Modeling interhospital stroke transfers for endovascular treatment in the four largest cities in the USA

Interhospital transfers of stroke patients to higher levels of care for endovascular treatment (EVT) are common. Optimizing transfer times may help improve overall reperfusion times. We sought to model and compare the ground travel times for health system-affiliated and nearest EVT-capable stroke center transfers and examine associated factors in the four largest cities in the USA: New York City (NYC), Los Angeles (LA), Chicago, and Houston.

We identified EVT-capable centers (defined as comprehensive, thrombectomy-capable, and primary plus stroke centers) and their health system-affiliated hospitals for each city. For each affiliated hospital, we used a Monte Carlo simulation to model ambulance travel times to its system EVT-capable center as well as the nearest EVT-capable center. We subsequently analyzed hospital service area variables associated with prolonged (extra travel time to affiliated hospital >20 min) transfers.

The total travel time (mean (SD), min) to an affiliated versus nearest EVT-capable center was 29.9 (10.2) versus 14.4 (2.0) (P<0.001), with only ~7% of the nearest EVT-capable centers sharing affiliation with the transferring center. The travel times were significantly greater in LA compared with Houston and NYC (P<0.001). Finally, hospital service areas with a greater proportion of patients over 50 years of age and teaching hospital status were associated with potentially reduced extra transfer time.

This study characterized theoretical stroke transfer networks in the four largest US cities and showed that affiliated-hospital system transfers may increase overall transfer times compared with transferring to the nearest EVT-capable center.