From the Editor
He survived the civil war. But nightmares haunted him and he rarely left his apartment. My patient had survived, yes, but his mental health problems, untreated in his home country, tormented him. In a world of war and violence, my patient’s experiences – the mental illness and the lack of care – are all too common.
With few psychiatrists and psychologists, could non-specialist providers help fill the treatment gap? Anushka R. Patel (of Harvard University) and her co-authors attempt to answer that question in an impressive new systematic review and meta-analysis published in Lancet Psychiatry. Drawing on a quarter century of studies and involving more than 10 000 survivors, they compared non-specialist care with a relatively inactive control. “Our findings support the overall effectiveness of non-specialist provider-delivered psychosocial interventions for violence survivors…” We consider the paper and its implications.
In the second selection, from JAMA, Dr. Ezekiel J. Emanuel and his co-authors write thoughtfully if controversially about artificial intelligence. While many assume that medicine will be reshaped by the collaboration of physicians and computers, they suggest otherwise, focusing on five cognitive medical tasks. “That in the near future AI alone may provide better patient care than physicians or physician-controlled hybrids at 5 fundamental cognitive medical tasks is unsettling but seems probable.”
Finally, we review the latest news with articles from The Toronto Star, The Wall Street Journal, and The Globe and Mail. The topics: addiction treatment, the bias of therapists, and the PTSD and research of Margaret McKinnon.
DG
Selection 1: “Non-specialist delivered psychosocial care after war and interpersonal violence: a systematic review and meta-analysis”
Anushka R. Patel, Ginger Ramirez, Luanna Fernandes, et al.
The Lancet Psychiatry, July 2026
More than 71% of the global population experiences lifetime exposure to traumatic events. Traumatic events exert a host of psychosocial consequences, most prominent among which are post-traumatic stress disorder (PTSD), depression, anxiety, and functional impairment. Disorder prevalence varies across settings, but rates are higher in conflict-affected settings… The gap between people needing treatment and those who have contact with any health-care provider remains large. In 2023, findings from the World Mental Health Surveys indicated approximately 20% of trauma-exposed individuals in low-income and middle-income countries (LMICs) who needed treatment received it.
The global scarcity of specialist mental health professionals is a major driver of this treatment gap. Task-sharing has been used successfully to address this barrier, wherein non-specialists (ie, people without specialist mental health qualifications) are trained to deliver interventions with training and close supervision from specialists. Non-specialist provider-delivered interventions for depression and anxiety are effective in both LMICs and HICs. Treatments that incorporate cognitive-behavioural and narrative approaches have shown effectiveness in HICs. Although some studies show similar effectiveness of cognitive-behavioural and narrative approaches for PTSD in LMICs, there is less robust evidence of their effectiveness when delivered by non-specialist providers; only four studies were identified in a review of non-specialist provider-led care…
So begins a paper by Patel et al.
Here’s what they did:
- They conducted a systematic review and meta-analysis, in accordance with PRISMA reporting guidelines.
- They searched several databases (including PubMed) for studies published between January 2000 and June 2025.
- They included randomized controlled trials if they involved adults exposed to potentially traumatic events who screened positive for psychological distress, and compared a psychosocial intervention delivered by a non-specialist provider “with a relatively inactive control (e.g., usual care or waitlist).”
- Non-specialists were defined as providers without formal clinical psychology/psychiatry training or prior clinical case-based training with trauma survivors; they included community health workers, volunteers, peers, and some primary-care providers.
- They focused on four primary outcomes: depression, post-traumatic stress disorder, anxiety, and impairment symptom severity.
Here’s what they found:
- They screened more than 54 000 abstracts and 45 RCTs were included, involving 9 431 adults exposed to war-related or interpersonal violence.
- Demographics, backgrounds, and settings. Participants had a mean age of 37.9 years. Most participants were refugees, and the interventions were delivered by non-specialists in community or medical settings.
- Treatment. Non-specialist-delivered psychosocial treatments were better than control conditions across all four outcomes: anxiety (SMD −0.44), depression (−0.41), PTSD (−0.34), and functional impairment (−0.34).
- Effects. Effects were generally small to moderate, with somewhat larger effects for depression and anxiety (moderate) than for PTSD and impairment. Importantly, the results remained statistically significant when the authors excluded studies judged to have a high risk of bias.
- Subgroups. Refugees appeared to benefit most. In contrast, among veterans, benefits weren’t statistically significant for any of the four outcomes.
- Treatment approaches. Benefits were observed across different approaches to delivering care. Community and medical settings both showed benefit, as did interventions delivered by community workers, medical workers, and peers. Transdiagnostic interventions were particularly encouraging, showing benefits across all four outcomes.
- Risk of bias. 44 of the 45 studies (98%) had some risk of bias.
A few thoughts:
1. This is an important study on a practical problem with relevance around the globe, published in a major journal.
2. The main finding in a sentence: “Small to moderate clinical benefits for non-specialist-delivered care were observed across a broad set of treatments for diverse populations exposed to violence worldwide.”
