Ashish Jha, Arnold M. Epstein
A. K. Jha and A. M. Epstein, Hospital Performance Improvement: Are Things Getting Better?, The Commonwealth Fund, April 2007
Commentary on the Fund reports Hospital Performance Improvement: Trends in Quality and Efficiency and Hospital Quality Improvement: Strategies and Lessons from U.S. Hospitals
By Ashish K. Jha, M.D., M.P.H., and Arnold M. Epstein, M.D., M.A., both of the Harvard School of Public Health. Watch a Commonwealth Fund E-forum, with synched slides and audio from presentations by Dr. Jha and others.
Interest in improving health care quality has increased palpably in recent years. We have seen this interest play out in numerous ways, including greater emphasis on collecting and publicly disseminating performance data. A consortium of organizations, including the Center for Medicare and Medicaid Services (CMS), the American Hospital Association, and the Joint Commission for the Accreditation of Healthcare Organizations, has initiated a national quality monitoring system, the Hospital Quality Alliance (HQA).
Congress and federal agencies have spurred change as well. In the Medicare Modernization Act of 2003, Congress established a small financial incentive for hospitals to report data on 10 HQA quality indicators for congestive heart failure, pneumonia, and acute myocardial infarction. In the Deficit Reduction Act of 2005, Congress increased the financial incentive for hospitals to report required quality metrics and authorized the Secretary of Health and Human Services to modify and expand the list of required quality indicators. And, later this year, CMS will phase in requirements for additional measures, such as patients' experiences of care.
In the private sector, more than half of the country's HMOs—together providing care for more than 80 percent of managed care plan enrollees—have established "pay-for-performance" (P4P) programs for physicians and/or hospitals. CMS has established a large, three-year demonstration program to evaluate the impact of P4P on the quality of hospital care, and Congress has mandated that CMS develop plans to introduce a pay-for-performance program in Medicare.
Policymakers have also expressed substantial interest in finding other ways to improve care, including promoting health information technology.
Despite these efforts, it is still unclear whether the attention paid to quality measurement and public reporting has begun to improve the care Americans receive and, if so, to what extent it has done so. The two reports by Kroch, Silow-Carroll, and colleagues provide complementary snapshots of the health care system, illustrating the dynamics of quality improvement both at the macro level and "on the ground" at individual hospitals.
Evidence Care Is Improving Overall
In Hospital Performance Improvements: Trends in Quality and Efficiency, Kroch et al., examine changes in the quality and costs of care in U.S. hospitals between 2001 and 2005. Using three different sets of data, the report investigates changes in mortality (as a surrogate for quality) and lengths of stay (as a surrogate for costs). The investigators find that both unadjusted as well as adjusted inpatient mortality rates improved substantially (with relative risk reduced by about 4 to 8 percent annually). They also found somewhat smaller (approximately 2 percent) reductions in lengths of stay in each of the databases.
All in all, these findings seem like good news for the health care system, or at least for hospital care. However, several findings and key limitations of the data should give us pause. In particular, critics of administrative data may worry that lumping together multiple conditions could obscure finer-grained patterns of care. In fact, the investigators did find that the number of secondary conditions increased nearly 20 percent from 2003 to 2005 in one of the databases they examined. This change could be due in part to hospitals improving their coding of co-morbid conditions, which would have reduced risk-adjusted mortality rates even without changes in care. However, the investigators also found that raw mortality rates declined over the same time period, suggesting that coding alone cannot account for the observed improvements.
Another concern is that the reduced mortality rate might reflect not improved quality of care, but rather an increasing focus on palliative care for the terminally ill that prompts some hospitals to discharge patients home or to other facilities instead of letting them die in the hospital. However, there is no evidence that this practice is widespread enough to account for the rather impressive changes in mortality rates. Therefore, it seems reasonable to accept the report's findings of better care and lower costs. Yet, the evidence is not adequate to reveal how much quality has improved or costs have been reduced.
The changes in length of stay are particularly remarkable since they extend declines going back more than two decades. Improvements in outpatient care have made it easier and even prudent to discharge patients earlier than was possible a decade ago. While the optimal length of stay for most conditions is unknown, the fact that the mean length of stay seems to be approximately three days for all patients and just four for the elderly makes one wonder when these trends will plateau, and whether further declines in length of stay would harm patients.
Triggers for Change
In the second report, Hospital Performance Improvements: Strategies and Lessons from U.S. Hospitals, Silow-Carroll et al. shed light on policies to improve care and specific triggers for change. The four case studies reviewed a diverse group of hospitals that seemed to improve care and control costs between the 2002-to-2004 study period. While the process of change in various settings had different features, common themes emerged: shared goals among clinicians and administrators, clinical champions who advocate improvement, and the establishment of quality improvement as a high priority. The work by Silow-Carroll and colleagues bolsters prior research and reminds us that, while improvement is slow and often challenging, celebrating short-terms gains and having patience for broader, larger achievement is critical to the success of quality improvement efforts.
The other major finding from the case studies is that there is no one prescription for improved performance, either in terms of motivational "triggers" or specific scenarios. For some hospitals, the impetus to improve might be bad publicity from high-profile errors. For others, it might be appointing a leader dedicated to change. The paths to high performance vary widely as well. One hospital created a board-level commission on quality of care while others focused on hiring hospitalists to assiduously provide high-quality care. Yet, for every institution studied, making quality a high priority and dedicating resources to improving care seemed to be critical.
New Insights for Researchers and Policymakers
The reports by Kroch, Silow-Carroll, and their teams at CareScience, Inc. and Health Management Associates provide insights into a key question facing policymakers: given the considerable efforts to measure quality, make data publicly available, and catalyze quality improvement, are things actually beginning to get better? The quantitative study provides hints (and these complement findings from other studies) that, indeed, the answer might be yes. Given the importance of this question and the limitations of the data, of course, further work will be essential to ensure that the differences in mortality are not due primarily to changes in coding of administrative data. Additional work will also be needed to elucidate ways in which patterns of care are changing for different clinical conditions. Still, at this point the first study provides enough credible data to suggest that real, broad-based improvements in hospital care are likely taking place.
The second study offers a different lesson for policymakers: While it is undoubtedly valuable to continue focusing on quality measurement and rewarding excellence, being very proscriptive about how hospitals improve (e.g., through adoption of health information technology) is likely unnecessary and perhaps even misguided. That is not to say that technical assistance or information about successful strategies is not valuable. However, given that right mix of financial and non-financial incentives, it seems clear that health care organizations will improve in ways that draw on their own local culture and traditions.
The views presented in this commentary are those of the authors and should not be attributed to The Commonwealth Fund or its directors, officers, or staff.