IHPME Course Spotlight: HAD5314H – Applied Bayesian Methods in Clinical Epidemiology and Health Care Research

By: Marielle Boutin

Applied Bayesian Methods in Clinical Epidemiology and Health Care Research, a course within IHPME’s Clinical Epidemiology and Health Care Research (CEHCR) program, introduces students to Bayesian data analysis, a different approach from the frequentist methods most have previously learned.

Although still less commonly used than frequentist approaches, Bayesian methods are gaining traction in clinical research and are increasingly being applied to estimate treatment effects, disease risk and other health outcomes by combining prior knowledge with new evidence.

“Bayesian methods introduce a different way of thinking about evidence and uncertainty,” says Dr. Kuan Liu, assistant professor at IHPME and instructor of the course. “We do not just teach another set of statistical models – we spend time discussing how Bayesian and frequentist approaches ask and answer questions differently, and when those differences matter in clinical research.”

According to Dr. Liu, the course offers students a unique opportunity to apply these statistical methods to realistic case studies drawn from clinical research. Students put this into practice by using R statistical software to fit Bayesian models for common health outcomes, choose priors, check convergence, evaluate and compare models and interpret results.

“That’s a big part of what makes the course appealing to CEHCR students,” says Dr. Liu.

She goes on to add that the course is not meant to show that one method is better than the other, but to help students understand both Bayesian and frequentist approaches so they can choose the one that best fits their research question and critically evaluate studies using either framework.

One area where Bayesian methods have become particularly useful, is adaptive trial design, where accumulating evidence can be used to make planned changes during a trial, such as stopping early when there is sufficient evidence, changing how participants are assigned to treatments, or dropping a treatment that is not performing well.

Ultimately, the course teaches students to think about how evidence is generated and how much confidence to place in it.

“Students learn to judge not only whether an analysis was done correctly, but whether the evidence is convincing and clinically meaningful,” says Dr. Liu.