implementation;⁹ we also used CFIR constructs in theme development. We conducted in-depth interviews with providers.

A priori, we estimated 25 interviews would be sufficient to include the purposive sample and achieve data saturation,¹⁰ which was reached after 31 interviews. Interviews were held in person or by telephone, at the convenience of the subject. All interviews were transcribed by a professional service. Transcriptions were reviewed against recordings with any mistakes corrected. Prior to each interview, we conducted a brief demographic survey.

Qualitative data were analyzed using a constant comparative analytic technique.¹¹ The investigative team met after reviewing the first 10 interviews and discussed emergent themes from these early transcripts, which led to the initial code list. Two investigators coded the transcripts. Reliability was evaluated by independent coding of a 20% subset of interviews. Differences were reviewed and discussed until consensus was reached. Final intercoder reliability was determined to be greater than 95%.¹² All investigators reviewed and refined the code list during the analysis phase. Codes were clustered into themes based on CFIR constructs.⁹ Analyses were performed using Atlas.ti v. 7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany).

We conducted interviews with 31 providers. Of these, 12 were on the medicine service, 12 were on the surgery or a surgical subspecialty service, and 7 were on other services; 11 were attending physicians, 12 were resident physicians, 5 were NPs, and 3 were PAs. Only 2 providers—an attending in medicine and a resident in surgery—had a specialty focus that was cardiac-related. Median time in current position was 4 years (Table 1). Seventeen of the interviews were in person, and 14 were conducted by telephone. The mean interview time was 20 minutes and ranged from 11 to 41 minutes.

We identified 5 main themes with 29 supporting codes (Table 2) describing provider attitudes towards the management of chronic disease for hospitalized patients. These themes, with related CFIR constructs, were: 1) perceived impact on patient outcomes (CFIR construct: intervention characteristics, relative advantage); 2) hospital structural characteristics (inner setting, structural characteristics); 3) provider knowledge and self-efficacy (characteristic of individual, knowledge and beliefs about the intervention and self-efficacy); 4) hospital priorities (inner setting, implementation climate, relative priority); and 5) continuity and communication (inner setting, networks and communications). For most themes, subjects described both positive and negative aspects of chronic disease management, as well as related facilitators and barriers to delivery of chronic disease care for hospitalized patients. Illustrative quotes for each theme are shown in Table 3.

Perceived Impact on Patient Outcomes

Perceived impact on patient outcomes was mixed. Most providers believed the management of chronic diseases could lead to improvement in important patient outcomes, including decreased length of stay (LOS), prevention of hospital complication, and decreased readmissions. Surgical providers focused particularly on the benefits of preventing surgical complications and noted that they were more likely to manage chronic conditions—primarily through use of specialist consultation—when they perceived a benefit to prevention of surgical outcomes or a fear that surgery may worsen a stable chronic condition:

“Most of the surgery I do is pretty stressful on the body and is very likely to induce acute on chronic exacerbations of heart failure. For someone with Class II or higher heart failure, I’m definitely gonna have cardiology on board or at least internal medicine on board right from the beginning.”

Hospital Structural Characteristics

For many providers, the hospital setting provides a unique opportunity for care of patients with chronic disease. First, a hospitalization is a time for a patient’s management to be reviewed by a new care team. The hospital team reviews the management plan for patients at admission, which is a time to reevaluate whether patients are on evidence-based therapies: “It’s helpful to have a new set of eyes on somebody, like fresh information.” According to providers, this reevaluation can overcome instances of therapeutic inertia by the outpatient physician. Second, the hospital has many resources, including readily available specialist services and diagnostic tests, which can allow a patient-centered approach that coordinates care in 1 place, as a surgery NP described: “I think the advantage for the patient is that they wind up stopping in for 1 thing but we wind up taking care of a few without requiring the need for him or her to go to all these different specialists on the outside. They’re mostly elderly and not able to get around.” Third, the high availability of services and frequent monitoring allows rapid titration of evidence-based medicines, as discussed by a medicine resident: “It’s easier and faster to titrate medication—they’re in a monitored setting; you can ensure compliance.”

Patients may also differ from their usual state while hospitalized, creating both risks and benefits. The hospital setting can provide an opportunity to educate patients on their chronic disease(s) because they are motivated: “They’re in an office visit and their sugars are out of whack or something, they may take it a little bit more seriously if they were just in the hospital even though it was on an unrelated issue. I think it probably just changes their perspective on their disease.” However, in the hospital, patients are in an unusual environment with a restricted diet and forced medication compliance. Furthermore, the acute condition can lead to changes in their chronic disease, as described by 1 medicine attending: “their sugar is high because they’re acutely ill.” Providers expressed concern that changing medications in this setting may lead to adverse events (AEs) when patients return to their usual environment.

Provider Knowledge and Self-Efficacy

Insufficient knowledge of treatments for chronic conditions was cited as a barrier to some providers’ ability to actively manage chronic disease for hospitalized patients. Some providers described management of conditions outside their area as less satisfying than their primary focus. For example, an orthopedic surgeon explained: “…it’s very simple. You see your bone is broken, you fix it, that’s it…it’s intellectually satisfying…managing chronic diseases is less like that.” Reliance on consultants was 1 approach to deal with knowledge gaps in areas outside a provider’s expertise.

For a number of providers, management of stable chronic disease is the responsibility of the outpatient provider. Providers expressed concern that inpatient management was a reach into the domain of the primary care provider (PCP) and might take “away from the primary focus” of the hospitalization. Nonetheless, some providers noted an “ethical responsibility to manage [a] patient correctly,” and some providers believed that engaging in chronic disease management in the hospital would present an opportunity to expand their own expertise.

