Hi friend,

Greetings from a very warm Nashville. I’m here for BHT this week and am looking forward to seeing many of you here.

It was a relatively quiet week last week. Many organisations would have been saving their big announcements for BHT. But there are still some very interesting news items and pieces of research for us to review.

Sword officially announced their acquisition of Headspace. Slingshot opened a new research partnership programme — which includes deploying Ash for free with partner organisations — and the UK’s AI healthcare commission published its recommendations for a new regulatory framework.

A paper in Nature Mental Health Research evaluated how AI performs on psychiatric assessments, and new research from Cross et al. explores the unbundling of therapy, specifically, which parts of psychological therapy AI may be able to perform and where human therapists remain important.

Let’s get into it!

Join our Session with Thomas Insel MD

I'm incredibly excited to have Thomas Insel MD join us for a Hemingway Session next month. Tom will be sharing the lessons he has learned from a decade in behavioural health start-ups (and several decades more working across the mental health ecosystem).

Sign up here 👉  https://luma.com/hemingway-1rv8

FYI, Sessions are a benefit of Hemingway Pro membership. To join our Sessions and access all of the other benefits of being a Hemingway Pro Member, click below.

What happened in mental health last week?

Industry News:

⏺️ Sword officially announced their acquisition of Headspace. After weeks of unconfirmed reporting on this deal, both companies officially announced the acquisition. Sword will add Headspace’s meditation, coaching, therapy and psychiatry services to its broader virtual-care business. The deal is expected to close by the beginning of Q4 2026. (link)

⏺️ Renew Health + Recovery bought the assets of Aware Recovery Care. BHB reported that the deal included many of Aware’s clinical assets and care teams, which provide in-home and virtual addiction treatment across 11 states. Aware had been facing solvency problems and entered an assignment for the benefit of creditors. The deal has an unusually complex structure, with BHB reporting that it remains unclear exactly which assets and operations sit with which entities. (link)

⏺️ Brightline opened its fifth New York mental-health clinic. The Albany clinic offers in-person and virtual therapy, psychiatry and psychological testing for children and teenagers and represents a further commitment to Brightline’s hybrid care strategy. Long live the clinic indeed. (link)

⏺️ Slingshot AI launched a research partnership programme for Ash. The programme will let health plans, primary-care groups, employers and other organisations pilot the company’s AI mental-health product with their populations, for free. Slingshot’s research team will work with participating organisations to study outcomes and how Ash performs when integrated into existing systems of care. (link)

⏺️ Spirit Tech launched a new platform to track the emerging sector. I wrote about Spirit Tech last year, exploring the emerging category of technologies designed to support spiritual experiences and practices. The new platform, launched by Eddy Vaisberg and colleagues, includes a directory of more than 500 organisations working across areas including meditation, AI, psychedelics, neurotechnology and spiritual development. If you’re interested in that space, it offers a good way to keep on top of what is happening. (link)

⏺️ SocialRx received NIH funding to further develop its social-prescribing platform. The US company received a Small Business Innovation Research award through the National Center for Complementary and Integrative Health, part of the National Institutes of Health. SocialRx connects patients with community-based services and activities intended to address social needs, including loneliness and isolation. The funding will be used for the further development of SocialRx’s member engagement platform for social prescriptions. (link)

⏺️ Mississippi awarded $104M to 167 rural healthcare projects. The state awarded 167 grants to rural healthcare providers, including mental-health practitioners, while separate projects will expand crisis diversion centres and mental telehealth services. Two crisis centres will receive funding to provide alternatives to emergency departments and jails for people experiencing acute mental-health or substance-use crises, with one expanding from eight to 16 beds. (link)

⏺️ Young Futures and Bukhman Philanthropies launched a £4M UK programme. Groundwork UK will fund organisations working across youth mental health, wellbeing, learning, belonging and young people’s experiences with AI. The multi-year programme will begin with research and consultation with young people before its first funding round opens in spring 2027. (link)

⏺️ The UK’s AI healthcare commission published its recommendations for a new regulatory framework. The 119-page report called for regulation of healthcare AI to become more proportionate, lifecycle-based and system-wide, with greater use of real-world evidence and ongoing monitoring after products reach the market. It also recommended clearer rules for which AI products count as medical devices, more flexible pathways for updating AI systems, and stronger guidance for NHS organisations adopting them. The recommendations are not yet government policy; a cross-government response will follow. NHS Resolution separately published guidance on indemnity and liability when AI is used in NHS settings. We cover this report and many other elements of AI regulation in our recent article with May Lee. (link)

Join more than 500 other mental health leaders who are Hemingway Pro members. As a Pro Member, you get access to deep industry insights, access to our private Slack group and invites to our member-only events.

