This is an exclusive BHB+ story
Momentum in the serious mental illness (SMI) sector of behavioral health is steering away from episodic and reactionary and more toward whole-person, comprehensive care.
Historically, SMIs have been a collection of mental health conditions that even trained clinicians veer from due to their inherent complexities. Yet, I believe there is a missed opportunity here. The SMI space is ripe for innovation. And it could pay off for providers willing to take the risk. The average per-patient lifetime cost burden of SMI is $1.85 million.
There are signs this paradigm could be changing, as some virtual companies begin to invest in the space. This year, two major players in the digital health space, Sword Health and Valera Health, both made moves to enter SMI care.
Valera Health is a digital psychiatry platform that offers telemental health care, individual and group therapy and medication management for patients. The company acquired suicide prevention startup Vita Health in June, deepening its swath of SMI offerings.
Following that, Sword Health, a digital-first company that has primarily provided care in the physical pain and injury sector, announced a move into mental health with SMI care offerings on the horizon after receiving a $4 billion valuation.
“Severe mental illness is a very difficult problem to solve, and is very complex,” Virgilio Bento, Sword Health’s CEO, told me previously. “I think AI tools may attempt to have a good impact on it in the short term.”
Bolting onto the momentum around SMI care offerings, Bento said that when it does add SMI to its wheelhouse, the company will likely do so by way of an acquisition.
Still, I think it’s important to remember that SMI care is highly complex and often requires working closely with Medicaid. Venture-backed SMI startup firsthand laid off most of its staff in five states after turmoil with Medicaid contracts earlier this year.
As Medicaid cuts from the One Big Beautiful Bill Act begin to take effect, providers in this space will need to pay close attention to novel solutions and techniques. This will allow them to make the most of care practices for this patient population as regulatory and reimbursement frameworks tighten.
Based on recent research and conversations, I have identified some key trends emerging within the SMI care business that are likely to solidify as mainstays in 2026 and beyond.
In this week’s exclusive, members-only BHB+ Update, I explore:
- How AI is adding value to SMI care, bringing efficiency and support for providers and patients
- The role that new devices like wearables might play in patient monitoring and care personalization
- How providers are rethinking patient engagement with treatment and medications
- The rise of collaborative and integrated care models for SMI patient populations
Realizing the value of AI
Just as Bento suggested, AI has the potential to impact SMI care for the better and recent research underscores that.
From hyper-personalization of treatment plans for SMI patients to the potential for earlier detection of SMI conditions using machine learning models, integrating AI’s versatility into clinical applications holds the potential to rethink treatment.
“Early diagnosis enabled by machine learning models could potentially reduce the duration of untreated illness in schizophrenia and bipolar disorder, leading to better prognoses and improved illness trajectories,” the authors of a study published in February in JAMA Psychiatry wrote.
The study evaluated whether machine learning tools could use notes and data from the electronic health records of patients who were already in some form of treatment for less severe mental disorders and use that information to predict the onset of bipolar disorder or schizophrenia. They found the AI-powered technology was successful in doing so, but was more capable of detecting and predicting schizophrenia than bipolar disorder.
Ultimately, the study author determined that the model still needs more research and training for accuracy before clinical deployment. However, this points to a direction AI could take SMI care in the future – sooner diagnosis, saving costs and improving outcomes.
Other AI-infused tools can even save clinicians time.
Dr. Katherine Hobbs, CEO of Author Health, a specialized mental health provider that serves older adults with SMI diagnoses, told me her company is using AI tools to reduce administrative burdens and support patients between clinical visits.
“There are some very clear skills and practices that you [can] implement in response to the need, and that’s where things like AI come in,” Hobbs said. “You don’t need a clinician to help you with these things. Patients can very much be guided through this practice of a tool, asking ‘How are you feeling today?’ [and they respond] ‘Oh, I’m feeling anxious.’ Then it responds, ‘These are the things you have identified that are helpful for you when you’re anxious. Let’s do a deep breathing exercise.’ Those kinds of things are ripe for technology. Having that assistance in between visits with the clinician and having an AI assistant do those things really helps patients adhere to their treatment and do that work in between their visits.”
However, standalone AI tools have been blamed in other scenarios for inducing psychosis, encouraging suicide and reinforcing dangerous delusions.
Preprint research released Sept. 13 concluded that “across 1,536 simulated conversation turns, all LLMs demonstrated psychogenic potential, showing a strong tendency to perpetuate rather than challenge delusions.”
Stas Sokolin, CEO and co-founder of Amae Health, a company that focuses on providing behavioral health services for patients with SMIs, explained that using AI with clinical oversight is even more important for this patient population.
