Modeling the Decay in Probability of Receiving Endovascular Thrombectomy on the Basis of Time From Stroke Onset

Abstract

Background American Heart Association guidelines specify infarct core volume as 1 determinant of eligibility for endovascular thrombectomy. Therefore, it is important to understand how time‐dependent infarct core growth translates to a patient's declining probability of thrombectomy eligibility. Modeling the probability that a patient with suspected large‐vessel occlusion would qualify for thrombectomy on the basis of their expected time from stroke onset to treatment can help inform the optimal prehospital emergency transport protocols, maximizing the likelihood of an excellent patient outcome. Methods We extended a published physiological model of infarct core growth to derive a decay curve of thrombectomy eligibility (based on a given infarct core volume threshold) as a function of time from stroke onset. We then adapted an existing model of the time‐dependent probability of an excellent outcome to incorporate this decay curve. Using the adapted model, we determined the optimal prehospital emergency transport protocols in Alberta, Canada, and compared these with the protocols that assumed all patients were thrombectomy eligible. Results The probability of qualifying for thrombectomy decays exponentially as time elapses from stroke onset. We found that the area where mothership is the optimal transport protocol increased by 18.6% after incorporating our decay curve of thrombectomy eligibility into the underlying optimization model. The benefit of mothership versus drip‐and‐ship also increased in the areas where mothership was favored, and in areas where drip‐and‐ship was favored, the benefit of drip‐and‐ship weakened. We also performed a number of sensitivity analyses to observe how these results change on the basis of our assumptions for model parameters. Conclusion This methodology provides a novel, physiology‐based approach to derive a thrombectomy eligibility curve. These models are necessary to better optimize prehospital transport decisions and consequently improve outcomes of patients with suspected large‐vessel occlusion.

Keywords

• acute stroke
• decision modeling
• emergency care prehospital
• patient outcome
• probability
• thrombectomy