How to Prepare for a Value-Based Behavioral Health Contract
A practical preparation guide for DSO operators and medical directors — what you need to have, what to build, what questions to ask, and what mistakes to avoid.
Who This Is For
This guide is for the people actually doing the work: the Chief Operating Officer managing multiple behavioral health locations, the Medical Director who will be accountable for the quality metrics, and the VP of Revenue who needs to understand the financial structure before anyone signs anything.
It is not a theoretical overview of value-based care. It is a practical preparation guide — what you need to have, what you need to build, what questions to ask, and what mistakes to avoid before you enter a value-based contract negotiation.
Value-based contracting in behavioral health is no longer early-adopter territory. CMS launched the Innovation in Behavioral Health (IBH) Model in January 2025, awarding cooperative agreements to Michigan, New York, and South Carolina. These Cohort I states are now in a three-year pre-implementation period (2025–2027), with the implementation period beginning in 2028. A second cohort of up to five additional states is open for application through June 2026, with awards expected fall 2026. Commercial payers are expanding outcome-based arrangements. Medicaid managed care organizations are building VBC requirements into network contracts. The timeline for “when do we need to be ready for this” has collapsed.
Groups that prepare now will negotiate from strength. Groups that wait will accept whatever terms payers offer.
Part 1 — Understanding What You’re Agreeing To
Before preparing for a VBC, you need to understand the range of arrangements that fall under that label. They are not equivalent. The risks, the data requirements, and the upside are different for each.
The Four Models You Will Encounter
Pay-for-Performance (P4P)
The most common entry point. Your base reimbursement stays fee-for-service, but you earn bonus payments if you hit specified quality thresholds — depression remission rates, HEDIS measure compliance, follow-up after hospitalization rates. You can lose the bonus but you cannot owe money back. Risk is low. Upside is modest. This is where most behavioral health groups start.
What it requires: Systematic outcome measurement. You cannot report what you haven’t tracked.
Shared Savings (Upside-Only)
You are measured against a baseline cost-per-member-per-month (PMPM) or total cost of care benchmark. If your attributed population generates lower total costs than the benchmark — fewer ER visits, fewer hospitalizations, better medication adherence — you receive a share of those savings, typically 30–50%. If costs exceed the benchmark, you keep your FFS rates and lose nothing. Risk is still low. Upside is meaningful.
What it requires: Population health capability — the ability to identify high-risk patients before they deteriorate, not after. This is operationally harder than P4P.
Shared Savings with Downside Risk (Two-Sided)
Same structure as shared savings, but if your attributed population’s costs exceed the benchmark, you owe a portion of the excess back to the payer. This is real financial risk. Some practices lose money under two-sided models, particularly in their first year before they’ve built the operational muscle to manage it.
What it requires: Everything shared savings requires, plus actuarial analysis of your population before you agree to a benchmark, and capital reserves to absorb a bad performance year. Do not enter a two-sided arrangement without both.
Capitation (Full or Partial)
You receive a fixed PMPM payment for a defined population and are responsible for delivering all covered behavioral health services within that budget. If your costs are below the capitated rate, you keep the difference. If above, you absorb the loss. This is the most mature form of VBC and requires the most infrastructure.
What it requires: Claims data on your attributed population, actuarial support, utilization management capability, and a strong network if you’re taking partial delegation of physical health costs. Not appropriate for most behavioral health groups in their first VBC. Start with P4P or shared savings.
The Rule Nobody Tells You
Payers will offer you what they think you can handle. A payer who sees you have no outcomes data will offer you P4P with easy benchmarks and minimal upside — it’s a safe bet for them and a low bar for you. A payer who sees you have 5 years of longitudinal PHQ-9 data, a structured population health program, and clean claims will offer you a shared savings arrangement with real upside because they believe you can perform. Your preparation determines your negotiating position, which determines your contract terms.