A few providers were worried about legal risk related to chronic disease management: “we don’t typically deal too much with managing some of these other medical issues for medical and legal reasons.” Providers again suggested that consults can help overcome this concern for risk, as discussed by 1 surgical attending: “We’re all not wanting to be sued, and we want to do the right thing. It costs me nothing to have a cardiologist on board, so like—why not.”

Hospital Priorities

Providers explained that the hospital has strong interests in early discharge and minimizing LOS. These priorities are based on goals of improving patient outcomes, increasing bed availability and hospital volume, and reducing costs. Providers perceive these hospital priorities as potential barriers to chronic disease management, which can increase LOS and costs through additional testing and treatment. As a medicine resident described: “The DBN philosophy, ‘discharge before noon’ philosophy, which is part of the hospital efficiency to get people in and out of the hospital as quickly as [is] safe, or maybe faster. And I think that there’s a culture where you’re encouraged to only focus on the acute issue and tend to defer everything else.”

Continuity and Communication

According to many providers, care continuity between the outpatient setting and the hospital played a major role in management of chronic disease. One barrier to starting a new evidence-based medication was lack of knowledge of patient history. As noted, providers expressed concern that a patient may not be on a given therapy because of an adverse reaction that was not documented in the hospital chart. This is particularly true because, as discussed by a surgery resident, patients with “PCPs outside the system [in which providers] don’t have access to the electronic medical record.” To overcome this barrier, providers attempt to communicate with the outpatient provider to confirm a lack of contraindications to therapies prior to any changes; notably, communication is easier if the inpatient provider has a relationship with the outpatient PCP.

Some providers were more likely to start chronic disease therapies if the patient had no prior outpatient care, because the provider was reassured that there was no rationale for missing therapies. One neurology attending noted that if a patient had newly documented “hypertension even if they were in for something else, I might start them on an antihypertensive, but then arrange for a close follow-up with a new PCP.”

Following hospitalization, providers wanted assurance that any changes to chronic disease management would be followed up by an outpatient physician. Any changes are relayed to the outpatient provider and the “level of communication…with the outpatient provider who’s gonna inherit” these changes can influence how aggressively the inpatient provider manages chronic diseases. Providers may be reluctant to start therapy for patients if they are concerned about outpatient follow up: “they have diabetes and they should really technically be on an ACE [angiotensin converting enzyme]inhibitor and aspirin, but they’re not. I might send them out on the aspirin but I might either start ACE inhibitor and have them follow up with their PCP in 2 weeks if I’m confident that they’ll do it or if I’m really confident that they’ll not follow up, I will help them get the appointment and then the discharge instruction is to the PCP is ‘Please start this patient on ACE inhibitor if they show up.’”

Providers frequently perceive benefit to chronic disease management in the hospital, including improvements in clinical outcomes. Notably, providers see opportunities to improve compliance with evidence-based care to overcome potential barriers to managing chronic disease in the outpatient setting, which can be limited by pressure for brief encounters,¹³ clinical inertia,¹⁴ difficulty with close monitoring of patients,¹⁵ and care fragmentation.¹⁶ Concurrently, inpatient providers are concerned about potential for patient harm related to chronic disease management, primarily related to AEs from medications. Similar to a case study about a patient with outpatient hypotension following aggressive inpatient hypertension management,⁷ providers fear that changing a patient’s chronic disease management in a hospital setting may cause harm when the patient returns home.

Although some clinicians have argued against aggressive in-hospital chronic disease management because of concerns for risk of AEs,⁷ our study and others⁸ have suggested that many clinicians perceive benefit. In some cases, such as smoking cessation counseling for all current smokers and prescribing an angiotensin converting enzyme inhibitor for patients with systolic heart failure, the perceived importance is so great that chronic disease management has been used as a national quality metric for hospitals. While these hospital metrics may be justified for short-term benefits after hospitalization, studies have demonstrated only weak improvement in short-term postdischarge outcomes related to chronic disease management.¹⁷ The true benefit is likely from improved processes of care in the short term that lead to long-term improvement in outcomes.^4,5,18 Thus, the advantage of starting a patient hospitalized for a stroke on blood pressure medication is the increased likelihood that the patient will continue the medication as an outpatient, which may reduce long-term mortality.

For hospital delivery systems that are concerned with such care process improvement through in-hospital chronic disease management, we identified a number of barriers and facilitators to delivering this care. One significant barrier was poor transitions between the inpatient and the outpatient settings. When a patient transitions into the hospital, providers need to understand prior management choices. Facilitators to help inpatient providers understand prior management included either knowing the outpatient provider, or understanding that there was a lack of regular outpatient care; in both these cases, inpatient providers felt more comfortable managing chronic diseases because they had insight into the outpatient plan, or lack thereof. However, these facilitators may not be practical to incorporate in interventions to improve chronic disease care, which should consider overcoming these communication barriers. Use of shared electronic health records or standardized telephone calls with well-documented care plans obtained through health information exchanges may facilitate an inpatient provider to manage appropriately chronic disease. Similarly, discontinuity between the inpatient provider and the outpatient provider is a barrier that must be overcome to ease concerns that any chronic disease management changes do not result in harm in the postdischarge period. These findings again point to the need for improved documentation and communication between inpatient and outpatient providers. Of course, the transitional care period is one of high risk, and improving communication between providers has been an area of ongoing work.¹⁹

Lack of comfort among inpatient providers with managing chronic diseases is another important barrier, which appears to be largely overcome through the use of consultation services. Ready availability of specialists, common in academic medical centers, can facilitate delivery of chronic disease management. Inpatient interventions designed to improve evidence-based care for a chronic disease may benefit from involvement or at least availability of specialists in the effort. Another major barrier relates to hospital priorities, which in our study were closely aligned with external factors such as payment models. As hospitalizations are typically paid based on the discharge diagnosis, hospitals have incentives to discharge quickly and not order extra diagnostic tests. As a result, there are disincentives for chronic disease management that may require additional testing or monitoring in the hospital. Conversely, as hospitals accept postdischarge financial risks through readmission penalties or postdischarge cost savings, hospitals may perceive that long-term benefits of chronic disease management may outweigh short-term costs.

