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

Alec Walker

Q&A With Alec Walker, MD, DrPH, Principal, World Health Information Science Consultants, LLC
Advisor to the Reagan-Udall Foundation

What is the biggest challenge today for post-market surveillance?
First we need to define clearly the types of adverse drug effects for which the current systems permit sensitive surveillance and ready exclusion of false-positive signals. Many common, immediate and severe reactions already fall into this class, as do less frequent, less dramatic or more delayed drug effects for which there are no counterparts among spontenously occurring diseases. And for these "easy-to-find" effects, we need to let the existing systems work. The next order of business is to dissect the harder-to-study effects, to find out where the existing systems need to be adapted, in order to bring the effects under scrutiny. Screening available data to identify highly susceptible populations may be a solution. Find groups in whom the adverse effects are easy to identify and for whom the products involved ought to be contraindicated. Perhaps hardest of all will be to lay down criteria for the kinds of adverse effects for which surveillance is unlikely to be successful, and for which experimental studies or aggressive diagnostic approaches will be important.

What is your landmark or pioneering pharmacoepi article that you like to refer to?
I'd name two papers. The first I wrote with Bob Wise of the FDA. (Walker AM, Wise RP. Precautions for proactive surveillance. Pharmacoepidemiol Drug Safety 2002;11:17-20 ) We explored the challenges to active drug safety surveillace that were then clearly on the horizon, and now are with us. The second, led by colleagues now at the Brigham and Women's Hospital in Boston, provided an instance of detecting an difficult-to-find adverse effect using rich data a gamut of statistical techniques. (Schneeweiss S, Seeger JD, Landon J, Walker AM. Aprotinin during coronary-artery bypass grafting and risk of death. N Engl J Med 2008;358:771¬83)

What excites you about the IMEDS-Methods Research Agenda?
We have a coherent program that responds both to science and to the FDA's expressed needs.

IMEDS-Methods is important to me because ... it welcomes contributions from every corner of the drug safety arena, from regulators and researchers, from industry and consumers, and it provides an open arena where ideas can contend and the most successful ones can prosper.