VIGIL is an intelligent relapse prevention system for individuals living with schizophrenia and serious mental illness. It transforms the critical window between early warning signs and psychotic relapse into an opportunity for collaborative action — equipping patients to recognize their own patterns, family members and care partners to identify early warning signs, clinicians to respond with precision, and everyone involved to work from the same shared understanding.
Research has established a critical fact about psychotic relapse: early warning signs appear approximately four weeks before full decompensation. That's a month-long window where intervention could change everything — preventing hospitalization, preserving functioning, protecting relationships, and avoiding the cascading losses that follow each relapse episode.
Yet in practice, this window is almost never used. Patients may not recognize their own warning signs. Clinicians see patients too infrequently to catch subtle changes. Family members notice something is different but don't know what to do. Despite decades of evidence that 30–40% of patients relapse within a year of discharge, mental health care still lacks practical tools for systematic early recognition and collaborative response.
VIGIL was built to close that gap — not with more medication monitoring, but with a behavioral system that teaches patients to read their own signals, trains family members and care partners to recognize escalation patterns, and gives clinicians the structured data they need to act. Because relapse prevention isn't a solo effort — it's a shared vigilance.
Sleep changes, social withdrawal, concentration difficulties, increased suspicion — the subtle shifts that precede relapse are highly individual. VIGIL helps each person identify and track their own unique pattern.
Family members are often the first to notice that something has changed — but they don't know what to look for or how to respond. VIGIL equips care partners with the knowledge to recognize escalation patterns and act before crisis, making them active participants in prevention rather than helpless bystanders.
Life stressors and environmental changes are among the strongest predictors of relapse. The system systematically assesses stress exposure and individual vulnerability, connecting triggers to prevention strategies.
Medication non-adherence is the single strongest predictor of relapse. VIGIL addresses adherence not through surveillance, but through understanding barriers, strengthening the therapeutic alliance, and shared problem-solving.
VIGIL integrates assessment, analysis, collaboration, and action planning into a seamless clinical workflow. Each component was designed for people with active cognitive symptoms — using video narration, simplified interfaces, response-adaptive logic, and touch-screen interaction that meets users where they are, without condescension.
Video-enhanced, narrated self-assessment covering five evidence-based risk factors
Two reports carrying the same information — one for the patient, one on the clinician’s tablet — built to be read side by side
Shared decision-support tool for real-time patient-provider problem-solving
Personalized relapse prevention action plan with concrete next steps
The self-assessment is completed before a routine appointment, at home or in the clinic. It evaluates early warning signs, life stressors, social support, therapeutic alliance and medication adherence — the five factors research identified as driving relapse — along with current symptoms, illness perception and relapse history. The response-adaptive logic presents only relevant questions based on prior answers, so no two assessments follow the same path. In the pilot it ran to roughly forty-five minutes, and people stayed with it: “I lost track of time” was a common remark.
VIGIL is fundamentally different from medication tracking apps and symptom diaries. It is a behavioral health intervention built on three psychological principles that drive its design at every level.
The assessment itself is an intervention. Patients learn the stress-vulnerability model, the importance of early warning signs, and the role of medication and social support through engaging with the tool — not through lecture.
By involving patients in identifying their own warning signs, stressors, and protective factors, the system shifts the locus of control from passive medication compliance to active relapse monitoring. Patients develop ownership of their recovery plan.
The patient, having the lived experience of the illness, is expert in the illness; the clinician is expert in treatment. Both bring their resources to bear on the same data, and the plan is what they agree on — not what one of them was handed.
Understanding relapse risk factors isn't a prerequisite — it's a natural consequence of using the system. The assessment process itself teaches patients to think about their illness in structured, actionable terms.
Before a line of the program was built, twenty project advisors were recruited from the clinic: ten clients, three peer providers and seven clinicians. Each was interviewed twice — once about their own experience, once while operating a working version of the system — and then brought together in focus groups to argue with one another about it. The design changed because of what they said.
The peer providers mattered more than anyone expected. People carrying a mental illness diagnosis who work in a supportive role to other clients, they spend hours in ordinary conversation on errands and in waiting rooms, and they knew things about the population that neither the clinicians nor the designers did. They were candid about who the program would not reach as well as who it would.
One finding from those interviews shaped the whole interface. Clients said they were willing to be more honest with the software than with a person in the room — as one put it, “you don’t get a bad vibe from a computer.” That is not a technology insight. It is a reason to treat the assessment as a place where something true can be said, and then carried into the appointment by the person who said it.
An 8-month pilot study, running from October 2013 to May 2014, evaluated the system with 24 patients and 10 clinicians at the University of Rochester Medical Center's Strong Ties Community Support Clinic, a Certified Community Behavioral Health Clinic. It assessed four areas across both groups: usability, satisfaction, engagement in relapse prevention, and overall helpfulness.
The people involved are worth stating plainly. Patients averaged 44 years of age, 96% carried a diagnosis of schizophrenia or schizoaffective disorder, and their illness had begun on average at 24. The clinicians — two psychiatrists, a nurse practitioner, a nurse therapist and six social work therapists — brought an average of sixteen years of practice each. The self-assessment ran forty-five minutes on a touch-screen tablet, and the consultation that followed ran about forty-five more.
