Behavioral health demand has surged but our coding system hasn’t caught up

What happens when a healthcare system built for lab values, imaging, and vital signs is tasked with adjudicating care that is far more narrative, contextual, and interpretive? Health plans and providers do not need to think too hard about it. They’re living with the friction it creates every day.

The spike in behavioral health claims was not a temporary COVID-era anomaly. This is the new normal, a fundamental reset in how people access the care they need. According to a report for LexisNexis Risk Solutions, behavioral health claims increased 83% between 2019 and 2023. That trend has been sustained in recent years, and as a result, there was a 45% rise in utilization of behavioral health services claims between January 2023 and December 2024. This makes behavioral health one of the fastest-growing cost categories in healthcare. Our healthcare system is ill-equipped for that shift.

The good news is that incredible progress has been made in this arena of care over the past few years. Telehealth reduced access barriers, stigma surrounding behavioral health plummeted, and patients have begun seeking care at record high volumes. Those are important gains, and they must be protected.

The problem is that our claims infrastructure was not built to interpret this kind of care correctly. That gap — between the care being delivered and the way it’s adjudicated — is wide, and it’s showing up in denied claims, rework, appeals, unrecoverable payments and mounting administrative burden for patients, plans and providers alike.

Why interpretation matters

Behavioral health must be treated with parity under federal and state law, which means health plans cannot apply more stringent review standards to behavioral health claims than they do for medical and surgical claims. That parity obligation extends to the review process itself. At the same time, HIPAA’s “minimum necessary” standard requires that the scope of any claims review be limited to the clinical documentation needed to validate coding, and psychotherapy notes carry separate, heightened protections that further constrain what reviewers may access.

Therein lies the rub. Intensity and necessity in med/surg are supported by objective markers like lab values, imaging, and vital signs, making the coding process far more standardized. Behavioral health documentation, by contrast, is largely narrative. It hinges on aspects like time, treatment purpose, modality, and clinical nuance that don’t translate well within a med/surg framework.

The coding philosophy is the same, but the evidence base is different: in behavioral health, the justification for reimbursement lives almost entirely in the quality and context of the narrative note. This requires reviewers to have technical coding knowledge as well as clinical judgment and experience.

Here’s an example: A note stating “patient is distressed” in a med/surg chart may correlate with measurable physiological indicators. But in a behavioral health setting, that same phrase could reflect anxiety, agitation, crying, or something else entirely. It’s easy for wires to cross, and when they do, it creates downstream impacts on adjudication.

In services like intensive outpatient programs, residential treatment, and telehealth therapy — where codes depend on duration, frequency, and cumulative weekly requirements — determining whether a behavioral health code is supported relies heavily on understanding the full narrative context of the patient.

The problem is equal parts technology and standardization. In med/surg, providers are equipped with decision-support tools, established billing workflows, and more consistent documentation structures. But behavioral health clinicians and billers have far less guidance on how narrative clinical notes translate into specific billing codes, creating a lot of interpretive burden that, at scale, can break the system.

That interpretive burden is further compounded by out-of-network utilization. Patients are 3.5 times more likely to go out of network for behavioral health clinician visits than for medical/surgical visits. For psychologists specifically, 18.2% of visits were out-of-network versus 1.7% for medical specialists. These out-of-network claims introduce additional variability and fewer documentation controls.

The path toward a better claim infrastructure for behavioral health begins with treating interpretation as a clinical discipline. Only then can we build the interpretive infrastructure required to meet the current level of behavioral health utilization.

What responsible behavioral health payment looks like

When done correctly, claims infrastructure should reduce waste and costs while improving accuracy for patients, providers, and payers alike. That’s not necessarily unique to behavioral health, but considering the preparedness of the healthcare system, it is doubly true.to interpret context and establish validation standards. Reviews should focus on documentation and coding alignment, confirming that the record supports the code billed. Done well, this process strengthens access to care by reducing denials, rework, and administrative friction for providers and patients.

In the ideal state, patients experience lower long-term costs and more sustainable coverage. Providers receive clear rationale and feedback so documentation improves over time. Payers see more accurate reimbursement the first time, fewer avoidable appeals and rework cycles, improved compliance, and more predictable behavioral health spend.

The healthcare industry succeeded in making behavioral healthcare more accessible. Now, it must build the infrastructure to support that care responsibly.

Lisa Pincher, MSN, RN, PHN, is senior vice president of complex payment solutions at Machinify.