The study findings should be interpreted in the context of its limitations. Findings of our study of providers from a single academic medical center may not be generalizable. Nearly half of our interviews were conducted by telephone, which limits our ability to capture nonverbal cues in communication. Providers may have had social desirability bias towards positive aspects of chronic disease management. We did not have the power to determine differences in response by provider characteristic because this was an exploratory qualitative study. Future studies with representative sampling, a larger sample size, and measures for constructs such as provider self-efficacy are needed to examine differences by specialty, provider type, and experience level.

In conclusion, inpatient providers believe that hospital chronic disease management has the potential to be beneficial for both process of care and clinical outcomes; providers also express concern about potential adverse consequences of managing chronic disease during acute hospitalizations. To maximize both quality of care and patient safety, overcoming communication barriers between inpatient and outpatient providers is needed. Both a supportive hospital environment and availability of specialty support can facilitate in-hospital chronic disease management. Interventions that incorporate these factors may be well-suited to improve chronic disease care and long-term outcomes.

Disclosures

This work was supported by the Agency for Healthcare Research and Quality (AHRQ) grant K08HS23683. The authors report no financial conflicts of interest.

Millions of individuals with chronic diseases are hospitalized annually in the United States. More than 90% of hospitalized adults have at least 1 chronic disease,¹ and almost half of Medicare beneficiaries in the hospital have 4 or more chronic conditions.² While many patients are admitted for worsening of a single chronic disease, patients are hospitalized more commonly for other causes. For instance, although acute heart failure is among the most frequent causes of hospitalizations among older adults, three-fourths of hospitalizations of patients with heart failure are for reasons other than acute heart failure.³

When a patient with a chronic disease is hospitalized, the inpatient provider must consider whether to actively or passively manage the chronic disease. Studies have suggested that intervening in chronic diseases during hospitalizations can lead to long-term improvement in treatment;^4-6 for instance, stroke patients who were started on antihypertensive therapy at discharge were more likely to have their blood pressure controlled in the next year.⁵ However, some authors have argued that aggressive hypertension management by inpatient providers may result in patient harm.⁷ One case-based survey suggested that hospitalists were mixed in their interest in participating in chronic disease management in the hospital.⁸ This study found that providers were less likely to participate in chronic disease management if it was unrelated to the reason for hospitalization.⁸ However, to our knowledge, no studies have broadly evaluated inpatient provider attitudes, motivating factors, or barriers to participation in chronic disease management.

The purpose of this study was to understand provider attitudes towards chronic disease management for patients who are hospitalized for other causes. We were particularly interested in perceptions of barriers and facilitators to delivery of inpatient chronic disease management. Ultimately, such findings can inform future interventions to improve inpatient care of chronic disease.

In this qualitative study, we conducted in-depth interviews with providers to understand attitudes, barriers, and facilitators towards inpatient management of chronic disease; this study was part of a larger study to implement an electronic health record-based clinical decision-support system intervention to improve quality of care for hospitalized patients with heart failure.

We included providers who care for and can write medication orders for hospitalized adult patients at New York University (NYU) Langone Medical Center, an urban academic medical center. As patients with chronic conditions are commonly hospitalized for many reasons, we sought to interview providers from a range of clinical services without consideration of factors such as frequency of caring for patients with heart failure. We used a purposive sampling framework: we invited participants to ensure a range of services, including medicine, surgery, and neurology, and provider types, including attending physicians, resident physicians, nurse practitioners, and physician assistants. Potential participants, therefore, included all providers for adult hospitalized patients.

We identified potential participants through study team members, referrals from department heads and prior interviewees, and e-mails to department list serves. We did not formally track declinations to being interviewed, although we estimate them as fewer than 20% of providers directly approached. While we focused on inpatient providers at New York University Langone Medical Center, many of the attending physicians and residents spend a portion of their time at the Manhattan Veterans Affairs Hospital and Bellevue Hospital, a safety-net city hospital; providers could have outpatient responsibilities as well.

All participants provided verbal consent to participate. The study was approved by the New York University Institutional Review Board, which granted a waiver of documentation of consent. Participants received a $25 gift card following the interview.

We used a semi-structured interview guide (Appendix) to elicit in-depth accounts of provider attitudes, experiences with, and barriers and facilitators towards chronic disease management in the hospital. The interview began by asking about chronic disease in general and then asked more specific questions about heart failure; we included responses to both groups of questions in the current study. The interview also included questions related to the clinical decision-support system being developed as part of the larger implementation study, although we do not report on these results in the current study. The semi-structured interview guide was informed by the consolidated framework for advancing implementation science (CFIR), which offers an overarching typology for delineating factors that influence guideline implementation;⁹ we also used CFIR constructs in theme development. We conducted in-depth interviews with providers.

A priori, we estimated 25 interviews would be sufficient to include the purposive sample and achieve data saturation,¹⁰ which was reached after 31 interviews. Interviews were held in person or by telephone, at the convenience of the subject. All interviews were transcribed by a professional service. Transcriptions were reviewed against recordings with any mistakes corrected. Prior to each interview, we conducted a brief demographic survey.

Qualitative data were analyzed using a constant comparative analytic technique.¹¹ The investigative team met after reviewing the first 10 interviews and discussed emergent themes from these early transcripts, which led to the initial code list. Two investigators coded the transcripts. Reliability was evaluated by independent coding of a 20% subset of interviews. Differences were reviewed and discussed until consensus was reached. Final intercoder reliability was determined to be greater than 95%.¹² All investigators reviewed and refined the code list during the analysis phase. Codes were clustered into themes based on CFIR constructs.⁹ Analyses were performed using Atlas.ti v. 7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany).

