Is digital technology on the brink of becoming a viable treatment option for schizophrenia? While many experts are optimistic — especially about its potential to address negative symptoms — opinions vary on how quickly these tools could move research into clinical practice. Questions about feasibility, scalability, cost, and access remain major barriers.
Until last year, no new schizophrenia medication had been approved by the FDA for decades. But after disappointing results of xanomeline/trospium chloride (Cobenfy, Bristol Myers Squibb), many clinicians are calling for additional treatment options and are becoming more open to exploring digital and technical solutions.
From mobile apps to virtual therapy platforms, a growing number of digital interventions show promise in managing schizophrenia — particularly its often intractable negative symptoms.
Clinicians and researchers see digital tools as uniquely positioned to engage patients with schizophrenia, offering low-burden support between appointments. Yet major hurdles remain: Most technologies aren’t yet available outside trials, few are covered by insurance, and evidence is often limited to small studies.
“People are frustrated. Treatment is still so limited that I think there’s real excitement about digital options,” said John Torous, MD, psychiatrist and director of Digital Psychiatry at Beth Israel Deaconess Medical Center in Boston.
So far, only one digital therapeutic (DTx) designed solely for schizophrenia has received Breakthrough Device Designation from the FDA, and none has received full marketing authorization in the US.
While many other DTx are under investigation, they remain confined to the research stage or are designed to treat related conditions. Others are marketed as “wellness” apps rather than “medical” apps in order to avoid FDA-clearance requirements.
Although the field isn’t quite there yet, Torous believes “we’re closer than people realize” to practice-changing schizophrenia-focused DTx, especially with the advancement of artificial intelligence (AI) and with at least one large trial currently underway.
“It’s tantalizingly close yet still far away,” Torous said.
App Gap
The World Health Organization estimates that approximately 24 million people worldwide have schizophrenia — about 1 in every 300 individuals.
The condition commonly includes “positive” symptoms, which involve abnormal experiences such as hallucinations, delusions, and disorganized speech or behavior, and/or “negative” symptoms characterized by a loss of normal functioning, including blunted affect, social withdrawal, decreased motivation, loss of pleasure, and cognitive deficits.
Current treatment mainly consists of antipsychotic medications combined with psychosocial therapy. Emerging approaches include avatar therapy for hallucinations, virtual reality (VR)-based cognitive-behavioral therapy (CBT), and a range of digital apps.
The Pew Research Center reported in November that 91% of all Americans own a smartphone, including 79% of adults over the age of 65 years.
“We have very good data that people with schizophrenia are comfortable using apps,” said Torous, who is also an assistant professor of psychiatry at Harvard Medical School, Boston.
“But there’s a gap where the apps built in the research world haven’t become accessible tools that you and I can download and give to someone. The FDA hasn’t fully cleared any yet, and of the ones that are direct-to-consumer, none of them are superb,” he said.
While digital tools hold a lot of promise, improvements are needed to make them accessible and scalable, Torous said. “We have really good theoretical evidence at this point that these can work. So it’s time to put them to a real test in the real world,” he added.
Higher Engagement With Schizophrenia Apps
In the US, CT-155 (Boehringer Ingelheim and Click Therapeutics) received Breakthrough Device Designation from the FDA last year. The mobile app-based software is designed to address negative symptoms of schizophrenia when used alongside standard pharmaceutical treatment.
In addition, the National Institute for Health and Care Excellence (NICE) released an “early value assessment” last year recommending three healthcare technologies as potential adjunctive treatments for schizophrenia in adults and youth within the UK’s National Health Service “while more evidence is generated.”
These included avatar therapy for managing auditory hallucinations; SlowMo, a digital therapeutic that delivers eight app-based therapy sessions aimed at slowing distressing thoughts or paranoia; and CareLoop, a smartphone app designed to monitor psychosis symptoms and help prevent relapse.
