Broader Use of Post-Marketing Surveillance — Discussion
The discussion was led by Dr. Michael Weintraub who began by pointing out that data gathered from post-marketing surveillance has its limitations but should not be underestimated; doctors are neither fools nor geniuses. The risk of ignoring utilization data is the perpetuation of errors although there are some questions which can only be answered by ‘naturalistic’ studies. The data from PMS studies are ‘dirty’ because of inexact diagnoses, variable ancillary treatments, lack of concurrent controls and imprecise ‘end-points’. In spite of this, they can provide important information. An illustration of the worth of imprecise end-points is the study on Valproic acid (JAMA, 244:785–788) which evaluated the anti-epileptic effect of the drug not only by counting the number of seizures but also by using social and familial criteria. The details of this vividly illustrate the value of PMS data.