Both groups scored it consistently, not just the patients. Clinicians rated usability 89.2% and satisfaction 93.3%; engagement in relapse prevention came in at 89.0% for patients and 88.1% for clinicians, and helpfulness at 92.3% and 91.3%. Ninety-two percent of patients said the program would help them talk about important things with their clinician, and 90% would recommend it to another patient.
These numbers matter because of who achieved them. Patients with active psychotic illness — a population often assumed to be unable or unwilling to engage with technology — rated the system highly usable and expressed clear intent to continue using it. Clinicians reported that the assessment data transformed their appointments: pre-visit documentation meant less time gathering information and more time on collaborative problem-solving.
Qualitative feedback revealed something the numbers alone don't capture: patients said the system helped them disclose concerns they hadn't raised before. The video-based peer narratives — hearing other individuals with lived experience describe their relapse journeys — decreased shame and increased willingness to discuss medication concerns, early warning signs, and stress vulnerabilities.
What the study does not show is worth saying as clearly as what it does. This was a cross-sectional test of usability and acceptability, not of outcomes. Whether RAPS reduces relapse rates, service costs or hospitalizations is an open question, and answering it needs a prospective randomized controlled trial against usual care. That was the report's own conclusion in 2014, and it still stands.
Following the pilot, the principal investigators recorded a professional video documenting their experience of the system and its impact on clinical practice. Dr. Steven Lamberti, Professor of Psychiatry at the University of Rochester Medical Center and Director of the Severe Mental Disorders Program, served as the study's Principal Investigator and Dr. Robert Weisman, Medical Director of Strong Ties at URMC, as Co-Investigator. Tom Conant was Co-Principal Investigator. Selected observations from the video appear below.
"In traditional biomedical science, studies are designed by doctors, by principal investigators. In this study, with the help of Collaborative Care Interactive, we had something called a user-centered design. It's the first time in my career that I actually worked in partnership with patients to develop research that they would be involved in. Otherwise, we end up with a pill that nobody will take."
Dr. Steven Lamberti — Principal Investigator, RAPS Study
"The use of this tool opened up and investigated things with the clients that they've had over the time that they never even touched. It allowed them to get deeper into some of the symptoms that really don't come up — by the triangle effect of using this third piece in the room."
Dr. Robert Weisman — Co-Investigator & Medical Director, Strong Ties
"The medication compliance part was incredibly helpful because it really did open a dialogue for patients to talk about how they are using medicine… The one person we never knew he was not taking his medicine as he was supposed to — and now that's a topic that he and I talk about at almost every session."
Clinician subject, RAPS pilot study
The clients who used it were less formal and, in places, more striking.
"It felt like I was in a support group with some very caring people."
Pilot participant, Strong Ties Community Support Program
"I felt that it was about me, that the people in it were like me, and the doctor really cared about me."
Pilot participant, Strong Ties Community Support Program
Most mental health technology is designed for the worried well — people who are already engaged, digitally fluent, and motivated to track their mood. VIGIL was designed for individuals living with the most severe and stigmatized mental illnesses, people whose cognitive symptoms, social isolation, and healthcare system experiences create barriers that generic wellness apps never address.
Built explicitly for psychotic disorders, with content and complexity calibrated to this population's cognitive and affective needs. Not a generic mental health platform stretched to fit.
Patients hear and see other individuals with lived experience discussing their relapse journeys, increasing psychological safety and normalizing conversations that shame would otherwise prevent.
The same assessment generates separate reports for patient empowerment and clinician efficiency — closing the gap between what providers know and what they can practically address in busy clinic settings.
Shared decision-making is widely recommended but rarely implemented. VIGIL provides concrete structure — a tablet-based consultation interface with assessment data, risk levels, and action planning tools.
VIGIL's development spans from evidence-based clinical research through real-world pilot validation to current AI-enhanced re-engineering. Each phase built on the last, refining both the clinical content and the technology platform.
Research and design of an evidence-based relapse prevention system grounded in five key risk factors: early warning signs, life stressors, social support, therapeutic alliance, and medication adherence. Clinical content developed in collaboration with psychiatric researchers at the University of Rochester Medical Center, Department of Psychiatry.
Full system development: video-enhanced self-assessment, dual-format risk profiles, shared decision-support tool, and personalized action planning. Pilot study launched at the University of Rochester Medical Center's Strong Ties Community Support Clinic, an award-winning Certified Community Behavioral Health Clinic, with support from Otsuka America Pharmaceuticals.
Eight-month pilot completed with 24 patients and 10 clinicians. Results demonstrated high usability (92%), satisfaction (90%), and clinician adoption intent (98%). Clinical reports and white paper documented findings and outlined a four-stage development roadmap.
Now being rebuilt with current AI capabilities: predictive analytics for relapse risk trajectories, natural language interaction, real-time monitoring and alert systems, and expansion beyond schizophrenia to other serious mental illnesses with relapse-remitting courses. The clinical foundation stays the same. The technology finally matches the vision.
Built under stated limits — what this will never do, and who owns what goes in.
Prevention is possible when people learn to read their own signals.
If you work in behavioral health and see the need for technology that supports real patient engagement in relapse prevention — or if you're interested in AI-enhanced approaches to serious mental illness — let's talk.
Start a Conversation