3. In a world marked by war and violence but constrained by limited mental health resources, this study helps point the way forward. As the authors note at the beginning of their paper, just one in five trauma-exposed individuals in low- and middle-income countries receives treatment. (!) There aren’t enough psychologists or psychiatrists, but training up non-specialists is possible and plausible.
4. Like all studies, there are limitations. The authors note several, including: “We did not use individual participant data for analyses, which precludes our ability to control for baseline symptom severity and to classify individual trauma exposures – or polytraumatic exposures – to specific individuals.” The risk of bias highlights the need for further research but, despite the limitations, the main finding is deeply promising and consistent with other studies on task sharing.
The full Lancet Psychiatry paper can be found here:
https://www.thelancet.com/journals/lanpsy/article/PIIS2215-0366(26)00123-9/fulltext
Selection 2: “Will Autonomous AI Exceed AI-Aided Physicians as the Best Medical Care?”
Ezekiel J. Emanuel, Abe Baker-Butler, Neal Khosla, Vinod Khosla
JAMA, 17 August 2026 Online First
The prevailing view of artificial intelligence (AI) in medicine is that it will support physician-led care. The American Medical Association regularly calls AI-augmented intelligence to focus on AI’s assistive role. Similarly, the American College of Physicians argues that AI ‘should be limited to a supportive role in clinical decision-making’ and ‘should not replace physician decision-making.’ In A Giant Leap: How AI Is Transforming Healthcare and What That Means for Our Future, Wachter argues that the highest tier of care will be AI-aided physicians, whereas AI-only care will be medicine’s ‘economy class.’
We disagree.
So begins a paper by Emanuel et al.
“Large language models (LLMs) were only publicly introduced in November 2022, and already generative AI rivals or outperforms licensed physicians at 5 fundamental cognitive medical tasks: (1) eliciting medically relevant information; (2) establishing a differential diagnosis; (3) specifying diagnostic testing; (4) prescribing guideline-concordant treatment; and (5) managing chronic diseases. The gap between physicians’ and LLMs’ performance will likely widen because AI is rapidly improving, whereas physicians’ skills are threatened by AI-induced deskilling.”
When Does AI Alone Exceed Physicians?
They review the literature:
- Histories. “AI gathers patient information as effectively as, if not more effectively than, physicians. Using 159 objective structured clinical examination simulated case scenarios covering multiple specialties, physicians judged Google’s Articulate Medical Intelligence Explorer as significantly better than physicians at eliciting patient actors’ complaints (97% vs 50% favorable), systems review (88% vs 35%), medical history (85% vs 50%), family history (50% vs 21%), and medication history (68% vs 45%)…”
- Diagnoses. “Many LLMs, including ChatGPT o3, Microsoft AI Diagnostic Orchestrator, and Google’s Articulate Medical Intelligence Explorer, produce more accurate diagnoses than physicians. ChatGPT o3 ranked the final diagnosis first in 60% of 377 real-world complex cases, whereas 20 internal medicine physicians did so in only 15.9% of a 302-case subset.”
- Guidelines. “Compared with physicians, AI alone excels at recommending guideline-concordant treatments. In the Google Articulate Medical Intelligence Explorer study, AI prescribed more appropriate treatment than licensed primary care physicians (90% vs 37% favorable…).”
When Do Humans Alone Equal or Exceed AI Alone?
“Excluding radiology, a domain in which AI alone does not reliably exceed physicians, only 9 studies published since January 1, 2024, concluded that physicians alone are equivalent or superior to AI alone at 5 fundamental cognitive medical tasks… One study showed humans alone exceeding AI alone at differential diagnosis, test ordering, and prescribing guideline-concordant treatment in 80 emergency and intensive care cases. However, the study excluded the best AI models – OpenAI’s and Google’s LLMs – underestimating AI’s abilities.”
How Does AI Alone Compare With Physician-AI Hybrids?
“When AI alone is better at a task than humans alone, human-AI hybrids actually degrade performance compared with AI alone. Thus, as AI improves, having humans in the loop will likely worsen patient care… A 2025 review of 52 clinical studies found that physician-AI hybrids ‘had a lower reliability level than the best performer [of either AI alone or human alone]’… The authors concluded that physician-AI hybrids ‘neither outperformed medical AI alone nor surpassed the best of clinicians or medical AI alone.’ Similarly, using real patient cases, OpenAI’s ChatGPT-4 alone had a median diagnostic reasoning score of 92%, whereas physicians using ChatGPT-4 scored 76%.”
Why Does AI Alone Exceed Physician-AI Hybrids?
