ESOC 2025 | The MAP-STROKE study: a pre-hospital personalized algorithm for stroke triage

The MAP-STROKE Project or study is a NIH-funded study in which we are trying to elucidate if we have a better way to triage patients, stroke patients to the right hospital. The rationale for that is in the current guidelines, triage patients to different closest hospitals based on time and some suspicion for a stroke. What we have developed as a first part of the study is a Bayesian framework algorithm in which we are able to really elucidate with the information at the field, what’s the highest suspicion for the type of a stroke and what is the best route, introducing some geo-mapping models to go with the best route that increases the chances for best neurological recovery based on the hospital capabilities. Patients who have a high suspicion for LVO are going to be triaged to the right hospital with mechanical thrombectomy capabilities. Patients who have a high suspicion for non-LVO stroke or stroke mimics, they might go to the closest hospital with thrombolysis capabilities in order to receive the best treatment and avoid delays. The first part is the development of the software in which we have applied different clinical trials data, over 11,000 patients, with the time-sensitive probability of good outcome for mechanical thrombectomy, as well as for IV tPA. Then we have developed an in silico model design in which we have over 300 million reiterations of strokes in the continental US. Then basically we replicate all these strokes and we really evaluate the best triage approach in order to estimate best outcomes as well as time delays and treatment delays. Then our study, main results of our study, demonstrate that by applying the algorithm compared with AHA current guidelines, patients who follow the algorithm might have a 3% higher rate of decreased disability measured by the ranking and modified ranking score less than 2. In addition to that, we decrease delays on mechanical thrombectomy for LVO patients when they follow this algorithm by more than 90 minutes. And more importantly, we also have shown that the probability of receiving the right treatment is much higher in rural areas in patients with LVO stroke. In those patients, the time reduction might be over two hours and the probability of better outcomes increases by 4% comparing with the current guidelines adherence. I think this is a very good first step. This is again a simulation study and the natural next steps are going to be validating our results through a clinical trial mechanism on the field. Then at this current moment, we are starting the grant application for a pre-hospital clinical trial in which there will be a randomization using the algorithm versus the current care that’s provided to these patients.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.