A secondary analysis investigating therapist-effects in a guided internet-based intervention for adults with symptoms of depression

Objective Studies consistently show that guided internet-based interventions (IBIs) are more effective than unguided IBIs for depression. Yet, little is known about whether average symptom change and rates of treatment discontinuation differ between therapists who guide clients through IBIs (IBI therapists). Method This secondary analysis draws on data from three randomized controlled trials (RCTs; RCT1: n = 543, number of therapists k = 40; RCT2: n = 2277, k = 21; RCT3: n = 1769, k = 10) that provided adults with depression access to the same IBI. IBI therapists provided written, module-wise feedback and were available for on-demand contact. We examined therapist effects in (i) treatment discontinuation rates, (ii) rates of meaningful symptom improvements (≥50% improvement), and (iii) symptom change. In addition, we tested the association of these outcomes with the number of previously treated clients and the length of feedback messages. Results Estimated therapist effects for treatment discontinuation were cVPC = 0.3 to 7.5% and cMOR = 1.081 to 1.617; for treatment response, cVPC = 0.4 to 3.1% and cMOR = 1.088 to 1.332; and for symptom change, conditional ICC = 0.2 to 0.7% and conditional median absolute difference = 0.121 to 0.275 PHQ-9 points. No association between the number of treated clients and any outcome emerged. However, longer feedback was associated with a reduced likelihood of treatment discontinuation. Conclusion The between-therapist differences in discontinuation and symptom change in guided IBIs for depression can be small.