Toronto’s Addiction Research Foundation and the 1960s LSD trial that said “not proven”

In 1962, a Toronto psychiatrist asked the province’s addiction agency to test a claim that was spreading from Saskatchewan: that one LSD session could help people stop drinking. The Addiction Research Foundation’s answer, published in 1966 and expanded into a book in 1967, was a careful “not proven” (Smart et al., 1967).

This article covers who ran the trial, how it was designed, what it found, why its authors thought earlier reports were too optimistic, and how later reviews treat it. For the Saskatchewan side, read our history of Humphry Osmond’s research. For today’s psilocybin trials, see psilocybin and alcohol use disorder trials.

What was the Addiction Research Foundation?

It was an agency of the Ontario government. Its own 1986–87 annual report says the province created the Alcoholism Research Foundation in 1949 to study alcoholism and to run a hospital for treating it. In 1961 other drugs were added to its mandate, and it became the Alcoholism and Drug Addiction Research Foundation, shortened to ARF (ARF annual report, 1986–87).

The foundation worked out of central Toronto. The 1966 paper gives its address as 24 Harbord Street, and by 1967 the team was writing from 344 Bloor Street West (reply, 1967). That address is a stop on our walking tour, along with Rochdale College, the free university that opened at 341 Bloor Street West in 1968. It also published its own book series, the Brookside Monographs, and the LSD study became number 6 (Smart et al., 1967).

Who ran the Toronto LSD trial?

Four people signed it. According to the monograph’s preface, Dr. Earle F. W. Baker approached the ARF in 1962 because of growing clinical reports that LSD helped alcoholics. The plan was worked out with the ARF’s clinical side, and Baker received a research grant. The ARF Research Division, through psychologist Reginald G. Smart, ran the assessments, follow-ups and analysis. Psychologist Thomas Storm helped plan the study and analyze the data. Baker and the psychiatrist Lionel Solursh gave the drug and did the psychiatric assessments (Smart et al., 1967).

Before running their own trial, Smart and Storm reviewed the existing evidence. Their 1964 paper, “The Efficacy of LSD in the Treatment of Alcoholism”, became chapter 3 of the later monograph. The monograph’s foreword, by the ARF’s advisory board chairman G. H. Ettinger, sets out its purpose: not to argue that LSD had no effect, but to show that “not proven” was the only verdict the evidence allowed (Smart et al., 1967).

What claims was Toronto testing?

The strongest claims came from Saskatchewan. In 1958 Colin M. Smith, of the province’s Department of Public Health in Saskatoon, published a paper calling hallucinogens a new adjunct to alcoholism treatment. In 1959 Nick Chwelos, Duncan Blewett, Smith and Abram Hoffer, working from Weyburn, Regina and Saskatoon, reported on LSD in alcoholism treatment.

The Toronto review tallied what those reports said. By its count, Smith reported 12 of 24 patients improved or much improved, and Chwelos and colleagues 15 of 16. Across the studies it reviewed, improvement rates averaged about 60 percent, and early reports claimed up to 94 percent, against about 40 percent for existing therapies (Smart et al., 1967).

The Toronto team’s problem was design. Of nine full reports, they found that only one used any control group, and in that one the comparison patients got a different kind of treatment altogether. None was blinded, so both patients and therapists knew LSD was being given. They also cited a 1958 survey of psychiatric journals: 83 percent of uncontrolled studies of new treatments reported success, against 25 percent of controlled ones (Smart et al., 1967).

Historian Erika Dyck describes the clash from the other side. Saskatchewan’s treatment had the backing of the provincial government, local chapters of Alcoholics Anonymous and the Bureau of Alcoholism. The ARF did not endorse it: as Canada’s leading addiction authority, it disputed the Saskatchewan findings and challenged the advocates to run properly controlled trials.

How was the ARF trial designed?

The monograph describes 30 patients, 28 men and 2 women, all in-patients or day patients at the ARF’s Toronto clinic. Their average age was 40. All had long histories of uncontrolled drinking and earlier treatment that had failed. People with major heart or liver disease, or early signs of psychosis, were excluded (Smart et al., 1967).

• Three groups of ten. Each patient was randomly assigned to LSD, to ephedrine (a stimulant with no hallucinogenic effect, chosen because its early physical effects can resemble LSD’s), or to a no-drug control group (Smart et al., 1967).
• Same clinic treatment for all. Every patient stayed in the ARF’s usual program of group meetings and therapy; the two drug groups added one drug session (Smart et al., 1967).
• A hospital session. The drug was given on the psychiatric ward of Toronto Western Hospital, in a single room, followed by a three-hour interview with a doctor and a nurse. Patients were secured to the bed with a belt, and no music or pictures were used (Smart et al., 1967).
• Partial blinding. Patients were told only of a “new drug” and that reactions vary, and clinic therapists did not know who got LSD. The LSD therapists did: they guessed the drug correctly in 19 of 20 cases (Smart et al., 1967).
• Blind follow-up. The follow-up interviewer never knew who got what, and the results were analyzed as groups A, B and C before the code was broken. All 30 patients were seen at about six months (Smart et al., 1967).

The authors admit the weak points. The control group knew it was getting nothing, and newspaper coverage led many patients to suspect a hallucinogen was involved. The authors called the result a “modified single-blind” design rather than a true double-blind one (Smart et al., 1967).

What did the Toronto trial find?

Everyone improved, and LSD made no measurable difference. In the six months after treatment, abstinence increased in all three groups: by an average of 33.7 percent in the LSD group, 31.5 percent with ephedrine and 19.6 percent with no drug. The authors’ statistical tests found no significant difference between the groups, either in total abstinence or in the longest sober stretch (Smart et al., 1967).

