Policy4 min read
Value-based care in behavioral health: what is actually paying for outcomes in 2026
For years, behavioral health was told value-based payment was coming. In 2026 a few programs actually tie money to PHQ-9 and GAD-7 results. This post covers which ones, how they pay, and the data an agency needs before it signs up.
Margent clinical team · Licensed clinicians and quality specialists
Model terms are from CMS and state documents as of publication. Some details come from secondary reports; those are flagged.
The Center for Health Care Strategies has noted that value-based payment has come to Medicaid behavioral health more slowly than to physical health. One common explanation is that outcomes are hard to measure. Validated instruments exist, so that isn't really the reason. The harder problem is that most agencies don't record scores consistently enough for anyone to pay on them. In 2026 several programs started paying on those scores, so that gap now costs money.
What is paying on outcomes now
Medicare: the ACCESS model's behavioral health track
CMS's ACCESS (Advancing Chronic Care with Effective, Scalable Solutions) model began on July 5, 2026 and runs for ten years, through June 30, 2036. It is voluntary and open to Original Medicare patients only. The behavioral health track covers depression and anxiety, with outcomes measured by the PHQ-9 and GAD-7. Participants receive outcome-aligned payments per patient for 12-month care periods. CMS bases payment on the share of patients who meet their outcome targets, compared with a minimum threshold that rises each year.
CMS's payment document lists $180 for the behavioral health track's initial period and $90 for the follow-on period, for July 5, 2026 through December 31, 2027. CMS reported 160 participating organizations at launch. Tracks for substance use disorder, heart failure, COPD and tobacco cessation begin April 1, 2027.
Medicaid: the Innovation in Behavioral Health (IBH) model
The CMS Innovation Center announced IBH in 2024, and the first states began their pilots in January 2025. Behavioral Health Business reported in January 2026 that New York, Michigan and South Carolina are among them. States get seven-year cooperative agreements. CMS describes payment moving over time from infrastructure and capacity-building funds toward per-member-per-month payments and outcomes-based components. A second cohort of up to five states could apply until June 3, 2026, with up to $7.5 million each. Selections are expected later in 2026.
State programs and CCBHCs
California's BH-CONNECT incentive program pays behavioral health plans for improvement on state-selected measures. Its technical manual lists 53 measures in 15 areas, with some moving to pay-for-performance between 2026 and 2029. CMS's October 2026 guidance on the new CCBHC state plan benefit explicitly lets states "incorporate considerations for quality" into CCBHC payment (see our CCBHC funding post).
Why measurement-based care is the foundation
Every model above pays on data from one practice: giving a validated instrument at intake and again at regular intervals, then acting on the result. That practice is measurement-based care (MBC). The Joint Commission's standard CTS.03.01.09 already requires accredited behavioral health organizations to use a standardized tool or instrument to monitor progress.
MBC is still uncommon. A 2019 review in JAMA Psychiatry by Lewis and colleagues described MBC as an evidence-based practice that is rarely used in routine care. The surveys it draws on found fewer than one in five practitioners using it routinely. The authors listed barriers at the patient, clinician, organization and system levels. One of the organizational barriers is that scores are collected but never reach the clinician at the point of care.
The data you need before you sign up
| Data element | Why the payer needs it | Where agencies usually fall short |
|---|---|---|
| Baseline score with a date | Defines the starting point for improvement or control | Given at intake but stored as a scanned form or in free text |
| Follow-up scores at set intervals | Shows change during the care period | Given when someone remembers, not on a schedule |
| Attribution: who counts as your patient | Sets the denominator | Program rosters don't match the payer's attribution list |
| Reason care ended | Separates completed treatment from dropout | Discharges recorded without a reason |
| Follow-up after ED visit or hospitalization | Utilization and Core Set measures | Admission and discharge dates reach the chart late or never |
| Engagement: visits in the period | Context for outcomes and a common gate for payment | Missed visits and outreach attempts aren't recorded |
Missing follow-up scores are the biggest risk. A patient without a follow-up score can't count as improved. Check how each model handles missing scores, and plan for most of them to count against you. The clinics that do well usually didn't have better clinical results. They had fewer patients with no second score.
- Measures
- PHQ-9 11 today (baseline 19 on 08/04; 15 at week 4). GAD-7 8 (baseline 14). Item 9 = 0.
- Interpretation
- 8-point PHQ-9 reduction from baseline; GAD-7 below 10. Client describes sleeping through the night 4 of 7 nights.
- Plan
- Continue CBT weekly; repeat PHQ-9/GAD-7 at week 12 to confirm trajectory before stepping down to biweekly.
Passes: Each score has a date and a baseline, and the next measurement is scheduled. The same data supports the note, the plan review and the outcome payment.
Getting started
- Pick a few instruments and use them everywhere. PHQ-9 and GAD-7 for adults cover what ACCESS pays on. Add a functional measure if your state or plan asks for one.
- Set a schedule. For example: at intake, every 30 days, and at discharge. Make it a system prompt, not something clinicians have to remember.
- Store scores as structured data. A score inside a paragraph can't be used for a measure.
- Show the trend to the clinician during the session. That is what makes MBC care rather than just data collection.
- Track your missing-score rate each month for every program. That is your early warning.
- Reconcile attribution with each value-based contract every quarter.
How Margent helps
Outcomes sends validated instruments between sessions on the schedule you set, scores them, and shows the trend in the note, the plan review and the program dashboard. Scores are stored as structured data, so outcome rates come from the clinical work. Follow-ups reach out to clients whose next score is overdue, with staff approval, before the gap affects your payments.
Sources
- 01CMS Innovation Center, ACCESS (Advancing Chronic Care with Effective, Scalable Solutions) Model (opens in a new tab)
- 02CMS, ACCESS Model Payment Amounts and Performance Targets (July 5, 2026 – Dec. 31, 2027) (opens in a new tab)
- 03CMS, ACCESS Technical Frequently Asked Questions (opens in a new tab)
- 04Behavioral Health Business, "Inside CMS pilot to expand VBC in mental health, integrated services" (Jan. 21, 2026) (opens in a new tab)
- 05CMS, Innovation in Behavioral Health (IBH) Model — Cohort II NOFO webinar slides (2026) (opens in a new tab)
- 06California DHCS, BH-CONNECT Incentive Program Technical Specifications Manual (opens in a new tab)
- 07CMS, State Medicaid Director Letter #26-004 (Oct. 6, 2026) (opens in a new tab)
- 08Lewis CC et al., "Implementing Measurement-Based Care in Behavioral Health: A Review," JAMA Psychiatry 2019;76(3):324–335 (opens in a new tab)
- 09Center for Health Care Strategies, Moving Toward Value-Based Payment for Medicaid Behavioral Health Services (opens in a new tab)
- 10The Joint Commission, CTS.03.01.09 (measurement-based care standard, behavioral health care and human services)