Association for Public Policy Analysis & Management

For-profit hospital presence has been increasing steadily over the past two decades, which may have a negative effect on patient outcomes and health care costs if profit motives influence for-profit hospitals to provide higher-cost and/or lower quality care than nonprofit and government hospitals. Anecdotal evidence suggests this may be true. Investigations by the Justice Department allege that some for-profit hospitals pressure physicians into admitting patients when it isn't medically necessary to increase profits. This paper investigates whether the empirical evidence is in line with the anecdotal evidence by investigating the relationship between hospital ownership and inpatient admission through the emergency department (ED). 

This paper uses data from four sources: The State Emergency Department Databases (SEDD), the State Inpatient Databases (SID), the American Hospital Association (AHA) Annual Survey of Hospitals, and the Bureau of Health Professions' Area Resource File (ARF). The SEDD and SID provide the universe of hospital discharge records for any patient who visited an ED in 8 states between 2005 and 2013. Hospital identifiers allow these data to be merged with the AHA survey and the ARF to give rich hospital and county characteristics.

The data are used to estimate linear probability models estimating the effect of hospital ownership on the probability of being admitted to the hospital. The model is estimated on a subset of patients who visited the ED for an injury. This is done to control for selection of patients into hospitals, as injuries are likely to be unanticipated, urgent, and non-preventable. Additionally, the severity of injury can be controlled for using the Injury Severity Score. The score takes into account the number, severity, and body region of a patient's injuries and was developed in the clinical literature to correlate linearly with mortality.

Results from the linear probability models indicate that conditional on the injury severity score and patient characteristics, patients visiting for-profit EDs are 1.2 percentage points more likely to be admitted to the hospital than patients visiting EDs in nonprofit hospitals. These results suggest that for-profit hospitals admit patients with less severe injuries.

The results from the linear probability models may be subject to selection bias and thus are not strong estimates of the causal relationship between hospital ownership and admission through the ED. To provide evidence of the causal effect, future analyses will take advantage of the variation in timing of hospitals converting to for-profit status to estimate difference-in-differences models. Additionally, differential effects by patient insurance status will be estimated to determine if for-profit hospitals are more likely to admit patients with private insurance than public insurance.