Health Policy and Health Services Research
Background: Optimizing regional systems of emergency stroke care requires understanding geospatial patterns in access to comprehensive stroke centers (CSCs). In Los Angeles County (LAC), emergency medical services (EMS) skip closer hospitals and transport patients with suspected large vessel stroke directly to a CSC only if the expected transport time is ≤ 30 minutes. In addition to areas which always or never have access to a CSC within 30 minutes, we hypothesized that this policy would result in areas which sometimes had access, depending on traffic.
Methods: We performed a prospective geospatial analysis of drive times to CSCs during various traffic conditions at the US census block group level. For each census block group, we identified the closest CSC in LAC and estimated drive time to it using the Google Maps Distance Matrix API. Drive times were sampled 12 times for each block group across morning and afternoon traffic peaks, between the peaks, and in the late evening. Each block group was then defined as sometimes, always, or never having access to a CSC depending on drive times ≤30 minutes. Comparisons between access groups were performed using ANOVA and multinomial logistic regression.<
Results: Among the 6,415 census block groups in LAC, the median straight-line distance to a CSC was 6.8Km (IQR 3.9-10.0Km) with an associated median driving time to this closest hospital across all traffic conditions of 14.6 minutes (IQR 9.8-19.5 minutes). 5,213 census block groups (79.6%) always had access to a CSC within 30 minutes of driving time, 225 (4.9%) never had access, and 977 (15.5%) sometimes had access. Geospatial analysis showed that the never group was mostly in rural LAC. The sometimes group occurred not only at the borders of the always and never groups, but also in large areas in south-central and eastern Los Angeles. Regression analysis showed the sometimes group as distinct in key demographic characteristics including insurance status, income, race, and age.