BALANCE Model Medicaid Participating States: 2026 Public-Record Tracker

Public-record snapshot by evidence stage

Source: The RX Index public-record compilation from CMS and official state-government records. Statuses are evidence stages created for this dataset, not labels assigned by CMS. See methodology.

Last verified: July 31, 2026. Scope: 50 states and the District of Columbia. Application deadline: July 31, 2026, 11:59 p.m. Eastern Daylight Time.

Jurisdictions with affirmative public evidence

No executed State Agreement was publicly verified in this snapshot. Indiana had the strongest affirmative record—a governor-directed commitment to participate—but its announcement said implementation details would follow as the agreement was finalized. Missouri and Rhode Island had disclosed intent notices. Maryland and Virginia had documented formal state-directed assessment or evaluation.

Source: The RX Index Research public-record compilation from CMS and the official state-government records linked in each row. Statuses are evidence stages created for this dataset, not labels assigned by CMS.

Why the answer is not simply “Indiana”

Indiana’s announcement is consequential, but a public commitment is not the same document as an executed CMS State Agreement. CMS’s State Medicaid request for applications says an accepted applicant becomes a model participant when it signs the State Agreement. The RFA also permits a state to withdraw before signing. The tracker therefore keeps commitment, application, agreement, implementation, and live coverage as separate milestones.

An intent notice is not a signed agreement

Missouri’s draft minutes and Rhode Island’s legislative response are useful because they establish real state action. They do not establish CMS acceptance, a signed State Agreement, or live coverage. That is why both remain at Stage 2.

An application is not a signed agreement

CMS accepts applications, then an accepted applicant is eligible to execute a State Agreement. The RFA permits withdrawal before execution. A public application confirmation would therefore move a state to Stage 3, not Stage 4.

A signed agreement is not automatically live coverage

Implementation can still require manufacturer supplemental rebate agreements, state legal authority or a state plan amendment where applicable, formulary and prior-authorization changes, and separate fee-for-service and managed-care dates. The tracker will not label coverage operational until an official source supplies an effective date and the population or delivery system covered.

What counts as participation in the CMS BALANCE Model?

For this tracker, formal participation is publicly verified only at the executed-State-Agreement stage. CMS materials establish that threshold. The earlier stages below are The RX Index’s editorial evidence classifications, built to prevent a notice, application, or political announcement from being mislabeled as a signed agreement.

Source: The participation threshold and model process come from the CMS State Medicaid RFA. Stage names and evidence rules are The RX Index methodology.

BALANCE Model status in every state and D.C.

The table below records the strongest qualifying affirmative action found for every jurisdiction. Stage 0 means no qualifying public action was located under the stated search method. It does not mean the state definitely did not submit a private application.

Source: The RX Index Research public-record compilation. Affirmative rows link to the strongest official record located. Stage 0 rows reflect the dated negative-search method described under Methodology, not a definitive claim about confidential state activity.

The permanent CSV URL serves the latest version. Dated snapshot files remain unchanged so later status changes do not erase what was verifiable on July 31, 2026.

When can BALANCE Medicaid coverage start?

CMS allows a participating state to select a model start date from May 1, 2026, through January 1, 2027. Fee-for-service and managed-care populations may start on different dates, so an agreement date alone cannot establish that coverage is live for every Medicaid member.

Source: CMS State Medicaid RFA and The RX Index publication rules.

Which GLP-1 drugs are included in BALANCE?

CMS currently lists Mounjaro, Ozempic, Rybelsus, Wegovy, the KwikPen formulation of Zepbound, and Foundayo in the model. Product eligibility, formulations, indications, and negotiated terms can change, so the list is dated to the page’s last verification rather than presented as permanent.

Source: CMS BALANCE Model page, verified July 31, 2026. A listed drug is not an individual coverage guarantee. State participation, implementation, model eligibility criteria, prior authorization, indication, and a member’s circumstances still matter.

How is BALANCE different from existing Medicaid GLP-1 coverage?

A state can already cover a GLP-1 for obesity without participating in BALANCE, and it can participate in BALANCE only after completing the model’s separate federal-state process. Existing formulary coverage cannot be used as proof that a state applied, signed a State Agreement, or launched BALANCE.

The broader state coverage question belongs in The RX Index’s separate GLP-1 Medicaid coverage by state tracker. This page answers the narrower question: what public evidence shows about participation in the CMS BALANCE Model specifically.

How is Medicaid BALANCE different from the Medicare GLP-1 Bridge?

Medicaid BALANCE is an opt-in model for state Medicaid agencies. The Medicare GLP-1 Bridge is a separate demonstration for eligible Medicare Part D beneficiaries. A state’s BALANCE status does not determine whether an eligible Medicare beneficiary can use the Bridge.

Source: CMS BALANCE Model page. See also: Medicare GLP-1 Bridge Tracker on this site.

What does this BALANCE dataset show and what does it not show?

It shows the strongest BALANCE participation evidence that could be verified in public federal and state-government records on the verification date. It does not disclose confidential applications, guarantee that every new document had already been indexed, estimate an individual’s eligibility, or prove that Stage 0 jurisdictions took no private action.

