From Stigma to Signal: The Executive Order, CNPV Vouchers, and What Comes Next for Psychedelics

Authored by

The recent Executive Order, followed by the FDA’s issuance of Commissioner’s National Priority Vouchers (CNPVs), changes the pace of psychedelic drug development without lowering the bar. By compressing review timelines from the traditional 10–12 months to as little as one to two months, the agency has made regulatory acceleration tangible.

Faster review does not make complex therapies simpler. It concentrates the evidentiary burden and raises the bar on how well these therapies are understood, studied, and executed before they ever reach submission.

At MMS, we have spent decades working at the intersection of neuroscience, complex trial design, and regulatory execution across pain, neurodegeneration, and psychiatry. One consistent reality has defined this space: the hardest diseases to study are often the ones that matter most.

The Executive Order reflects a broader shift in how complex therapies are evaluated. Prioritization of breakthrough-designated programs, expanded access pathways, and greater reliance on real-world data point to a system that is adapting, not relaxing.

In CNS, the constraint has never been interest. It has always been execution and the ability to adapt regulatory approaches to inherently complex conditions.

Heterogeneity Is the Core Challenge: Implications for Development, Access, and Outcomes

Mental health disorders are not uniform diseases. They are heterogeneous syndromes shaped by biology, environment, and lived experience. Even within treatment-resistant depression, definitions vary widely across trials, including number of failed treatments, duration, and response thresholds. Remission itself is inconsistently defined, limiting comparability across studies. (McAllister-Williams et al., 2020, British Journal of Psychiatry).

Yet clinical trials often exclude these realities, and that decision has direct consequences across the product lifecycle. When study populations fail to reflect the heterogeneity and comorbidity seen in real-world patients, early efficacy signals may appear cleaner, but they often come at the cost of delayed time to market as regulators request additional data to address gaps in generalizability, safety, and durability. The impact extends further into reimbursement, where payers increasingly demand evidence of effectiveness in the populations they actually serve, including those with concomitant conditions, polypharmacy, and functional impairment. Without this, coverage becomes more restrictive, access is limited, and adoption slows.

Ultimately, what begins as a scientific simplification becomes a commercial constraint, creating misalignment between trial outcomes and real-world performance. In CNS, where variability is inherent, designing trials that better reflect clinical reality is not just good science. It is essential for regulatory efficiency, payer acceptance, and long-term success.

Comorbidity Defines the Real Patient Population: A Regulatory Challenge for Generalizability and Decision-Making

In CNS development, comorbidity is not a confounding variable. It is the clinical reality regulators must ultimately evaluate.

Patients entering real-world care pathways rarely present with isolated conditions. PTSD coexists with substance use disorder. Depression overlaps with chronic pain. Neurodegenerative diseases carry layered psychiatric burden. TBI introduces cognitive, behavioral, and emotional complexity. These interactions directly influence treatment response, safety profiles, and durability of effect.

Yet clinical development programs often narrow eligibility to reduce variability and improve internal validity. While methodologically defensible, this creates a fundamental tension at the point of regulatory review. Traditional regulatory paradigms are built around relatively well-defined populations, prior treatment exposure, and clear lines of treatment failure. In CNS, those boundaries are far less distinct and the evaluation becomes a shared responsibility between sponsors and regulators.

Regulators are not in a position to impose overly rigid models on inherently heterogeneous conditions. Instead, they must interpret data generated in controlled settings and extend those findings to populations that are broader, more complex, and less predictable. A systematic review showed that PTSD trials frequently exclude patients with substance use disorder or suicidality, effectively limiting generalizability before enrollment even begins. (Varker et al., 2021, Journal of Affective Disorders Reports).

The consequence is a more iterative and cautious decision-making process. Agencies may require additional data, narrow initial labels, or rely more heavily on post-marketing evidence to bridge the gap between trial populations and real-world use. In many cases, this also translates into the implementation of Risk Evaluation and Mitigation Strategies (REMS), particularly for therapies with complex safety profiles or delivery models. REMS programs can include certified sites, trained providers, controlled administration settings, and ongoing patient monitoring, effectively becoming an extension of the clinical trial framework into the commercial setting. While necessary to ensure safe use, these requirements introduce additional operational complexity and can influence scalability, access, and adoption if not anticipated early in development.

