Do hospitals cross-subsidize?

Abstract

Despite its salience as a regulatory tool to ensure the delivery of unprofitable medical services, cross-subsidization of services within hospital systems has been notoriously difficult to detect and quantify. We use repeated shocks to a profitable service in the market for hospital-based medical care to test for cross-subsidization of unprofitable services. Using patient-level data from general short-term hospitals in Arizona and Colorado before and after entry by cardiac specialty hospitals, we study how incumbent hospitals adjusted their provision of three uncontested services that are widely considered to be unprofitable. We estimate that the hospitals most exposed to entry reduced their provision of psychiatric, substance-abuse, and trauma care services at a rate of about one uncontested-service admission for every four cardiac admissions they stood to lose. Although entry by single-specialty hospitals may adversely affect the provision of unprofitable uncontested services, these findings warrant further evaluation of service-line cross-subsidization as a means to finance them.

Keywords: Cross-subsidies; Hospital markets; Specialty hospitals.

Fig. 1

Median net revenue and margin, by exposure level. Notes: Median-band plot using Stata 12.0. Sample limited to hospitals that report in every year.

Fig. 2

Predicted change in uncontested admissions due to exposure. Notes: Predictions of conditional logit model of patient’s choice of hospital with hospital*drive time fixed effects; drive time interacted with the patient characteristics listed in Table 1 and a dummy variable indicating a partial year report of one hospital. Cubic specification of exposure in psychiatry, trauma, and neurosurgery samples and a quadratic specification is used for substance abuse. Parameter estimates in Table A3.