«Evidence-based» as a criterion for psychotherapy selection? Concerning a good concept – and its misuse

Abstract

This paper first points out that «evidence-based» is a good and useful concept but is frequently misused today due to reduction to RCTs. While the RCT design is absolutely an outstanding model for studying experimental research questions, this only applies if it is possible to isolate clearly definable causes and effects to a sufficient degree and if these provide a usable representation of the reality of the investigation. The greater the scope for designing the causes (interventions) and the more complex the relevant effects are, the less this approach is useful. We will show that RCT designs are suitable for psychotherapy research in particular if manualized programs are applied. This is primarily the case for behavioral therapy. For psychotherapeutic approaches that are mainly founded in the process-specific development of principles, the great variability of the detail causes (interventions) bursts the required bulk-statistical homogeneity for an experimental design. The requirement of proving effectiveness mainly with RCT designs amounts to methodical artifacts in this case or imposes a certain psychotherapeutic preference, i. e. for behavioral therapy, onto other approaches by means of scientific methodology. Finally, we discuss a few additional reductions and misinterpretations in connection with RCT and metastudy designs.