For emerging modalities such as psychedelic therapies, this challenge is amplified. These are not simply pharmacologic interventions. They are integrated treatment models where prior treatment history, patient context, and comorbidity all influence outcome.

Regulators must be brought along with the clinical reality, not forced into legacy frameworks. Development strategies that anticipate this by incorporating broader populations, generating complementary real-world evidence, and addressing variability upfront are better positioned to support both regulatory confidence and successful translation into practice.

Endpoints: A Persistent Structural Gap with Regulatory and Commercial Implications

Unlike many other therapeutic areas, CNS continues to operate without consistent, universally accepted endpoints across several key indications. This is not a technical limitation. It is a structural challenge that affects how therapies are evaluated, approved, reimbursed, and ultimately adopted in practice.

In opioid use disorder, the absence of a single, accepted primary endpoint illustrates the issue clearly. Clinical trials variably rely on abstinence, percent-negative urine drug screens, retention in treatment, time to relapse, or broader functional outcomes. Each captures a different dimension of disease, but none fully represents recovery. The NIDA Clinical Trials Network has explicitly highlighted this heterogeneity and the lack of a gold-standard endpoint, noting that outcome selection remains inconsistent across studies and stages of the treatment cascade (Kiluk et al., 2019, Journal of Substance Abuse Treatment). For regulators, this creates ambiguity in interpreting clinical benefit. For payers, it complicates value assessment. And for sponsors, it introduces risk in both trial design and commercial strategy.

A similar challenge exists in PTSD, where endpoint selection reflects a tension between clinical rigor and real-world relevance. The Clinician-Administered PTSD Scale, or CAPS-5, remains the gold standard for regulatory trials, offering structured, clinician-driven assessment. However, it is resource-intensive, subject to inter-rater variability, and difficult to scale outside controlled settings. In contrast, the PTSD Checklist, or PCL-5, is widely used in practice and captures patient-reported outcomes more efficiently, but it introduces greater subjectivity and variability. Trials often incorporate both measures, but discordance between clinician-rated and patient-reported outcomes can complicate interpretation of treatment effect, particularly in therapies where expectancy and experiential factors play a role (Weathers et al., 2018, National Center for PTSD; Bovin et al., 2016, Psychological Assessment. This is especially relevant for emerging modalities such as psychedelic-assisted therapies, where functional unblinding and patient perception are inherent challenges.

More broadly across CNS, endpoint fragmentation reflects a deeper issue: the difficulty of defining meaningful, durable change in complex neuropsychiatric conditions. Symptom reduction alone may not capture functional recovery, quality of life, or long-term stability. As a result, regulators are increasingly required to interpret multidimensional datasets that include symptom scales, functional measures, and safety signals, often without clear consensus on which should carry the greatest weight. This ultimately converges on a fundamental dilemma: how to assess risk-benefit when benefit itself is variably defined, measured across imperfect instruments, and not always aligned with real-world outcomes.

Endpoint variability introduces friction across regulatory review, limits cross-trial comparability, complicates payer decision-making, and creates uncertainty in clinical adoption. It also reinforces the need for complementary evidence strategies, including real-world data and longitudinal follow-up, to contextualize trial outcomes.

In this environment, endpoint strategy cannot be deferred. It is a defining decision in development. Programs that rely on precedent or attempt to force legacy endpoint frameworks onto inherently heterogeneous CNS conditions risk generating evidence that is technically acceptable but fundamentally incomplete. In many cases, precedent is not just insufficient. It is misleading.

Regulators are not evaluating these therapies against historical constructs alone. They are assessing whether the evidence reflects how these diseases actually present and progress in the real world. Endpoint strategies that fail to account for this reality create avoidable friction at review and limit the ability to demonstrate meaningful benefit.

The path forward is not incremental adjustment. It requires intentional, adaptable design. Programs that align endpoints with clinical reality, regulatory expectations, and real-world applicability from the outset are the ones that will generate evidence that is not only approvable, but credible, defensible, and ultimately usable in practice.

Retention: Interpreting the Signal Correctly

In psychiatric trials, attrition is not noise but part of the signal. A meta-analysis of PTSD treatments reported average dropout rates of approximately 21%, with even greater variability in more complex and comorbid populations (Varker et al., 2021, Journal of Affective Disorders Reports). More than just an operational challenge it is reflective of disease burden, treatment tolerability, and the realities of sustained engagement in psychiatric conditions.

