Economic evaluation of internet-based psychological interventions: a scoping review of methodological choices.

Abstract

Objectives: Internet-based psychological interventions hold promise for cost-effectiveness, yet their evaluation lacks standardisation, potentially leading to methodological discrepancies and inconclusive results. This study aims to conduct a scoping review of economic methods used when evaluating these interventions.

Methods: Articles published between January 2015 and December 2020 were retrieved from PubMed, Embase, Cochrane Library, CINAHL, ECONLIT, and PsychINFO. Two reviewers independently screened titles, abstracts, and full texts of relevant publications, and extracted prespecified data. Outcomes related to intervention characteristics, comparators, perspective, time horizon, costs, benefits, economic endpoints, and uncertainty analysis methods were retrieved and synthetised narratively.

Results: We identified 703 references and included 85. Seventy-four included a cost-utility analysis and 58 a cost-effectiveness analysis, with 47 carrying out both. The comparator was treatment as usual in 52 studies (61.2%) but varied widely across studies. A societal perspective was adopted in 60 studies, supplemented by a healthcare perspective in half. Time horizon was one year or less in 68 articles (80.0%). Intervention costs (71/85 studies) predominantly covered delivery costs (45/71 studies), while development and promotional costs were infrequently considered (respectively 14 and 5/71 studies). Interventions' reach, opportunity costs, user engagement and equity issues were rarely addressed. Key factors influencing cost-effectiveness included perspective, time horizon, costs included, and methods for handling missing data.

Conclusions: Assessment of cost-effectiveness in internet-based psychological interventions shows variability, potentially affecting efficiency evidence. Conventional methods are often favoured overlooking digital tools' specificities. Tailored guidelines for such evaluations could be helpful for standardised and reliable evidence.