There have been 13 high‐profile product withdrawals in the United States during the past 4 years. These withdrawals have been due to a limited range of adverse event types, usually associated with relevant antecedent information. Lesser degrees of regulatory action (e.g. specific post‐marketing surveillance requests, labelled warnings, black boxes, and ‘Dear Doctor’ letters) have often preceded product withdrawal. Criticism of regulators has been misdirected because there is no evidence for a chronological trend in the need to withdraw products, nor in the marketed lifespan. The existence of antecedent information is not necessarily a basis for allegations of earlier, prospective grounds for NDA rejection or product withdrawal: In some cases, adverse events have arisen in patients for whom warnings against prescription already existed. Nonetheless, there appears to be scope for improving this situation. Prescribers need to be more familiar with product labels, and to be able to perceive clinically significant adverse event reports from among the mass of information provided. Black boxes are also currently used ambiguously in the United States. Non‐professional constituencies need to understand that relative rather than absolute judgments are required for all drug approvals and withdrawals.