Cost-effectiveness of psilocybin-assisted therapy for severe depression: exploratory findings from a decision analytic model

Major depressive disorder (MDD) causes substantial personal and societal burden and a sizeable minority of patients do not achieve sustained recovery after standard treatments. Interest has resurged in psychedelic-assisted therapies, and psilocybin-assisted psychotherapy (PAP) has shown promising results in several small, short-term studies for moderate to severe depression, including some work in populations with comorbid medical or psychiatric conditions. PAP differs from most pharmacotherapies because it is typically administered only once or twice per episode and is delivered together with intensive psychotherapeutic support; both the staffing intensity and the regulatory status of psilocybin (Schedule 1 in the UK) affect its likely cost profile. Mccrone and colleagues set out to provide a first exploratory economic evaluation of PAP versus three comparators: conventional antidepressant medication (escitalopram), cognitive behavioural therapy (CBT), and the combination of medication plus CBT. The study used a short-term decision analytic model to estimate healthcare and societal costs and quality-adjusted life years (QALYs) over six months, and applied NICE-style incremental cost-effectiveness thresholds (£20 000–£30 000 per QALY) to assess whether PAP could be considered cost-effective under plausible scenarios of drug price, therapist support intensity, and clinical effectiveness.

The investigators constructed a decision-tree model in Microsoft Excel with a six-month time horizon, comprising a six-week acute phase followed by a 20-week follow-up. The simulated cohort had difficult-to-treat depression and could receive one of four options: PAP, antidepressant medication alone (escitalopram), CBT alone, or combined antidepressant plus CBT. After initiation, patients could complete or drop out of treatment; completers could achieve remission, response without remission, or no response. During follow-up, those improved could relapse or remain improved. Clinical probabilities (completion, remission, response, dropout) were largely taken from an existing trial comparing psilocybin with an SSRI (Carhart-Harris et al.), supplemented by relative risks for CBT and combination therapy from other sources and relapse rates from guideline and case-note data. Psilocybin dosing in the base case was 25 mg orally on two occasions; escitalopram was modelled at 10 mg/day through follow-up (the trial had dose escalation but this was judged to have minor cost impact). CBT was represented as a ten-session course delivered by an accredited therapist. Costs included the drug price (varied in the model from £400 to £2000 per person in £100 increments), therapist time for PAP (38 therapist hours in the trial multiplied by a unit cost based on a Band 7 clinical psychologist: £58 per hour plus 30 minutes non-direct time giving £87 per hour), CBT (£1050 per course), and drug costs from the BNF. Healthcare resource use associated with remission, relapse and other states was adapted from a UK case-note review; lost employment costs were included in the societal perspective by applying a multiplier of two to healthcare costs, based on a prior study of treatment-resistant depression. Utility values for health states were taken from published sources cited by the authors. The analysis compared PAP against each comparator separately, reporting incremental costs, incremental QALYs and, where appropriate, incremental cost-effectiveness ratios (ICERs). Deterministic sensitivity analyses varied key parameters including therapist support (±25% and ±50%), psilocybin price, remission probabilities, relapse rates, CBT costs, and the lost-work multiplier. Probabilistic sensitivity analysis was not performed.

Over six months, the model produced the highest expected QALYs for PAP (0.310), followed by combined medication plus CBT (0.287), CBT alone (0.283), and medication alone (0.276). Expected healthcare costs for PAP ranged approximately from £6255 to £7775 across the modeled psilocybin price range (£400–£2000 per person). Comparator healthcare costs in the results section were reported as roughly £3700 for medication alone, £4405 for CBT alone, and £4351 for combined therapy. Under the baseline assumption of high therapist support (the trial-based 38 hours at Band 7 cost), PAP was more costly than comparators and generated ICERs above £30 000 per QALY for all comparisons across the psilocybin price range. Including lost-employment costs (societal perspective) improved PAP’s apparent value: at a psilocybin cost of £400 per person the ICERs fell below £30 000 when PAP was compared with medication alone or CBT alone, and at £600 per person the ICER versus medication alone also fell under £30 000. Therapist-support costs were a key driver. Reducing therapist costs by 50% produced ICERs below £30 000 in several comparisons (for example, PAP versus combined therapy when psilocybin cost = £600; versus medication alone when psilocybin cost = £400; versus CBT alone when psilocybin cost = £800). From a societal perspective and with therapist support reduced by 50%, the authors report that PAP dominated other options (lower costs and more QALYs) up to a psilocybin cost of £1200 per person; a 25% reduction in therapist cost made PAP dominant at a £400 drug price. Changing the assumed remission probability had marked effects: increasing the psilocybin remission probability by 50% (from 0.57 to 0.86 in the extracted text) improved cost-effectiveness substantially, with the model indicating much higher price thresholds under which PAP remained under £30 000 per QALY (the extracted text reports an upper threshold of up to £8000 per person for one comparison). Sensitivity analyses otherwise indicated limited sensitivity to most parameters except therapist support, remission probability and the perspective taken (healthcare versus societal). The model used deterministic two-way sensitivity analyses rather than probabilistic methods.

Mccrone and colleagues interpret the findings as indicating that PAP has potential to be cost-effective for severe depression under specific conditions, notably when therapist-support costs are substantially lower than those used in the trial setting and when the price charged for psilocybin per patient is at the lower end of the plausible range. They note that the current trial evidence base is small and short-term, and that PAP delivery in routine practice would likely use fewer therapist hours and possibly lower-grade staff, which would materially improve cost-effectiveness. The authors emphasise the importance of perspective: from a narrow healthcare viewpoint there are few scenarios in which PAP is cost-effective at trial-level support costs, whereas including productivity (lost employment) makes PAP substantially more favourable. They also highlight that drug price and the true sustained remission rate after PAP are influential; higher remission and lower relapse rates would improve value considerably. Practical and regulatory factors are discussed: Schedule 1 status raises current research and handling costs, and commercial pricing could be higher than research costs cited in the model. Key limitations acknowledged by the investigators include heavy reliance on a small number of trials with short follow-up and mixed patient populations, use of some dated cost inputs, the simplicity of a decision-tree model versus potentially more appropriate long-term models (Markov or discrete-event simulation) given depression’s chronicity, and the absence of probabilistic sensitivity analysis. They also note that adverse events were not explicitly costed, direct trial comparisons with CBT or combined therapy were lacking, and several parameter estimates were uncertain or derived from non-ideal sources. The authors therefore call for longer-term follow-up data, trial-based economic evaluations comparing different levels of therapist support, and improved evidence on who should be prioritised to receive PAP (for example, treatment-resistant groups) before firm policy conclusions can be drawn.

This exploratory modelling study provides preliminary evidence that psilocybin-assisted psychotherapy could be cost-effective for severe depression in some scenarios, particularly from a societal perspective and when therapist support requirements and per-patient drug prices are reduced. The authors conclude that better data on long-term outcomes, optimal levels of psychological support, and the target patient groups for PAP are needed to strengthen the economic evidence base and inform policy or commissioning decisions.