We conducted interviews with 31 providers. Of these, 12 were on the medicine service, 12 were on the surgery or a surgical subspecialty service, and 7 were on other services; 11 were attending physicians, 12 were resident physicians, 5 were NPs, and 3 were PAs. Only 2 providers—an attending in medicine and a resident in surgery—had a specialty focus that was cardiac-related. Median time in current position was 4 years (Table 1). Seventeen of the interviews were in person, and 14 were conducted by telephone. The mean interview time was 20 minutes and ranged from 11 to 41 minutes.

We identified 5 main themes with 29 supporting codes (Table 2) describing provider attitudes towards the management of chronic disease for hospitalized patients. These themes, with related CFIR constructs, were: 1) perceived impact on patient outcomes (CFIR construct: intervention characteristics, relative advantage); 2) hospital structural characteristics (inner setting, structural characteristics); 3) provider knowledge and self-efficacy (characteristic of individual, knowledge and beliefs about the intervention and self-efficacy); 4) hospital priorities (inner setting, implementation climate, relative priority); and 5) continuity and communication (inner setting, networks and communications). For most themes, subjects described both positive and negative aspects of chronic disease management, as well as related facilitators and barriers to delivery of chronic disease care for hospitalized patients. Illustrative quotes for each theme are shown in Table 3.

Perceived Impact on Patient Outcomes

Perceived impact on patient outcomes was mixed. Most providers believed the management of chronic diseases could lead to improvement in important patient outcomes, including decreased length of stay (LOS), prevention of hospital complication, and decreased readmissions. Surgical providers focused particularly on the benefits of preventing surgical complications and noted that they were more likely to manage chronic conditions—primarily through use of specialist consultation—when they perceived a benefit to prevention of surgical outcomes or a fear that surgery may worsen a stable chronic condition:

“Most of the surgery I do is pretty stressful on the body and is very likely to induce acute on chronic exacerbations of heart failure. For someone with Class II or higher heart failure, I’m definitely gonna have cardiology on board or at least internal medicine on board right from the beginning.”

Hospital Structural Characteristics

For many providers, the hospital setting provides a unique opportunity for care of patients with chronic disease. First, a hospitalization is a time for a patient’s management to be reviewed by a new care team. The hospital team reviews the management plan for patients at admission, which is a time to reevaluate whether patients are on evidence-based therapies: “It’s helpful to have a new set of eyes on somebody, like fresh information.” According to providers, this reevaluation can overcome instances of therapeutic inertia by the outpatient physician. Second, the hospital has many resources, including readily available specialist services and diagnostic tests, which can allow a patient-centered approach that coordinates care in 1 place, as a surgery NP described: “I think the advantage for the patient is that they wind up stopping in for 1 thing but we wind up taking care of a few without requiring the need for him or her to go to all these different specialists on the outside. They’re mostly elderly and not able to get around.” Third, the high availability of services and frequent monitoring allows rapid titration of evidence-based medicines, as discussed by a medicine resident: “It’s easier and faster to titrate medication—they’re in a monitored setting; you can ensure compliance.”

Patients may also differ from their usual state while hospitalized, creating both risks and benefits. The hospital setting can provide an opportunity to educate patients on their chronic disease(s) because they are motivated: “They’re in an office visit and their sugars are out of whack or something, they may take it a little bit more seriously if they were just in the hospital even though it was on an unrelated issue. I think it probably just changes their perspective on their disease.” However, in the hospital, patients are in an unusual environment with a restricted diet and forced medication compliance. Furthermore, the acute condition can lead to changes in their chronic disease, as described by 1 medicine attending: “their sugar is high because they’re acutely ill.” Providers expressed concern that changing medications in this setting may lead to adverse events (AEs) when patients return to their usual environment.

Provider Knowledge and Self-Efficacy

Insufficient knowledge of treatments for chronic conditions was cited as a barrier to some providers’ ability to actively manage chronic disease for hospitalized patients. Some providers described management of conditions outside their area as less satisfying than their primary focus. For example, an orthopedic surgeon explained: “…it’s very simple. You see your bone is broken, you fix it, that’s it…it’s intellectually satisfying…managing chronic diseases is less like that.” Reliance on consultants was 1 approach to deal with knowledge gaps in areas outside a provider’s expertise.

For a number of providers, management of stable chronic disease is the responsibility of the outpatient provider. Providers expressed concern that inpatient management was a reach into the domain of the primary care provider (PCP) and might take “away from the primary focus” of the hospitalization. Nonetheless, some providers noted an “ethical responsibility to manage [a] patient correctly,” and some providers believed that engaging in chronic disease management in the hospital would present an opportunity to expand their own expertise.

A few providers were worried about legal risk related to chronic disease management: “we don’t typically deal too much with managing some of these other medical issues for medical and legal reasons.” Providers again suggested that consults can help overcome this concern for risk, as discussed by 1 surgical attending: “We’re all not wanting to be sued, and we want to do the right thing. It costs me nothing to have a cardiologist on board, so like—why not.”

Hospital Priorities

Providers explained that the hospital has strong interests in early discharge and minimizing LOS. These priorities are based on goals of improving patient outcomes, increasing bed availability and hospital volume, and reducing costs. Providers perceive these hospital priorities as potential barriers to chronic disease management, which can increase LOS and costs through additional testing and treatment. As a medicine resident described: “The DBN philosophy, ‘discharge before noon’ philosophy, which is part of the hospital efficiency to get people in and out of the hospital as quickly as [is] safe, or maybe faster. And I think that there’s a culture where you’re encouraged to only focus on the acute issue and tend to defer everything else.”

