Q&A With Susan Gruber, PhD, MPH, MS
IMEDS Senior Director for Methods Research
What are the statistical challenges in using electronic health records (EHR) to conduct post-market surveillance?
The fundamental challenge is that estimating causal effects from EHR data requires adjusting for sources of bias. However, full adjustment is not always possible, particularly when there is unmeasured confounding. Despite this, these observational data are often the best source of information for monitoring the safety of a drug or medical device after its been approved by the FDA. How can we reliably identify true safety signals when effect estimates are biased? How can we improve upon current statistical and computational methodology? IMEDS is playing a pivotal role in defining and answering these important open questions.
How did you come to be interested in pharmacoepidemiology?
Over the course of my career I have worked as a chemist, a teacher, a computer scientist, and a biostatistician. Pharmacoepidemiology is a field that draws on every aspect of my training, and allows me to work with experts in epidemiology, pharmacology, medicine, and public health to improve public health and safety.
What excites you about the IMEDS Research Laboratory?
The lab provides unprecedented access to data and software for users from academia, industry, the FDA, non-profit agencies, and other institutions. This common ground provides the foundation for open reproducible research.
IMEDS-Methods is important to me because… our targeted research agenda focuses on the most important issues in safety surveillance, as defined by leaders in industry and the FDA, while at the same time fostering the interdisciplinary collaborations that are vital to producing the most reliable evidence possible from disparate, distributed data sources.
Previous Q&A Members