top of page

The Lancet Psychiatry Review Finds No Evidence for Cannabis in Mental Health Treatment, But Real-World Data Tells a More Complex Story

  • chadwalkaden
  • 4 days ago
  • 7 min read

A study published in The Lancet Psychiatry has concluded there is no evidence that medicinal cannabis is effective for treating depression, anxiety, or post-traumatic stress disorder. But critics say the review's methodology may be missing what's happening in doctors' offices every day.


The systematic review and meta-analysis, led by Dr. Jack Wilson from the University of Sydney's Matilda Centre, examined 54 randomised controlled trials conducted over 45 years (1980-2025) involving 2,477 participants [1]. It represents the largest-ever assessment of cannabinoids for mental health and substance use disorders.



The findings come at a moment of explosive growth in medicinal cannabis use. In Australia alone, more than one million prescriptions have been approved, with sales tripling over the past four years [2]. In the United States and Canada, 27 percent of people aged 16–65 have used cannabis for medical purposes, with about half using it to manage mental health conditions [3].






What the Study Found


The review found no significant benefit for several conditions commonly cited as reasons for medical cannabis use, including anxiety, depression, PTSD, and Autism [1].


"Some people may experience legitimate benefits, and that's great," Wilson said. "But when we look at the evidence as a whole, we just don't see that the evidence is quite there for the routine use of these medicines" [4].


The researchers also noted a striking gap: they found no randomised controlled trials evaluating cannabinoids for depression at all [1].


Wilson raised concerns about potential harms. "Though our paper didn't specifically look at this, the routine use of medicinal cannabis could be doing more harm than good by worsening mental health outcomes, for example a greater risk of psychotic symptoms and developing cannabis use disorder, and delaying the use of more effective treatments" [4].


The study did find limited evidence suggesting potential benefits for certain conditions. A combination of CBD and THC was associated with reduced cannabis withdrawal symptoms and lower cannabis consumption among people with cannabis-use disorder [1]. Cannabinoids were also linked to reductions in tic severity in people with Tourette's syndrome, and researchers found some reduction in autistic traits among those with autism spectrum disorder [1]. But Wilson emphasised the overall quality of evidence for autism and insomnia was low [4].


The Methodological Debate


But some researchers argue that randomised controlled trials alone may not tell the full story, particularly for a medicine as complex as cannabis.


Professor Iain McGregor, academic director of the University of Sydney's Lambert Initiative for Cannabinoid Therapeutics, warned that policy decisions based solely on this review could be problematic. "The danger of taking a review like this to set policy is that you confuse an absence of evidence with a conclusion that we shouldn't prescribe because cannabis doesn't work," he said [5].


The review's own findings acknowledge this limitation. A separate scoping review of RCT evidence published in December 2025 found substantial heterogeneity in study design, with sample sizes ranging from 6 to 150 participants and follow-up durations from 1 day to 13 weeks [6]. The median follow-up was just six weeks [6].


"No trial demonstrated long-term efficacy," that review concluded [6].


What Real-World Data Shows


While RCTs remain the gold standard for evidence, real-world data from clinical practice tells a different story about patient outcomes.


A Canadian real-world evidence study published in the Canadian Journal of Pain followed patients over 24 weeks and found measurable improvements. Under physician guidance, patients using Health Canada-verified products showed decreases in General Anxiety Disorder-7 scores of 2.24 points and Patient Health Questionnaire-9 depression scores of 2.79 points from baseline [7].


The study, which included 141 participants, also found improvements in pain interference and quality of life measures [7]. While 85 per cent of patients indicated pain, the improvements in mental health outcomes were statistically significant [7].


Another large study published in PLOS ONE in April 2025 analysed 2,353 participants and found "clinically meaningful improvements" in health-related quality of life, fatigue, and sleep disturbance over 12 months [8]. For participants diagnosed with mental health conditions, there was "significant improvement in Depression Anxiety and Stress Scale anxiety and depression" [8].



The Health Collective's Real-World Analysis


The Health Collective, a research initiative powered by the OnTracka platform, has been building one of the largest real-world medicinal cannabis datasets [9]. Their white paper "From Data to Decision-Making" analysed over 1,600 tracked doses across 317 patients [9].


The findings provide granular detail about how patients actually respond to treatment:


Depression and nausea: Both symptoms improved at steady but modest rates [9].


Product associations: Full-extract products were more often linked with improvements in pain and depression, while CBD-inclusive regimens were associated with reduced fatigue. THC and CBD combinations were more commonly linked to improvements in nausea [9].


Treatment timeline insights: Most adverse effects, such as sedation and cognitive fog occurred during the first five doses. By dose 10, side effects declined and improvements became more consistent. Between doses 20 and 40, patients consistently reported symptom relief with fewer side effects [9].


This sequencing provides a framework for clinical follow-ups that simply doesn't exist in short-term RCTs [9].



An Outdated Approach to Guarding Healthcare


The Lancet Psychiatry review represents a specific kind of evidence: the randomised controlled trial. For decades, RCTs have been positioned as the only reliable gatekeeper for healthcare decisions. But that gatekeeping model assumes patients, clinicians, and the public lack the tools to interpret research for themselves.


