RT Journal Article
SR Electronic
T1 Predicting and preventing relapse of depression in primary care: a mixed methods study
JF The Annals of Family Medicine
JO Ann Fam Med
FD American Academy of Family Physicians
SP 4660
DO 10.1370/afm.22.s1.4660
VO 21
IS Supplement 3
A1 Moriarty, Andrew
A1 Archer, Lucinda
A1 Snell, Kym
A1 Meader, Nick
A1 Paton, Lewis
A1 Riley, Richard
A1 McMillan, Dean
A1 Chew-Graham, Carolyn
YR 2023
UL http://www.annfammed.org/content/21/Supplement_3/4660.abstract
AB Context: Relapse (re-emergence of depression symptoms) is common after depression. We lack evidence-based approaches for identifying people at higher risk of relapse. It is uncertain how widely, and in what form, relapse prevention strategies are provided in primary care.Objectives: Identify and appraise prediction models for relapse of depression; Develop a primary care-based prediction model to support primary care with risk stratification; Understand the views of General Practitioners (GPs) and people with lived experience of depression (PWLE) around relapse prediction and prevention in practice.Study Design and Analysis: Mixed methods study. Systematic review of prediction models. Derivation (multilevel logistic regression analysis) and validation (internal-external cross-validation) of a model to predict relapse within 6-8 months. Concurrent qualitative work-stream with semi-structured interviews exploring perspectives of GPs and PWLE on relapse risk and prevention. Thematic analysis used principles of constant comparison. A Patient Advisory Group contributed to all stages of the study.Setting or Dataset: Primary care - individual participant data (n=1244) from 6 randomised controlled trials and a cohort study. Recruitment of GPs and PWLE from the UK NHS.Population Studied: Adults with experience of depression. Practising GPs.Intervention/Instrument: NA. Outcome Measures:Remission and relapse (PHQ-9).Results: 12 prediction model studies identified, none can be implemented in primary care. We developed a novel model with sub-optimal predictive performance (calibration slope 0.95 (0.54-1.35), C-statistic 0.62 (0.57-0.67)). 22 GPs and 23 PWLE were interviewed. Thematic analysis of the qualitative data generated three over-arching themes: perceived determinants of depression course - social, personal and environmental; relationships and communication; importance of relapse but limited discussion in practice. We developed a framework to guide the ongoing care of people with depression in general practice.Conclusions: Relapse risk prediction for depression is challenging with no existing models suitable for use in primary care. Discussion of relapse risk would be useful but is not routinely offered. There would be benefits to relapse prevention for depression being embedded within primary care consultations, although additional resource would be required. Barriers and facilitators to implementing relapse prevention in primary care would need to be addressed.