Why Value Models Miss It

Behavioral health does not fall out of value-based models by accident. It falls out through four specific mechanisms, each of which can be identified in a contract and each of which can be fixed.

The carve-out

Many payers contract behavioral health separately, to a specialized managed behavioral health organization, and exclude that spending from the accountable organization’s total cost of care. Module 3 examines the arrangement in detail.

The consequence is a clean misalignment. An ACO that invests in integrated behavioral health bears the cost of the care managers and the psychiatric consultant, and the savings appear as reduced admissions and emergency visits on the medical side, which it does keep. But if behavioral spending is carved out, the ACO cannot see whether its investment increased behavioral utilization, and the behavioral vendor has no stake in the medical savings. Neither party sees the whole trade.

Attribution

Attribution generally runs through primary care visits. A patient whose only regular clinician is a psychiatrist or a community mental health center may attribute to nobody, or attribute to a primary care practice they rarely see.

This produces a particularly unhelpful outcome: the patients whose care is most fragmented, and who therefore have the most to gain from coordination, are the ones an ACO is least likely to be responsible for.

Risk adjustment

Risk models built primarily on diagnosis codes will underweight behavioral health when those diagnoses are not reliably coded. Behavioral diagnoses are undercoded for real reasons, including stigma, provider reluctance to label, and the fact that treatment delivered outside the medical record never generates a claim the model can see.

The consequence is that an organization taking on a population with significant behavioral health burden may be paid as though it had not. The introductory course’s risk adjustment module explained why that produces adverse selection; here it produces the specific outcome that organizations serving people with serious mental illness are systematically underfunded.

Measure sets

Quality measures determine what gets managed. Behavioral health measures exist and are typically a small minority of any measure set, weighted lightly, and often process-based, such as whether follow-up occurred after a hospitalization for mental illness rather than whether the patient improved.

Worth remembering: each of these four is a specific, checkable feature of an arrangement rather than a cultural failing. Before concluding that an organization neglects behavioral health, read its contract and ask four questions. Is behavioral spending in the total cost of care? How are patients without a primary care relationship attributed? Does the risk model reflect behavioral burden? What share of the quality score is behavioral, and does any of it measure improvement? An organization answering badly on all four is not indifferent, it is responding rationally to an arrangement that made behavioral health someone else’s problem. Fixing the arrangement is more productive than exhorting the organization.

Key takeaways

• Carve-outs separate the party that invests in behavioral care from the party whose medical spending falls.
• Attribution through primary care visits misses patients whose main clinician is behavioral.
• Undercoded behavioral diagnoses cause risk models to underpay for the populations that need most.
• Behavioral measures are usually few, lightly weighted, and process-based.

Sources

• KFF, 5 Key Facts About Medicaid Coverage for Adults with Mental Illness (opens in a new tab)
• Carlo et al., Association of Alternative Payment and Delivery Models With Outcomes for Mental Health and Substance Use Disorders: A Systematic Review, JAMA Network Open 2020 (opens in a new tab)