Using point allocation to identify outcomes that matter: results from two cognitive behavioural therapy trials for depression

BACKGROUND: Depression trials typically use symptom-based outcomes to determine treatment efficacy. Qualitative work with service users and clinicians indicates that other outcomes may be equally or more important. Quantitative investigations are needed to assess the relative importance of outcomes to key stakeholders. AIMS: To examine how stakeholders (service users with minimal symptoms (Beck Depression Inventory Second Edition (BDI-II) score ≤18), those who were currently symptomatic (BDI-II score ≥19) and clinicians working with people with depression) prioritise outcome domains drawn from two completed trials of cognitive behavioural therapy (CBT) for depression. This was from a new sample of stakeholders in order to explore similarities, differences and predictors of domain importance. METHOD: An online survey using a novel point-allocation method examined stakeholder priorities for outcome domains from two completed UK trials of CBT for depression. Stakeholder free-text responses explaining point allocations were thematically analysed. Multivariate analysis of covariance explored differences in domain importance across groups, and exploratory multivariate general linear models investigated predictive variables. RESULTS: In total, 229 participants completed the survey (57 with minimal symptoms, 108 who were currently symptomatic and 64 clinicians). Median point allocations suggest that all three participant groups assigned similar importance to depression symptoms, quality of life, anxiety symptoms and impact on activities. Sociodemographic factors significantly predicted domain importance within the currently symptomatic group, and age was a predictor both across and within this subgroup. CONCLUSIONS: Stakeholders valued several outcome domains with similar importance. While depression symptoms are not unimportant to stakeholders, they are balanced with other outcomes, such as quality of life, that are rarely measured in depression trials. These findings demonstrate the need to incorporate stakeholder-important outcomes into trial design, with stakeholder-weighted composite outcomes offering a potential method to capture this broader judgement of benefit.