Mental disorders (MDs) represent a significant global health burden, impacting one in five individuals and causing substantial economic costs (estimated at US$2.5 trillion in 2010 globally and projected to double by 2030 in the EU). Despite effective interventions, a large treatment gap exists due to attitudinal and structural barriers. Internet- and mobile-based interventions (IMIs) offer a promising solution, providing anonymous, accessible, and effective treatment. While IMIs may offer lower overall expenditure due to low marginal costs and reduced travel/therapist time, existing evidence on their cost-effectiveness is inconclusive and outdated, with previous reviews including only a small number of studies and often focusing only on iCBT. This systematic review aims to provide a comprehensive and updated overview of the cost-effectiveness of IMIs across a range of mental health disorders and symptoms, considering methodological quality and employing a rigorous analysis of full health economic evaluations.

Several previous systematic reviews have explored the cost-effectiveness of IMIs for MDs. However, these reviews suffered from limitations including small sample sizes (e.g., n=3, n=4, n=12, n=1, n=5), obsolete data (with the latest primary study dating back to 2016), broad definitions of IMIs, and lack of full health economic evaluations. Furthermore, research on economic evaluations for common treatment options (various psychotherapies and pharmacological interventions) for depression and anxiety disorders is limited. This lack of comprehensive and high-quality evidence necessitates this current review to provide a more robust assessment of IMIs' cost-effectiveness.

This systematic review followed PRISMA and guidelines for systematic reviews of economic evaluations. A comprehensive search across MEDLINE, PsycINFO, CENTRAL, PSYNDEX, and the NHS Economic Evaluations Database was conducted for trial-based economic evaluations published before May 10, 2021. Inclusion criteria included studies with participants diagnosed with MDs or symptoms, psychological interventions delivered online, a control group (another psychological intervention, TAU, WLC, or AC), economic evaluation estimates (CEA, CUA, CBA, or CMA), and RCT design, with full texts accessible in English or German. Two independent reviewers screened titles and abstracts, then full texts for eligibility. The Consensus on Health Economic Criteria (CHEC) checklist and Cochrane Collaboration's tool for assessing risk of bias were used to assess quality and risk of bias. Data on participants, study design, intervention, economic outcome measures, cost-effectiveness estimates, and currency conversion to 2020 Pound Sterling were extracted. IMIs were considered cost-effective if dominant, costs per QALY < £30,000, or probability of cost-effectiveness at a WTP of £0 was ≥80% for disease-specific outcomes.

The review included 36 economic evaluations from 32 studies (N=10,083). Studies targeted various disorders including depression (MDD and depressive symptoms, n=15), anxiety (n=7), OCD (n=4), sleep disorders (n=2), stress (n=2), PTSD (n=1), and suicidal ideation (n=1). Most studies (n=16) conducted both CEA and CUA; others focused solely on CUAs (n=10) or CEAs (n=4). Three also conducted CBA. The majority (n=24) evaluated guided IMIs, primarily based on iCBT. In 14 CEAs, IMIs were the dominant treatment option, with five showing a high probability (≥80%) of cost-effectiveness at a WTP of zero. Across all CEAs, the probability range at WTP zero varied from 0% to 95%. For cost-utility analyses, many interventions were cost-effective, either dominant or below the £30,000 per QALY threshold. Applying an 80% probability threshold at £30,000, 11 IMIs were deemed cost-effective. Cost-benefit analyses from the employer's perspective (n=3) showed positive net benefits. Overall, CHEC quality was good in most studies (average 85%, range 56–100%), with risk of bias assessment predominantly good (n=22). Guided IMIs for MDD and anxiety were likely cost-effective. Cost-effectiveness of unguided interventions and interventions for OCD, PTSD, sleep disorders and stress was less clear.

This review provides updated and more robust evidence on the cost-effectiveness of IMIs for mental health, expanding upon previous literature. The findings support the cost-effectiveness of guided IMIs for depression and anxiety, consistent with prior research. Importantly, this review also includes data on under-researched disorders such as OCD, PTSD, sleep disorders and stress. However, the high heterogeneity across studies hinders strong conclusions. Methodological differences in study design, populations, comparators, and cost analysis impact comparability. The inconsistent operationalization of societal and intervention costs potentially led to underestimation of the latter. The lack of universally accepted WTP thresholds for disease-specific measures and the use of QALYs introduces further challenges in interpreting cost-effectiveness. While many interventions appear cost-effective based on QALYs, additional criteria should be considered, such as disease burden and implementation feasibility.

Guided iCBTs for anxiety and MDD are likely cost-effective. IMIs for insomnia, suicidal ideation, and stress show potential, but evidence for OCD is weaker. High heterogeneity across studies prevents robust conclusions. Future research should focus on unguided and preventive IMIs, under-researched disorders, longer follow-up periods, active control groups, diverse geographical locations, and should adhere to established health-economic guidelines for improved comparability and decision-making in healthcare.

High heterogeneity across studies limits the strength of conclusions. Methodological differences in study designs, populations, comparators, and cost analysis affected comparability. The inconsistent operationalization of societal and intervention costs may have led to underestimation of intervention costs. The lack of universally accepted willingness-to-pay thresholds for disease-specific measures and the arbitrary threshold for QALYs present challenges in interpretation. Generalizability is limited by the focus on Western, high-income countries, and the inclusion of only English and German language studies.