[The generalisability of depression trials to clinical practice]

Evidence-based therapies for major depression, as described in the clinical guidelines, are based on results from randomised controlled trials (RCTs). So far, it is not known to what extent results of RCTs on major depression can be generalised to 'real life' clinical practice.

To compare treatment results for major depression from RCTs (efficacy) and results from daily practice (effectiveness); furthermore, to assess to what extent eligibility criteria and (un)intended selection by recruitment procedures influences treatment outcomes in daily practice.

In a 'real life' patient population (n=1653) suffering from major depression (established by the MINIplus) and assessed in routine outcome monitoring at baseline, we explored how many patients met the eligibility criteria for antidepressant and psychotherapy efficacy trials. Furthermore we explored to what extent RCT participants differed in socio-demographic and socio-economic status from 'daily practice' patients. 626 of the ROM patients had at least one follow-up assessment. In this follow-up group we compared the treatment outcome (assessed by the MADRS and BDI-II) to the results of 15 meta-analyses of RCTs. We also explored to what extent patient selection based on eligibility criteria and socio-demographic/socio-economic status influenced treatment outcome.

Remission percentages (21-27% in ROM versus 34-58% in RCTs) and effect sizes (0.85 in ROM versus 1.71 in RCTs, within-group data) were lower in daily practice than in RCTs. ROM patients differed from RCT participants in many disease-specific and socio-economic features. These differences are due to patient selection in RCTs. However, the influence of patient selection based on eligibility criteria and socio-demographic differences in treatment outcome were very modest (explained variances 1-11%).

Treatment success for major depression is lower in daily practice than in RCTs and 'real life' patients differ in many features from RCT participants. However, these differences cannot explain the difference between efficacy and effectiveness. The generalisability of the results of depression trials to daily practice might not be jeopardised by the use of eligibility criteria and recruitment procedures to the extent suggested in earlier research.