Patients with the highest clinical burden are often the least likely to complete trials, introducing systematic bias into observed outcomes. As a result, the dataset that remains is often enriched for those who are more stable, more adherent, or more responsive, creating an incomplete picture of treatment effect. At the same time, attrition complicates interpretation of durability. When patients disengage early, it becomes difficult to determine whether outcomes reflect true lack of effect, tolerability limitations, or external factors influencing adherence. This uncertainty carries directly into regulatory review and payer evaluation.

In other words, retention is inseparable from efficacy. It is a composite reflection of whether a therapy can be sustained in practice, not just demonstrated under controlled conditions. High attrition raises questions about the robustness and generalizability of the evidence base and increases the risk that real-world performance will diverge from trial results. Addressing retention through study design, patient support strategies, and longitudinal follow-up becomes imperative. Retention becomes a critical signal of clinical efficacy and directly informs the interpretation of risk-benefit.

Remission and Durability: The Illusion of Early Signal

Many therapies work for a while. Short-term improvements are common across psychiatric and neurologic indications, particularly in controlled trial environments. Placebo response, patient expectation, site engagement, and structured follow-up all contribute to early signal. But sustaining that effect over time remains far less predictable. Across depression and other CNS disorders, relapse rates remain high even after initial response, and durability of effect is often inconsistent outside tightly managed trial settings (Rush et al., 2006, American Journal of Psychiatry; Pigott et al., 2010, Psychotherapy and Psychosomatics).

This creates a fundamental challenge for regulators. Approval decisions are often anchored in relatively short-duration studies, yet the diseases themselves are chronic, relapsing, and influenced by comorbidity, adherence, and environmental factors. The question is not whether a therapy works, but whether it continues to work in a way that is meaningful over time. Without clear, consistent definitions of remission and sustained response, regulators are left to interpret durability through incomplete or variable data.

For emerging therapies, including psychedelic-assisted interventions, this question becomes even more important. Early data may show substantial improvements, but the durability of those effects and the need for re-treatment, maintenance strategies, or adjunctive care remain areas of active uncertainty. This often translates into increased regulatory scrutiny, post-marketing requirements, and a greater reliance on real-world evidence to confirm long-term benefit.

The implication is clear. Demonstrating initial efficacy is necessary but not sufficient. Durability of effect is where therapies ultimately succeed or fail. Until durability is consistently defined, measured, and demonstrated, it will remain a central hurdle in both regulatory decision-making and real-world adoption.

Why Psychedelics, Why Now

The renewed focus on psychedelic therapies reflects both unmet need and emerging data, but it also reflects a broader shift in how mental health is understood and treated. For decades, these therapies sat at the intersection of scientific curiosity and social stigma. That stigma is now eroding, driven by a combination of patient need, advancing neuroscience, and a growing recognition that existing treatment paradigms are not sufficient for many individuals with serious psychiatric conditions.

At the same time, use of these therapies has not been absent. It has largely existed outside formal healthcare systems. Patients have sought access through underground or unregulated settings, often without standardized protocols, trained oversight, or safety monitoring. Bringing these therapies into regulated clinical development is therefore not just about innovation but about enforcing safety, consistency, and accountability. Moving from informal use to structured, evidence-based delivery models allows for controlled dosing, defined therapeutic frameworks, and systematic evaluation of both benefit and risk.

From a data perspective, late-stage programs have begun to demonstrate meaningful clinical signals. Psilocybin-based therapies have shown benefit in treatment-resistant depression, and MDMA-assisted therapy has demonstrated significant improvements in PTSD symptoms and functional outcomes, with a substantial proportion of patients no longer meeting diagnostic criteria (Mitchell et al., 2021; 2023, Nature Medicine). These findings are notable not only for their magnitude, but for the populations studied, many of whom have not responded to existing treatments.

These therapies also represent a fundamental shift in mechanism and delivery. They are not purely pharmacologic interventions. They combine drug effect with structured therapeutic engagement, targeting underlying cognitive and emotional processing rather than symptom suppression alone. This introduces both opportunity and complexity.

