Medication noncompliance is highly problematic in pharmaceutical studies. Without direct observation of pill-taking patterns, investigators - much like clinicians - are often misled by reported adherence, resulting in false interpretation of a protocol's true efficacy. Pill counts fail to reveal timing of intake, lost or hoarded medications, and """"""""white coat"""""""" compliance before appointments. Applied researchers have relied on recall questionnaires for collecting self-reported summary data and the use of simple paper and pencil diaries for self-monitoring intake in naturalistic contexts. Such retrospectively captured data suffer from autobiographical and recall biases; low frequency of measurement, which may result in poor reliability; and inability to capture fine-grained, process data that may shift and change over time. For pharmaceuticals, improved data quality can lead to rapid time to market for a new therapy, and with it, cost savings and alleviation of suffering. More reliable data collection methods can reduce the period a drug needs to be under evaluation, its digitalization time, and number of specialized personnel before it can be commercialized. Various electronic tools have emerged that provide part of the solution, but none that measure pill-taking in its temporal context for multiple medications at once. This proposal calls for an integrated system that provides passively monitored, real-time data capture in naturalistic settings, and collects responses to study questions, while it also tracks real-time intake of medications. Medication adherence is problematic in pharmaceutical studies. This proposal addresses an integrated system that passively monitors real-time meds-taking in naturalistic settings, collects responses to study questions, communicates usage data via auto-dialed telephone uploads, and wirelessly connects with remote biosensors and uploads their health status data daily. ? ? ?
