Market Response to Food and Drug Administration\u27s Safety Warnings: a Case Study Using an Interrupted Time Series Analysis of the Medicare Database for 2006-2008

Objectives: To evaluate the impact of a safety warning on the utilization rates of thiazolidinedione oral anti-diabetes medication using an interrupted time series analysis. Methods: We extracted data from a five percent national sample of Medicare Part D beneficiaries for the periods between January 2006 and December 2008. Beneficiaries with Part D claims for thiazolidinediones were classified into appropriate-use, at risk, and contraindicated groups based on certain comorbid conditions. We assessed the effects of the May 2007 FDA safety warning about an ongoing safety review of rosiglitazone\u27s potential to increase cardiovascular risks on thiazolidinedione utilization rates using an interrupted time series consisting of 32 data points (13 months before and 19 months after the safety warning). Results: There was an increasing trend in the total utilization rates of thiazolidinediones before the safety warning. Significant decline in drug utilization rates were observed at the end of the study period for all patient groups on rosiglitazone (relative difference -74.78%, -79.93%, and -90.21% respectively in appropriate-use, at risk and contraindicated patient groups). The intervention did not have significant immediate effects on the post-intervention utilization rates of pioglitazone. Following the intervention, we observed a general decline in utilization for thiazolidinediones. Conclusions: The initial safety warning about rosiglitazone\u27s cardiovascular safety was effective in decreasing rosiglitazone\u27s utilization in the targeted population and hence its Medicare market share. The safety warning also had spillover effects by reducing utilization of drugs in other patient cohorts not targeted by the warning