“I think it’s a much better idea to have personalized AI therapists where it is working hand-in-hand with your provider,” Sokolin told me.
Likening its use case to a physical therapist giving a patient homework and reporting back, Sokolin said AI can help SMI patients engage in care practices and coping techniques outside of treatment or therapy sessions and it can be used to adjust the course as necessary.
“Imagine there was a way of actually tracking how well a patient does on the homework? It’s a bit different with verbal versus physical, but I view that as a really good next step,” Sokolin said. “That’s something we have initially built and are going to pilot at some point soon, where it’s an adjunct to your clinical care team. It is not replacing anyone. It is personalized.”
I’ll be closely watching to see if the hype behind AI in SMI care pays off and if the technology is able to improve clinician efficiency and yield cost savings. If AI can improve early detection of SMIs, this can pay off enormously in the space. Research has demonstrated that patients who receive early interventions for psychosis have fewer hospitalizations and more years of employment.
If AI could be a tool in detecting SMI early, supporting patients between treatment sessions and reducing clinician burden, this has the potential to improve lives and save significant amounts of money.
Wearables and digital monitoring
In the vein of technology advancements, wearable devices are becoming more commonly used alongside therapeutic techniques, promising to become a “stethoscope” of sorts for SMI care.
Sonia Garcia, co-founder and chief growth officer of Amae Health, previously told me these tools are on the rise. As the sensors on wearable devices become advanced, they could become “the first part of a stethoscope for measuring mental health.”
The company began using these devices after receiving a grant to study how adherence to the ketogenic diet affected patients with SMI. The devices became helpful for care teams to see sleep patterns, activity, stress levels and heart rate variability – which can signal a rise in SMI symptoms.
“The ketogenic intervention had phenomenal engagement and outcomes, but it was great for people to see and want to try new things with that,” Garcia said. “We have deployed wearables within our clinics to also analyze activity data and get to know more of their physical health and embed that into treatment recommendations.”
Yet, a preparedness from clinicians to act upon real-time data from patient wearables and how to use that information to ultimately guide care plans is still in question to a degree across the field.
But with the push from U.S. Health and Human Services (HHS) Secretary Robert F. Kennedy Jr. to have every American use a wearable device for health purposes within four years, it’s likely that these advancements will continue, allowing for even further care personalization.
While industry insiders have discussed the potential benefits of wearables, I could see patients, especially those with negative health care interactions, such as involuntary commitment, being skeptical of the technology. For wearables to fully become a mainstay in SMI care, there will likely need to be a lot of trust between the patient and provider.
Emerging methods for engagement
One of the major challenges in caring for SMI patient populations is treatment engagement and medication adherence, but Hobbs at Author Health believes technology advancements can also play a role here.
Among the roughly 14.6 million U.S. adults who have an SMI, around 70.8% received mental health treatment in the last year, yet almost 3 in 10 did not receive any treatment, according to the National Survey on Drug Use and Health.
The primary reasons for not engaging with treatment among adults with SMI diagnoses ranged from not knowing how or where to receive treatment, not being ready to begin treatment, not having enough time for treatment or not finding a program or clinician they trusted to get treatment from.
“There are several companies that are focused on engagement and a lot of the major pieces are leading with relationship building,” Hobbs said. “If we can be thoughtful in how we address their needs, build trust and relationship over time, as the patient is interested and needs care, then we’re there to offer it, if that makes sense. … You can imagine that’s very resource-intensive, so we’re also thinking about how you can use technology to drive that relationship building and that engagement.”
Amae Health has seen care retention rates of as high as 72% at the 90-day treatment mark across its patient population, which Garcia and Sokolin attribute to the company’s long-term approach to care rather than episodic like many programs are.
“I think most programs are structured to be a short-term program lasting eight, 10 or 12 weeks,” Sokolin said. “Schizophrenia or bipolar is not going to go away in eight to 12 weeks. It might get better, but you still need some level of ongoing support, and that’s also been a very key, critical part of our program.”
There has also been a shift in the mindset of patients, Garcia has seen, since mental health has become less stigmatized in recent years.
“It’s not just are you engaging? You’re staying in your care, Garcia said. “I think from the patient perspective, there’s a lot more acknowledgement about ‘I’m going to have to manage this over the long haul to really keep it under control.’”
Across the SMI patient population, medication adherence can be as low as 16.1% for patients presenting with early episodes or around 52.5% for those recently discharged after an episode. These patients also change medications fairly often – making care continuity and pharmaceutical outcomes challenging.