Continuity and Communication

According to many providers, care continuity between the outpatient setting and the hospital played a major role in management of chronic disease. One barrier to starting a new evidence-based medication was lack of knowledge of patient history. As noted, providers expressed concern that a patient may not be on a given therapy because of an adverse reaction that was not documented in the hospital chart. This is particularly true because, as discussed by a surgery resident, patients with “PCPs outside the system [in which providers] don’t have access to the electronic medical record.” To overcome this barrier, providers attempt to communicate with the outpatient provider to confirm a lack of contraindications to therapies prior to any changes; notably, communication is easier if the inpatient provider has a relationship with the outpatient PCP.

Some providers were more likely to start chronic disease therapies if the patient had no prior outpatient care, because the provider was reassured that there was no rationale for missing therapies. One neurology attending noted that if a patient had newly documented “hypertension even if they were in for something else, I might start them on an antihypertensive, but then arrange for a close follow-up with a new PCP.”

Following hospitalization, providers wanted assurance that any changes to chronic disease management would be followed up by an outpatient physician. Any changes are relayed to the outpatient provider and the “level of communication…with the outpatient provider who’s gonna inherit” these changes can influence how aggressively the inpatient provider manages chronic diseases. Providers may be reluctant to start therapy for patients if they are concerned about outpatient follow up: “they have diabetes and they should really technically be on an ACE [angiotensin converting enzyme]inhibitor and aspirin, but they’re not. I might send them out on the aspirin but I might either start ACE inhibitor and have them follow up with their PCP in 2 weeks if I’m confident that they’ll do it or if I’m really confident that they’ll not follow up, I will help them get the appointment and then the discharge instruction is to the PCP is ‘Please start this patient on ACE inhibitor if they show up.’”

Providers frequently perceive benefit to chronic disease management in the hospital, including improvements in clinical outcomes. Notably, providers see opportunities to improve compliance with evidence-based care to overcome potential barriers to managing chronic disease in the outpatient setting, which can be limited by pressure for brief encounters,¹³ clinical inertia,¹⁴ difficulty with close monitoring of patients,¹⁵ and care fragmentation.¹⁶ Concurrently, inpatient providers are concerned about potential for patient harm related to chronic disease management, primarily related to AEs from medications. Similar to a case study about a patient with outpatient hypotension following aggressive inpatient hypertension management,⁷ providers fear that changing a patient’s chronic disease management in a hospital setting may cause harm when the patient returns home.

Although some clinicians have argued against aggressive in-hospital chronic disease management because of concerns for risk of AEs,⁷ our study and others⁸ have suggested that many clinicians perceive benefit. In some cases, such as smoking cessation counseling for all current smokers and prescribing an angiotensin converting enzyme inhibitor for patients with systolic heart failure, the perceived importance is so great that chronic disease management has been used as a national quality metric for hospitals. While these hospital metrics may be justified for short-term benefits after hospitalization, studies have demonstrated only weak improvement in short-term postdischarge outcomes related to chronic disease management.¹⁷ The true benefit is likely from improved processes of care in the short term that lead to long-term improvement in outcomes.^4,5,18 Thus, the advantage of starting a patient hospitalized for a stroke on blood pressure medication is the increased likelihood that the patient will continue the medication as an outpatient, which may reduce long-term mortality.

For hospital delivery systems that are concerned with such care process improvement through in-hospital chronic disease management, we identified a number of barriers and facilitators to delivering this care. One significant barrier was poor transitions between the inpatient and the outpatient settings. When a patient transitions into the hospital, providers need to understand prior management choices. Facilitators to help inpatient providers understand prior management included either knowing the outpatient provider, or understanding that there was a lack of regular outpatient care; in both these cases, inpatient providers felt more comfortable managing chronic diseases because they had insight into the outpatient plan, or lack thereof. However, these facilitators may not be practical to incorporate in interventions to improve chronic disease care, which should consider overcoming these communication barriers. Use of shared electronic health records or standardized telephone calls with well-documented care plans obtained through health information exchanges may facilitate an inpatient provider to manage appropriately chronic disease. Similarly, discontinuity between the inpatient provider and the outpatient provider is a barrier that must be overcome to ease concerns that any chronic disease management changes do not result in harm in the postdischarge period. These findings again point to the need for improved documentation and communication between inpatient and outpatient providers. Of course, the transitional care period is one of high risk, and improving communication between providers has been an area of ongoing work.¹⁹

Lack of comfort among inpatient providers with managing chronic diseases is another important barrier, which appears to be largely overcome through the use of consultation services. Ready availability of specialists, common in academic medical centers, can facilitate delivery of chronic disease management. Inpatient interventions designed to improve evidence-based care for a chronic disease may benefit from involvement or at least availability of specialists in the effort. Another major barrier relates to hospital priorities, which in our study were closely aligned with external factors such as payment models. As hospitalizations are typically paid based on the discharge diagnosis, hospitals have incentives to discharge quickly and not order extra diagnostic tests. As a result, there are disincentives for chronic disease management that may require additional testing or monitoring in the hospital. Conversely, as hospitals accept postdischarge financial risks through readmission penalties or postdischarge cost savings, hospitals may perceive that long-term benefits of chronic disease management may outweigh short-term costs.

The study findings should be interpreted in the context of its limitations. Findings of our study of providers from a single academic medical center may not be generalizable. Nearly half of our interviews were conducted by telephone, which limits our ability to capture nonverbal cues in communication. Providers may have had social desirability bias towards positive aspects of chronic disease management. We did not have the power to determine differences in response by provider characteristic because this was an exploratory qualitative study. Future studies with representative sampling, a larger sample size, and measures for constructs such as provider self-efficacy are needed to examine differences by specialty, provider type, and experience level.