The opportunity lies in potentially transformative outcomes. The complexity lies in the need to standardize delivery, manage variability, and generate evidence that captures both the pharmacologic and experiential components of treatment. This is further compounded by the regulatory reality that agencies approve products, not the practice of medicine. While regulators can define labeling, safety frameworks, and in some cases REMS requirements, they do not control how therapies are ultimately delivered in clinical practice. For integrated treatments such as these, where outcomes are influenced as much by the therapeutic context as the molecule itself, this creates an inherent challenge.

As a result, the burden shifts to development programs to anticipate real-world variability and design evidence packages that are robust enough to withstand both regulatory scrutiny and clinical implementation. The challenge is not only to demonstrate efficacy under controlled conditions, but to ensure that those outcomes can be reproduced safely and consistently once the therapy moves beyond specialized settings.

The timing, therefore, is not accidental. Scientific signal, patient demand, and policy momentum are finally aligning. The challenge now is to translate that momentum into rigorous, scalable, and safe therapeutic models that can withstand regulatory scrutiny and deliver consistent outcomes in real-world practice.

The Regulatory Reality: Raising the Bar

The executive order should be viewed as an accelerator, not a conclusion. It signals urgency and intent, but it does not change the fundamental requirements for approval. If anything, it sharpens the focus on what constitutes acceptable evidence in a space that is already under close scrutiny.

Recent regulatory interactions have made this clear. Promising efficacy signals, even when clinically meaningful, are not sufficient on their own. They must be supported by data that is reproducible, interpretable, and consistent across sites and populations. Challenges such as functional unblinding, variability in therapist delivery, safety interpretation, and overall data integrity are not unique to psychedelic therapies, but they are amplified in this class because of the integrated nature of the intervention. Regulators are therefore not just evaluating whether these therapies work, but whether the evidence can withstand variability in how they are delivered and experienced.

This is further complicated by a shift in where innovation is coming from. Much of the progress in this space is being driven by emerging biotech rather than traditional large pharmaceutical companies. These organizations are advancing novel mechanisms alongside therapy-assisted delivery models and integrated treatment paradigms. That changes the development equation. Success is no longer defined by the molecule alone, but by the consistency of execution across therapist training, site performance, protocol fidelity, and the integrity of data generated across both behavioral and pharmacologic components. This is a different risk requiring a different level of discipline.

At the same time, traditional clinical trial models are being pushed to their limits. Static designs are often insufficient to capture the variability inherent in CNS conditions and integrated interventions. We are seeing a shift toward more adaptive approaches, including the use of synthetic and external controls, real-time data capture through digital endpoints, and Bayesian or platform trial designs that allow for flexibility in heterogeneous populations. Increasingly, real-world evidence is being embedded earlier in development, not as a retrospective exercise, but as part of a continuous evidence generation strategy. These approaches are becoming necessary to generate data that reflects both controlled and real-world settings.

And that evidence does not end at approval. It extends into how these therapies are controlled, delivered, and monitored in practice. For psychedelic therapies in particular, approval is only the beginning. REMS will likely play a central role, with requirements for certified sites, trained providers, and controlled administration environments. Post-marketing commitments will be essential to establish long-term safety, durability of effect, and performance in broader populations. In this context, real-world evidence becomes infrastructure, not an adjunct, supporting label expansion, payer acceptance, and ongoing assessment of benefit and risk.

This raises the bar in a very practical way. In psychiatry, effect size alone is no longer the differentiator. Regulators are looking for durability of response, consistency across diverse patient populations, and evidence that outcomes can be replicated beyond highly controlled environments. They are also assessing whether the delivery model itself can scale safely, particularly when treatment depends on trained providers, structured settings, and patient engagement over time. This is where many programs encounter challenges, not because the signal is absent, but because the evidence does not fully capture the complexity of real-world use.

Addressing this requires a different level of discipline in development. Study design, site training, protocol adherence, and oversight are not mere operational details. They are central to the evidentiary package. At MMS, we have seen this firsthand. Strengthening execution frameworks, standardizing delivery, and proactively addressing boundary-related risks are essential to ensuring that the data generated is credible, defensible and withstands scrutiny to translate into consistent outcomes in practice.

The pathway forward for psychedelic therapies, and for CNS more broadly, will not be defined by initial signal alone. It will be defined by the ability to generate rigorous, reproducible evidence that holds up across settings, populations, and time. That is the standard regulators are applying. And it is the standard the field and sponsors will need to meet even as broader structural constraints such as scheduling continue to shape how quickly this science can translate into practice. This is a topic I will return to in my next blog.