Catching Crisis Earlier: What Continuous Signals Reveal Before a Safety Event

The warning that lands between visits

Most people picture a crisis as something that arrives without warning. In practice, the warning is often there. It just lands in the weeks between appointments, in a stretch of time when no one is scheduled to be watching. A person can leave a visit stable and reach a point of acute risk long before the next one, and the signs that something changed rarely wait for the calendar.

This is the quiet reality that clinical and safety leaders live with. Care is organized around scheduled contact, but risk does not keep that schedule. The result is a blind window, and closing it is one of the hardest and most important problems in behavioral health.

The blind window

Consider what care actually looks like from the patient’s side. Someone at elevated risk might see a provider once a month, sometimes less. Between those visits, life keeps moving. A relationship ends, sleep unravels, a loss reopens an old wound, and the internal state shifts in ways that no one on the care team can see until the next appointment, or until a crisis forces contact.

The people carrying this weight are clinical and quality leaders and the frontline clinicians responsible for at-risk patients. They know the between-visit stretch is where risk concentrates, and they know that a monthly snapshot cannot capture a fast-moving change. The problem is not effort or attention. It is that the standard rhythm of care leaves long unwatched intervals, and warning signs tend to appear inside them.

That is what makes this a patient-safety priority rather than a scheduling inconvenience. When the window is unwatched, the earliest and most actionable signs of deterioration are the ones most likely to be missed.

Why the between-visit window stays unwatched

The gap persists for a reason that has nothing to do with anyone doing their job poorly. Episodic check-ins were built for continuity of care, not for continuous risk detection, and those are different jobs. A visit every few weeks is a reasonable cadence for adjusting a treatment plan. It is a poor instrument for catching a change that develops over days.

There is a second reason the window stays dark. People in distress often do not reach out in the moment, and they may not report what they are feeling even when someone asks directly. That is not evasion; it is part of the clinical picture. Yet the same people frequently do have recent contact with the health system. The evidence here is striking and sobering: more than 40% of people who die by suicide were seen by a primary care provider in the month before their death, and contact with mental health services, while lower, is far from rare. The touchpoint often exists. The signal simply did not surface at the moment it could have changed the response.

What continuous signals add

If the problem is an unwatched window, the answer is not more appointments layered onto already stretched clinicians. It is a way to see meaningful change during the interval, without asking the patient to schedule their distress or the clinician to work around the clock.

Continuous behavioral signals do that. Structured, between-visit measures, gathered through brief and regular check-ins, give the care team a read on how someone is actually doing over time rather than only on the afternoon of a visit. In practice, that can look like a short check-in a person completes on their own device every few days, a couple of minutes responding to prompts by voice or video from wherever they are, rather than a new appointment to attend. When those measures shift in a direction that matters, the system can flag it and route it to a clinician. The clinician then decides what to do with a fuller picture and more time to act.

The framing matters here, and it is worth being precise. The technology surfaces signal and puts it in front of a person who is trained to interpret it; it does not make the clinical decision. Videra Health builds this to keep the clinician in the loop on every decision, because judgment about risk belongs with the clinician, not the model. What changes is not who decides. What changes is how early they can see enough to decide well.

What it looks like in practice

This is not only a concept. A behavioral health services network in Upstate New York serving more than 4,000 individuals put continuous, AI-powered monitoring in place to identify people at risk before escalation. Care coordinators received early, structured signal from regular check-ins and could reach out proactively rather than waiting for a call or an emergency-room arrival.

The organization saw a 64% reduction in crisis alerts, meaning fewer people reached the point of acute decompensation, not fewer people watched. Among those actively monitored, the network reported zero behavioral-health-related emergency department visits and zero police interventions, while depression and anxiety scores (PHQ-9 and GAD-7) improved by half within two weeks. Staff time that would have gone to crisis response shifted toward earlier, preventive outreach. The result reads less like a dramatic intervention and more like the quiet effect of seeing sooner: someone noticed the change while there was still room to help. It is evidence of what earlier visibility makes possible, not a promise of any single outcome.

A visibility problem, and a solvable one

The encouraging part of this problem is that it is a visibility problem, and visibility is something that can be improved. For a long time, the only realistic way to check on someone between visits was to wait for the next visit. That is no longer the only option. Continuous, clinician-guided signal gives care teams a chance to meet a person in the window where warning signs actually appear, and to do it with support rather than alarm.

There is a version of behavioral health care where the weeks between appointments are no longer a blind spot, and it is closer than it used to be. For a closer look at how continuous monitoring played out in a real behavioral health network, see Videra Health’s case study on AI-powered crisis monitoring at BHSN.

If you or someone you know is struggling or in crisis, help is available. You can call or text the 988 Suicide and Crisis Lifeline, any time, to reach trained counselors who are there to listen.