In conclusion, inpatient providers believe that hospital chronic disease management has the potential to be beneficial for both process of care and clinical outcomes; providers also express concern about potential adverse consequences of managing chronic disease during acute hospitalizations. To maximize both quality of care and patient safety, overcoming communication barriers between inpatient and outpatient providers is needed. Both a supportive hospital environment and availability of specialty support can facilitate in-hospital chronic disease management. Interventions that incorporate these factors may be well-suited to improve chronic disease care and long-term outcomes.

Disclosures

This work was supported by the Agency for Healthcare Research and Quality (AHRQ) grant K08HS23683. The authors report no financial conflicts of interest.

Millions of individuals with chronic diseases are hospitalized annually in the United States. More than 90% of hospitalized adults have at least 1 chronic disease,¹ and almost half of Medicare beneficiaries in the hospital have 4 or more chronic conditions.² While many patients are admitted for worsening of a single chronic disease, patients are hospitalized more commonly for other causes. For instance, although acute heart failure is among the most frequent causes of hospitalizations among older adults, three-fourths of hospitalizations of patients with heart failure are for reasons other than acute heart failure.³

When a patient with a chronic disease is hospitalized, the inpatient provider must consider whether to actively or passively manage the chronic disease. Studies have suggested that intervening in chronic diseases during hospitalizations can lead to long-term improvement in treatment;^4-6 for instance, stroke patients who were started on antihypertensive therapy at discharge were more likely to have their blood pressure controlled in the next year.⁵ However, some authors have argued that aggressive hypertension management by inpatient providers may result in patient harm.⁷ One case-based survey suggested that hospitalists were mixed in their interest in participating in chronic disease management in the hospital.⁸ This study found that providers were less likely to participate in chronic disease management if it was unrelated to the reason for hospitalization.⁸ However, to our knowledge, no studies have broadly evaluated inpatient provider attitudes, motivating factors, or barriers to participation in chronic disease management.

The purpose of this study was to understand provider attitudes towards chronic disease management for patients who are hospitalized for other causes. We were particularly interested in perceptions of barriers and facilitators to delivery of inpatient chronic disease management. Ultimately, such findings can inform future interventions to improve inpatient care of chronic disease.

In this qualitative study, we conducted in-depth interviews with providers to understand attitudes, barriers, and facilitators towards inpatient management of chronic disease; this study was part of a larger study to implement an electronic health record-based clinical decision-support system intervention to improve quality of care for hospitalized patients with heart failure.

We included providers who care for and can write medication orders for hospitalized adult patients at New York University (NYU) Langone Medical Center, an urban academic medical center. As patients with chronic conditions are commonly hospitalized for many reasons, we sought to interview providers from a range of clinical services without consideration of factors such as frequency of caring for patients with heart failure. We used a purposive sampling framework: we invited participants to ensure a range of services, including medicine, surgery, and neurology, and provider types, including attending physicians, resident physicians, nurse practitioners, and physician assistants. Potential participants, therefore, included all providers for adult hospitalized patients.

We identified potential participants through study team members, referrals from department heads and prior interviewees, and e-mails to department list serves. We did not formally track declinations to being interviewed, although we estimate them as fewer than 20% of providers directly approached. While we focused on inpatient providers at New York University Langone Medical Center, many of the attending physicians and residents spend a portion of their time at the Manhattan Veterans Affairs Hospital and Bellevue Hospital, a safety-net city hospital; providers could have outpatient responsibilities as well.

All participants provided verbal consent to participate. The study was approved by the New York University Institutional Review Board, which granted a waiver of documentation of consent. Participants received a $25 gift card following the interview.

We used a semi-structured interview guide (Appendix) to elicit in-depth accounts of provider attitudes, experiences with, and barriers and facilitators towards chronic disease management in the hospital. The interview began by asking about chronic disease in general and then asked more specific questions about heart failure; we included responses to both groups of questions in the current study. The interview also included questions related to the clinical decision-support system being developed as part of the larger implementation study, although we do not report on these results in the current study. The semi-structured interview guide was informed by the consolidated framework for advancing implementation science (CFIR), which offers an overarching typology for delineating factors that influence guideline implementation;⁹ we also used CFIR constructs in theme development. We conducted in-depth interviews with providers.

A priori, we estimated 25 interviews would be sufficient to include the purposive sample and achieve data saturation,¹⁰ which was reached after 31 interviews. Interviews were held in person or by telephone, at the convenience of the subject. All interviews were transcribed by a professional service. Transcriptions were reviewed against recordings with any mistakes corrected. Prior to each interview, we conducted a brief demographic survey.

Qualitative data were analyzed using a constant comparative analytic technique.¹¹ The investigative team met after reviewing the first 10 interviews and discussed emergent themes from these early transcripts, which led to the initial code list. Two investigators coded the transcripts. Reliability was evaluated by independent coding of a 20% subset of interviews. Differences were reviewed and discussed until consensus was reached. Final intercoder reliability was determined to be greater than 95%.¹² All investigators reviewed and refined the code list during the analysis phase. Codes were clustered into themes based on CFIR constructs.⁹ Analyses were performed using Atlas.ti v. 7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany).

We conducted interviews with 31 providers. Of these, 12 were on the medicine service, 12 were on the surgery or a surgical subspecialty service, and 7 were on other services; 11 were attending physicians, 12 were resident physicians, 5 were NPs, and 3 were PAs. Only 2 providers—an attending in medicine and a resident in surgery—had a specialty focus that was cardiac-related. Median time in current position was 4 years (Table 1). Seventeen of the interviews were in person, and 14 were conducted by telephone. The mean interview time was 20 minutes and ranged from 11 to 41 minutes.

