Drug withdrawals and the utilization of therapeutic substitutes: The case of Vioxx

The lack of research on how the 2004 safety-related withdrawal of the drug Vioxx affected consumer drug utilization or outcomes for competitors is a missed opportunity to learn from the largest drug withdrawal event in history. Our study fills this void using state-level repeated cross section data from the Medicaid State Drug Utilization (SDU) database of fee-for-service Medicaid claims for prescription drugs, and individual-level panel data from the Medical Expenditure Panel Survey (MEPS) which is a nationally representative survey that contains information on medication use across two years. We find that the withdrawal of Vioxx had both positive and negative effects for specific substitute drugs in its own class (COX-2s), and that it led to an overall increase in the usage of both its most direct competitor class (NSAIDs) and in a class of older similar therapy (Analgesics). We argue these shifts in drug usage represent what could be viewed as an appropriate response to the events. However, aggregate use of drugs in the COX-2 and related classes declined overall, suggesting that some consumers may have over-reacted to the withdrawal events in ways that lessened the health benefits they could receive from this family of drugs. These findings about medication utilization changes in response to negative information are highly relevant for policy design and for determining thresholds for regulatory interventions.