Modeling with Real World Evidence

HSPH Adjunct Professor Uwe Siebert spoke at a CHDS lunchtime seminar on the potential bias that can accompany the use of “real world” evidence in modeling. Using a case study in ovarian cancer treatment, Siebert identified a range of methodological challenges for evaluating causality in this context, including confounding, missing/misclassified data, lack of clear treatment assignment, dynamic treatment regimens, and switching. Siebert explained a method of obtaining valid using the “target trial concept” along with casual methods adjusting for time-dependent confounding.

Learn more: Explore decision analytic modeling here.

Related news: RDS286 with Uwe Siebert
Related news: Seminar with Sheffield Scholar