Costs and cost-effectiveness of interventions based on third generation therapy models to treat symptoms of depression and anxiety: A systematic review and meta-analysis

<b>Background: </b>Depression and anxiety are the most prevalent mental health problems globally, and while there are currently a number of effective interventions for both mental health problems, information on the cost-effectiveness and costs associated with these interventions is limited. Especially, in recent interventions such as third generation therapies, there is limited information on the cost-effectiveness and costs associated with these interventions.<b>Objective: </b>This study aimed to evaluate the costs and cost-effectiveness of third-generation therapies for treating depressive and anxiety symptoms.<b>Methods: </b>We conducted a systematic review and meta-analysis, searching SCOPUS, Web of Science, PsycINFO, EMBASE, and MEDLINE from 1980 to March 12, 2024. Randomized controlled trials reporting costs or cost-effectiveness of third-generation therapies for depression or anxiety were included. Two independent reviewers screened studies, extracted data, and assessed risk of bias using the ECOBIAS checklist. Meta-analysis of quality-adjusted life years (QALYs) was conducted using Hedges' g and random-effects models. Costs were standardized to US dollars and adjusted for inflation.<b>Results: </b>Twenty studies met inclusion criteria, with 10 suitable for meta-analysis. Meta-analysis showed no significant difference in QALY improvement between third-generation therapies and control conditions (Hedges' g = 0.015, 95% CI: -0.001 to 0.032; P=0.070). Moderate heterogeneity was observed (I²=47.8%, P=0.032). Subgroup analyses by therapy type (Mindfulness-Based Cognitive Therapy, Behavioral Activation, Acceptance and Commitment Therapy, and Problem-Solving Therapy) also yielded non-significant results. High risk of bias was observed in 50% of studies due to narrow perspective, in 45% due to omission of cost measurement, and in 46% due to insufficient information on discounting. Publication bias was detected (Egger's test: P = 0.011). Our study found no significant difference in the average cost per person between receiving an intervention based on third-generation therapies (M = $2,494.6; SD = $7,090.9) and receiving a control condition (M = $1,912.0; SD = $2,081.3) (t = 0.253; p=0.802).<b>Conclusions:</b> Our study found that interventions based on third-generation therapies do not entail significant additional costs compared to control conditions. However, the interventions based on third-generation therapies did not improve the QALYs of the participants more than the control group.<br>