During the early 2000's, a number of rigorous observational studies from several pharmaco-epi research groups, including ours, identified a higher risk of cardiovascular disease in patients taking rofecoxib compared to other NSAIDs, after controlling for possible confounders. Skeptics (including the drug's manufacturer) denied that this risk or any others could be rigorously demonstrated by population-based observational studies. In late 2004, a placebo-controlled RCT proved that the drug caused a more than two-fold elevation in the risk of myocardial infarction or stroke compared to placebo. Once one of the best-selling drugs in the world and used by over 20 million Americans and millions more worldwide over its five years on the market, Vioxx was immediately withdrawn by its manufacturer. Its trajectory led to a worldwide reassessment of how we track adverse events, the potential use of large health care datasets, and legislation to prevent a recurrence of another such unanticipated large-scale debacle of preventable adverse effects.