iiinova — where your digital self keeps evolving

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 07/10/2026

Someone asked the professor, “will the test be focused on facts or theory?” His answer: Facts without theory is trivia. Theory without facts is bullshit. Still haunts me.

110

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Psychedelics are moving rapidly toward approval, adoption, & scaling. If we scale first and only later figure out how much benefit is pharmacologic vs belief-dependent, we may waste resources and time and put patients at unnecessary risk. We should find out now.

000

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Reflecting on a sham-controlled trial that challenged a widely adopted treatment: “Results of ORBITA show (once again) why regulatory agencies, the medical profession, and the public must demand high-quality studies before the approval and adoption of new therapies.”

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

The goal isn’t UMBRAA. The goal is finding out what these treatments actually do.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Maybe UMBRAA will work. Maybe it won’t. “Masking is impossible” should be a conclusion of serious attempts, not the premise for avoiding them. If someone has a better way to credibly mask psychedelic trials, even better.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

"Umbra" is Latin for “shadow.” In an eclipse, the umbra is the region of complete shadow. Complete obscuration of treatment assignment may be aspirational in psychedelic trials. But we should at least try harder to get there.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Put those ideas together and you get UMBRAA: Unidentifiable Multidrug Blinding with Reduced, Authorized Awareness. Yes, we tortured the English language a little to get the acronym.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

The multidrug comparator makes the EXPERIENCE harder to identify. Incomplete disclosure makes the experience less informative about ASSIGNMENT. Together, they attack unmasking from both directions.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

We also tell them: “We’re not going to tell you exactly how many treatment arms there are or how the trial is structured.” They agree to that uncertainty in advance. But the one thing we will NEVER do is not tell them (the possible risks in the study) and then take them to the cabin.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

G.O.B may have accidentally invented the double-masked RCT. We tell participants: “We’re not going to tell you whether you’re getting active drug or placebo.” Authorized incomplete disclosure just takes this one step further.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

But pharmacology is only half the problem. What participants KNOW about a trial determines what they can infer from what they FEEL. If the only possibilities are 25 mg psilocybin or placebo, any psychedelic experience effectively reveals assignment.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Maybe the first combination won’t work. Change the ingredients. Change the doses. Change the timing. Maybe it still won’t work. But first, actually try.

static.klipy.com

ALT:

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Better masking may require some comparator risk. That can be ethical when needed to answer an important question. An outpatient-tolerable comparator would be ideal. But if mild comparators can’t mask 25 mg psilocybin, why insist on one?

static.klipy.com

Justice Democracy

ALT: Justice Democracy

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Multiple drugs give us multiple knobs to turn. Agents, doses, & timing can be adjusted to increase both experiential intensity and breadth, duration, etc... moving the comparator closer to the psychedelic experience. Treat masking as an optimization problem.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

We propose combining drugs from different mechanistic classes, e.g., a stimulant, sedative, anticholinergic and/or cannabinoid, to create an intense, broad & unfamiliar experience without reproducing psilocybin’s primary pharmacology. We don’t know the right combination yet. That’s OK.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Our proposed solutions: First, no single comparator gets close to psilocybin on both experiential INTENSITY and BREADTH. That’s the design space we need to optimize. Why restrict ourselves to one drug?

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Yet a kind of “masking nihilism” has taken hold: psychedelics are too distinctive, successful masking is impossible, so why bother trying? Maybe. But “impossible” is a strong conclusion to reach before we’ve made serious attempts to do it.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

None of this shows that psychedelic benefits are “just placebo.” That’s the point: we don’t know how much of the observed effect is pharmacologic and how much is belief-dependent. An unmasked efficacy estimate can’t cleanly answer that question.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

But who cares about masking if patients get better?” Because improvement doesn’t tell us what produced it. Psychedelic therapy is expensive, resource-intensive & not risk-free. We have cheaper, safer ways to generate hope, expectancy & therapeutic attention. We need to know what the drug adds.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Then SPACE randomized patients to opioid or nonopioid treatment for a year. Opioids weren’t better for function. Pain intensity actually favored nonopioids. And opioids caused more medication-related symptoms. Confidence is not a control condition. doi.org/10.1001/jama...

doi.org

Opioid vs Nonopioid Medications on Pain-Related Function

This randomized clinical trial compares the effects of opioid vs nonopioid medications over 12 months on pain-related function, pain intensity, and adverse effects among Veterans Affairs patients with moderate to severe chronic back pain or hip or knee osteoarthritis pain.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Another example: Long-term opioids became standard treatment for chronic musculoskeletal pain despite remarkably little evidence that they were better than nonopioid treatment.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

