Person-generated Health Data: Is Pharmacoepidemiology Ready for this New Frontier?

Background: ‎
Much of pharmacoepidemiology focuses on analysis of structured data and ‎increasingly unstructured data e.g., the clinical narratives in electronic health record ‎databases. Person-generated health data captured directly from patients through ‎digital health technologies (e.g., apps, wearables) is an emerging source of important ‎new data for pharmacoepidemiology. These new streams can provide huge amounts ‎of data in near real-time and provide patient insights not easily gleaned from clinical ‎databases (e.g., relevance of data elements). For example, more continuous ‎measurement allows us to explore potential changes in disease trajectory and the ‎precise timing of medically significant events. This is coupled with an increase in ‎Software as Medical Device (SaMD), more combinations of drug-device and device-‎software, and the growth of medical device interventions to treat diseases. These ‎trends create a critical need for pharmacoepidemiologists to understand the role of ‎digital data and devices in the real world.‎

Objective: ‎
Pharmacoepidemiology studies will increasingly have to consider devices and digitally ‎derived data, directly or indirectly in their studies. This session will assess the ‎opportunities and challenges in using digital data to ascertain study exposures, ‎outcomes, covariates or interventions in epidemiologic studies.‎

Description: ‎
The session will cover 1) an introduction to a range of novel and emerging person-‎generated health data sources that can complement more traditional healthcare ‎provider data (Dixon), 2) a regulatory perspective on the relevance and reliability of ‎medical device real-world evidence generation, with examples (Ritchey), 3) lessons ‎from NEST the National Evaluation System for health Technology and using digital data ‎from surgeries for machine learning models to predict patient benefits from surgeries ‎machine learning models to improve patients’ surgical outcomes (Coplan), 4) ‎traditional pharmacoepidemiology design and inference challenges when applied to ‎these emerging data sources (Platt) and 5) an interactive discussion on case studies, ‎next steps and opportunities for the ISPE community to develop evidence from these ‎growing sources of real world data (all).‎