Part 2 — What Payers Actually Want
Understanding the payer’s objective is the single most important mental model shift in value-based contracting. Payers are not charities trying to reward good care. They are businesses trying to reduce medical loss ratio (MLR) — the percentage of premium revenue paid out as claims. Every VBC conversation is, at its core, a conversation about whether your clinical program will reduce their MLR.
That said, payers care about five specific things:
1. PMPM cost reduction. Will your program reduce emergency department visits, psychiatric hospitalizations, and high-cost crisis episodes for your attributed population? Averted psychiatric crises represent significant claims savings for payers. Your clinical outcomes need to map to a financial story. “Our PHQ-9 remission rate is 32%” is a clinical outcome. “Patients who achieve remission have significantly lower downstream utilization” is a payer argument.
2. HEDIS and STAR quality measure performance. For commercial plans, HEDIS. For Medicare Advantage, STAR ratings. These are the public scorecards payers are evaluated on. When you improve their HEDIS scores — particularly DRR-E (Depression Remission or Response), FUH/FUM (Follow-Up after Hospitalization for Mental Illness), and AMM (Antidepressant Medication Management) — you are solving a problem they report to regulators annually.
3. Member persistency and satisfaction. Payers want their members to stay with the plan and report high satisfaction. Members who get effective behavioral health treatment are less likely to churn off the plan. This is a commercial consideration worth raising explicitly in negotiations.
4. Network adequacy compliance. Regulators require payers to demonstrate adequate access to behavioral health services. Payers will pay enhanced rates for providers who solve their network adequacy problem in underserved geographies, for specific populations (adolescents, geriatric), or for specific modalities (TMS, intensive outpatient, SPRAVATO).
5. ROI they can report internally. Payers have internal teams who approve VBC arrangements by calculating expected ROI. Come to the negotiation with a preliminary financial model showing what a reduction in ED utilization among your attributed population would mean in dollars. It signals you understand their business.
Part 3 — The Readiness Assessment
Before you approach a payer or respond to a payer’s VBC proposal, complete this readiness assessment honestly. The gaps you find here are your preparation roadmap.
3.1 — Outcomes Data Readiness
This is the highest-leverage readiness dimension. Outcome data is the currency of VBC negotiations.
| Question | Ready | Partial | Not Ready |
|---|---|---|---|
| Do you collect PHQ-9 at intake for all depression patients? | |||
| Do you collect PHQ-9 at regular intervals throughout treatment? | |||
| Do you collect a PHQ-9 at discharge or end of episode? | |||
| Can you calculate response rate (≥50% PHQ-9 reduction) for a cohort? | |||
| Can you calculate remission rate (PHQ-9 < 5) for a cohort? | |||
| Do you collect GAD-7 for anxiety patients? | |||
| Do you collect C-SSRS for suicide risk? | |||
| Are scores structured in your EHR (not just in notes)? | |||
| Can you aggregate scores across providers and locations? | |||
| Do you have at least 12 months of longitudinal data? |
Scoring: If more than 3 answers are “Not Ready,” your outcomes data infrastructure needs to be the first investment you make before any VBC conversation. A payer will ask to see your outcomes data in the first meeting. Walking in without it is like applying for a loan without a credit history.
3.2 — Population Health Capability
P4P requires reporting outcomes after the fact. Shared savings and capitation require managing outcomes proactively — identifying who is at risk before they become expensive.
| Question | Ready | Partial | Not Ready |
|---|---|---|---|
| Can you identify your highest-risk patients by PHQ-9 trajectory? | |||
| Do you have a systematic process for outreach to patients who miss appointments? | |||
| Can you identify patients who are not responding to treatment at session 10? | |||
| Do you have a care management or case management function? | |||
| Can you track medication adherence for your med management patients? | |||
| Do you have a process for coordinating with PCPs on shared patients? | |||
| Can you identify patients with 2+ ER visits in the last 6 months? |
If you scored mostly “Not Ready” here: You may still be able to enter a P4P arrangement but should not take on shared savings with downside risk until these capabilities are built.
