Predicting and preventing relapse of depression in primary care: a mixed methods study

Background Most people with depression are managed in primary care. Relapse (re-emergence of depression symptoms after improvement) is common and contributes to the burden and morbidity associated with depression. There is a lack of evidence-based approaches for risk-stratifying people according to risk of relapse and for preventing relapse in primary care. Methods In this mixed methods study, I initially reviewed studies looking to predict relapse of depression across all settings. I then attempted to derive and validate a prognostic model to predict relapse within 6-8 months in a primary care setting, using multilevel logistic regression analysis on individual participant data from seven studies (n=1244). Concurrently, a qualitative workstream, using thematic analysis, explored the perspectives of general practitioners (GPs) and people with lived experience of depression around relapse risk and prevention in practice. Results The systematic review identified eleven models; none could currently be implemented in a primary care setting. The prognostic model developed in this study had inadequate predictive performance on internal validation (C-statistic 0.60; calibration slope 0.81). I carried out twenty-two semi-structured interviews with GPs and twenty-three with people with lived experience of depression. People with lived experience of depression and GPs reflected that a discussion around relapse would be useful but was not routinely offered. Both participant groups felt there would be benefits to relapse prevention for depression being embedded within primary care. Conclusions We are currently unable to accurately predict an individual’s risk of depression relapse. The longer-term care of people with depression in general practice could be improved by enabling continuity of care, increased consistency and clarity around follow-up arrangements, and focussed discussions around relapse risk and prevention. Scalable, brief relapse prevention interventions are needed, which would require policy change and additional resource. We need to better understand existing interventions and barriers to implementation in practice.