Instructor Spotlight: Sam Berchuck, PhD

Sam Berchuck, PhD teaches BIOSTAT 725: Bayesian Health Data Science in Duke’s Master of Biostatistics program, bringing his expertise in Bayesian methods and patient-centered research to inspire students with real-world applications.

Could you share a bit about your background and what led you to the field of biostatistics/bioinformatics?

I come from a family of doctors, and while I never wanted to be a physician, I knew I wanted to improve people’s health. I discovered a love of statistics in high school AP Stats and quickly knew I wanted to major in statistics at Duke. Biostatistics was the perfect way to combine mathematical thinking with meaningful impact on patient care.

What is your current research or professional focus outside of teaching?

My research focuses on developing statistical models to improve patient care, particularly around symptoms like pain and distress. In glaucoma, for example, we’re creating Electronic Health Record-based algorithms to automatically screen patients for distress and developing a mobile health app (VISION-ACT) to deliver interventions. I’m also working in breast cancer to identify patients at high risk for persistent pain and tailor interventions for precision pain management.

Which course(s) do you teach in the program?

BIOSTAT 725: Bayesian Health Data Science.

What do you most enjoy about teaching this course?

Bayesian statistics is what I use in my own research, and I love sharing that passion with students. It’s exciting to teach something so actionable. Students leave with real skills they can apply immediately.

What are the key skills or concepts you hope students will take away from BIOSTAT 725?

The goal is to make Bayesian statistics truly usable. Students learn probabilistic programming in Stan, giving them the tools to apply Bayesian methods in research and practice, not just theory.

Are there particular projects, examples, or applications that students tend to find especially meaningful or exciting?

Yes, I bring in real-world data from my research, such as EHRs and glaucoma visual field data. Students get hands-on experience with messy, complex health data and learn how to build models that matter for patient outcomes.

What do you enjoy most about working with Duke MB students?

They’re motivated, curious, and engaged. Their questions push me to think deeply about the material, which makes teaching especially rewarding.

How do you see students grow or change as they move through your course and the program?

Students come in with a strong foundation in traditional statistics, and in this course, they discover how Bayesian methods open a wide range of flexible modeling strategies. It’s rewarding to see them realize they can focus on the scientific question and choose models that fit the problem, rather than forcing the problem into a narrow framework.

What advice do you give students for getting the most out of your class?

Engage actively with the material and practice applying the tools, the more you work with real problems, the more confident you’ll become in using Bayesian methods in your own research.

What is something students might be surprised to learn about you?

I was actually born at Duke Hospital and have spent my whole life here. I graduated from Duke as an undergrad, went to UNC-Chapel Hill for graduate school, returned to Duke for my postdoc, and now I’m faculty. Duke has always been home, and I’m passionate about being part of this community.

What do you enjoy doing outside of teaching and research?

Most of all, I enjoy spending time with my wife and two daughters, which often includes hiking, cooking, and gardening.