Accessing the Generalizability of Prevalence Estimates from the All of Us Research Program

This event is part of the All of Us Data Use Case Webinar Series hosted by the All of Us Researcher Academy. The three-part series focuses on methodological challenges and innovations using the All of Us Researcher Workbench.

This webinar will describe a new framework for evaluating how well electronic health records–based prevalence estimates from All of Us generalize to other estimates from the U.S. population. We will present the weighting methods using case studies of dementia, type 2 diabetes, and depression compared against National Health Interview Survey benchmarks.

Attendees will gain practical guidance and tools to implement the weighting methods in their own work and to recognize when weighting improves generalizability—and when it can unintentionally amplify bias.

Presenter: Barrett Montgomery, PhD, RTI International

Panelists: Daniel Brannock, MS, and Mahmoud Elkasabi, PhD, RTI International; Tamara Litwin, PhD, former NIH epidemiologist