3.3 — Data Infrastructure
| Requirement | Status |
|---|---|
| Certified EHR with structured data fields | |
| Ability to receive and process payer-provided claims data | |
| Ability to export quality measure data in standard formats (HEDIS) | |
| Connection to a health information exchange (HIE) or health information network | |
| HIPAA-compliant data sharing agreements and BAAs in place | |
| Staff or vendor capable of running quality measure reports |
Note on the CMS IBH Model: If you are in Michigan, New York, or South Carolina (Cohort I) or a state entering Cohort II, IBH-participating practices are required to adopt ONC-Certified HIT. Federal infrastructure funding is available to support this transition.
3.4 — Financial Readiness
| Question | Yes | No |
|---|---|---|
| Do you know your cost per visit / cost per episode of care? | ||
| Do you know your ER utilization rate for your patient population? | ||
| Do you have 2 years of claims history with your major payers? | ||
| Do you have operating reserves sufficient to absorb a 6-month payment delay? | ||
| Do you have actuarial or financial modeling support? | ||
| Do you know your current HEDIS performance for DRR-E? |
If you’re missing the last item: Request your HEDIS performance data from each payer before entering any VBC discussion. This is the baseline you’ll be measured against.
Part 4 — Building Your Outcomes Infrastructure (12-Month Roadmap)
If your readiness assessment reveals gaps — and most honest assessments will — here is a sequenced plan.
Months 1–2: Baseline and Standardize
Step 1: Audit your current outcome measurement rate. Pull the percentage of patients who have both an intake PHQ-9 and at least one follow-up PHQ-9. For most practices, this number is lower than expected.
Step 2: Pick your instruments and commit.
- PHQ-9 for all patients with a depression diagnosis — non-negotiable for any VBC
- GAD-7 for anxiety — widely required in commercial VBC
- PCL-5 for PTSD — increasingly specified in specialty contracts
- C-SSRS for suicide risk — required for IBH Model participation
You do not need all four on day one. Start with PHQ-9. Add GAD-7. Add the others as you build measurement cadence.
Step 3: Define your measurement cadence and make it operational. Common protocols:
- Session 1 (intake): baseline
- Session 6: mid-point check
- Session 12 or final session: outcome
- Every session for high-risk patients
The cadence you choose matters less than that it is followed consistently. Inconsistent measurement produces data that payers will not accept.
Step 4: Integrate collection into the clinical workflow. The single most common reason outcome data is missing is that collection requires an extra step no one has time for. Embed it: patient completes PHQ-9 on a tablet or patient portal before the session, score is auto-calculated and appears in the chart, clinician reviews it at session start. If your EHR doesn’t support this natively, platforms like OutcomesAI connect to your existing system and handle the collection layer.
Months 3–4: Calculate Your Baseline
Once you have 60+ days of consistent measurement, calculate:
- Your current response rate: What percentage of patients who started with PHQ-9 > 9 and completed at least 8 sessions have shown ≥50% PHQ-9 reduction?
- Your current remission rate: Of completed episodes, what percentage ended with PHQ-9 < 5?
- Your measurement compliance rate: What percentage of patients have a baseline and at least one follow-up score?
- Your no-show and dropout rate: What percentage of patients attend fewer than 4 sessions?
These four numbers are your negotiating baseline. You cannot negotiate without them.
Months 5–8: Identify and Address Variation
Run your response and remission rates by:
- Provider — You will find meaningful variation. Providers with the lowest response rates are not necessarily the least skilled; they may have the most complex patient panels. But you need to understand the source.
- Location — Multi-site groups often see significant location-level variation.
- Payer — Some payers send sicker patients. Understanding your outcomes by payer helps you set appropriate benchmarks.
- Diagnosis — PHQ-9 remission rates for first-episode depression look different from treatment-resistant cases.
