# GLP-1 Medicaid policy review — data dictionary and methodology

Publisher: The RX Index Research  
Release: `2026-09-28-review-1`  
Review as of: **2026-09-28**  
Page: https://therxindex.com/research/glp1-medicaid-coverage-by-state/

## What this release measures

This is a source-linked editorial review of Medicaid adult obesity/weight-management GLP-1 policy and public BALANCE status. It contains one record for each of the 50 states and a separate D.C. record. It is **not a complete current national coverage census**, a clinical eligibility calculator, a formulary of every product, an approval-rate dataset, or a list of every active BALANCE participant.

It separates standard obesity-only coverage from non-obesity indications, under-21 EPSDT review, program-specific exceptions and publicly announced future changes. A preferred-drug listing by itself is not treated as an obesity-coverage finding. Explicit obesity criteria, exclusion language, implementation notices and appropriately scoped regulations carry more evidentiary weight.

Sources include current or currently hosted Medicaid agency documents and official pharmacy contractors. Some hosted documents are older; some mutable links showed conflicting editions. The review date records when the stated finding was inspected. It does not prove that an old rule remained unchanged across all later plan contracts or products. Those limitations remain in the records.

## Reading the files

`state-policy-review.csv` has exactly the seven visible table fields. It is UTF-8 with a byte-order mark for spreadsheet compatibility. Multiple source URLs or date labels appear in one quoted cell, separated by semicolons. It has 51 data rows, including D.C.; do not count D.C. as a state.

`state-policy-review.json` contains release metadata and the same 51 records with structured source records, program scope, separate verification fields, classification labels and limitations. `sources.json` contains a deduplicated registry of the inspected source URLs. A source marked `classification_supported: false` is a locator/directory, not proof of the state's coverage classification.

The downloadable citation files identify this release, not whatever future version may appear at the live page. Source documents remain the authority for the policy; this dataset is the comparative synthesis.

## Seven public fields

| Field | Meaning and rules |
|---|---|
| State | U.S. state name; D.C. is a separate jurisdiction identified in JSON. |
| Current obesity GLP-1 status | Bounded finding as of the review, not a guarantee of individual coverage. Read the product/program limitations. “Not established” is not “Excluded.” |
| BALANCE public status | Strongest inspected public-record stage: announcement, recommendation or authorized evaluation where supported; otherwise not verified. Not verified is not non-participation. |
| Effective/start date | Labeled policy date, event date or document edition where known. A form revision, announcement date or review date is never silently treated as a coverage start. |
| Primary source | Agency or official contractor documents supporting the stated finding. A locator is labeled as such. A blank-evidence record explicitly says no primary policy was verified. |
| Last verified | Separate date for the bounded policy finding and the BALANCE finding. A dated source review does not establish every product, plan or later amendment. |
| Notes | Program scope, product/indication boundaries, scheduled changes, exceptions and unresolved evidence. These notes are part of the data, not optional decoration. |

## JSON fields

`record_id` is a stable state anchor (`state-indiana`, for example). `record_url` points to the canonical page and anchor. `state_code` is the postal abbreviation. `jurisdiction_type` distinguishes state/district. `as_of` is the release review date.

`coverage_classification` is an editorial evidence category, not an insurance eligibility code:

- `covered_pa`: a reviewed obesity authorization pathway exists within the stated scope.
- `restricted_pa`: a reviewed obesity pathway exists with unusually restrictive criteria; read the source rather than infer a uniform clinical threshold.
- `excluded`: the reviewed policy excludes obesity/weight-loss use within the stated scope; separate indications and exceptions are not ruled out.
- `excluded_with_exceptions`: the reviewed source expressly identifies an exception process alongside the standard exclusion.
- `excluded_specific_products`: the reviewed named-product criteria do not support obesity-only use. This is not a completed all-brand/formulation exclusion census.
- `scheduled_exclusion`: a future exclusion is documented but not effective at the review date.
- `implementation_unverified`: a public expansion announcement exists but operative implementation was not established.
- `limited_evidence`: narrower, older, proposal-stage or indication-specific material was inspected but does not resolve a uniform current statewide benefit.
- `unresolved`: the current obesity-only finding was not established from the retrieved primary evidence.

`balance_stage` in this release is `announced`, `recommendation`, `evaluation_authorized`, or `unverified`. Future releases may use `applied`, `agreement_executed`, `operational`, or `declined` only with primary evidence of that exact stage. CMS's RFA distinguishes a state agreement from preliminary intent. State entry into a model and an individual drug approval are separate events.

`coverage_last_verified` and `balance_last_verified` are ISO dates or null. Null means not established, not “never.” Source `checked_on` describes the document inspection; it does not make that document a complete current policy. `source_version` preserves the edition/date visible in the evidence, including a descriptive conflict when appropriate. `locator` identifies the supporting page/section.

## Important limits of this release

Eight states and D.C. remain unresolved: Alaska, Arkansas, Colorado, Hawaii, Idaho, Maine, Oregon, Vermont, and District of Columbia. Other records also have explicitly limited product/program/adoption evidence. The record count is not a coverage count.

Only inspected public BALANCE records are positively classified. This was not an exhaustive verification that every other state had made no announcement. Do not write “only Indiana is participating,” “all other states declined,” or a count of operational participants from these data.

Rhode Island's October 1 change is **scheduled** in this September 28 observation. Do not automatically alter that record when the calendar passes October 1; a new source check belongs in a new release. Indiana's July 30 announcement is not a verified coverage start.

The RX Index July 10, 2026 snapshot and KFF January 2026 benchmark are historical context with different source/review scopes. They are not silently revised or treated as one newly verified trend series. Enrollment reach estimates remain withdrawn until the underlying data and calculation can be reproduced. This dataset contains no patient counts, enrollment reach, net spending, savings, or individual approval probability.

## Versioning and correction policy

Before first publication, the publisher reviews and approves the prepared release. Once published, the versioned CSV/JSON files must remain unchanged. A material correction creates a new release ID and change-log entry describing the affected state, field, old finding, new finding and evidence. The stable article may point to the newest release while continuing to link to older releases.

Policy date, source publication/version, inspection date, page update date and dataset release date are different concepts. Do not refresh them all with a single `new Date()` or replace them with a page-build timestamp.

Free access is intended. Cite The RX Index, release ID, state record and original source. No DOI has been assigned. No blanket license is asserted over third-party source documents, and no unapproved Creative Commons license is claimed for this compilation.