Artificial intelligence has changed that assumption.


AI tools now allow anyone to parse complex studies, identify methodological flaws, and cross-reference findings against live patient data. When you apply that lens to this review, the questions it asks start to look limited.


The review asks: Does cannabis work in short, fixed-dose trials with homogenous patient groups?


It does not ask: How do patients actually use these medicines in practice? What happens when dosing is individualised? How do different formulations affect different symptom clusters? What does improvement look like over months, not weeks?


These are the questions patients are asking in clinics every day. And they now have the tools to find answers.


The same technology that powers platforms like The Health Collective also empowers individual patients. When someone reads a headline declaring cannabis ineffective for anxiety, they can now access real-world datasets showing thousands of patients reporting symptom improvement. They can see dosing timelines, adverse effect patterns, and product-specific outcomes.


This isn't an anecdote. It's structured data captured at scale.


Clinicians are using these tools too. Faced with a patient who hasn't responded to conventional treatment, they can now reference real-world evidence from patients with similar profiles. They can see what formulations correlated with improvement and what side effects to monitor.


This shifts the balance of power. When only a handful of researchers could interpret clinical trial data, they controlled the narrative. Now that patients and clinicians can analyse real-world outcomes themselves, the narrative is being rewritten in real time.



A New Model for Healthcare Decisions


The choice isn't between RCTs and real-world data. Both matter. But healthcare systems have spent decades prioritising one at the expense of the other.


RCTs tell us whether a medicine can work under ideal, controlled conditions. Real-world data tells us whether it does work in the community, across diverse populations, over meaningful timeframes.


For conditions like depression and anxiety, where the current standard of care leaves millions without relief, this distinction matters. Patients can't wait 45 more years for RCTs to catch up to their lived experience.


Biotech and pharmaceutical companies are already pivoting. The ones leading in this space are designing hybrid evidence models: they run RCTs for regulatory approval while simultaneously building real-world datasets that show how their products perform once they reach patients.


Regulators are starting to move too. The TGA's recent consultation on medicinal cannabis oversight signals recognition that the old frameworks don't fit [10]. But the pace of regulatory change lags far behind the pace of technological change.


Reconciling the Evidence


The discord between randomised controlled trial evidence and patient-reported outcomes presents a challenge for clinicians, regulators, and patients.


Traditional clinical trials take years to complete and often struggle to capture the complexity of cannabis prescribing. Unlike pharmaceuticals developed in a laboratory, cannabis is already widely prescribed, forcing researchers to work in reverse to validate what's happening in real-world care.


For clinicians and patients making treatment decisions today, the evidence remains mixed. RCTs show no proof of efficacy for mental health conditions, but real-world data consistently shows patients reporting improvement.


"Statistically significant and clinically meaningful improvements in overall HRQL, fatigue, and sleep disturbance were maintained over 12 months in patients prescribed medical cannabis for chronic health conditions," the PLOS ONE study authors concluded [8].


The healthcare system faces a choice. It can continue treating RCTs as the only valid form of evidence, accepting that this will leave large gaps in what we know about treatments patients are already using. Or it can embrace a broader definition of evidence, one that includes structured real-world data collected at scale.


The first path keeps control in the hands of a few researchers running slow, expensive trials. The second path distributes the ability to generate and interpret evidence to anyone with access to the right tools.


Patients have already chosen. They're using real-world data to make decisions about their care, with or without regulatory approval. The question is whether the systems designed to protect them will catch up.


The data exists. The tools exist. The only thing outdated is the approach.



---

References:


1. The Lancet Psychiatry: Cannabinoids for mental health and substance use disorders (2026) - https://www.thelancet.com/journals/lanpsy/article/PIIS2215-0366(26)00015-5/fulltext


2. Regulators come together as one million Australians turn to medicinal cannabis treatments (2024) - https://www.ahpra.gov.au/News/2024-02-20-medical-cannabis-treatment.aspx



4. University of Sydney: No evidence to suggest medicinal cannabis is effective for depression, anxiety or PTSD (2026) - https://www.sydney.edu.au/news-opinion/news/2026/03/17/no-evidence-to-suggest-medicinal-cannabis-is-effective-for-depre.html



6. PubMed: Randomised Controlled Trial Evidence on Medicinal Cannabis for Treatment of Mental Health and Substance Use Disorders: A Scoping Review (2025) - https://pmc.ncbi.nlm.nih.gov/articles/PMC12815972/


7. Canadian Journal of Pain: Real-world cannabis outcomes (2025) - https://www.tandfonline.com/doi/full/10.1080/24740527.2025.2593253


8. PLOS ONE: Improvements in health-related quality of life are maintained long-term in patients prescribed medicinal cannabis in Australia: The QUEST Initiative 12-month follow-up observational study (2025) - https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0320756


9. The Health Collective: From Data to Decision-Making White Paper (2025) - https://researchers.mq.edu.au/en/publications/from-data-to-decision-making-an-analysis-of-symptom-changes-and-a/


10. TGA: Medicinal cannabis framework consultation (2025) - https://consultations.tga.gov.au/medicines-regulation-division/test-soms/



 
 
 

Comments

bottom of page