Why Scaling Behavioral Health and ADHD Care Requires More Than a Faster Front Door

Demand for behavioral healthcare has never been higher, and the pressure on health system leaders to meet that demand has never been more intense. For many organizations, AI has become the answer, driving real improvements in access by automating the administrative work that has long slowed care delivery. Scheduling, prior authorizations, billing, and the operational friction that causes patients to disengage before treatment can work are increasingly being handled without burdening the clinical team. More patients are getting through the door, and getting there faster.

But access is only half of the equation. The harder question, the one that does not get asked often enough, is whether the care those patients receive once they are in the door is accurate.

The Access Problem Is Only the Beginning

The barriers to behavioral healthcare in this country are well-documented. Research published in JAMA Psychiatry found that nearly half of psychiatrists do not accept private insurance, compared to roughly 90% of other physicians. For those who do, the average wait time for a new patient appointment is 67 days, and for patients on Medicare or Medicaid, the wait is typically longer. 

The operational friction compounds the problem. Patients who do secure an appointment often drop out of care prematurely, and the reasons are rarely clinical. Prior authorizations stall. Pharmacy pickups fail. Providers buried in administrative work have less time for the patient in front of them. The experience of getting care ends up feeling harder than the condition it is supposed to treat. When patients with depression or anxiety abandon treatment mid-course, there is no way to know whether the treatment would have worked.

Removing that friction matters, and technology has made real progress in doing so. The harder problem sits underneath. Even when patients get in the door quickly and remain in care, the diagnostic process itself is prone to failure in ways that faster access will never fix.

Patient Demand Is Not the Same as Clinical Need

ADHD illustrates this tension more clearly than almost any other condition. Demand for ADHD care has climbed steadily, and patients arrive with strong beliefs about what is driving their symptoms. At my practice, we see this firsthand. In March 48.8% of patients self-reported seeking care specifically for ADHD, while another 32.4% mentioned ADHD alongside other concerns. Rather than treating the reported demand as a diagnosis, we use a structured clinical workflow — intake data, medical and prescription history, validated measures, and FDA-cleared objective assessment tools to confirm or rule out ADHD before any treatment decisions are made.

That gap is not a system failure. It is what rigorous clinical evaluation is supposed to produce. ADHD symptoms overlap significantly with anxiety, depression, trauma, sleep disorders, and a range of other conditions. Nearly 78% of children with ADHD have at least one co-occurring condition, and adult rates are similarly high, meaning that even a confirmed diagnosis rarely tells the full clinical story. Patient-reported concerns signal that someone needs help, and on their own, they are not a reliable basis for diagnosis or treatment.

The industry has largely responded by moving to one of two extremes. Done Global’s founder and clinical president were convicted in connection with a scheme involving over 40 million stimulant pills distributed for what federal prosecutors alleged were largely illegitimate medical purposes. Cerebral agreed to pay more than $3.6 million to settle DOJ and DEA allegations that it used internal metrics to drive higher stimulant prescription rates regardless of clinical appropriateness. Other providers have swung in the opposite direction, refusing to engage with ADHD at all because the regulatory and reputational risk feels too high. Both responses fail patients, and neither builds the clinical infrastructure required to tell the difference between a patient who has ADHD and a patient who has something else that looks like it.

What Rigorous Clinical Workflows Require

Rigorous ADHD care depends on structured, multi-data-point protocols rather than individual provider judgment applied inconsistently. For every patient presenting with suspected ADHD, that means pulling complete medical history through the health information exchange, retrieving prescription history, conducting a thorough intake, and requiring objective measurement of attention, activity, and impulsivity using validated assessment tools. Before a patient is initiated on a stimulant medication, cardiac screening and labs should confirm medical safety. For treatment-naive patients, a non-stimulant medication is the appropriate starting point.

Each of those steps serves a distinct clinical purpose. Objective assessment gives providers a data point independent of what the patient reports, something grounded in measured performance rather than self-perception. ADHD is frequently confused with anxiety and other conditions, and that data gives providers diagnostic confidence that interview-based evaluation alone cannot supply. American Academy of Pediatrics guidelines recommend using multiple sources of information and standardized measures in ADHD evaluation for exactly this reason. Objective, FDA-cleared assessment tools help clinicians distinguish ADHD from conditions that mimic it.

There is also a secondary effect worth naming plainly. Patients who are seeking medications for reasons other than clinical need tend to disengage from a process this thorough, which is exactly what a well-designed protocol should accomplish.

AI strengthens this workflow when it is integrated meaningfully rather than added on top. A well-built platform synthesizes patient history from every available source before a provider enters the encounter, surfacing issues that might otherwise go unnoticed, like a contraindicated medication based on a condition buried deep in a patient’s records. Post-visit audits that compare what was actually done against what the clinical evidence would predict create a continuous improvement loop for both the clinical model and provider practice over time.

Measuring What Actually Matters

None of these matters if it does not produce better patient outcomes. Standardized measures like the GAD-7 and PHQ-9 provide an accountability layer that is most commonly available in behavioral health today. In our patient population, 67% report a statistically significant improvement on those measures, roughly three times the national average.

But standardized questionnaires only capture what patients can articulate about themselves. Objective measurement reveals what self-report misses. Using FDA-cleared assessment tools, we can now examine symptom severity by age and by gender, and what the data shows is clinically significant. Sixty percent of our patients who tested for ADHD were female; this alone reflects a population shift worth paying attention to. More importantly, female patients presented with higher overall symptom severity scores than male patients, with the largest differentials in impulsivity and reaction time.

This matters because ADHD in women is one of the most persistently underdiagnosed conditions in behavioral health. Symptoms present differently, are more frequently attributed to anxiety or mood disorders, and are often missed entirely by evaluation approaches that rely on patient self-report and provider judgment alone. Objective measurement doesn’t eliminate that gap, but it gives clinicians a data point that is independent of perception — their own and their patient’s.

At a population level, this kind of data creates something the field has rarely had: outcomes evidence disaggregated by gender that health systems and payers can actually use. Retention over time is a meaningful proxy as well, given that patients who are stable and improving tend to stay in care.

The field is moving toward something more continuous. Longitudinal analysis across visit transcripts, patient communications, and in-between-session check-ins will eventually allow providers to track how individuals are progressing in ways that standardized questionnaires alone cannot capture.

What the Industry Has to Get Right

The behavioral health access crisis is real, and the urgency to address it is justified. Organizations that deploy AI primarily to move patients faster through an intake workflow, without the clinical structure to ensure those patients are being assessed accurately, are solving the wrong problem.

The organizations that will get this right understand that clinical rigor and scalable access reinforce each other. AI should be helping providers make better decisions, surfacing what they might otherwise miss, and becoming more precise with every patient encounter. Adult ADHD affects an estimated 6% of the U.S. population, and those patients deserve diagnosis and treatment grounded in evidence.

When healthcare leaders evaluate AI-enabled care models, the most important question is whether those models improve the quality of clinical decision-making at scale. The gap between the patients who present with ADHD concerns and the patients who actually have ADHD is not a number to be optimized away. Finding that gap is what rigorous clinical care is designed to do.