“Although integrating humans in the loop catches some AI errors and adds insights, it also introduces errors when humans contribute errors and incorrectly overrule AI. Humans perform poorly at deciding when to trust AI, especially when AI exceeds them. This poor human decision-making is partly driven by algorithm aversion (distrust of AI). Algorithm aversion is evident in 75% of studies on algorithmic advice and persists even when humans know they are less accurate than AI. In high-stakes tasks with high-expertise decision-makers, such as expert physicians, algorithm aversion seems consistently greatest. In medicine, algorithm aversion causes significant inaccuracy because humans have lower medical accuracy than AI and struggle at integrating cross-specialty medical knowledge, which AI does easily. As AI improves and exceeds humans, physicians’ error-catching and insight-adding benefits dwindle, whereas physician-introduced errors remain constant or increase. This problem may worsen because physicians are likely to become increasingly deskilled.”
Caveats and Additional Research
“Most studies comparing AI alone with physicians alone and physician-AI hybrids are simulations of discrete cognitive medical tasks, not analyses of real clinical interactions. More assessments of AI in real-life clinical encounters are essential and seem to be expanding.”
A few thoughts:
1. This is an amazing paper, reviewing a vast literature and offering big conclusions.
2. It’s also sobering.
3. If you read one AI paper in 2026, it should be this paper.
4. The argument about physician-AI hybrids being inferior to AI is fascinating and thoughtful. Is it persuasive? Readers can draw their own conclusions.
5. Playing devil’s advocate: are the authors too quick to take small studies and make big predictions? Remember when doctors warned of a link between coffee and pancreatic cancer? Is this the AI equivalent?
The full JAMA paper can be found here:
https://jamanetwork.com/journals/jama/article-abstract/2852952
In the News
Part of an occasional series.
“Mayors want Doug Ford’s government to consider involuntary addiction treatment. Why the controversial option may be ‘having a moment’”
Mahdis Habibinia and Omar Mosleh
The Toronto Star, 18 August 2026
“The mayors of some of Ontario’s largest cities are pleading with the province to explore options for involuntary addiction treatment as a tool to fight the current homelessness and drug crisis.
“‘We need to not be afraid of having the difficult conversations,’ said Burlington Mayor Marianne Meed Ward, also chair of the Ontario’s Big City Mayors caucus, which represents 29 of the province’s largest municipalities… ‘We are ready to have a conversation without prescribing what the outcome of that conversation will be,’ she added…”
The article details efforts by several provinces to mandate treatment for substance misuse, although neither Alberta nor Saskatchewan have implemented this legislation. The authors describe the controversy surrounding the proposal, and cite a joint statement from several mental health organizations that the evidence “does not suggest involuntary treatment is effective, let alone more effective than voluntary treatment.” Some, nevertheless, support the proposal; Oshawa’s mayor describes it as compassionate.
Readers may agree or disagree with that assessment. On this, we can all agree: across North America, involuntary treatment is having a moment.
“Is Your Therapist Biased? Personal Agendas Are Creeping Into the Therapy Room”
Elizabeth Bernstein
The Wall Street Journal, 22 August 2026
“Is your therapist biased?
“Mental-health professionals are supposed to keep their personal experiences, values and opinions out of the therapy room. Yet in a culture sharply divided along political and personal identity lines, more appear to be falling short, experts say.
“Some push their own ideologies, encourage patients toward a path that worked for them, and even judge family members they have never met. And the results can be dangerous.”
The article discusses whether some therapists are biased. Bernstein notes that bias may affect the quality of the therapy. She explores the possible influence of politics and social issues, and the broadening of terms like harassment and abuse. The article focuses on a woman who sought care for anxiety but was told to end her relationship with her boyfriend. It closes with suggestions for patients including setting goals and establishing boundaries.
“Air Transat Flight 236 passenger on why she decided to study memory and trauma after near-crash landing”
Lana Hall
The Globe and Mail, 24 August 2026
“On Aug. 24, 2001, Margaret McKinnon and her new husband, John Baljkas, boarded Air Transat Flight 236 from Toronto to Lisbon for their honeymoon. About four hours into the flight, the Airbus A330’s right engine began leaking fuel into the Atlantic Ocean. Eventually, both engines would flame out, and captain Robert Piché would execute what remains the longest glide with no engine power ever for a passenger aircraft – 120 kilometres – into the tiny Lajes military air base in the Azores.
“Even before Ms. McKinnon boarded that ill-fated flight, she was interested in the way people’s brains work. She earned a master’s degree in psychology from the University of Toronto and was working toward a PhD with a focus on aging. But the flight, and the unlikely survival of all 306 people on board, changed the course of her research focus entirely.”
The article includes an interview with McKinnon, who is now the Associate Chair (Research) of the Department of Psychiatry and Behavioural Neurosciences at McMaster University. Her story reads like a page-turner: a near-death experience and PTSD, a recovery, and an extraordinary career focused on those with that disorder. Reflecting on her research and on her opportunities to work with those affected by trauma, McKinnon comments: “I don’t regret what happened.” Amazing.
Reading of the Week. Every week I pick articles and papers from the world of Psychiatry.
Recent Comments