Editorials18 December 2012Hospital Readmission Rates: Are We Measuring the Right Thing?Eugene Z. Oddone, MD, MHSc and Morris Weinberger, PhDEugene Z. Oddone, MD, MHScFrom Center for Health Services Research in Primary Care, Durham Veterans Affairs Medical Center and Duke University School of Medicine, Durham, NC 27705; and Center for Health Services Research in Primary Care, Durham Veterans Affairs Medical Center and Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599.Search for more papers by this author and Morris Weinberger, PhDFrom Center for Health Services Research in Primary Care, Durham Veterans Affairs Medical Center and Duke University School of Medicine, Durham, NC 27705; and Center for Health Services Research in Primary Care, Durham Veterans Affairs Medical Center and Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599.Search for more papers by this authorAuthor, Article, and Disclosure Informationhttps://doi.org/10.7326/0003-4819-157-12-201212180-00013 SectionsAboutVisual AbstractPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinkedInRedditEmail In this issue, Kaboli and colleagues (1) examined the relationship between hospital length of stay (LOS), hospital readmission rates, and mortality for more than 4 million veterans hospitalized in 1 of the 129 Veterans Affairs (VA) acute care facilities. Over a 14-year period, they showed that hospital LOS decreased approximately 27%, 30-day hospital readmissions rates decreased by 25%, and 90-day mortality rates actually improved. This study builds on their earlier work, which showed that an average hospital LOS was longer in VA facilities than in non-VA facilities (2). In the past decade, it seems that the VA has closed the LOS gap with non-VA facilities without sacrificing higher readmission rates or worsening outcomes for patients. Drilling down further, Kaboli and colleagues examined these same rates within separate diagnostic categories, some of which will be targeted by Medicare for penalty should patients hospitalized for the condition be readmitted within 30 days: heart failure, chronic obstructive pulmonary disease, community-acquired pneumonia, acute myocardial infarction, and gastrointestinal hemorrhage. Similar trends were found in each subset of patients as those that were reported in the total sample. However, less than 20% of the sample of older, chronically ill veterans fell into one of these groups, meaning that taking a disease-specific approach in an attempt to further reduce hospital readmission rates is not likely to lower the overall rate to a large extent. In this study, no single diagnostic category accounted for more than 5% of hospitalizations.Hospital readmission within 30 days of discharge has long been used as a quality metric. After all, it is relatively easy and inexpensive to measure, does not require subjective judgments, and may reflect poor quality of care during, or immediately after, discharge. After decades of research, what have we learned about hospital readmission as a quality indicator? First, hospital readmission rates may be a poor measure of quality of care because of the complexity of factors that cause them and the poor correlation among those factors. It's not only the quality of care during the index hospitalization or the quality of the handoff to postdischarge care that influences readmission rates. Rather, many important factors affect when and how often patients are hospitalized, including access to postdischarge care, ability to purchase evidence-based medications or services prescribed at discharge, disease and disease severity, socioeconomic status, community resources, and social support (3). So, if we focus intensely on a few of these factors, will readmission rates decrease? Not always. Sixteen years ago, we showed that improving the handoff from hospital to primary care and enhancing veterans' access to primary care systematically increased 30-day hospital readmission rates by 26%, but patients were much more satisfied with their care (4). Despite that finding, the VA continued an impressive health care system redesign, including greatly expanding access to primary care for its veterans and using some of the very elements from the readmission study cited, which resulted in tremendous improvements in quality and outcomes of care surpassing all other U.S. health care systems (5). If administrators at the VA had focused only on hospital readmission rates, they would have missed this mark, and the system would not be what it is today.Second, overall hospital readmission is, at best, a crude indicator of quality. There are clearly times when it is in patients' best interests to receive care in a hospital. It is preventable hospitalizations that should be our focus, but operationally defining preventable hospital readmissions is fiendishly difficult. To date, no discharge diagnosis–based strategy can effectively uncover preventable hospitalizations with sufficient reliability for anything except research purposes. The explicit review of the hospital medical record, although better than discharge diagnoses, is painfully slow and not feasible on a large scale (6). Finally, focusing on hospital readmission is a reactive approach to improving quality. Even when hospitals learn about readmissions in a timely fashion, this knowledge comes after the fact. Thus, hospitals can only speculate on what they may have done differently, but they will often do so with incomplete knowledge (for example, details about postdischarge care or prescription fill rates). This may partially explain why hospital-based continuous quality improvement strategies aimed at reducing readmissions are often not effective (7).Does that mean we should give up and erase hospital readmission rates from all score cards? Probably not. Certainly, in a patient-centered approach, patients generally prefer to spend time at home rather than the hospital, when it can be done safely. Just because we cannot effectively distinguish the preventability of hospital readmission today does not mean that we should stop trying. Better defining the unique factors that explain hospital readmission, and chipping away at them, must be a goal of effective health care systems. However, we must understand that hospital readmissions are not solely, or even minimally, the fault of the hospitals. Therefore, punishing hospitals alone will not necessarily lower the rates, and it may lead to potentially perverse practices (for example, funneling sick patients to short-stay units when higher levels of care are needed). As the United States moves toward accountable care organizations as the appropriate structure of care, where someone receives care or who is at fault for the readmission may become less important. The article by Kaboli and colleagues (1) shows that, over a 14-year period, reducing LOS in VA hospitals decreased mortality and hospital readmissions. The authors acknowledge that a limitation of the study is that it was conducted in a single health care system (VA). That statement is included in all articles from VA investigators. Perhaps it is time to embrace the VA as the largest U.S. accountable care organization. Let the VA serve as an example of how to enhance both efficiency (reduced LOS) and quality (reduced readmission rates and mortality). Understanding the effect of the transformation of the VA health care system provides valuable lessons as hospitals and provider organizations partner to share responsibility for the health of defined populations.Eugene Z. Oddone, MD, MHScCenter for Health Services Research in Primary Care, Durham Veterans Affairs Medical Center and Duke University School of MedicineDurham, NC 27705Morris Weinberger, PhDCenter for Health Services Research in Primary Care, Durham Veterans Affairs Medical Center and Gillings School of Global Public Health, University of North Carolina at Chapel HillChapel Hill, NC 27599