Opportunities and Challenges for Using Networks of Observational Health Care Data for Medical Product Safety Surveillance

• Activity Number: 500
• Type: Invited
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Jesse Aaron Berlin*+ and Patrick Ryan and David Madigan and Martijn Schuemie
• Companies: Johnson & Johnson and Johson & Johnson and Columbia University and Erasmus University and Janssen Research & Development
• Keywords: pharmaceutical safety ; observational data ; Bayesian methods ; pharmacoepidemiology

Abstract

The Affordable Healthcare Act has helped focus the healthcare community on the need for data comparing effectiveness and safety of different interventions to treat or prevent the same clinical conditions. The Observational Medical Outcomes Partnership (OMOP)has been conducting methodological research to empirically evaluate the performance of various analytical methods on their ability to identify true associations and avoid false findings and establishing a shared resource so that the broader research community can collaboratively advance the science. This talk will present recent research findings by OMOP researchers, focusing on two areas of active research: calibration of p-values, using an empirical null distribution, in order to take into account the biases in database studies, and calculation of the posterior probability of an association, given a prior belief and the observed data, showing essentially that p< 0.05 doesn't guarantee a true effect exists and that p>0.05 doesn't guarantee no effect is present.