“Digital health technologies offer another option for people with psychosis who may otherwise not have psychological interventions,” the NICE report authors wrote.
In an earlier report, NICE also recommended gameChange VR for agoraphobia avoidance. This automated CBT product guides users through simulated social situations.
Scores of other digital approaches are being investigated, and most show effectiveness, especially for negative symptoms. At this year’s American Psychiatric Association (APA) annual meeting, nearly every research poster session featured presentations on apps targeting conditions such as insomnia, substance use, anxiety, and schizophrenia.
Interestingly, patient engagement is higher for schizophrenia-focused apps than for those targeting other mental health disorders, Torous noted.
“I think that’s because schizophrenia apps are built really well. There’s a lot of concern regarding harms for these patients, so there’s a real effort not to cut corners. It could be that because more eyes are [monitoring] these apps, there’s more attention to detail,” he said.
A Catch-22
So why haven’t more of these research-backed treatments been approved for clinical use? Torous explained that to date most studies have been relatively small, and the research is stuck in a catch-22: Sponsors are hesitant to fund large trials without clear evidence of feasibility — yet that level of evidence can’t be generated without larger studies.
In addition, key questions remain around how to deliver timely, cost-effective training and how to scale digital therapies for broad implementation. Another major concern: Will insurance ever cover these treatments?
Torous noted that the large phase 3 CONVOKE trial could possibly answer a lot of these questions. It has pharmaceutical backing and, if proven effective, could roll out quickly, he said.
Launched in May 2023, CONVOKE is a 16-week randomized controlled trial evaluating the safety and efficacy of CT-155 or a second digital therapeutic smartphone app when used alongside standard antipsychotic treatment for negative symptoms of schizophrenia. The study aims to enroll 432 participants across 54 US sites.
The development company already received FDA marketing authorization in April for one of its other DTx smartphone apps (CT-132) for the prevention of episodic migraine. However, the product is not yet available to prescribe or purchase.
While not specifically designed for schizophrenia, several currently accessible apps target-related conditions, particularly depression. Widely available AI-powered chatbots are also showing promise in supporting mental health care.
ChatBots an ‘Emotional Sanctuary’?
In a small 2024 study by Torous and Steven Siddals, who was with King’s College London, London, England, at the time of publication, participants reported high engagement and positive effects from using commercial mental health-related generative AI chatbots such as ChatGPT, while emphasizing the need for safety “guardrails.”
“What surprised us the most was the depth of emotional support people described. Participants told us it felt like a kind of ‘emotional sanctuary’ — a safe space where they could be vulnerable without fear of judgment,” Siddals said in a release.
“It really was phenomenal how deeply understood people felt talking to [the chatbot], even knowing it wasn’t human,” he told Medscape Medical News. It was also used by some participants as a “practice session” in order to work out feelings before going to in-person appointments.
Siddals emphasized the value of future research to better understand what contributed to the intervention’s success. He noted that AI therapy offers a unique opportunity for in-depth analysis, since — with patient consent — researchers can access complete transcripts of interactions.
Building on these promising early findings, results from a larger 200-participant randomized controlled trial published in April demonstrated that a generative AI-powered therapy chatbot known as Therabot was linked to significant reductions in major depressive disorder, generalized anxiety disorder, and eating disorders — though schizophrenia was not among the conditions assessed in the study.
“The effect sizes weren’t just significant; they were huge and clinically meaningful — and mirrored what you’d see in a gold standard dose of evidence-based treatment delivered by humans over a longer period of time,” senior study author Nicholas Jacobson, PhD, Dartmouth College’s Geisel School of Medicine, Hanover, New Hampshire, told Medscape Medical News at that time.
However, in a small pilot study presented at this year’s APA meeting, a human therapist outperformed an AI model (ChatGPT-3.5) for feedback, collaboration, and guided discovery in the delivery of CBT. Still, the AI did perform well in providing a structured therapeutic approach, the investigators reported.
