Risk Evaluation Examples
Effects of Risk Communication on Physician Prescribing Behavior
1. The Center for Devices Survey Following a Public Health Advisory
Issue: On March 20, 1998, the Center for Devices distributed an FDA Public Health Advisory of interference between digital TV transmissions and medical telemetry systems. FDA wanted to evaluate the format and content of the Advisory to assess the effectiveness of the communication.
FDA Action: The Center for Devices mailed a survey questionnaire to a randomly selected sample of 308 advisory recipients. To maximize response rates, two additional mailings of the survey were directed at nonrespondents.
Outcome: FDA received a total return of 190 questionnaires, of which 183 were determined to be valid completions, for a response rate of 59.4 percent. Of the respondents, approximately 98 percent felt that the problem addressed in the advisory was clearly identified and easily understood. The respondents stated that the information was both timely (93 percent) and useful (86 percent). They also felt the actions for reducing risk were clearly explained (97 percent). About 36 percent of the respondents stated they were aware of the interference problem prior to receiving the advisory. Approximately one-third of the respondents first became aware of the problem from a professional bulletin. About 37 percent of the sample group reported they had taken actions to eliminate or reduce the risk as a result of the advisory. The most commonly reported reason for not taking action was that the subject matter of the advisory was not applicable. Most respondents would prefer to receive future alerts as they are currently distributed -- printed and mailed (75 percent). Only 10 percent of the sample group preferred to receive future bulletins electronically (Internet mailing list).
The Center for Devices has found such surveys to be helpful in determining the effectiveness of Public Health Advisory risk communications and whether improvements in the process need to be addressed.
2. Population-Based Data Resource for Serevent
The effect of risk communication or other regulatory intervention on physician prescribing behavior can be measured using population-based data resources. As described in Appendix C, FDA uses IMS health products and services to monitor the effect of regulatory actions by the Center for Drugs and a variety of other users. Measuring physician prescribing behavior can help FDA evaluate its communication methods and make changes in the process.
Issue: Communicating risks may not always have the intended effect. Following its market introduction, FDA received a number of reports of fatal asthma attacks in patients using Serevent (salmeterol), a long-acting beta-agonist approved for the treatment of stable asthma. FDA was concerned that the product was being used to control patients with unstable asthma or was being substituted for other asthma medications inappropriately. A Dear Healthcare Professional letter was sent informing practitioners that Serevent should not be used for acutely worsening asthma, that it was not a replacement for short-acting bronchodialators, and was not a substitute for anti-inflammatory medications.
FDA Action: FDA used data from New York Medicaid to measure the effect of this letter. The computerized pharmacy claims records of each patient who filled a prescription for salmeterol within the database were examined for the presence of an albuterol claim in the 3-month period prior to their first salmeterol prescription.
Outcome: The proportion of patients without prior albuterol use was compared for the time interval before and after the letter. Prior to the Dear Healthcare Professional letter, 75 percent of new salmeterol users had no prescriptions for albuterol during the 3-month interval preceding their salmeterol use. After the letter, this estimate fell to 67 percent, suggesting that physician prescribing behavior within this Medicaid database had not changed in a clinically meaningful way.
Although the evidence did not clearly demonstrate an on physician prescribing patterns, FDA has noticed that the number of reports of fatal asthma attacks fell to below the number of reports prior to the Dear Healthcare Professional letter and has remained low. In addition to monitoring the number of events, FDA discussed its evaluation results and concerns with the product sponsor. The sponsor also followed up with physicians to ascertain that the information was appropriately affecting prescribing habits and is currently conducting a large phase-4 controlled trial to address the issue of mortality risk in patients who use the product in accordance with the label.
Summary: This example illustrates that continued efforts are needed to further evaluate FDA's risk assessment tools and subsequent interventions to ensure effective communication efforts and processes. As part of this continued effort, FDA is evaluating other interventions and is currently performing a study within the Center for Drugs' Cooperative Agreement Program to measure the effect of a series of Dear Healthcare Professional letters relating to liver transaminase monitoring in patients treated with troglitazone (Rezulin). This study will assess the performance of baseline and monthly transaminase monitoring using computerized billing claims for laboratory testing. The results from this study should contribute to regulatory policy relating to this product.