Other measures pointed the same way. Drunkenness fell for the three groups combined, but not more in the LSD group. LSD did not reduce morning drinking, blackouts or other drinking symptoms, and it did not increase contact with the clinic. Convictions for drunkenness fell in both control groups but rose from four to six in the LSD group (Smart et al., 1967).

A second follow-up, about two years after the study began, reached 24 of the 27 surviving patients. In the no-drug group, three were much improved, four somewhat improved and three unimproved. In the LSD group, none was much improved. Three patients had died: two from causes linked to alcoholism and one, in the LSD group, by suicide about a year after the session. The authors wrote that LSD’s role, if any, was unknown and that a close connection seemed unlikely (Smart et al., 1967). A 2025 review notes that no psychosis was reported among the trial’s participants (Sabé et al., 2025).

The authors also recorded the few findings that leaned toward LSD. More people in the LSD group had jobs after treatment, and there were non-significant trends toward a healthier self-concept. Their conclusion was that the drinking results were consistent: “no effect” (Smart et al., 1967).

Why did Toronto and Saskatchewan disagree?

The Toronto authors argued the gap was partly an illusion. Eight of their ten LSD patients counted as improved at six months, close to the success rates reported elsewhere. The difference was that Toronto also had control groups, and those patients improved too. Earlier reports, they concluded, may have looked strong because they had nothing to compare against (Smart et al., 1967).

The defenders of LSD therapy disagreed, and the journal printed their objections. In 1967 the Quarterly Journal of Studies on Alcohol ran a comment by J. Ross MacLean and W. E. Wilby, “with special reference to issues of responsibility in research reporting”, another by James Fadiman, and a reply from the Toronto authors. We read only the records of these three items. A 2019 systematic review summarizes the main objection: Toronto’s methods were far from those that LSD therapists believed produced benefit (Fuentes et al., 2019).

The differences were real. Saskatchewan workers such as Chwelos used music, family photographs, flowers and prepared questions during sessions; Toronto used none, and its authors said they didn’t know whether those things mattered (Smart et al., 1967). The 2019 review contrasts Toronto’s approach, with no attempt at psychotherapy during the session and a waist belt, with later studies that paid close attention to setting (Fuentes et al., 2019). Our guide to set and setting traces that idea.

Dyck adds that, by the late 1960s, LSD’s link with the counterculture further eroded its clinical standing, whatever the trials showed. The monograph itself mentions two other early controlled studies, by Van Dusen and colleagues and by Johnson, whose preliminary reports also found no LSD effect (Smart et al., 1967).

What do later meta-analyses make of the 1960s trials?

They pool the trials, and pooled they lean toward LSD. In 2012 Teri Krebs and Pål-Ørjan Johansen combined six randomized trials of LSD for alcoholism, with 536 participants. They reported an odds ratio of 1.96 (95% confidence interval 1.36 to 2.84) in favor of LSD for reducing alcohol misuse, with negligible variation between trials (Krebs and Johansen, 2012). We could read only the abstract, so we can’t say which six trials it used.

A 2025 meta-analysis of psychedelics for substance use did include the Toronto trial. It pooled seven LSD studies with 555 people and found a small-to-moderate overall effect for LSD (Hedges’ g 0.36). The Toronto trial carried the least weight of the seven, and its own estimate (g 0.21) had a confidence interval running from −0.55 to 0.99, which includes no effect (Keighley et al., 2025).

Quality ratings tend to favor Toronto’s design. The 2019 systematic review of 11 randomized LSD trials rated the ARF study at low risk of bias for randomization, allocation concealment, outcome assessment, missing data and selective reporting, with blinding of participants and staff unclear (Fuentes et al., 2019). A 2025 review is more skeptical of the pooled result itself: the comparison treatments ranged from stimulants to lower LSD doses to usual care, and outcome measures were loose and inconsistent.

So the two readings can both hold. The Toronto trial, on its own, was too small to detect a modest effect, as its authors acknowledged. Pooled with other trials, the 1960s evidence points toward some benefit, but it rests on old studies with different methods. None of this concerns psilocybin, which these trials did not test.

What happened to the ARF after the LSD trial?

Its LSD work shifted from treatment to harms. In 1967 Smart and K. Bateman published a review of unfavourable reactions to LSD in the Canadian Medical Association Journal. A 2025 review describes it as drawing on 21 case reports covering 225 adverse reactions (Sabé et al., 2025). The ARF also studied Toronto’s street scene in the late 1960s; that story is told in Psychedelic Toronto and our Toronto psychedelic stories.

The foundation kept growing as a research body. Its 1987–88 annual report describes it as an agency of the Province of Ontario since 1949, with a mandate expanded in 1961 and an official designation as a World Health Organization Collaborating Centre.

In 1998 it stopped existing as a separate body. CAMH says it was created that year by merging the Queen Street Mental Health Centre, the Addiction Research Foundation, the Clarke Institute of Psychiatry and the Donwood Institute, and that it became Canada’s largest mental health teaching hospital. For what CAMH and others are studying now, see psilocybin clinical trials in Toronto.

What we could not confirm

• Which six trials Krebs and Johansen pooled. Their full text is closed, and we read only the abstract.
• The exact wording of the 1966 paper. We read the 1967 monograph, which its authors say revised and enlarged the paper’s chapters on methods and results.
• What the 1967 comments and the authors’ reply actually said. We read their bibliographic records and one later summary.
• How the trial was received inside the ARF beyond the monograph’s foreword.

Shroomverse sells mushroom products. This article is general information about the history of addiction research in Toronto. It is not medical, legal or purchasing advice, it does not encourage anyone to use drugs, and it does not recommend any product.