Geospatial Analysis to Determine Optimal Distribution of Mobile Stroke Units

BACKGROUND: Mobile stroke units (MSUs) provide faster stroke treatment with improved outcomes, but are expensive, and their urban and rural deployment differs. Geospatial analysis may be useful for planning optimal MSU distribution. METHODS: We geo-coded Texas state-designated level I or II stroke centers that did not overlap catchment areas and mapped 30-, 60-, 120-, and 180-minute drive time buffers around each center, superimposing them on the distribution of patients with stroke in the state, including estimates of rural, vulnerable, and minority populations within each buffer. We assumed that an MSU deployed from these MSU centers could rendezvous with emergency medical services units halfway between a rural stroke location and the destination stroke center. For each buffer, we compared the number of patients potentially served by the MSU to a base case estimate of emergency medical services transport represented by a 30-minute drive time buffer surrounding all nonoverlapping level I, II, III, or IV stroke centers. RESULTS: We identified 11 level I and 3 level II potential MSU stroke centers. A 180-minute buffer around each of these (MSU emergency medical services rendezvous 90 minutes from the stroke center) resulted in 741 852 patients with stroke potentially receiving thrombolysis within 3 hours of stroke onset representing 99.1% adult patients with stroke in the state; a net increase of 105 522 (16.6%) patients compared with base case and a 279% increase in patients from rural areas. A 120-minute buffer increased total and rural treatments by 12.3% and 232%. A 60-minute buffer resulted in no net increase in treated patients, though 600 101 more would receive faster care by MSUs. CONCLUSIONS: When distributed using geospatial analysis, MSUs can provide faster acute stroke treatment and potentially better outcomes to virtually the entire state of Texas, with a particular increase in rural populations that are not currently reached by emergency medical services. Our findings might be useful to health care planners in any state.