And medicine has quite a collection of these: Mammary artery ligation. Meniscectomy. Shoulder decompression. Vertebroplasty. Coronary stenting. Gabapentin for back pain. It is known. Until it isn’t.

static.klipy.com

ALT:

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

An example: Knee arthroscopy was widely performed for osteoarthritis. Patients and surgeons knew it worked. Then Moseley et al. randomized patients to real or sham surgery. Over 2 years, real surgery was no better than sham for pain or function.

doi.org

A Controlled Trial of Arthroscopic Surgery for Osteoarthritis of the Knee | NEJM

Many patients report symptomatic relief after undergoing arthroscopy of the knee for osteoarthritis, but it is unclear how the procedure achieves this result. We conducted a randomized, placebo-con...

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Their pain wasn’t fake. Their inference about what caused it was wrong. That distinction is the whole reason masking matters. Feeling dramatically better after a treatment doesn’t tell us how much of the improvement was caused by the treatment.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

How powerful can that be? Bayer et al. connected volunteers to electrodes & turned up the dial. 50% reported pain; 25% requested pain medication. But the electrodes weren’t connected. “Pain can be produced in the absence of peripheral stimulation.” www.sciencedirect.com/science/arti...

static.klipy.com

ALT:

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

But psychedelics for psychiatric disorders aren’t like parachutes for Involuntary Deceleration Events. Symptoms fluctuate & respond to attention, context, expectancy. & outcomes can be influenced by what patients believe they received. That’s when knowing treatment assignment becomes a problem.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

To be clear: not everything requires a double-masked RCT. We know smoking causes lung cancer. We know insulin works in type 1 diabetes. And I wouldn't sign up for a placebo-controlled trial of parachutes: www.bmj.com/content/327/... Sometimes the signal overwhelms plausible bias.

bmj.com

Parachute use to prevent death and major trauma related to gravitational challenge: systematic review of randomised controlled trials

Objectives To determine whether parachutes are effective in preventing major trauma related to gravitational challenge. Design Systematic review of randomised controlled trials. Data sources: Medlin...

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

One response: sure, the trials are unmasked. But the effects are WAY too big to be explained by expectancy. Balazs and I tested that argument. They aren’t. Unmasking bias in other psychiatric trials can be large enough to produce effects of this magnitude.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

A clue comes from work led by the brilliant Balázs Szigeti, a member of our group and one of the sharpest thinkers on expectancy in psychedelic research. We have a paper on this hopefully coming soon: Placebo responses in psychedelic trials are smaller than in in conventional antidepressant trials.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

And the problem isn’t subtle. In high-dose psychedelic trials, >90% of participants acutely & correctly identify what they received. That doesn’t mean the drugs don’t work. It means pharmacology and belief are tangled together in estimates of how much they work.

100

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

Huge thanks to Kevin Yang, @theborislab.bsky.social, & Andrew Krystal for thinking this through with me. The masking problem isn’t new. Neither are proposals to address it. But masking has become a bit like the weather: everybody talks about it, but nobody does anything about it.

132

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 17/09/2026

I’m especially proud of our new @jamapsychiatry.com Viewpoint. The psychedelic field seems to have largely given up on masking before really trying. We think that’s a mistake. We call for trying harder and offer concrete ideas for how to do it: UMBRAA. @ucsfpsych.bsky.social 1/36

doi.org

Validate User

132

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 24/07/2026

Strong concur. With my name maybe all my posts are bleets?

010

Reposted by Josh Woolley MD/PhD

christine liu / two photon art @christineliu.art · 10/07/2026

I joined the Sohal lab with one major hypothesis: psychedelic drugs can be powerful psychiatric treatments, depending on underlying neurobiology. Indeed, psilocybin restored cognitive flexibility in one circuit model of dysfunction, but not another! Then, we had to find out how! 🧠🍄🧪

36221

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 20/05/2026

UCSF is the place to be for psychedelic research!

000

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 26/04/2026

Are you sure? I heard something different.

130

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 25/04/2026

Right? Like if it was really no therapy why not have a nurse rn instead of a licensed mental health professional? Rn is a bachelor or associate degree. That’s what they do in the icu where 1 rn per 2 patients is typical.

140

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 29/03/2026

Are we making the classic blunder of getting involved in a land war in Asia?

010

Josh Woolley MD/PhD @trpr-ucsf.bsky.social · 06/12/2025

“Now that ICOs will not rely on funding paylines when developing their pay plans, it should be clearer for applicants to know the award decision was not made only based on overall impact score, without necessarily considering the additional valuable information provided by peer review.” 🤔 huh?

120