Behavioral Health: An Honest Scorecard

The capstone: we know what works clinically, we know payment exists, and the field has mostly measured spending instead of whether anyone got better.

Behavioral health is the clearest case in this curriculum where the field knows what to do and has mostly not done it, and where the standard measure of success may be measuring the wrong thing.

The scorecard

Question	Answer
Does an effective delivery model exist?	Yes, with randomized evidence and replication
Is there a payment pathway?	Yes, Medicare since 2017, with many payers following
Has adoption been widespread?	No
Under risk, did spending fall?	Yes
Under risk, did outcomes improve?	Mostly unmeasured, in fewer than a third of models
Did organizations avoid expensive patients?	Found in three of eight models where it was examined
Is the data barrier gone?	The legal one largely is, as of February 2026
Does parity require equal access?	The statute does; the strongest rule is currently unenforced

The three findings that matter most

Collaborative care works. 45 percent of intervention patients achieved a 50 percent symptom reduction against 19 percent in usual care, replicated across a Cochrane review for both depression and anxiety. This is a stronger evidence base than most of what this curriculum covers.

It is paid for. Since 2017, through dedicated codes billed by the treating clinician. The barrier is operational, not financial.

Accountability has mostly measured the wrong thing. Across 27 studies, alternative payment models reduced behavioral utilization and spending, and clinical outcomes were measured in only five of seventeen models.

The inversion

The single most important idea in this course is that behavioral health inverts the usual heuristic. Everywhere else, lower utilization with stable quality suggests value. Here, where roughly four in ten adults with mental illness receive no treatment at all, lower utilization is at least as likely to be a symptom as a solution.

A model rewarding reduced behavioral spending in a chronically undertreated population may simply be paying for the status quo. And the same organizations reporting those savings were, where anyone checked, sometimes avoiding the patients who cost the most.

Worth remembering: an arrangement that puts an organization at risk for behavioral health spending without measuring whether patients improved is a budget cap, not a value-based model. That sentence is the practical test to apply to any behavioral health arrangement you encounter. It is not a high bar, and most arrangements to date have not cleared it. The evidence to fix this is available, the instruments are brief and free, the payment exists, and the legal barrier to the data has largely expired. What remains is the decision to measure the thing that matters rather than the thing that is easy to count.

What to do about it

For anyone building or evaluating an arrangement that includes behavioral health:

Put behavioral spending inside total cost of care, and pair it with outcome measures that carry real weight.
Treat declining behavioral utilization as a signal to investigate rather than a result to report.
Test attribution and risk adjustment specifically against adverse selection, since it was found wherever it was sought.
Build capacity through integration rather than referring into a network that cannot absorb the referrals.
Ask the exchange partner for the behavioral data now, because the legal answer that used to end that conversation has expired.

Key takeaways

Collaborative care has randomized, replicated evidence and an established Medicare payment pathway.
Adoption remains limited for operational rather than financial reasons.
Risk arrangements reduced behavioral utilization and spending while rarely measuring outcomes.
Undertreatment is the baseline, so falling utilization is ambiguous and requires investigation.
Risk without outcome measurement is a budget cap, which is the test to apply to any arrangement.

Sources

Check your understanding

What is the most defensible summary of behavioral health in value-based care?