Association for Public Policy Analysis & Management

Medicaid reimbursement for health care services is often lower than reimbursement from private insurers, which can incentivize hospitals to select against treating Medicaid beneficiaries. Selection by providers, combined with patient preferences based on travel distance, trust, and other factors, may result in the concentration of care for Medicaid beneficiaries within specific hospitals. The implications of this concentration for patient outcomes are unclear: concentration of care may limit access to certain advanced forms of specialty care, but it may also direct Medicaid beneficiaries to organizations that have the most experience treating vulnerable populations. In this study, I used data on the approximate universe of inpatient hospital discharges for 16 states to first describe geographic and temporal variation in the segregation of hospital care for Medicaid relative to privately insured beneficiaries. I then used longitudinal discharge data for five states to estimate the causal effect of de-concentrating care on patient outcomes. To estimate this causal effect, I leveraged the closure of safety-net hospitals as a natural experiment: some patients previously treated at a safety-net hospital will obtain care from similar safety-net institutions following the closure, whereas others will obtain care from non-safety-net institutions. I conducted a difference-in-differences analysis, comparing outcomes among patients treated at the same safety-net hospital who, following closure of that hospital, received care from an institution with a higher Medicaid (safety-net) focus, relative to those treated at a hospital with a lower Medicaid focus. I limited the sample to patients receiving emergency or urgent care, for whom the location of hospital treatment following closure of their original safety-net hospital was likely to be plausibly exogenous.

Based on preliminary analyses, I found that about one-third of Medicaid discharges would have to occur in a different hospital for the distribution of discharges to be uniform. Segregated markets — those with a higher concentration in care for Medicaid beneficiaries — had a higher prevalence of both for-profit and teaching hospitals. Difference-in-differences estimates involving safety-net hospital closures indicated that Medicaid beneficiaries admitted to hospitals with a greater focus on the Medicaid population witnessed significantly lower spending per hospital episode and a lower prevalence of adverse medical events.

These findings suggest that safety-net hospitals may develop some returns to specialization in the Medicaid population, which allows them to deliver high-quality care at a lower cost. Some heterogeneity in the results across different types of medical conditions implies that a hospital’s specialization in care for a specific population may not compensate for a lack of clinical specialization in care for certain types of conditions. State-level network adequacy standards — and enforcement of those standards — must deftly balance the efficiency gains that may arise from narrowly curated hospital choice sets with the ability for patients to seek out (with appropriate guidance) the best-matched hospital location for treatment.