Research News:

⏺️ A study in Nature Mental Health Research found AI could perform parts of a psychiatric assessment, but struggled with symptoms requiring clinical reasoning. Researchers compared a multimodal AI model with two independent panels of psychiatrists as they assessed simulated patient videos using the Mental Status Examination. An examination which clinicians use to evaluate areas such as speech, mood, thought processes, delusions and perceptions. Across 396 assessments, psychiatrists showed substantial agreement with each other (ICC 0.87), while agreement between the AI and psychiatrists was moderate (ICC 0.70–0.72). The AI tended to over-identify problems it could directly observe, such as abnormalities in speech and affect, while under-detecting symptoms that required interpreting what a patient was experiencing, such as delusions. Its agreement with psychiatrists also declined as symptoms became more severe. (link)

⏺️ A new framework mapped which parts of psychological therapy AI may be able to perform and where human therapists remain important. The authors broke therapy into its component functions and assessed how readily each might be performed by AI. They argued that AI may increasingly handle tasks such as providing information, practising skills and monitoring progress, but identified three areas where human involvement remains particularly important: building and repairing therapeutic relationships, adapting treatment to the individual, and providing clinical accountability. The paper proposes a framework for thinking about how AI could be incorporated into therapy rather than simply testing whether AI can safely or effectively replace therapists. This supports the “unbundling” of therapy concept, which is a helpful framework for understanding how care models may change as AI develops. (link)

⏺️ A review found that despite decades of research, few psychiatric biomarkers are ready to help match patients to treatments. An American College of Neuropsychopharmacology position paper examined biomarkers ranging from brain imaging and blood tests to genetics, EEG and digital measures. The authors found that despite substantial research, few biomarkers can reliably identify which patients are most likely to respond to particular treatments or help select patients for clinical trials. They argue that progress has been held back by problems including inconsistent methods, limited replication and validation, fragmented datasets and a lack of coordination between researchers, drug developers and regulators. The paper also sets out a roadmap for overcoming these barriers. (link)

⏺️ An analysis of five schizophrenia trials found stopping paliperidone did not significantly increase the likelihood of rapid relapse. Researchers studied 417 people who relapsed after either continuing paliperidone or switching to placebo. In trials of long-acting injections, 20% of people who stopped treatment experienced rapid relapse compared with 11% who continued; with oral medication, the figures were 27% and 26%. Stopping antipsychotics still increases the overall risk of relapse, but the findings did not support the idea that discontinuing paliperidone commonly causes a distinct rapid-relapse effect. (link)

⏺️ A major review of 221 potential risk factors for psychosis found only six had convincing evidence. Researchers brought together 69 systematic reviews covering 970 studies and graded the evidence linking factors ranging from pregnancy and birth characteristics to ethnicity, migration, cannabis use and other health conditions with later psychotic disorders. Although 126 factors had statistically significant associations, only six met the authors’ highest standard of evidence: Black-African ethnicity in England, South Asian ethnicity in England, having a father younger than 20, two measures of low birthweight, and a mother having had at least three previous pregnancies. The authors stress that these are associations rather than evidence that the factors themselves cause psychosis, and that findings involving ethnicity are likely to capture underlying social and environmental exposures. (link)

That’s all for this week.

Keep fighting the good fight!

Steve Duke

Founder of Hemingway

P.S. If you find these updates valuable, consider becoming a Hemingway Pro Member. As a Pro Member, you will get access to even deeper industry insights and will be invited to join our community of over 500 mental health leaders.