Cost-effectiveness of genetic and clinical predictors for choosing combined psychotherapy and pharmacotherapy in major depression
AbstractBackgroundPredictors of treatment outcome in major depressive disorder (MDD) could contribute to evidence-based therapeutic choices. This study tested the cost-effectiveness of a pharmacogenetic and clinical predictive model (PGx-CL-R) vs a clinical risk (CL-R) predictive model in guiding the assignment of combined pharmacotherapy and psychotherapy vs pharmacotherapy in MDD.MethodsWe hypothesized that the prescription of combined treatment, a strategy with evidence of increased efficacy vs pharmacotherapy, may be optimized based on the baseline risk of pharmacotherapy resistance, estimated through PGx-CL-R or CL-R predictive models. Both strategies were compared to standard care (ST, pharmacotherapy to all subjects). Treatment effects, costs and utilities (quality adjusted life years, QALYs) were based on the literature and included in a three-years Markov model.ResultsCL-R was cost-effective compared to PGx-CL-R, with ICER (incremental cost effect ratio) of £2341 (CL-R) and £3937 (PGx-CL-R) per QALY compared to ST. PGx-CL-R had similar or better ICER compared to ST only when 1) the cost of genotyping was £100 per subject or less or 2) the sensitivity of the PGx-CL-R test was at least 0.90 and the specificity at least 0.85. CL-R had ICER of £3664 and of £4110 when the CL-R model was tested in two independent samples.Limitationslack of validation in clinical trial.ConclusionsPrediction of pharmacotherapy resistance according to clinical risk might be a cost-effective strategy if confirmed on large samples from the general population. Combined treatment with drugs having a very good tolerability profile could be a cheaper alternative to psychotherapy.