Health Affairs Study: Wide Variation in Emergency MD Decision-Making Influences Admissions
Do variations in inpatient admissions through emergency department care result at least in part from the divergent clinical decision-making processes of the emergency physicians evaluating and caring for patients in EDs? The answer seems to be yes. A team of healthcare policy analysts has performed a study on the subject, which appears in the February issue of Health Affairs.
In “Variation In Emergency Department Admission Rates Among Medicare Patients: Does The Physician Matter?” Peter B. Smulowitz, M.D., A. James O’Malley, Ph.D., Lawrence Zaborski, J. Michael McWilliams, M.D., and Bruce E. Landon, M.D. analyzed Medicare fee-for-service claims for a 20 percent random sample of beneficiaries from January 1, 2012, through September 30, 2015, to identify all ED visits made by traditional Medicare beneficiaries.
Smulowitz is an assistant professor of emergency medicine at Beth Israel Deaconess Medical Center (Boston); O’Malley is a professor of biomedical data science at Dartmouth University (Hanover, N.H.); Zaborski is a senior statistical programmer in the Department of Health Care Policy at Harvard Medical School (Boston); McWilliams is a professor of healthcare policy at Harvard Medical School an a general internist at Brigham and Women’s Hospital (Boston); Landon is a professor of medicine and practicing internist at Beth Israel.
The authors write that “Hospitalizations account for the largest share of health care spending. New payment models increasingly encourage health care providers to reduce hospital admissions. Although emergency department (ED) physicians play a major role in the decision to admit a patient, the extent to which admission rates vary among ED physicians even within the same hospital remains poorly understood. In this study we examined physician-level variation in ED admission rates for Medicare patients. We found meaningful variation in admission rates: The mean physician-level adjusted admission rate was 38.9 percent and ranged from 32.2 percent to 45.6 percent for physicians at the tenth and ninetieth percentiles, respectively, of the estimated distribution within the same hospital. In contrast, the predicted risk for admission based on patient characteristics varied little among these physicians, suggesting that the variation in admission rates was not due to differences in patients seen. Our results suggest that strategies targeting physician decision making could modify (by either increasing or decreasing when appropriate) rates of admissions.”
The authors state that, “Across the large clinical groupings, we found moderate-to-high correlation in admission rates at the physician level across clinical conditions. The magnitude of these correlations ranged from 0.59 to 0.96 (for example, the correlation between admission tendency for gastrointestinal and pulmonary conditions was 0.81). That is, physicians generally had consistently higher or lower tendencies to admit (relative to other physicians in the same hospital) across conditions.”
They further note that “A particular strength of our study is that we were able to demonstrate that patients are indeed naturally randomly assigned to ED physicians. Unlike many studies outside of emergency medicine, our study can take advantage of the fact that patients largely do not select which physician they see in an ED. Our results differ from those of a recent study that found evidence of nonrandom sorting of patients to ED physicians, but that study was limited to a single academic ED and did not account statistically for the contribution of sampling error to physician-level differences in patient characteristics.18 Our study suggests that nonrandom selection of patients by physicians accounts for only a small proportion of patient allocations in EDs nationally.”
All of this provides quite fertile ground for policy-based intervention. “That there is significant variation among physicians even within the same hospital suggests that there is an opportunity to devise interventions targeted at physician decision making and that such interventions might be effective in hospitals, with some physicians having low rates and other physicians having high rates of admission,” the researchers write. “Such interventions, such as clinical pathways for specific conditions or feedback of physician admission metrics, would seek to better support ED physicians’ decision making by supplying them with more information about patients’ need for admission. This information would assist physicians in employing alternative strategies to achieve the same clinical outcomes without admission or in choosing an admission when it is the appropriate disposition. This balance again highlights that for many conditions the ‘right’ level of admissions is unknown.”
And, they conclude that “Reducing unnecessary hospital admissions from the ED and their associated costs first requires an understanding of the factors driving these admissions. Efforts to ensure that patients who could benefit from hospitalization are admitted also require a similar understanding. The wide variation in ED physicians’ admission rates seen in our study suggests that physician decision making contributes considerably to whether a patient in the ED is admitted and might therefore be a fruitful target for interventions.”