We identified 5 main themes with 29 supporting codes (Table 2) describing provider attitudes towards the management of chronic disease for hospitalized patients. These themes, with related CFIR constructs, were: 1) perceived impact on patient outcomes (CFIR construct: intervention characteristics, relative advantage); 2) hospital structural characteristics (inner setting, structural characteristics); 3) provider knowledge and self-efficacy (characteristic of individual, knowledge and beliefs about the intervention and self-efficacy); 4) hospital priorities (inner setting, implementation climate, relative priority); and 5) continuity and communication (inner setting, networks and communications). For most themes, subjects described both positive and negative aspects of chronic disease management, as well as related facilitators and barriers to delivery of chronic disease care for hospitalized patients. Illustrative quotes for each theme are shown in Table 3.

Perceived Impact on Patient Outcomes

Perceived impact on patient outcomes was mixed. Most providers believed the management of chronic diseases could lead to improvement in important patient outcomes, including decreased length of stay (LOS), prevention of hospital complication, and decreased readmissions. Surgical providers focused particularly on the benefits of preventing surgical complications and noted that they were more likely to manage chronic conditions—primarily through use of specialist consultation—when they perceived a benefit to prevention of surgical outcomes or a fear that surgery may worsen a stable chronic condition:

“Most of the surgery I do is pretty stressful on the body and is very likely to induce acute on chronic exacerbations of heart failure. For someone with Class II or higher heart failure, I’m definitely gonna have cardiology on board or at least internal medicine on board right from the beginning.”

Hospital Structural Characteristics

For many providers, the hospital setting provides a unique opportunity for care of patients with chronic disease. First, a hospitalization is a time for a patient’s management to be reviewed by a new care team. The hospital team reviews the management plan for patients at admission, which is a time to reevaluate whether patients are on evidence-based therapies: “It’s helpful to have a new set of eyes on somebody, like fresh information.” According to providers, this reevaluation can overcome instances of therapeutic inertia by the outpatient physician. Second, the hospital has many resources, including readily available specialist services and diagnostic tests, which can allow a patient-centered approach that coordinates care in 1 place, as a surgery NP described: “I think the advantage for the patient is that they wind up stopping in for 1 thing but we wind up taking care of a few without requiring the need for him or her to go to all these different specialists on the outside. They’re mostly elderly and not able to get around.” Third, the high availability of services and frequent monitoring allows rapid titration of evidence-based medicines, as discussed by a medicine resident: “It’s easier and faster to titrate medication—they’re in a monitored setting; you can ensure compliance.”

Patients may also differ from their usual state while hospitalized, creating both risks and benefits. The hospital setting can provide an opportunity to educate patients on their chronic disease(s) because they are motivated: “They’re in an office visit and their sugars are out of whack or something, they may take it a little bit more seriously if they were just in the hospital even though it was on an unrelated issue. I think it probably just changes their perspective on their disease.” However, in the hospital, patients are in an unusual environment with a restricted diet and forced medication compliance. Furthermore, the acute condition can lead to changes in their chronic disease, as described by 1 medicine attending: “their sugar is high because they’re acutely ill.” Providers expressed concern that changing medications in this setting may lead to adverse events (AEs) when patients return to their usual environment.

Provider Knowledge and Self-Efficacy

Insufficient knowledge of treatments for chronic conditions was cited as a barrier to some providers’ ability to actively manage chronic disease for hospitalized patients. Some providers described management of conditions outside their area as less satisfying than their primary focus. For example, an orthopedic surgeon explained: “…it’s very simple. You see your bone is broken, you fix it, that’s it…it’s intellectually satisfying…managing chronic diseases is less like that.” Reliance on consultants was 1 approach to deal with knowledge gaps in areas outside a provider’s expertise.

For a number of providers, management of stable chronic disease is the responsibility of the outpatient provider. Providers expressed concern that inpatient management was a reach into the domain of the primary care provider (PCP) and might take “away from the primary focus” of the hospitalization. Nonetheless, some providers noted an “ethical responsibility to manage [a] patient correctly,” and some providers believed that engaging in chronic disease management in the hospital would present an opportunity to expand their own expertise.

A few providers were worried about legal risk related to chronic disease management: “we don’t typically deal too much with managing some of these other medical issues for medical and legal reasons.” Providers again suggested that consults can help overcome this concern for risk, as discussed by 1 surgical attending: “We’re all not wanting to be sued, and we want to do the right thing. It costs me nothing to have a cardiologist on board, so like—why not.”

Hospital Priorities

Providers explained that the hospital has strong interests in early discharge and minimizing LOS. These priorities are based on goals of improving patient outcomes, increasing bed availability and hospital volume, and reducing costs. Providers perceive these hospital priorities as potential barriers to chronic disease management, which can increase LOS and costs through additional testing and treatment. As a medicine resident described: “The DBN philosophy, ‘discharge before noon’ philosophy, which is part of the hospital efficiency to get people in and out of the hospital as quickly as [is] safe, or maybe faster. And I think that there’s a culture where you’re encouraged to only focus on the acute issue and tend to defer everything else.”

Continuity and Communication

According to many providers, care continuity between the outpatient setting and the hospital played a major role in management of chronic disease. One barrier to starting a new evidence-based medication was lack of knowledge of patient history. As noted, providers expressed concern that a patient may not be on a given therapy because of an adverse reaction that was not documented in the hospital chart. This is particularly true because, as discussed by a surgery resident, patients with “PCPs outside the system [in which providers] don’t have access to the electronic medical record.” To overcome this barrier, providers attempt to communicate with the outpatient provider to confirm a lack of contraindications to therapies prior to any changes; notably, communication is easier if the inpatient provider has a relationship with the outpatient PCP.

