Association for Public Policy Analysis & Management

Hospitals are differentiated by ownership, mission, and economic circumstances. While much attention has been paid to price and quality competition across hospitals, much less research has focused on market interactions around the supply of social goods. These interactions are especially important in money-losing psychiatric care markets where government-owned and subsidized (i.e. safety-net) providers have a large market share. On the one hand, having a safety-net provider in an area may ‘crowd out’ private provision of psychiatric care to low-income individuals. On the other hand, having a strong safety net may allow private providers the space to serve less needy patients and thus increase access to care overall. In this research, I use quasi-experimental methods to look at spillovers of safety-net hospitals’ provision of inpatient psychiatric care on private hospitals’ provision of care.

Using administrative data on psychiatric discharges and hospital finances in California from 2002-2014, I relate the availability of inpatient psychiatric beds and services at private hospitals to nearby closures of psychiatric units at safety-net hospitals. I examine how psychiatric unit closures at safety-net hospitals within five, ten, and fifteen miles of a given provider affect that provider’s provision of inpatient psychiatric care. I study outcomes related to overall capacity in psychiatric units, the composition of patients in emergency departments, psychiatric units, and other inpatient units, and hospital and psychiatric unit profitability. To isolate changes due to plausibly exogenous financial events, I identify the stated reason for closure in county reports and local newspaper stories. Over my study period, 51 hospitals in CA closed their psychiatric units due to financial losses (closures account for ~30% of safety-net hospitals offering inpatient psych), and 26 hospitals opened psychiatric units.

I find that when safety-net hospitals close their psych units, nearby hospitals with psychiatric units absorb just over a third of the closing hospital’s capacity. There is clear selection in the types of patients neighboring hospitals absorb. While neighboring safety-net providers primarily absorb low-income and severely ill patients, private hospitals absorb higher paying and less difficult to treat patients on average. Evidence suggests private hospitals positively select patients by altering their admissions criteria and the specific psychiatric services they offer. Forthcoming analyses will test for additional underlying mechanisms and assess the implications for patient care and outcomes. The degree to which different classes of providers serve as strategic complements or substitutes in markets for low-income care has important implications for optimal government payment and subsidy design.