Association of hospital closures with changes in Medicare?covered ambulance trips among rural emergency medical services agencies

The emergency medical services (EMS) system plays a critical role in the U.S. health care safety net, performing over 16 million transports per year.1 Although often treated as publicly funded good, only two-thirds of EMS agencies are run by governments or municipal fire departments2 and many rely on volunteers. As result, recruitment retention personnel has become more difficult time, especially rural areas3 where report increasing stress decreasing satisfaction.4 Closures 120 hospitals past decade5 could worsen strain providers. Existing research shows that closures negatively impact patients through increased time ambulances, but may too. If increase trip distance, with fixed capacity have less to prep vehicles be unable respond some calls, particular nonemergent ones. Shifting away from trips negative financial implications for agencies, such provide dependable revenue stream. We estimated association hospital agency miles, trips, emergent status, stratifying whether were privately supported. linked annual, EMS-provider-level data ambulance covered fee-for-service Medicare Physician Other Supplier Public Use File (PUF) between 2012 2018 University North Carolina Sheps Center (SC). PUF contain annual count service providers year identified Healthcare Common Procedural Codes. used these create our primary dependent variables: EMS-agency-level measures total ground miles average trip, share “nonemergent” based billing codes Basic Life Support, Advanced Support 1, specialty transport (see Daa Supplement S1, Appendix S1 [available supporting information online version this paper, which is available at http://onlinelibrary.wiley.com/doi/10.1111/acem.14273/full]. For each we also extracted agency's zip code name. Housing Urban Development zip-to-county crosswalk assign counties. applied text-based algorithm names probable ownership type S2 details validation algorithm). SC independent variable: time-varying, county-level indicator general acute closures, set one well all subsequent years. counties multiple earliest date closure excluded sample located experienced 2007 2011, 5 years before began. County-level controls included poverty rates, uninsured population, demographic characteristics, morality rates injuries substance abuse (controls described S2). After being limited counties, contained 104 53 affected closure, 5,338 1,117 not (N = 32,644; S3 describes sample.) Our difference-in-differences analysis compared outcomes closed change among did close. multivariable, ordinary least squares regression binary effects year-level effects. Total logged because skewed nature data. Nonemergent expressed trips. All analyses performed STATA 16, statistical significance was judged 5% level. Sensitivity defining areas larger than county redefining geographic county; clustering standard errors level; reestimating model using binomial account nonlinear outcomes6 including state-by-year differential Medicaid expansion across states time; winsorizing minimize influence outliers; regressions event studies, separately assess results due specific trending differently occurred. (Sensitivity provided S4.) Figure presents unadjusted means 95% confidence intervals (CIs) main any point period those without closure. 800 while number traveled 400 same period. decreased 34 20 Average 3 fell percentage points (PP) remained affected. Table 1 difference-in-difference analysis. closes within its county, 16% (95% CI 8% 24%) relative Results similar (15%, 23%). do –10% 3%). unaffected 19% 15% (either non-emergency interfacility transfers) PP –7 –3 PP), controls. Relative (16%) estimate represents 31% (5 PP/16%) decrease Among assigned public support algorithm, 22% 18% 27%) versus agencies. 10% supported 3% 17%). control –4 –2 PP) no PP). consistent sensitivity except alternate errors. suggest result longer surprising given distance next ranges approximately 17 minutes.7 Increased both economic operational depreciation vehicles, limit stocking cleaning, technicians paramedic staffing costs who must remain duty periods. Longer delays time-sensitive care. Failing timely acutely ill “moral injury” workers, experience psychological distress perceived patient suffering.8 shift toward predictable sources revenue. Given an important facilitating access other transportation options limited, unavailable altogether, previously unexplored aspect patients. Finally, operated been Privately owned constitute 20% nationally2 17% 24% study sample. In light provider burnout, reimbursement cuts agencies9 expected decline state local budgets coronavirus pandemic,10 policy intervention warranted specifically had several limitations. First, relied claims data, capture use bill nor serve Advantage beneficiaries. Second, private novel, produce measurement error, biasing zero. Third, characteristics capabilities, mix, vehicles. Fourth, focused exclusively include air EMS. close, agencies—especially ones—must travel farther higher authors thank Hannah Geressu, Ruolin Lu, India Pungarcher assistance. Please note: publisher responsible content functionality supplied authors. Any queries (other missing content) should directed corresponding author article.