Data sources and methods for ascertaining human exposure to drugs.

Abstract

Estimates of population exposure based on drug use data are critical elements in the post marketing surveillance of drugs and provide a context for assessing the various risks and benefits associated with drug treatment. Such information is important in predicting morbidity and planning public health protection strategies, indepth studies, and regulatory actions. Knowledge that a population of one thousand instead of one million may potentially be exposed to a drug can help determine how a particular regulatory problem will be handled and would obviously be a major determinant in designing a case-control or cohort study. National estimates of drug use give an overview of the most commonly used drug therapies in current practice. They also furnish valuable comparison data for specific studies of drug use limited to one group of drugs, one geographic region, or one medical care setting. The FDA has access to several different national drug use data bases, each measuring a different point in the drug distribution channels. None covers the entire spectrum of drug exposures. The major "holes" in this patchwork of data bases are the inability to measure OTC drug use with any accuracy and the lack of qualitative information on drug use in hospitals. In addition, there is no patient linkage with the data. The data can only show trends in drug use. They impart no sense of the longitudinal use of drugs for individual patients. There is no direct connection between the different data bases, all of which have their own sampling frames and their own projection methodologies. The market research companies have complete control over these methodologies and they are subject to periodic changes, a situation not entirely satisfactory for epidemiologic research. Sometimes it is a struggle to keep up with these changes. Over the past two years, every one of these data bases has undergone some type of sampling or projection methodology change. One important limitation to the use of all of these data bases is that they are subscription data bases, that is, the FDA purchases the data under contract to the marketing research companies and by doing so assumes certain contract obligations. Anytime the FDA wants to release any data outside of the Agency, it must first notify the company in sufficient time for review and approval. Subscribing to these data bases is costly, but the subscription cost is insignificant, compared to the estimated cost of duplicating these services. In spite of all of the limitations of these systems, there are obvious advantages.(ABSTRACT TRUNCATED AT 400 WORDS)