Cost-effectiveness of midomafetamine-assisted therapy (MDMA-AT) in chronic and treatment-resistant post-traumatic stress disorder of moderate or higher severity: A health-economic model

PTSD is a common and burdensome psychiatric condition arising after exposure to traumatic events; prevalence estimates in the US reach up to 9.1% among civilians and 50.2% among military personnel. The condition is heterogeneous in onset and course, with certain subgroups (for example, females, those with childhood adversity, comorbid mental illness, low socioeconomic status or poor social support) at higher risk. Chronic PTSD carries substantial clinical comorbidity (depression, anxiety, substance use), increased premature mortality, and a sizeable economic burden: an incremental per-patient cost in the US was estimated at about $19.6k in 2018, higher among military cohorts. Stanicic and colleagues note that existing pharmacotherapies and psychotherapies have important limitations, including need for long-term use, adverse events, treatment discontinuation and substantial non-response. Midomafetamine-assisted therapy (MDMA-AT) combines manually delivered psychotherapy with MDMA sessions and has shown efficacy in Phase III randomized trials versus placebo with therapy. The present study therefore aims to evaluate the cost-effectiveness of MDMA-AT compared to placebo with therapy for adults with chronic PTSD of moderate or higher severity in a US payer context, using trial and real-world economic data to populate a state-transition model.

The investigators developed a health state-transition (Markov-style) economic model and reported according to CHEERS guidance. Model coding and calculations were performed in Microsoft Excel and trial data were analysed with IBM SPSS. Licensed, de-identified individual participant data from two Phase III trials (MAPP1 and MAPP2) and their long-term follow-up study (MPLONG) were used alongside a real-world burden-of-illness (BOI) study to derive inputs for costs, utilities and utilisation. The model population comprised adults with chronic PTSD of moderate or higher severity. The clinical intervention mirrored the trial protocol: three 90-minute preparation psychotherapy sessions, three 8-hour interventional sessions (MDMA or placebo administered with two therapists), and nine 90-minute integration psychotherapy sessions, with interventional sessions spaced about four weeks apart. MDMA dosing was modelled as split doses: 80 mg + 40 mg (120 mg total) in the first session and 120 mg + 60 mg (180 mg total) in sessions two and three. The comparator arm had identical psychotherapy but placebo during the interventional sessions. Health states were No PTSD (loss of DSM-5 diagnosis measured by CAPS-5), Non-severe PTSD (treatment responders who still meet PTSD criteria or patients with moderate PTSD who retain diagnosis), Severe PTSD (non-responders or withdrawals), and death. Treatment response was defined as a ≥10-point decrease in CAPS-5. The structural model time frame was up to 25 years with annual cycles, but the base-case economic horizon reported in the analysis was 5 years. Durability and transition probabilities beyond the 4-month trial endpoint were informed by MPLONG data (mean follow-up 17 months) and extrapolated to 1-year and subsequent cycles using a best-fit (log) approach; where necessary, proportions transitioning were allocated between Non-severe and Severe PTSD states as described. Cost inputs comprised three components: one-time intervention costs (MDMA medication plus therapist time for MDMA-AT, or psychotherapy-only for comparator), ongoing PTSD-related healthcare visit costs (outpatient, inpatient, ED) from the BOI study, and regular PTSD treatments (psychotherapy utilisation from MPLONG, and medication costs from BOI). Unit costs for psychotherapy were scaled from a 90-minute session cost to estimate an 8-hour session cost and adjusted for two therapists per session. Utilities were derived from EQ-5D-5L data collected in MAPP1, MAPP2 and MPLONG and aggregated per health state and arm to produce QALY estimates. Base-case assumptions included a payer perspective, a 5-year horizon, 3.5% annual discounting of costs and utilities, and a willingness-to-pay (WTP) threshold of $150,000 per QALY. Uncertainty was explored with one-way sensitivity analyses (OWSA) varying parameters ±10% and probabilistic sensitivity analysis (PSA) with 5,000 Monte Carlo iterations. A scenario/goal-seek analysis calculated the maximum MDMA medication cost per session consistent with specified WTP thresholds. The primary outcome was the incremental cost-effectiveness ratio (ICER), calculated as incremental cost divided by incremental QALYs between MDMA-AT and placebo with therapy.

