Impact of safety warnings on drug utilization: marketplace life span of cisapride and troglitazone.

Study Objective: To evaluate the impact of safety alerts on the volume of cisapride and troglitazone usage.

Design: Retrospective database analysis.

Setting: University research center.

Material: Idaho Medicaid claims data from January 1994--July 2000.

Measurements And Main Results: Monthly counts of total and new prescriptions filled for cisapride and troglitazone were analyzed graphically over time as a function of all prescriptions. New prescriptions were defined as those filled by patients who had not received the drug within the previous year. A binomial comparison of the 5 months before and after each safety alert was conducted by Poisson distribution. Overall and new cisapride usage increased after the first alert, which occurred in February 1995 (p<0.05). After the second alert, in September 1995, growth in new prescriptions ended but total prescriptions continued to grow (p<0.05). After the third alert, in June 1998, growth in total use ended and the number of new prescriptions declined (p<0.05). The final two alerts (June 1999 and January 2000) were met with significant declines (p<0.05 for both). Troglitazone was the subject of two alerts in October and December 1997. After these, overall usage increased (p<0.05), whereas the number o new prescriptions decreased (p<0.05). The third alert, in July 1998, caused no change as total prescription use continued to grow (p<0.05), whereas the number of new prescriptions decreased (p<0.05). A fourth alert, in June 1999, resulted in a decrease of overall usage and new prescriptions (p<0.05 for both).

Conclusion: Numerous safety alerts were required for each drug before drug usage declined. The decline in overall use was slower than the decline in new prescriptions, possibly indicating a need for increased assessment of refilled prescriptions after the release of new safety data.