Post-marketing studies can make important contributions to medical knowledge

My experience in collaborating with various drug companies in the design, execution, analysis, and publication of post-marketing surveillance (PMS) studies differs greatly from that described in your recent article.1 The general consensus was that PMS studies serve a scientific purpose. Studies typically had predefined scientific questions and were powered to answer them. Some studies had been recommended by the regulatory authorities, such as ones investigating tolerability in high risk subpopulations,2 underlying pathophysiology of disease,3 and disease assessment tools.4

Properly designed PMS studies are not scientifically inferior to randomised controlled trials—they just serve a different purpose. Being closer to real life, they often reflect a broader spectrum of patients, and because per patient costs are lower, can often recruit larger numbers of patients than would be possible in trials. Their main disadvantage is the lack of a control group, but that can be partly overcome.5 This results in lower internal validity but greater external validity than with trials. Although they cannot be used to make claims about the absolute efficacy of a given treatment, the often large numbers of patients can allow the analysis of subpopulations or the application of multiple regression models to analyse potentially related variables with adequate statistical power.

I do not claim that in the past PMS studies have not been used as marketing tools disguised by poor science. Investigators, drug companies, and journal editors are all responsible for ensuring that only PMS studies with relevant scientific questions and methods are performed and published. Such PMS studies can make important contributions to medical knowledge.