The Impact of Rural Hospital Closures on Ambulance Service Times

Since the 1980’s, rural hospitals have struggled to maintain financial viability while providing critically needed services to local communities.(Kaufman et al., 2016; McDermott, Cornia, & Parsons, 1991; Mick & Morlock, 1990; Ricketts & Heaphy, 2000) Despite the increased attention to the plight of rural hospitals, the rate of rural hospital closures continues to increase in recent years. (Kaufman et al., 2016) Closures have been found to have negative income and employment effects in a community (Holmes et al 2006) and may also delay timely access to urgent care for members of the surrounding community. The objective of this analysis is to examine the impact of rural hospital closures on emergency medical services (EMS) times. We hypothesize closures are associated with increased EMS times.

Using the National EMS Information System (NEMSIS, 2010-2016), we identify EMS times, patient demographics, and primary complaint. We use publically available Area Health Resource Files to identify county level characteristics and the Centers for Medicare and Medicaid Services: Provider of Services Current Files to identify the ZIP codes of every acute hospital in the United States. We define a closure as ZIP codes experiencing an increase of ≥2 miles in distance to the nearest ZIP code with a hospital compared to the year prior. The sample includes patients transported by emergent 9-1-1 calls (requiring lights and sirens) in rural or wilderness areas.

Using a retrospective cohort study with matched comparison group by nearest-neighbor matching with replacement (2:1), difference-in-difference and quantile regressions were used to analyze the effect of hospital closures on system response time (unit notification of 9-1-1 call to arrival at incident location), transport time (departure of incident location to destination arrival), and total activation time (time from 9-1-1 call to unit returns to service). Characteristics for propensity matching comparison group included county level indicators predicting hospital closures. We verified consistent time trends between counties experiencing a closure and not experiencing a closure in the periods prior to closure, as required by the difference-in-difference approach.

In bivariate analyses, the change in average minute times in closure counties and their matched controls, respectively, was (a) system response time (+0.77, +0.15), (b) transport time (+1.67, -0.57) and (c) total activation times (+3.7, -1.82). In adjusted analyses, rural hospital closures increased mean EMS transport times by 2.58 minutes (P=0.09) and total activation time by 7.17 minutes (P=0.02), but had no effect on mean system response times. We also found closures had heterogeneous effects across the distribution of EMS times. We found closures had larger effects for those with longer baseline system response and transport times; for example, the increase in the 70th percentile of transport was 7.00 minutes (P<0.05). We also found at the median of total activation times, a closure increased total activation times by 13.82 minutes (P<0.05).

Rural hospital closures are associated with an increase in timely access to EMS resulting in increased mean transport, scene to patient and total activation times with larger effects for those ZIPs with longer baseline response times.