The practical limits are straightforward:

• A signed document can exist before it appears on a public website.
• Fee-for-service and managed-care implementation can begin on different dates.
• A state may sign an agreement before it publishes beneficiary criteria.
• A model drug’s inclusion does not guarantee that a particular person qualifies.
• State status can change quickly near the application and agreement deadlines.

When a source establishes only intent, the row says intent. When it establishes only evaluation, the row says evaluation. The tracker does not fill the gaps with inference.

Methodology

How did The RX Index verify the 51-jurisdiction BALANCE dataset?

The classification starts with CMS’s own participation rule, then records only what an official public record can prove. Every state and the District of Columbia was checked for state-directed assessment, intent, application confirmation, public commitment, agreement execution, implementation instruments, and operational coverage. Secondary reporting can help locate a document, but it cannot by itself move a state above Stage 0.

Sources reviewed

For each jurisdiction, the review covered the state Medicaid agency, governor or executive branch, legislature, enacted budget materials, pharmacy and therapeutics records, drug-utilization-review records, provider notices, state plan materials, and official testimony. Federal checks covered the CMS BALANCE model page, State Medicaid RFA, model updates, participant announcements, and publicly available state plan records.

Search terms

The state scan used the exact model name and common variants, including:

• “BALANCE Model”
• “CMMI BALANCE”
• “Better Approaches to Lifestyle and Nutrition”
• “BALANCE” Medicaid GLP-1
• “BALANCE” State Agreement
• “BALANCE” Supplemental Rebate Agreement
• The state name paired with: application, agreement, opt in, participation, implementation, and start date

Inclusion rule

A jurisdiction moved above Stage 0 only when a federal or state-government source documented an affirmative state action. Routine discussion in a committee, a general model briefing, manufacturer advocacy, an outside organization’s recommendation, or ordinary Medicaid GLP-1 coverage did not count.

Stage 1 required a state-directed participation review, assessment, conditional implementation instruction, or formal recommendation. This prevents a passing mention of BALANCE from inflating the affirmative-state count.

Stage 4 required either the executed State Agreement itself or unequivocal confirmation from CMS or the state that the agreement had been signed.

Negative-search rule

“No qualifying affirmative public action located” is deliberately narrow. BALANCE applications are submitted through a nonpublic CMS Qualtrics portal, and a state document may be published or indexed after the action occurred. A Stage 0 row therefore reports the result of the dated public-record scan. It is not evidence that a private application does not exist.

Reproducibility and version control

Each dataset row carries a jurisdiction, evidence stage, plain-language status, strongest evidence, source URL, evidence date, unresolved milestone, and last-verification date. The permanent CSV URL serves the latest version. Dated snapshot files remain unchanged so later status changes do not erase what was publicly verifiable on July 31, 2026.

What changed in the July 31, 2026 update?

Indiana moved to Stage 3 after the governor publicly directed the state’s Medicaid agency to participate. It was not moved to Stage 4 because the announcement said details would follow as the agreement was finalized, and no executed State Agreement was located in the public scan.

The same audit also made three evidence-quality corrections: Missouri’s official draft minutes support Stage 2, with the draft status stated plainly; Maryland’s legislative record supports Stage 1 as a formal assessment; and Virginia’s enacted budget supports Stage 1 as an evaluation and conditional implementation direction, not a completed opt-in.

Corrections are logged by date, original entry, corrected entry, reason, source, and whether the headline count changed. Prior snapshot files remain unchanged.

Frequently asked questions

How to cite this page

Primary sources

1. [1] Centers for Medicare & Medicaid Services. BALANCE Model. Verified July 31, 2026.
2. [2] Centers for Medicare & Medicaid Services. BALANCE Model State Medicaid Request for Applications. Verified July 31, 2026.
3. [3] State of Indiana. Governor Braun announces state Medicaid program will now cover GLP-1 treatments for eligible beneficiaries. July 30, 2026.
4. [4] Missouri Department of Social Services. MO HealthNet Drug Prior Authorization Committee draft minutes. January 20, 2026.
5. [5] Rhode Island Executive Office of Health and Human Services. Medical Assistance responses submitted to the General Assembly. April 27, 2026.
6. [6] Maryland Department of Legislative Services. Fiscal and Policy Note for Senate Bill 496. 2026 legislative session.
7. [7] Commonwealth of Virginia. 2026–2028 budget, Item 291. 2026.

Related research

• GLP-1 Medicaid Coverage by State — Broader tracker of which states cover GLP-1s under Medicaid, independent of BALANCE
• Medicare GLP-1 Bridge Tracker — Separate Medicare Part D demonstration program data
• GLP-1 Statistics 2026 — Usage, cost, and coverage statistics from institutional sources
• GLP-1 Employer Coverage Tracker — 2026 employer plan data
• GLP-1 Price & Access Tracker — Monthly self-pay cost tracking
• The RX Index Research Hub — All datasets and trackers