Some providers were more likely to start chronic disease therapies if the patient had no prior outpatient care, because the provider was reassured that there was no rationale for missing therapies. One neurology attending noted that if a patient had newly documented “hypertension even if they were in for something else, I might start them on an antihypertensive, but then arrange for a close follow-up with a new PCP.”

Following hospitalization, providers wanted assurance that any changes to chronic disease management would be followed up by an outpatient physician. Any changes are relayed to the outpatient provider and the “level of communication…with the outpatient provider who’s gonna inherit” these changes can influence how aggressively the inpatient provider manages chronic diseases. Providers may be reluctant to start therapy for patients if they are concerned about outpatient follow up: “they have diabetes and they should really technically be on an ACE [angiotensin converting enzyme]inhibitor and aspirin, but they’re not. I might send them out on the aspirin but I might either start ACE inhibitor and have them follow up with their PCP in 2 weeks if I’m confident that they’ll do it or if I’m really confident that they’ll not follow up, I will help them get the appointment and then the discharge instruction is to the PCP is ‘Please start this patient on ACE inhibitor if they show up.’”

Providers frequently perceive benefit to chronic disease management in the hospital, including improvements in clinical outcomes. Notably, providers see opportunities to improve compliance with evidence-based care to overcome potential barriers to managing chronic disease in the outpatient setting, which can be limited by pressure for brief encounters,¹³ clinical inertia,¹⁴ difficulty with close monitoring of patients,¹⁵ and care fragmentation.¹⁶ Concurrently, inpatient providers are concerned about potential for patient harm related to chronic disease management, primarily related to AEs from medications. Similar to a case study about a patient with outpatient hypotension following aggressive inpatient hypertension management,⁷ providers fear that changing a patient’s chronic disease management in a hospital setting may cause harm when the patient returns home.

Although some clinicians have argued against aggressive in-hospital chronic disease management because of concerns for risk of AEs,⁷ our study and others⁸ have suggested that many clinicians perceive benefit. In some cases, such as smoking cessation counseling for all current smokers and prescribing an angiotensin converting enzyme inhibitor for patients with systolic heart failure, the perceived importance is so great that chronic disease management has been used as a national quality metric for hospitals. While these hospital metrics may be justified for short-term benefits after hospitalization, studies have demonstrated only weak improvement in short-term postdischarge outcomes related to chronic disease management.¹⁷ The true benefit is likely from improved processes of care in the short term that lead to long-term improvement in outcomes.^4,5,18 Thus, the advantage of starting a patient hospitalized for a stroke on blood pressure medication is the increased likelihood that the patient will continue the medication as an outpatient, which may reduce long-term mortality.

For hospital delivery systems that are concerned with such care process improvement through in-hospital chronic disease management, we identified a number of barriers and facilitators to delivering this care. One significant barrier was poor transitions between the inpatient and the outpatient settings. When a patient transitions into the hospital, providers need to understand prior management choices. Facilitators to help inpatient providers understand prior management included either knowing the outpatient provider, or understanding that there was a lack of regular outpatient care; in both these cases, inpatient providers felt more comfortable managing chronic diseases because they had insight into the outpatient plan, or lack thereof. However, these facilitators may not be practical to incorporate in interventions to improve chronic disease care, which should consider overcoming these communication barriers. Use of shared electronic health records or standardized telephone calls with well-documented care plans obtained through health information exchanges may facilitate an inpatient provider to manage appropriately chronic disease. Similarly, discontinuity between the inpatient provider and the outpatient provider is a barrier that must be overcome to ease concerns that any chronic disease management changes do not result in harm in the postdischarge period. These findings again point to the need for improved documentation and communication between inpatient and outpatient providers. Of course, the transitional care period is one of high risk, and improving communication between providers has been an area of ongoing work.¹⁹

Lack of comfort among inpatient providers with managing chronic diseases is another important barrier, which appears to be largely overcome through the use of consultation services. Ready availability of specialists, common in academic medical centers, can facilitate delivery of chronic disease management. Inpatient interventions designed to improve evidence-based care for a chronic disease may benefit from involvement or at least availability of specialists in the effort. Another major barrier relates to hospital priorities, which in our study were closely aligned with external factors such as payment models. As hospitalizations are typically paid based on the discharge diagnosis, hospitals have incentives to discharge quickly and not order extra diagnostic tests. As a result, there are disincentives for chronic disease management that may require additional testing or monitoring in the hospital. Conversely, as hospitals accept postdischarge financial risks through readmission penalties or postdischarge cost savings, hospitals may perceive that long-term benefits of chronic disease management may outweigh short-term costs.

The study findings should be interpreted in the context of its limitations. Findings of our study of providers from a single academic medical center may not be generalizable. Nearly half of our interviews were conducted by telephone, which limits our ability to capture nonverbal cues in communication. Providers may have had social desirability bias towards positive aspects of chronic disease management. We did not have the power to determine differences in response by provider characteristic because this was an exploratory qualitative study. Future studies with representative sampling, a larger sample size, and measures for constructs such as provider self-efficacy are needed to examine differences by specialty, provider type, and experience level.

In conclusion, inpatient providers believe that hospital chronic disease management has the potential to be beneficial for both process of care and clinical outcomes; providers also express concern about potential adverse consequences of managing chronic disease during acute hospitalizations. To maximize both quality of care and patient safety, overcoming communication barriers between inpatient and outpatient providers is needed. Both a supportive hospital environment and availability of specialty support can facilitate in-hospital chronic disease management. Interventions that incorporate these factors may be well-suited to improve chronic disease care and long-term outcomes.

Disclosures

This work was supported by the Agency for Healthcare Research and Quality (AHRQ) grant K08HS23683. The authors report no financial conflicts of interest.