From Signal to System: The Evidence Questions Psychedelic Therapies Must Answer

Authored by

Psilocybin for Treatment-Resistant Depression: The Evidence Questions That Decide Approval

In my previous blog, Psychedelic Therapy: Why a Signal Isn’t Enough, I explored why promising efficacy data alone will not determine the future of psychedelic therapies. Demonstrating a rapid clinical response is important, but regulators, payers, clinicians, and patients ultimately need confidence that those benefits are durable, reproducible, and meaningful in the real world. 

As the field continues to mature, the conversation is shifting beyond efficacy and toward a broader set of regulatory, operational, and healthcare-system challenges. Scientific momentum combined with growing patient demand, and evolving policy support have helped move psychedelic therapies from the margins of medicine into serious clinical development. Yet progress brings new expectations. Success will depend not only on the molecule itself, but on the quality of the evidence, the consistency of delivery, and the ability to translate promising trial results into safe, scalable care models.  

In this next blog, I examine how psychedelic development is evolving from stigma to signal and why generating that signal may ultimately be the easiest part of the journey. The greater challenge lies in proving that these therapies can deliver reliable, long-term benefit in the complex realities of clinical practice.  

Speed Matters, but Durability Matters More

For a patient who has lived with severe depression for years, rapid improvement can be profound. Waiting six or eight weeks for another antidepressant to fail is not a minor inconvenience. It is another six or eight weeks of lost function, disrupted relationships, fear, and exhaustion. 

But speed is not the full story. A rapid response is not necessarily remission. Remission is not necessarily recovery. A patient’s Montgomery-Asberg Depression Rating Scale (MADRS) score may improve while the person remains unable to work, care for children, maintain relationships, or return to ordinary life. An early response also tells us little about what happens months later. Depression may clear, return, intensify, or arrive from a different direction.  

Compass Pathways has described 26-week findings from its second Phase 3 trial as showing a rapid, durable profile for psilocybin in treatment-resistant depression. The company reported that 39% of patients from a higher dosage group achieved a clinically meaningful MADRS reduction by Week 6 and maintained benefit, on average, through at least Week 26. These findings are important because durability is central to the company’s proposed clinical value, not just an added follow-up measure, for a treatment designed to move beyond daily or frequent administration. The next question is what happens after six months. 

Which patients should receive another administration? Should re-treatment occur at a fixed interval, at the first sign of symptom return, after crossing a defined threshold, or solely through clinical judgment? 

Does the patient who partially responded receive another dose sooner? Does a patient who achieved full remission but later relapsed follow the same strategy? Does repeated exposure change the safety profile? Does every administration require the same preparation, staffing, observation, and follow-up? 

Clinical trials prefer clean algorithms. Patients rarely follow them. 

Re-treatment should not automatically be viewed as evidence that the therapy failed. Maintenance is part of nearly every serious depression-treatment strategy. The real regulatory and clinical question is whether re-treatment represents occasional reinforcement, predictable maintenance, or the beginning of another chronic treatment cycle. 

That distinction affects the benefit-risk assessment, healthcare capacity, cost, patient burden, and commercial model. 

A two-dose treatment is one proposition. A treatment that ultimately requires administration every few months inside a certified clinical setting is a very different one. The most important outcome may not be how quickly depression improves. It may be how long the patient can experience benefit before the disease returns, and whether the healthcare system can recognize that return early enough to respond. 

Functional Unblinding Cannot Be Treated as a Footnote

Traditional randomized trials rely on blinding to reduce bias. Psychedelic trials challenge that assumption almost immediately. 

Participants often know, or strongly suspect, whether they received active treatment. Clinicians may infer assignment from the patient’s behavior, language, physical response, or obvious psychoactive effects. It is difficult to hide a thunderstorm by closing the curtains. Once treatment assignment is suspected, expectations can influence symptom reporting, rater interactions, retention, rescue treatment, and how every later change is interpreted. 

A low-dose comparator may preserve the appearance of blinding without preserving a credible blind. An active comparator may create some sensations but also introduce its own pharmacologic effects. Central raters can reduce certain sources of bias, but the patient still carries the treatment experience into the assessment. 

The FDA’s final guidance recognizes this challenge and recommends assessing subjects’ expectations about potential drug effects to improve data interpretability. 

