PP71 Long-Term Evaluation Of Broad Mental Health Interventions: A Review

Abstract

Introduction:

Interventions and services for people with mental health problems can have broad remits: they are often designed to treat people with a variety of diagnoses. Furthermore, addressing mental health problems can have long-term implications for economic, social, and health outcomes. This represents a challenge for health technology assessment, for which long-term trial data can be lacking. In this review, we sought to identify how analysts have tackled this problem. We reviewed the methods used to extrapolate costs and outcomes for the purpose of economic evaluation where long-term trial data are not available.

Methods:

We conducted a systematic review of the medical and economic literature evaluating long-term costs and outcomes for mental health interventions and services designed to treat or prevent more than two mental health conditions. We searched key databases including MEDLINE, Embase, PsycINFO, CINAHL, and EconLit. Two authors independently screened citations. Articles were excluded if they reported within-trial analyses or employed a time horizon of less than 5 years.

Results:

The search identified 829 unique records. No papers could be included in the review.

Conclusions:

This review highlights the lack of research and understanding available to inform the appraisal of broad mental health interventions. In light of our findings, we consider the reasons for this lack of information and review relevant literature on the subject. Potential barriers to research in this context include: (i) challenges in understanding the value of broad mental health services, such as the mental and physical health nexus, intersectoral costs and benefits, and interpersonal impacts, (ii) methodological difficulties, such as data availability, patient heterogeneity, and the challenge of extrapolation, and (iii) parity of esteem. We make recommendations for resolving this problem with regard to funding, data collection, modelling methods, and outcome measurement.