Integrating Temporal Pattern Mining in Ischemic Stroke Prediction and Treatment Pathway Discovery for Atrial Fibrillation

Atrial fibrillation (AF) is associated with an increased risk of acute ischemic stroke (AIS). Accurately predicting AIS and planning effective treatment pathways for AIS prevention are crucial for AF patients. Because of the temporality of patients' disease progressions, sequential disease and treatment patterns have the potential to improve risk prediction performance and contribute to effective treatment pathways. This paper integrates temporal pattern mining into the AF study of AIS prediction and treatment pathway discovery. We combine temporal pattern mining with feature selection to identify temporal risk factors that have predictive ability, and integrate temporal pattern mining with treatment efficacy analysis to discover temporal treatment patterns that are statistically effective. Results show that our approach has identified new potential temporal risk factors for AIS that can improve the prediction performance, and has discovered treatment pathway patterns that are statistically effective to prevent AIS for AF patients.