A questionnaire can help show that the blind weakened, but it cannot put the toothpaste back into the tube. 

This in itself does not invalidate psychedelic trials. It means functional unblinding must be built into the design, analysis, and interpretation rather than acknowledged briefly in the limitations section. 

The relevant question is not merely whether the blind held. It is whether the estimated treatment effect remains credible after we accept that it probably did not and that requires careful evaluation of expectancy, participant beliefs, site behavior, missing data, concomitant treatment, rescue medication, and sensitivity analyses. 

The Molecule Does Not Arrive Alone

Psychedelic therapies are not being studied as simple prescriptions taken at home. 

The clinical programs generally include preparation, supervised administration, psychological support in many cases, observation, and follow-up. The patient enters not only a dosing session, but a deliberately constructed administration and care environment. 

The room, the preparation, the words used before dosing, the skill of the clinician, the patient’s expectations, and the response to distress may influence safety and outcome. 

FDA regulates the approval, labeling, manufacturing, and promotion of drug products, but it does not generally regulate the practice of medicine. Decisions regarding the therapeutic relationship, clinical setting, and manner in which psychotherapy is delivered remain within professional practice, subject to applicable state law and standards of care. This creates one of the field’s most difficult tensions. 

We need enough standardization to understand what was actually studied, but not so much rigidity that the treatment becomes impossible to reproduce outside a small group of highly specialized centers. 

Too little structure could lead to every site creating its own standard. Too much structure and the delivery model becomes a bespoke suit that fits the trial but no one in routine practice. 

Sponsors need to identify what is essential. How should patients be prepared? Who must be present? What training is required? How should panic, agitation, fear, or behavioral distress be handled? What follow-up is necessary? How are professional boundaries protected? What happens if suicidality worsens after discharge? 

These are way beyond operational questions for a site manual. They affect interpretation of efficacy. 

What exactly is the intervention? 

Is it a psychedelic or empathogen delivered within a defined supportive framework? Is support provided primarily for safety? Does it also contribute to treatment effect? If post-approval practice provides less preparation or fewer follow-up interactions, should we expect the same outcomes? 

The field must define the therapeutic environment clearly enough to protect patients and preserve reproducibility. But we should not build a cathedral when a safe, well-designed clinic will do. 

Positive Does Not Automatically Mean Benign

Psychedelic safety assessment also creates a language problem. 

A participant may describe feeling euphoric, connected, spiritually transformed, detached from the body, emotionally released, frightened, confused, or overwhelmed. 

The same experience may be expected, meaningful, therapeutic, adverse, or several of these simultaneously. 

That ambiguity does not disappear because the participant liked the experience. 

In pain and CNS development, we have long understood that descriptions such as “felt good,” “felt relaxed,” or “felt wonderful” may require review for euphoria and abuse-related effects. Psychedelic trials make this more complicated because the subjective experience may be closely connected to the treatment and still require disciplined safety characterization.  

Expected does not mean irrelevant and deserves accurate coding, medical review, and interpretation. MedDRA does not always provide a clean one-to-one term for complex subjective experiences. Two sites may hear similar language and document it differently. One may document euphoria. Another may record perceptual disturbance. A third may conclude that the event was anticipated and not record it with the same clinical significance or as a TEAE after all. 

By the time those inconsistencies are found, the database can resemble a kitchen where every cook followed a different recipe. Sponsors need prospective coding conventions, medical-review standards, case definitions, and cross-study consistency. Safety interpretation should not depend on how an individual site happens to hear the patient’s words. 

Why Approval Is Only Part of the Path Forward

For psychedelic therapies, the path from promising clinical signal to real-world treatment will depend on more than statistical significance. Sponsors must show that benefits are durable, outcomes are interpretable, safety is consistently characterized, and the care model can be delivered with enough structure to protect patients without making access impractical. These questions will shape how regulators, clinicians, payers, and healthcare systems judge whether the field is ready to move from controlled trials into routine care.  

The third and final blog in this series will examine what happens after FDA review, including why approval alone does not make psychedelic therapies immediately available, how DEA scheduling and rescheduling under the Controlled Substances Act may affect prescribing and distribution, and why federal, state, manufacturing, distribution, and potential REMS requirements could determine how widely these treatments can reach patients, particularly in underserved or rural communities.