Katie Quinn
Postdoctoral Scholar
Shah Lab, Stanford University

Abstract:

Concurrent use of multiple prescription drugs is widespread, with over 10% of Americans currently prescribed 5 or more drugs. Adverse drug reactions result in significant morbidity, and a substantial proportion of these are known to be caused by drug-drug interactions. Whereas premarket safety trials of all common drug combinations are infeasible, large-patient observational health data may enable postmarket surveillance of multi-drug safety.

We aim to identify associations between multi-drug combinations and health outcomes using the Truven Health MarketScanⓇ Databases, which document health coverage for approximately 100 million Americans over a median of 2 years. We first rank the incidence of concurrent drug prescriptions, to discover common multi-drug combinations. We then focus on a study of associations between these common multi-drug combinations and Emergency Department visits. We explore simple both disproportionality analysis and self-controlled case series study designs, and methods for empirical calibration in the absence of gold-standard controls.