• Even large clinical trials contain narrow subset of patients
• In populations who will use drugs, heterogeneity, e.g. age
– Adverse events occurring at low frequency would not be observed (e.g. 1 event in 1,000,000 cases)
– New adverse effects can occur in diverse pop.
• Active Postmarketing Surveillance
– More difficult than clinical trials
• Continue to track adverse events
• Aggregate data from millions of patients from data collected by insurers, government agencies