A patient comes in for their annual physical and screen negative for depression and anxiety. Six months later, they lose a job, a parent gets sick, and symptoms creep that no one on their care team will see until the next scheduled visit.

This is the first and most consequential breakdown in behavioral health referral leakage: patients with real behavioral health needs aren’t identified in a timely way. In our last post (link), we shared that of the 23% of patients who meet the diagnostic criteria for a behavioral health condition, about half are never diagnosed. Before health system leaders can improve referrals, follow-ups, or closing the loop, they must answer a more basic question, “How do we know a patient needs help at all?”

The Issue with Annual Check-ins

For most health systems, behavioral health screening happens at the annual primary care visit, if it happens at all. That cadence assumes a patient’s mental health today will look roughly the same eleven months from now, which is rarely the reality. 

Depression, anxiety, and substance use don’t wait for a scheduled physical. A patient can leave their annual visit in good standing and be in crisis by the following spring, and a once-a-year model is unable to catch that.

Real Identification Should Be Population Wide

Even when screening happens on a better cadence, it still depends on catching the right patient at the right moment and asking them directly. Real identification has to be population-wide, not limited to patients already flagged in a chart or already suspected by a provider. That approach will often miss the majority of people with unmet needs, because it depends on staff noticing a problem before any data confirms one exists. NeuroFlow CEO Chris Molaro put it plainly in an open letter to healthcare leaders: 59.3 million U.S. adults have a mental health need, and nearly half don’t receive treatment. That figure doesn’t even include the millions who go undiagnosed. Health systems can’t manage a risk they can’t see.

But population-wide screening has a ceiling too. Not every patient will respond, and not every care team has the bandwidth to follow up with everyone who does. That’s where AI-powered analysis of EHR data becomes a force multiplier. By surfacing patients who show physical and medical signals of an underlying behavioral health condition well before a formal diagnosis, the analysis creates a targeted chase list: the patients who most need to be reached, screened, and engaged. Limited resources get directed where they’re most likely to make a difference.

Patient Identification Before Screening

Rather than waiting for a patient to be screened or to disclose symptoms, NeuroFlow’s BHIQ analyzes the clinical data that health systems already have, such as diagnosis codes, procedures, medications, encounter and utilization patterns, demographics, and social risk signals drawn from the longitudinal EHR. This flags patients whose clinical pattern closely resembles people with a known, confirmed behavioral health need, even if that patient has never been screened or diagnosed. No survey or manual intake is required to generate the flag.

Each flagged patient comes with a risk tier and a recommended next step sized to match: the highest-risk patients get direct outreach from a care coordinator, moderate-risk patients get a remote screener, and lower-risk patients get a screen folded into their next scheduled primary care visit. That turns identification from a blanket, resource intensive exercise into a prioritized system. Care teams spend their limited outreach capacity on the patients most likely to need it.

What Population-Wide Screening Actually Looks Like 

Bergen New Bridge Medical Center, the largest hospital in New Jersey and a Medicaid safety-net provider, faced this problem directly. Rather than relying on occasional or annual screening, Bergen introduced seamless digital screening for depression and substance use during ambulatory visits, reaching patients as a part of routine care rather than waiting for a flag to go up.

As a result, Substance use and mental health screening rates jumped from 5.8% to 81.6%, a 14x increase, in a matter of months. In just a three-month window, the expanded screening identified 36 patients at risk for suicide, who likely not have been identified under the previous model. The difference population-wide identification makes is not a marginal improvement, but an entirely different population of patients coming into view. 

Closing the Identification Gap with NeuroFlow

NeuroFlow’s BHIQ addresses the identification problem directly. Its AI-powered data models analyze EHR data to surface patients with suspected behavioral health needs — before a diagnosis exists. The result is a prioritized chase list, ranked by severity, so care teams can direct limited resources to the patients most likely to benefit from outreach. Then, health systems can deploy NeuroFlow’s IntegrateBH platform to close the loop.

The technology deploys screenings through email, SMS and other digital channels, automating regular data collection so that providers are alerted as soon as patients indicate behavioral health symptoms. Patients with moderately severe behavioral health conditions receive intelligent behavioral health referrals, personalized to their specific needs and preferences. The technology also closes the loop, ensuring patients actually complete their referral and sharing those insights with the referring provider. In a peer-reviewed study, regional health systems leveraging IntegrateBH increased the likelihood of patients receiving outpatient behavioral health care 68% and increased speed to care 36%.

Together, BHIQ and IntegrateBH turn behavioral health identification from a reactive, moment-in-time exercise into a continuous, population-wide system — one that finds patients who would otherwise go unnoticed and ensures they get connected to care.

Schedule a personalized assessment to understand where patients are falling out of your behavioral health screening process and what closing that gap could mean for clinical outcomes and financial results.

Learn more:

• IntegrateBH
• NeuroFlow Expands IntegrateBH Solution: Specialized Technology Infrastructure Embedding Behavioral Health Across the Care Continuum

Related posts:

NeuroFlow Named to Fierce Healthcare’s 2026 Fierce15 Why Owl And NeuroFlow Have Joined Forces to Improve Patient Monitoring & Arm Providers With the Data They Need Key Benefits of Mental Health Treatment in Primary Care Proven Suicide Prevention Strategies Exist. Implementation at Scale Is The Challenge