Introduction: The effect of mechanical thrombectomy (MT) on functional outcomes after acute ischemic stroke is time-dependent. Interhospital transfer (IHT) introduces delays between stroke onset and MT.
Methods: We created models of two IHT network designs in New York City involving 26 Department of Health designated stroke centers without MT capabilities (spokes) and 14 hospitals able to perform MT (hubs). In Model A, all patients were transferred from the initial hospital without MT capabilities to the closest hub irrespective of hospital affiliation. In Model B, all patients were transferred to the closest affiliated hub. We calculated the number of patients eligible for MT presenting to each spoke annually using publicly available data. With Google Traffic API software, we estimated travel times between spoke and hub hospitals under constant conditions that approximated ambulance travel. We determined the effect of transfer time on MT eligibility based on a reported 2.5% reduction in odds of MT per minute of travel delay beyond a minimum threshold. Rates of a good functional outcome (modified Rankin Scale score 0-2) were predicted based on MT trial data. Last, we calculated the efficiency advantage needed among affiliated hospitals (Model B) to offset increased travel times.
Results: In our models, 371 patients (interquartile range [IQR], 229-628) with a stroke amenable to MT present to New York City spoke hospitals without MT capability each year. The mean travel time to the closest hub was 15.9 (±5.7) minutes in Model A versus 23.0 (±9.5) minutes to the closest affiliated hub in Model B. Transferring patients to the closest hub irrespective of affiliation (Model A) resulted in 71 (IQR, 44-120) additional patients being eligible for MT upon arrival and thereby an additional 14 (IQR, 8-23) patients achieving functional independence per year. To attenuate the differences in outcomes between Model A and Model B, a 10.9-minute decrease was required in the combined door-in to door-out time at the spoke plus the door-in to MT-start time at the affiliated hub.
Conclusions: Optimizing IHT networks requires balancing the efficiency gains of transferring to affiliated hospitals with the travel delays resulting from bypassing nearby MT-capable centers.