The Costs and Health Benefits of Expanded Access to MDMA-assisted Therapy for Chronic and Severe PTSD in the USA: A Modeling Study

Post-traumatic stress disorder (PTSD) is a common, disabling psychiatric condition associated with substantial morbidity, elevated healthcare utilisation, and increased risk of suicidal behaviour. While many people remit with or without standard treatments, a sizeable fraction develop chronic and severe PTSD that is treatment-resistant; approved pharmacotherapies have limited effectiveness and tolerability. Recent Phase III trial evidence indicates that MDMA-assisted therapy (MDMA-AT) substantially reduces PTSD symptoms compared with psychotherapy plus placebo, without increasing suicidal ideation or major adverse events in this population. Avanceña and colleagues set out to estimate the potential population-level health gains and economic consequences of expanding access to MDMA-AT for US adults with chronic and severe PTSD. Specifically, the study models three 10-year scale-up targets (25%, 50%, 75% of eligible patients) and compares MDMA-AT to providing standard of care (SoC) to the same number of patients, reporting deaths averted, quality-adjusted life years (QALYs) gained, and incremental costs from a payer perspective.

The investigators developed a decision-analytic, state-transition (Markov) model to simulate a closed cohort of adult patients with chronic and severe PTSD who would be eligible for MDMA-AT per the Phase III trial criteria. The base-case cohort size was approximately 3.52 million eligible patients (range 2.11–4.74 million), derived from national prevalence estimates, an assumed 50% proportion with chronic and severe PTSD, and an exclusion fraction for comorbidities. Three scale-up scenarios were modelled in which the cumulative proportion treated doubles year-to-year until the 10-year target (25%, 50%, or 75%) is reached. The model used annual cycles, a 30-year time horizon (extending beyond the 10-year scale-up), a payer perspective, and a 3% annual discount rate for both costs and QALYs. Programming was done in Microsoft Excel. Health states in the model were asymptomatic, mild, moderate, severe, extreme PTSD, and death, and transition probabilities and long-term trajectories were taken from a previously published Markov trace developed by the study team using Phase III trial data. Health utilities were calculated from EQ-5D-5L scores of Phase III participants and converted to US preference weights. The principal outcomes were deaths averted and QALYs gained; QALYs combine life expectancy and health-related quality of life. Costs included a one-time delivery cost for MDMA-AT and annual medical care costs that vary by PTSD severity. The MDMA-AT microcost estimate was $11,537 (range $8,076–$14,998) per patient and encompassed three 90-minute preparatory psychotherapy sessions, three 8-hour MDMA sessions, and nine 90-minute integration sessions, each delivered one-to-one with two clinicians. Medical care costs for different severity levels were taken from the literature; the base-case assumed SoC costs are embedded in those annual medical care costs. For each coverage target the model computed incremental costs and benefits of MDMA-AT versus SoC over 30 years and calculated incremental cost-effectiveness ratios (ICERs) where applicable. The authors used $100,000 per QALY as the willingness-to-pay threshold and also reported net monetary value (NMV), defined as (QALYs gained × threshold) minus total costs. Uncertainty was explored through one-way sensitivity analyses (varying each parameter across reported ranges), scenario analyses (including more effective SoC and a scenario where MDMA-AT effectiveness declines by 10% per year after five years), and probabilistic sensitivity analysis (PSA) with 10,000 Monte Carlo iterations using pre-specified parameter distributions.

Base-case projections indicated that between about 780,000 and 1.41 million patients would receive MDMA-AT under the 25%, 50%, and 75% targets. Over a 10-year period, the 25% coverage scenario was estimated to avert 43,618 deaths and gain 3.3 million discounted QALYs, the 50% scenario to avert 78,768 deaths and gain over 6.0 million QALYs, and the 75% scenario to avert 106,932 deaths and gain 8.2 million QALYs. For all three targets, the model found that the total costs of expanding access to MDMA-AT were lower than the costs of providing SoC to the same number of patients, meaning each target was “dominant” (i.e., cost-saving while producing greater health benefits). One-way sensitivity analyses showed that varying individual inputs across their reported ranges did not change the central finding of dominance. The parameters with greatest influence on NMV were the model time horizon (examined between 10 and 40 years) and the assumed number of eligible patients. Scenario analyses—reducing MDMA-AT effectiveness, increasing SoC effectiveness, or combining both assumptions—also did not overturn the dominance result. In probabilistic sensitivity analysis across 10,000 simulations, MDMA-AT remained dominant versus SoC on average and was the optimal choice (highest NMV) in 100% of simulations across cost-effectiveness thresholds considered. The extracted text reports the one-time per-patient MDMA-AT cost ($11,537) and that scale-up would require substantial upfront investment in clinician training and service capacity to deliver the intensive protocol.

Avanceña and colleagues interpret their modelled results as indicating that expanding access to MDMA-AT for chronic and severe PTSD could yield substantial population health gains and net savings to the healthcare payer over the modeled horizon. They place these gains in context by comparing the projected deaths averted and QALYs to those attributed to other public health interventions (for example, noting that the projected deaths averted exceed an estimate of suicides prevented by fluoxetine availability and that QALYs gained exceed certain historical estimates for early antiretroviral therapy programs). The authors note that their findings are consistent with prior economic evaluations that used Phase II and Phase III data and likewise found MDMA-AT to be cost-saving or cost-effective under conservative assumptions. At the same time, they emphasise practical and policy prerequisites for realising the modelled savings: substantial initial expenditure to train and certify thousands of therapists given the labour-intensive protocol (15 sessions with two therapists in the Phase III protocol), regulatory actions such as FDA approval and potential descheduling of MDMA, and payer decisions to cover the therapy. The discussion also highlights demand-side considerations—this analysis assumed no financial barriers, high care-seeking, and good adherence—factors that may not hold in practice because of stigma and other access barriers. The authors acknowledge several limitations. Key uncertainties include the true number of eligible patients with chronic and severe PTSD, potential underreporting of PTSD in population surveys, and the lack of precise prevalence estimates for comorbidities that would render patients ineligible. Modelling a closed prevalent cohort likely underestimates total benefits because incident cases were excluded. The payer perspective omitted societal costs and benefits such as productivity, caregiver burden, travel, and patient time. The model relied substantially on inputs and structural assumptions from a prior cost-effectiveness study, including an assumption that patients receiving SoC experience no long-term improvement; while supported by evidence on treatment resistance in chronic PTSD, this remains uncertain. Durability of MDMA-AT benefit is another acknowledged uncertainty—limited follow-up suggests sustained benefit for several years but longer-term effectiveness is unknown. Finally, generalisability is constrained by the Phase III trial population and its short-term efficacy data.

The study concludes that MDMA-assisted therapy has the potential to meaningfully improve care for people with chronic and severe PTSD in the USA: expanded access could save lives, increase quality-adjusted life years, and reduce healthcare costs under the modelled assumptions. The authors emphasise that the precise magnitude of benefits and savings depends on the number of eligible patients, long-term effectiveness, and implementation factors including regulatory approval, workforce capacity, and payer coverage.