2005 HSR&D National Meeting Abstract
3011 — What Effect Does Physician "Profiling" Have on Inpatient Physician Satisfaction and Length of Stay?
Zemencuk JK (VA Center for Practice Management and Outcomes Research)
Hofer TP (VA Center for Practice Management and Outcomes Research)
Hayward RA (VA Center for Practice Management and Outcomes Research)
Moseley RH (Medical Service, Ann Arbor VA Medical Center and Department of Internal Medicine, The University of Michigan Medical School)
Saint S (VA Center for Practice Management and Outcomes Research)
Physician profiling, comparing physicians’ practice patterns with that of their peers, is a proposed method for reducing length of stay (LOS). Many physicians have a negative view of profiling. We examined whether profiling affects physician satisfaction and if it is associated with a change in LOS.
LOS data was collected at an intervention and six control hospitals in the same VA network over 4 years (2 years before and 1 year after a profiling year). During the profiling year, physicians at the intervention hospital were told their patient LOS would be compared to their peers during their ward month. During the pre-profiling and profiling years, physicians at the intervention hospital were surveyed, assessing their ward month experience. We tested for differences in physician satisfaction and change in LOS, occurring beyond ecologic trends, adjusting for patient demographics, diagnoses, and comorbidities.
74 physicians completed questionnaires (response rate=90%). 34% of non-profiled physicians agreed being profiled would cause most physicians to discharge patients early, while just 8% of profiled physicians indicated the same (p<.05) and 86% disagreed that being profiled caused them to discharge patients earlier than they would have normally. More profiled physicians than non-profiled physicians felt indifferent toward profiling (56% vs. 26%, p<.05); however, 46% of profiled physicians reported more involvement than usual in patient care. During the profiling year, the profiled site experienced a significant decrease in LOS (-0.32 days; CI = -0.49 to -0.16; p<.001), relative to the non-profiled sites.
While a substantial number of profiled physicians reported more involvement than usual in patient care, they were less likely than those not profiled to believe profiling would cause physicians to modify their patient care practice and more likely to report indifference toward profiling. Although survey responses suggested profiling would not cause physicians to discharge patients early, taking into account background trend and relative to non-profiled sites, LOS at the profiled site decreased by 1/3 day in the profiling year.
Hospitals have substantial financial incentives for implementing cost-effective measures to reduce length of stay. It appears that physician profiling may be an effective method of significantly reducing LOS without adversely affecting physician satisfaction.