Abstract
Background: Treatment for depression in England utilises a trial-and-error approach with "watchful waiting" to see if an intervention is efficacious. The antidepressant treatment response (ATR) index is an electroencephalography derived biomarker for predicting a clinical response to pharmacotherapy for depression. Objective: The purpose of this study was to systematically review and quantitatively combine studies evaluating the ATR Index, and assess the cost-effectiveness of using the ATR Index in the management of first-line treatments for depression in primary care in England, compared to the current treatment paradigm for depression. Methods: Two meta-analyses of the ATR Index were conducted using a hierarchical regression model and a random-effects model. These results were then entered into a multifaceted decision analytic model with a 2-year time horizon. Two primary paths were developed: one for treatment as usual, and the second utilising the ATR Index to assist in determining treatments. Beginning with a clinical diagnosis of depression, the primary path selects appropriate treatment options based on treatment resistance (or the ATR Index), and then future health states transitions were modelled across 6 Markov models. Model inputs included probabilities derived from meta-analyses of intervention efficacy, survival analysis models for time-to- relapse or remission, and published literature on health care and pharmaceutical resource use and costs. Outcomes were measured in quality-adjusted life years (QALYs) and overall costs of treatment (reported in 2015 GBP). Analyses were performed from the perspective of the NHS (the primary payer and provider of care in England). Probabilistic sensitivity and value of information (VoI) analyses were conducted to test the robustness of results. Results: The expected health gains from standard treatment was 1.15 QALYs compared to 1.17 QALYs with use of the ATR Index. The expected cost per patient was £3 995 for standard care and £3 109 to £3 176 for ATR-based treatment, depending on the meta-analysis model used. Differences in costs were significant, giving an expected incremental cost-effectiveness ratio of less than - £40 910/QALY. Use of the ATR was shown to be cost-effective over current treatment practices with a probability of 97%, though VoI analyses indicate a cost of uncertainty within the model parameters. Conclusions: In this analysis, the use of the ATR Index in the treatment and management of depression appeared to be a cost-effective alternative to current treatment practices, leading to better clinical outcomes and cost savings. However, some uncertainty still exists and further research should be conducted prior to the adoption of the ATR Index into clinical practice.