Months 9–12: Build Your Contract-Ready Data Package
By month 9, you should be assembling:
- Outcomes summary: Response rate, remission rate, measurement compliance rate, average PHQ-9 change score — for the trailing 12 months
- Population profile: Number of attributed patients by diagnosis, severity at intake, payer mix, and demographics
- Utilization data: ER visits, psychiatric hospitalizations, crisis episodes — if available
- HEDIS performance: Your DRR-E rates by payer, if you can get them
- Methodology note: How you defined response and remission, what instruments you used, how scores were collected
This package is what makes you different from the 95% of behavioral health groups that walk into a payer conversation with nothing.
Part 5 — The Payer Conversation
Who to Approach First
- Medicaid MCOs in IBH Model states — Michigan, New York, South Carolina now; second cohort states by 2027. These MCOs are required to recruit practices and have infrastructure support from CMS.
- Regional commercial plans — More likely to have local decision-making authority. Blue Cross Blue Shield regional plans have been active in VBC behavioral health.
- Medicare Advantage plans — Scored on STAR ratings that include behavioral health measures. Strong financial incentives.
- Large national commercial plans — VBC programs exist but behavioral health VBC contracting is less mature. Expect slower processes.
- Self-insured employers via TPA — Direct employer contracts for behavioral health are growing.
What to Say in the First Meeting
The first meeting with a payer is not a sales call. It is intelligence-gathering. Your goals:
- Understand what VBC arrangements they have or are developing for behavioral health
- Learn what quality measures matter most to them
- Find out what data they can provide to you
- Understand their timeline and decision process
- Leave them with your outcomes data package
The questions that signal sophistication to a payer:
- “What is our current DRR-E performance in your data?”
- “What is the PMPM cost baseline for our attributed population?”
- “Are you looking at upside-only shared savings or two-sided risk?”
- “What data exchange format do you use for attributed member lists?”
- “Which HEDIS measures are you prioritizing for VBC incentive design?”
Most behavioral health providers they meet cannot ask these questions.
What to Avoid
Avoid vague quality commitments. “We provide high-quality, evidence-based care” is not a contract term. Every quality commitment should be tied to a measurable metric with a defined measurement method and baseline.
Avoid accepting the payer’s benchmark without analysis. Payers set benchmarks based on their data. Their data may not reflect your actual population. Always ask: how was this benchmark calculated, over what time period, and for what patient population?
Avoid two-sided risk in year one. The first year of a VBC is a learning year. Take those lessons in an upside-only arrangement.
Avoid signing without data sharing guarantees. A VBC without payer-provided claims data is a trap. You need attributed member lists, utilization data, and pharmacy data to manage your population.
Part 6 — Negotiating the Contract
The Five Terms That Matter Most
1. Benchmark methodology. How is the baseline calculated? Is it risk-adjusted? Is it rebased annually? The ratchet effect — where your success in year one makes year two’s benchmark harder — is real. Insist on multi-year benchmark stability or an explicit anti-ratchet provision.
2. Attribution methodology. Which patients are attributed to your practice? Patients attributed to you who you haven’t seen in 18 months can still generate claims that count against your benchmark.
3. Data sharing timing and format. Quarterly attributed member lists at minimum. Monthly is better. Missing data should pause the performance measurement period, not count against you.
4. Quality measure definitions. Define completely — the denominator, the numerator, the measurement period, the data source, and how conflicts between your data and the payer’s data are resolved.
5. Performance payment timing. Negotiate for quarterly reconciliation payments where possible.
Walk-Away Conditions
Decline or significantly renegotiate if:
- The payer will not commit to providing attributed population claims data
- The benchmark was set with no access to your historical data
- Quality targets are defined vaguely with payer-controlled measurement
- The contract is two-sided risk in your first year
- Performance payment timing is annual with no interim reconciliation
- There is no dispute resolution process
Part 7 — Common Mistakes
Mistake 1: Negotiating on faith instead of data. Practices that enter negotiations without outcomes data consistently accept worse terms.
