The association between adherence and outcome in an Internet intervention for depression

Background Adherence to Internet interventions is often reported to be rather low and this might adversely impact the effectiveness of these interventions. We investigated if patient characteristics are associated with adherence, and if adherence is associated with treatment outcome in a large RCT of an Internet intervention for depression, the EVIDENT trial. Methods Patients were randomized to either care as usual (CAU) or CAU plus the Internet intervention Deprexis. A total of 509 participants with mild to moderate depressive symptoms were included in the intervention group and of interest for the present study. We assessed depression symptoms pre and post intervention (12 weeks). Patient characteristics, a self-rating screening for mental disorders, attitudes towards online interventions, and quality of life were assessed before randomization. Results Adherence in this study was good with on average seven hours of usage time and eight number of sessions spent with the intervention. Some of the patient characteristics (age, sex, depressive symptoms, and confidence in the effectiveness of the program) predicted higher number of sessions in different models (explaining in total between 15 and 25% of variance). Older age (β = .16) and higher depressive symptoms (β = .15) were associated with higher usage duration. Higher adherence to the program predicted a greater symptom reduction in depressive symptoms over 12 weeks (number of sessions: β = .13, usage duration: β = .14), however, this prediction could mostly be explained by receiving guidance (β = .27 and .26). Limitations Receiving guidance and symptom severity at baseline were confounded since only participants with a moderate symptom severity at baseline received e-mail support. Therefore no firm conclusions can be drawn from the association we observed between baseline symptom severity and usage intensity. Conclusions We conclude that older age was associated with adherence and adherence was positively associated with outcome. The effects we have found were small however suggesting that adherence might also be influenced by further variables.