Automated Feedback After Internet-Based Depression Screening: Cost-Effectiveness Analysis of a Randomized Controlled Trial
Background: The clinical and cost-related consequences of internet-based depression screening, in combination with automated feedback, have been rarely investigated. We aimed to conduct a cost-effectiveness analysis (CEA) of DISCOVER, a three-armed, observer-masked, randomized controlled trial that focused on two versions of automated feedback interventions after internet-based depression screening. Objective: This study aims to evaluate the cost-effectiveness of automated non-tailored and tailored feedback interventions, after internet-based depression screening, from a societal perspective. Methods: Participants were recruited from the general population via traditional and social media. Participants who were undiagnosed but screened positive for depression on an online version of the Patient Health Questionnaire-9 (PHQ-9; ≥10 points) were randomized to automatically receive either no feedback, non-tailored feedback, or tailored feedback. The feedback interventions included the depression screening result, a recommendation to seek professional advice, and brief general information about depression. The tailored feedback was additionally framed as per the participants’ symptom profiles, treatment preferences, as well as health insurance plan, and local residency. The time horizon was six months. The main outcome was the incremental cost-effectiveness ratio (ICER) from a societal perspective using quality-adjusted life-years (QALY) based on the EQ-5D-5L. Cost-effectiveness acceptability curves (CEAC) were constructed. Furthermore, several sensitivity analyses and explorative subgroup analyses were conducted. Results: 1,012 participants (no feedback: 343; non-tailored feedback: 338; tailored feedback: 331) were included. Differences in costs and effects were not statistically significant. Still, ICER results indicated that both no feedback and tailored feedback exhibited dominance over non-tailored feedback. The ICER of tailored feedback compared to no feedback was €109,730/QALY, whereas both costs and QALYs were lower in tailored feedback. The cost-effectiveness probability of tailored feedback compared to no feedback ranged between 41% and 80%. Sensitivity analyses exhibited similar trends. Conclusions: Six months post-intervention, feedback interventions had no statistically significant effect on costs from a societal perspective nor on QALYs. Tailored feedback was associated with moderate cost-effectiveness probabilities compared to no feedback. Explorative subgroup analyses revealed subpopulations for which the interventions might be cost-effective. Clinical Trial: ClinicalTrials.gov: NCT04633096
