For more than a decade, behavioral health had a clear problem to solve: access. Employers and health plans responded by expanding EAPs, adding virtual therapy, and creating more ways for people to enter care.

However, access also became the organizing principle of the market. Soon, success came to mean more people entering care, more sessions, and more utilization.

The problem facing buyers today is different. Behavioral health spending is rising and clinician capacity is finite, yet very different levels of need are still routed toward the same familiar answer: the one-to-one therapy session. And too often, contracts reward activity, not whether people get better.

Five assumptions that still shape the market — and why they no longer hold:

1. Access is still the problem to solve.

Access can still be a bottleneck, but part of the constraint is how clinician time gets allocated. People with less intensive needs are often routed toward the same scarce resource as people with serious clinical needs.

Imagine three very different employees: one dealing with temporary work stress, another managing chronic illness and spiraling depression, and a young parent hiding a serious substance use disorder. Why should all three receive the same intervention simply because they entered through the same front door?

A benefit can expand access and still misallocate care.

2. Therapy is the right default.

Much of behavioral health still follows a familiar path: enter the system, connect with a therapist, receive a set number of sessions, and refer elsewhere when the need exceeds what that model can handle. This approach creates mismatches in both directions.

People with less intensive needs can consume scarce clinician capacity unnecessarily, and people with serious or specialty needs can enter a general therapy pathway only to be referred again.

Across other areas of medicine, assessment determines treatment intensity—behavioral health should work the same way.

3. More engagement means more value.

Behavioral health has become very good at counting activity in the form of registrations, sessions, utilization, and patient satisfaction. While those metrics are important, they only tell buyers about usage and experience, but they don’t establish whether anyone improved.

A program can meet its engagement targets while symptoms, function, downstream healthcare use, and total cost remain unchanged. Reporting can also look favorable if it focuses on people who complete care while failing to account for those who disengaged or never connected in the first place.

4. More clinicians will solve the mismatch.

When demand rises, adding supply feels like the obvious answer. But a system that uses licensed-clinician time as the default unit of care sends every level of need toward the same scarce resource. The real fix is allocation.

Lower-intensity needs increasingly have credible digital and AI-enabled options that don’t require a clinician hour at every touchpoint. Rising needs should trigger more intensive support, and specialist capacity should be preserved for people whose clinical needs require it.

The evidence for that model predates today’s AI. In one randomized stepped-care trial, matching treatment intensity to need reduced total costs by 22% and hospital costs by 50%.

5. Payment is just a contracting detail.

How care is paid for may be the single most important factor shaping how care gets delivered.

Under a fixed per-member model, higher utilization can make the benefit more expensive for the vendor to operate. Under session-based reimbursement, the incentive moves in the opposite direction: revenue rises with every additional session.

That contradiction is clearest in session-based care. A provider that resolves a need faster, avoids an unnecessary visit, or routes someone to a less intensive intervention can actually earn less for taking the clinically appropriate action.

It’s the economic reality of a system in which incentives are misaligned. More measurement, access, and clinical capacity won’t solve the problem if the economics still reward the same behavior.

What the next standard requires

The first era of behavioral health benefits focused on getting more people through the door; the next must focus on what happens after they enter.

In a 2025 assessment of virtual solutions for depression and anxiety, the Peterson Health Technology Institute called for better triage and stepped-care models that match patients to the most clinically appropriate and cost-effective level of support.

Those recommendations point to a new standard: match care to need, measure whether people get better, and align payment with results. Technology can help make that model possible at scale. Realizing that potential requires changing both how care is delivered and what the system rewards.

‍

___

Yusuf Sherwani, MD, is CEO of Pelago Health.