Ehsan Karim, PhD, M.Sc.
Scientist, Biostatistician, CHÉOS
The use of retrospective healthcare claims datasets are frequently criticised for the lack of complete information on potential confounders. The “high-dimensional propensity score” algorithm proposes to harness the information from a large number of proxy variables to reduce bias in such secondary database analyses. In this talk, Dr. Karim will compare the performance of this algorithm with a few popular machine learning approaches suitable for handling big-data problems.
Work in Progress (WiP) presentations take place at St. Paul’s Hospital in the Hurlburt Auditorium on alternating Wednesdays from 12:00–1:00 PM. These seminars provide investigators with an opportunity to present ongoing research, obtain feedback from colleagues and peers, and make new connections for their projects. Talks are open, and a light lunch is served.