Time trends in inpatient physician spending

Abstract

Objective: Despite falling admissions and declining lengths of stay, Medicare expenditures for inpatient physician services have continued to climb; this article seeks to understand this trend by examining the expenditures on a per admission basis.

Data sources and study setting: One hundred percent Medicare claims data were available from nine states for the 1985-1988 time period.

Study design: Because Medicare's prospective payment system encourages hospitals to shift some services outside the inpatient setting, we examined trends in episodes of care, encompassing some time both before and after the inpatient stay itself. Trends were also examined at the individual DRG level in order to partially control for case-mix shifts and increased surgical use. Allowed charges were purged of both Medicare fee updates and geographic price variation in order to derive estimates of real spending growth.

Data collection/extraction methods: Hospital and physician claims were merged to form inpatient episodes that included seven days prior to admission as well as 30 days following discharge.

Principal findings: Physician spending per episode increased 27 percent just over this four-year time period, but with considerable variation by DRG ranging from only 2 percent for transurethral prostatectomies (TURPs) to 56 percent for uncomplicated acute myocardial infarctions (AMIs). Changes in case severity and hospital and physician characteristics were all found to be important contributors to the increase in physician inpatient spending. Most important seemed to be the growth in the number of physicians associated with the inpatient stay (and the subsequent increase in diagnostic tests and other procedures).

Conclusions: The findings suggest that control of technology and control of the number of physicians involved in the care of a patient are both critical to constraining the rate of increase in physician inpatient expenditures.