In the era before large automated data sets became available for pharmacoepidemiology research, a number of ad hoc studies were mounted to look at the safety aspects of specific drugs. These studies had undoubted problems, and were generally expensive to mount and to conduct. Nonetheless they served to provide quantitation for several interesting risks, refute others, and they also helped to improve our understanding of methodology in this arena. They were however reported as being unhelpful to members of the UK Medicines Control Agency in their periodic safety assessments of licenced medicines. In a paper published in the British Medical Journal (Waller et al, 1992; 304: 1470-1472) individuals from the MCA and the Committee on Safety of Medicines effectively published their obituary. It was clear that from then on no future developments would be seen in that area of study, and such has proven to be the case. Whether it was wise of a regulatory agency to take this course of action is debatable, particularly when the details of the criticisms of individual studies was notably lacking in the article cited. This author believes that regulators should generally remain aloof from issuing guidance on methodology until such time as the issues are clear-cut and generally accepted by experts within the field. We are now left with the main source of information in this area being the multipurpose databases.
Was this article helpful?