Mistake 2: Treating VBC as a billing change. It’s a clinical and operational transformation. DSOs that hand VBC to their billing teams underperform.
Mistake 3: Underestimating attribution complexity. A patient attributed to you may also see other providers whose costs count against your benchmark.
Mistake 4: Ignoring the measurement burden on clinicians. PHQ-9 tracking needs to be normalized as a clinical tool, not experienced as an administrative burden. The Medical Director’s role is clinical culture change.
Mistake 5: Starting with too complex a model. Start with P4P. Build infrastructure. Move to shared savings. Get two years of experience before considering capitation.
Mistake 6: Neglecting the ratchet effect. Your success in year one makes year two harder. Negotiate multi-year benchmark stability from the start.
Part 8 — Regulatory Context
CMS Innovation in Behavioral Health (IBH) Model
Launched January 2025 in Michigan, New York, and South Carolina. Cohort I states are in a three-year pre-implementation period (2025–2027), with the five-year implementation period beginning 2028. A second cohort of up to five states has applications open through June 2026.
IBH is an upside-only model. It pays a monthly Integrated Support Payment (ISP) of $200–220 PBPM. Performance-based payments of up to 5% of ISP are available based on practice-level quality measures, shifting from pay-for-reporting to pay-for-performance starting in model year five.
MHPAEA 2024 Final Rule
The Mental Health Parity and Addiction Equity Act’s 2024 final rule expanded scrutiny of non-quantitative treatment limitations (NQTLs) — prior authorization, step therapy, and other restrictions that may make behavioral health benefits harder to access than medical/surgical benefits. Note: As of May 2025, enforcement of the 2024 final rule has been paused pending litigation. The underlying parity requirements of MHPAEA remain in effect, but the expanded NQTL compliance requirements are not currently being enforced. Monitor developments as this may change.
MIPS Measure #370 — Depression Remission at 12 Months
PHQ-9 remission at 12 months. Getting your MIPS performance data on this measure is another way to understand your baseline before payer negotiations.
Where OutcomesAI Fits
The preparation steps in this guide have one consistent dependency: outcomes data. Specifically, longitudinal PHQ-9 and GAD-7 data, structured, aggregated, and reportable across providers and locations.
OutcomesAI integrates with your existing EHR to automate the collection of PHQ-9, GAD-7, PCL-5, and C-SSRS as part of your existing clinical workflow. Scores are structured, timestamped, and linked to treatment episodes. Response rates, remission rates, and HEDIS-aligned quality measures are calculated automatically.
The groups entering VBC negotiations with the strongest positions are the ones who’ve been measuring outcomes systematically for years. This guide exists to help you become one of them.
Related reading:
- Psychiatry Outcomes Benchmarks — the PHQ-9 response and remission rates to benchmark against
- Every EHR Shows You a No-Show Rate. None Show You What Happened Next. — schedule performance as an operational readiness indicator
- Building a Data-Driven Behavioral Health Practice — the foundational data strategy
Glossary
- AMM — Antidepressant Medication Management (HEDIS measure)
- DRR-E — Depression Remission or Response for Adolescents and Adults (NCQA HEDIS measure)
- FFS — Fee-for-Service
- FUH/FUM — Follow-Up after Hospitalization for Mental Illness (HEDIS measure)
- HEDIS — Healthcare Effectiveness Data and Information Set
- IBH Model — CMS Innovation in Behavioral Health Model (launched January 2025)
- MCO — Managed Care Organization
- MHPAEA — Mental Health Parity and Addiction Equity Act
- MIPS — Merit-based Incentive Payment System
- MLR — Medical Loss Ratio
- NQTL — Non-Quantitative Treatment Limitation
- P4P — Pay-for-Performance
- PMPM — Per Member Per Month
- VBC — Value-Based Care / Value-Based Contract
This guide is intended for informational purposes and does not constitute legal, actuarial, or financial advice. Consult qualified counsel before entering any value-based contract arrangement.