Date of Award


Document Type

Open Access Thesis

Degree Name

Master of Public Health (MPH)


School of Public Health

First Advisor

Makuch Robert

Second Advisor

Zuoheng Wang


Drug post marketing safety surveillance is a very important part of pharmacovigilance, which is overseen by the Food and Drug Administration (FDA) and other regulatory agencies throughout the world. These studies are generally performed by pharmaceutical sponsors and academicians, although it is occasionally the case that a regulatory agency will either fund the research or have large databases which is analyzed by the regulators themselves. One example of agency surveillance is the FDA Adverse Event Reporting System (AERS). Generally, preapproval studies only involve several hundred to several thousand patients, so all possible side effects of a drug cannot be explored thoroughly because it is not possible to have a high probability of detecting rare, but important, adverse events. In addition, the population of patients in pre-approval studies are more restricted and better controlled than the wider population of patients who take a compound once it has been approved.. In order to monitor adverse events and serious/severe adverse reactions, FDA maintains a system of post-marketing surveillance programs to identify adverse events that did not appear during the drug approval process .

In this thesis, we will examine some regulatory/analytic issues that arise in the evaluation of pharmaco-epidemiologic data analysis using propensity score analysis from drug post-marketing safety surveillance. It is my desire to show the rationale and preference of using propensity score in post-marketing safety surveillance studies. We also examined several independent cases that used propensity score inappropriately, and concluded the FDA’s comments on them. We propose some advices on propensity score use in the future statistical analysis for reference.

Open Access

This Article is Open Access