Blurred Lines
There is also a proliferation of so-called “wellness” apps. In an editorial published online in Schizophrenia Research last year, Torous and colleagues wrote that many digital mental health products “skate the line between medical and wellness classification, a distinction often perceived as a blurred line from the wider public.” This means patients are using wellness apps they believe meet their needs — if they are able to find them.
At the time the editorial was published, only 10 direct-to-consumer apps specifically designed for schizophrenia were publicly available, with several since abandoned by their developers.
Torous pointed out that while it typically takes about 10 years to bring a new medication to market, digital apps require significant updates within that same timeframe due to the rapid pace of technological change.
Reports also show there’s a substantial plummet in engagement by the general population, with a 30-day retention rate of just 5.5%. This drop-off is even greater for those with psychiatric disorders.
Still, as patients increasingly turn to a variety of apps, clinicians should inquire about which patients are already using and consider integrating elements of those apps into treatment, Torous noted. For example, apps can serve as a form of “homework,” where patients use them at home and then discuss the results during clinical visits.
“Sometimes what I tell patients we treat is: ‘Even though there’s not a perfect path for your condition, you can still build a toolkit of different apps that provide things you like,’” Torous said.
Pros, Cons, and Ethics
Positives for technology-focused psychiatry treatment include that it usually has a better side-effect profile and provides easier and quicker access to care such as CBT.
Caveats include that these tools should only be used in collaboration with a trained professional — and, whenever possible, following a clinical assessment to determine what is best for the patient, Torous said. Additionally, goals and methods for monitoring progress should be established from the outset, along with clear protocols for sharing information about any potential adverse events.
But are there other concerns, red flags, or even ethical issues?
Paul Appelbaum, MD, practicing psychiatrist and professor of psychiatry at Columbia University, New York City, told Medscape Medical News a big question with any type of new technology is whether it really makes a difference in outcomes.
He noted that passive monitoring technologies being developed for the detection of early relapse are promising. “But the rubber hits the road when we can demonstrate that outcomes, not just in controlled studies but in real-world settings, actually improve. And we’re not there yet,” said Appelbaum, who is also director of the Division of Law, Ethics, and Psychiatry at Columbia.
Regarding chatbots, Appelbaum said there are several “obvious” ethical considerations. First, it is imperative that individuals are told exactly who (or what) they’re interacting with upfront and what the degree of human oversight might be. For example: If a chatbot responds inappropriately, how quickly can that be detected and then followed up with human intervention?
He also pointed out some privacy concerns. “Who exactly has access to the interactions and to what extent the privacy of the patients is protected are extremely important issues,” he said.
Appelbaum said he believes researchers are very careful with this, but as commercial apps become readily available, “the degree of care about protecting patient privacy may not be the same.”
In fact, the American Psychological Association sent a letter to the Federal Trade Commission in December about concerns it has with “underregulated development and deceptive deployment of generative AI or enabled technologies,” especially chatbots that represent themselves as therapists. It added that safeguards need to be put in place.
Advice for Clinicians: Get Ready
Still, new digital options are coming — and in some cases, are already here. The takeaway message for clinicians is that they should “get prepared” for what’s coming down the research pike by reading up on everything they can and trying out various small apps in order to become comfortable with the overall technology, Torous said.
“Medication and therapy can be effective. But if we could give patients the full range of evidence-based tools, including technology, that would be a big win,” he said.
Siddals agreed that clinicians should get familiar with these tools because “it’s coming, and I think it’s going to get much more prevalent.”
He added that there shouldn’t be a worry that AI or other tech will take away jobs but should be viewed instead as a possible assistant that is available 24/7, if needed.
“Would I ask patients about these tools? Yes. Would I recommend any of them? That’s harder because we don’t yet have strong evidence on effectiveness [for schizophrenia] or on safety. We don’t know enough about what the parameters should be,” Siddals concluded.
Torous, Siddals, and Appelbaum reported having no relevant financial relationships.