Multimodal Sensing and Modeling of Endocrine Therapy Adherence in Breast Cancer Survivors
Abstract
Endocrine therapy dramatically lowers recurrence risk for breast cancer survivors, yet its strict daily regimen is difficult to maintain. We present a multimodal sensing study that combines smartphone and wearable context, ecological momentary assessments, and patient-reported outcomes to characterize adherence behaviors and develop predictive models that surface moments of risk for just-in-time intervention.
We capture longitudinal multimodal streams from breast cancer survivors, fuse them with adherence diaries, and train interpretable temporal models that flag non-adherence risk windows. The results highlight psychosocial and behavioral indicators that precede missed doses and point toward personalized digital health interventions.
Fangxu Yuan
MS Student in Systems Engineering
My research interests include ubiquitous computing, mobile computing and digital health.