[The influence of study results on therapeutic decisions - experience in 35years]

This is a critical review on the influence of large multicenter studies on therapeutic decisions in rheumatology. Every prescription in rheumatology relies on such studies, because study results have to be presented for the approval of each new drug. Large randomized trials are rated highest within the evidence based medicine (EBM). They are nearly exclusively performed by pharmaceutical companies aiming at approval of drugs with the potential to gain high rewards. Old (and cheap) drugs known to be effective based on clinical experience are not included into modern trials. Therefore, they cannot meet the EBM standards, are no longer mentioned by authorities in the field of rheumatology and are rarely prescribed. Large patient numbers required by statisticians can only be reached by including inappropriate patients and centers more interested in the (very good) honorarium than in the scientific problem rendering questionable results because the quality of clinical investigation cannot be controlled. Statistical evaluation, study report, preparing the manuscript and publication are supervised by the sponsor. Negative studies are not published. Often studies are not comparable due to differences in inclusion criteria, patient population, disease activity, drug dosage, withdrawal rates and many other factors. Even similarly designed studies may come to contradictory results. Parameters measured in clinical trials may be unreliable indicators of the true longterm outcome. Metaanalyses, also highly ranked within the EBM, may come to controversial conclusions on the same subject, because different studies are selected, complex studies have to be simplified and condensed and pooled with other studies that cannot even be compared. Very often the true performance of a drug becomes evident only after broad clinical use (i.e. Auranofin) or is known due to clinical experience long before randomised studies were performed (i.e. MTX). Open longterm observational studies are undispensable in the field of rheumatology and may render more important knowledge than randomised studies.