EPA-0834 – Documentation of Therapeutic Investigator Allegiance in Meta-Analyses of Randomised Controlled Trials

Abstract

Introduction:

Investigators allegiance is widely discussed as a risk of bias in psychotherapy research.

Objective:

To sum up the best available data concerning allegiance effect.

Aim:

We systematically investigated whether meta-analyses and their included randomised controlled trials (RCTs) of psychotherapeutic treatments were reported and assessed allegiance effect.

Method:

We searched meta-analyses of RCTs of various types of psychotherapies in Medline from 1977 to 2012 and the Cochrane Database of Systematic Reviews (Last update on December 2012). We considered only meta-analyses of RCTs with at least 1 study with allegiance of the experimenter published in journals with impact factor higher than 5 and in Cochrane Database.

Results:

Of 146 meta-analyses reviewed which included 2727 RCTs, only 15 meta-analyses (10.3%) reported RCTs allegiance. Of 1198 metaanalyzed RCTs only 1 (0.1%) was controlled for allegiance and 25 of 1198 (2.8%) were reported allegiance. In all meta-analyses reviewed, 66.2% of primary included RCTs were allegiant studies. Even we found a median of 10 [interquantile range (IQR) 7–15] allegiance RCTs per metaanalysis, only 6 (4.3%) of them used a method to controlling its effects.

Conclusions:

The majority of meta-analyses of psychological interventions published in high-impact specialty psychiatric/psychological, general medical journals as well as Cochrane Database was rarely reported and evaluated allegiance effect. The results of the present study highlight a major gap in this information in meta-analyses of psychotherapeutic interventions.