In the base-case (payer perspective, 5-year horizon, 3.5% discounting), the ICER for MDMA-AT versus placebo with therapy was $83,845 per QALY, indicating cost-effectiveness at the predefined $150,000 WTP threshold. The per-patient MDMA-AT treatment course cost was $48,376, composed of $36,000 for MDMA medication (assumed $12,000 per MDMA session) and $12,376 for psychotherapy components. The per-session breakdown reported a single MDMA-AT session cost of $16,125 ($12,000 MDMA, $2,357 for the 8-hour psychotherapy component, and $1,768 corresponding to preparation and integration sessions). The comparator intervention cost was $12,376. Total direct PTSD-related costs over 5 years were $64,745 for the MDMA-AT arm versus $33,132 for placebo with therapy, an incremental cost of $31,612 driven largely by the higher intervention cost in the MDMA-AT arm. However, MDMA-AT incurred lower downstream costs: PTSD visit costs were $4,831 versus $7,342 in the comparator, and PTSD treatment expenditures were $11,538 versus $13,414. Utility outcomes favoured MDMA-AT: total QALYs were 3.691 versus 3.314 (increment 0.376 QALY). One-way sensitivity analysis identified utility inputs, particularly the utility assigned to the No PTSD state in the MDMA-AT arm over multiple cycles, as the most influential parameters. For example, decreasing the MDMA-AT utility in the No PTSD category by 10% across cycles 2–25 increased the ICER to $247,773 per QALY, while a 10% increase lowered the ICER to $50,460 per QALY. Varying the same utility only in the first cycle produced ICERs of $101,243 (-10%) and $71,549 (+10%). Probabilistic sensitivity analysis (5,000 iterations) produced mean direct PTSD expenditures of $64,805 (MDMA-AT) and $33,106 (placebo), a mean cost increment of $31,699 (±$5,378) and a mean QALY increment of 0.376 (±0.305), yielding an estimated mean ICER of $84,240 per QALY. MDMA-AT’s ICER was below the $150,000 threshold in 72.12% of PSA iterations. Scenario analysis found that a maximum MDMA medication price of $20,314 per session would yield an ICER of approximately $149,998 per QALY; the corresponding maximums to meet $100,000 and $50,000 WTP thresholds were $14,030 and $7,746 per session, respectively.

Stanicic and colleagues interpret their findings as indicating that MDMA-AT is cost-effective relative to placebo with therapy for adults with chronic PTSD of moderate or higher severity under the base-case assumptions, with an ICER of $83,845 per QALY. They emphasise that model inputs combined individual patient data from two Phase III trials and their long-term follow-up with a real-world BOI analysis to capture intervention costs, downstream healthcare use and quality-of-life changes, aiming to reduce reliance on assumptions used in earlier economic models. The reported MDMA-AT course cost ($48,376) and the QALY gains produced the principal drivers of the favourable ICER. The authors situate their results within prior economic evaluations of MDMA-AT and other PTSD interventions. Earlier models (for example, Marseille et al. and analyses using Phase II data) also suggested MDMA-AT could be cost-effective or dominant over standard care at longer horizons, but those studies used different model structures, smaller samples, more assumptions and limited long-term data. The present study’s use of larger Phase III trial data and BOI inputs is cited as a comparative strength. The investigators also reference broader PTSD economic evaluations which have shown heterogeneity in methods and quality, underscoring the need for rigorous, data-driven economic studies. Key limitations acknowledged include dependence on the quality and generalisability of the BOI study and clinical trial inputs, potential lack of representativeness given sample sizes and heterogeneity of patient characteristics, and the choice of comparator: trials used placebo with therapy rather than an active, real-world standard-of-care comparator. The authors also note the mismatch between the model’s long (25-year) structural horizon and the average MPLONG follow-up of 17 months, which constrains certainty about very long-term outcomes despite reporting a shorter 5-year base-case to improve reliability. Sensitivity analyses showed that quality-of-life (utility) gains were the most influential drivers of ICER estimates, and PSA results were broadly consistent with base-case conclusions. The authors suggest that, given these caveats, MDMA-AT may represent a novel therapeutic option that could reduce disease and economic burden after a single treatment course, while calling for continued data collection to refine long-term estimates.

The study concludes that MDMA-AT was more cost-effective than placebo with therapy for adults with chronic PTSD of moderate or higher severity under the modelled assumptions. Quality-of-life improvements were the inputs with the largest impact on cost-effectiveness, and results were robust across 5,000 probabilistic simulations. The authors suggest MDMA-AT could reduce PTSD-related disease and economic burden following a single course of treatment, while recognising the need for further data to address uncertainties identified in sensitivity analyses and limitations of input sources.