“To me, that’s the most pressing issue is, how do we get better uptake and adoption of these treatments that actually work?” Hobbs said. “It’s engagement and making sure that people actually feel connected and can trust the system, so that they actually want to be engaged.”
Addressing the workforce shortage of those who can prescribe SMI medications across the field and better payment models will build a better ecosystem, she said, “where these treatments are actually being used, and people can actually get to them.”
Engagement is probably the hardest question to solve for in SMI care and there is no easy fix. The decline in stigma and improvements in the clinical alliance will likely help boost engagement.
The company firsthand has an innovative model, which uses peers to engage patients, though it is heavily reliant on Medicaid contracts. I’ll be closely watching to see which engagement models work best for this notoriously difficult-to-reach patient population.
Collaborative, value-based, integrated care
Further integration of physical health and social components into treatment for SMI patients is also becoming a more common thread across the behavioral health providers who treat these conditions.
Nutrition, sleep and exercise are often the last components on a patient’s mind when they are in the middle of a psychosis, but are core to getting patients all the way stabilized, Garcia explained. Creating ways to incorporate physical care with SMI treatment and monitoring both physical and mental health progress is going to be a focus of future care models, she hopes.
“Right now, there’s no area to even consider how you should be really regulating all of these things, but it only exacerbates the issue if they’re not under control,” Garcia said. “Integrated care matters. … You need all those stabilizing factors to really come into play before you can even metabolize the antipsychotic meds that you are on, which are usually very heavy, hardcore medications that alter a lot of your physiology in very profound ways.”
Building on that, the capacity for the collection of AI, new technologies and wearable devices to present concrete data that providers can use to show value and outcomes may give way for an increased number of value-based arrangements to emerge in SMI care.
“Treating having adequate payment is a critical piece of actually being able to deliver the robust, robust care that these patients need,” Hobbs said. “Then being able to track with data and being able to share that data back is a critical component.”
While it is promising, Sokolin thinks a wave of value-based contracts might still be wishful thinking for many SMI providers.
As patients are becoming younger and younger when they experience their first psychosis, Sokolin explained that some organizations may attempt to do too many things at once and not spend enough energy where it matters for these complex patients, which could prolong this.
What is more likely in the interim, Sokolin and Garcia predict, is a growth of collaborative care models for SMI patients between providers, health systems and hospitals.
“A lot more folks are being hospitalized for cardiac conditions, oncology, other things, but they also have an SMI diagnosis, and they are not responding very well to their oncology treatments or other components, because their schizophrenia and bipolar disorders are not managed right at all,” Sokolin said. “It’s not solely patients being released from inpatient psychiatry, but also, what about a patient who has a heart attack and has schizophrenia? How do we manage that patient together? They can do the cardiology, but we can work in tandem to treat the SMI.”
Patients with SMIs incur much higher medical costs on average than those without. I think integrating physical health services into the fold of care could improve overall health outcomes and cut costs. Still, providers coordinating these services will need to have robust integration plans and hammer out the reimbursement pathway in order for these services to be sustainable.
The swath of new technologies, modalities of care and novel techniques for patient engagement and collaborative care are certainly positioned to move the needle toward more comprehensive wellbeing for patients with SMI, but change does take time. I anticipate we’ll see growth in this space begin with the adoption of technologies first, before value-based care arrangements come more into play.
The elephant in the room
While technological advancements may be a major lever in improving care for people with SMI, insurance coverage and reimbursement pathways remain a major sticking point.
More than 2.3 million Medicaid enrollees have been diagnosed with a serious mental illness, according to data from the Kaiser Family Foundation. Major cuts to the Medicaid program could threaten coverage for individuals with SMI.
As a result, providers may continue to provide services for SMI patients that they don’t get reimbursed for – further constricting time and resources.
While individuals with a “disabling” mental health condition are exempt from Medicaid work requirements under OBBA, the additional paperwork burden could lead many to lose coverage.
This could disproportionately impact providers offering SMI care, including major players such as Universal Health Services (NYSE: UHS) and Acadia Healthcare (Nasdaq: ACHC).
These cuts have yet to take effect, but some behavioral health providers worry they could be disastrous for safety-net clinicians and have major implications for Medicaid innovation – which could slow down progress on SMI care overall.
Before innovation really takes off for SMI care, first providers must figure out how to weather the implications these cuts may bring to their practice. I feel like new funding models and collaborative care could bridge some gaps, but the ecosystem as a whole might take a hit with layoffs, higher demand and more scrutiny.
Companies featured in this article:
