Physicians' Response to Length-of-Stay Profiling

• Published in: Medical care Vol. 33; no. 11; pp. 1106 - 1119
• Main Authors: Evans, John H., Hwang, Yuhchang, Nagarajan, Nandu
• Format: Journal Article
• Language: English
• Published: United States J. B. Lippincott-Raven Publishers 01-11-1995; Lippincott-Raven Publishers
• Subjects: Arithmetic mean; Cost efficiency; Diagnosis related groups; Economic impact analysis; Health care costs; Health outcomes; Health Resources - statistics & numerical data; Hospital costs; Humans; Length of Stay; Physician Incentive Plans; Physicians; Severity of Illness Index; Statistical significance; United States; Utilization Review
• DOI: 10.1097/00005650-199511000-00004
• EISSN: 1537-1948
• ISSN: 0025-7079

One of the techniques adopted recently by certain hospitals to meet the competitive pressure for reducing costs is physician profiling. Profiling produces periodic reports that compare a physician's resource consumption to a benchmark figure. This study analyzes the effectiveness and implications of one hospital's introduction of physician patient length-of-stay profiling. Data for 24,000 patients treated by 400 physicians in 450 diagnosis-related groups over 42 months were analyzed, including both preprofiling and post-profiling periods. Statistical tests examined whether more physicians achieved the length-of-stay benchmark after profiling was introduced, controlling for physician, disease category (diagnosis-related group), and patient severity level. First, the results establish a significant increase in the percentage of physicians who achieve the length-of-stay benchmark after the introduction of profiling. Second, it was found that physicians who had initially failed to meet the benchmark reduced their patients' average length of stay much more than those physicians who initially achieved the benchmark. Further, reductions occurred primarily at intermediate severity levels, and in diagnosis-related groups with a large economic impact for the hospital. Although the profiling program did achieve the objective of reducing patient length of stay, further improvement may be possible. First, providing different benchmarks or targets for different physicians may extend the improvement to a greater percentage of all physicians involved. Second, an analysis of monthly data on total weighted procedures reveals that the reduction in length of stay resulted in an increase in the number of procedures performed per patient day. This finding suggests that to achieve a reduction in hospital costs and charges, profiling programs should be combined with process improvement initiatives.