Coding-set version registry
This content is for 2026. Switch to the latest version for up-to-date documentation.
The coding-set version registry is the governance surface for classification families used by pricing-year manifests and validation rules. It keeps the version story explicit for AR-DRG, AECC, UDG, Tier 2, AMHCC, ICD-10-AM, ACHI, and ACS without treating licensed products as public package content.
For AR-DRG specifically, the registry tracks the versioned relationship between ICD-10-AM, ACHI, ACS, and the licensed grouping tables or software used to derive the DRG output. The registry records provenance and compatibility, not a proprietary grouper reimplementation.
Use the registry to answer four questions:
- Which coding-set family is in scope for a pricing year?
- Which version, release date, and implementation date apply?
- Is the artifact public metadata, or is it a restricted licensed product?
- Which stream-specific validator should reject an incompatible version?
What belongs in the registry
Section titled “What belongs in the registry”Registry entries should record the minimum metadata needed to make version compatibility testable and reviewable.
| Registry field | Purpose | Public? |
|---|---|---|
| Family name | Identifies the coding-set family in a pricing-year manifest | Yes |
| Version | Pins the exact released version used by the year | Yes |
| Release date | Shows when the version became available | Yes |
| Implementation date | Shows when the version becomes effective in a pricing year | Yes |
| Pricing-year scope | Lists the years or transitions that depend on the version | Yes |
| License or redistribution note | Explains whether the version can be redistributed | Yes |
| Source reference | Points to the authoritative publication or catalog entry | Yes |
| Local asset reference | Names the checked-in or user-supplied artifact path | No, if it exposes restricted content |
The registry should stay machine-readable so pricing-year manifests and validators can resolve expected versions without guessing.
Public metadata vs restricted licensed products
Section titled “Public metadata vs restricted licensed products”Public metadata can describe a coding set without redistributing the code list or grouper itself. Examples include family names, version numbers, effective dates, applicability notes, and license caveats.
Restricted licensed products include the actual code lists, grouper binaries, mapping tables, or product packages that cannot be redistributed unless the license explicitly permits it.
Keep these boundaries clear:
- Public docs may describe that a licensed product exists and how it is used.
- Public docs may name the version and the pricing-year impact.
- Public docs must not copy restricted code tables or attached product files.
- A restricted product should remain outside the public docs tree unless the license permits redistribution.
- Public docs may record the versioned ICD-10-AM/ACHI/ACS inputs and the presence of licensed AR-DRG grouping tables or software.
- Public docs must not claim to reproduce the proprietary AR-DRG grouper.
Local user-supplied groupers and code lists
Section titled “Local user-supplied groupers and code lists”Some workflows require users to supply their own licensed groupers or code lists locally.
That workflow should be described as a user-managed boundary, not as a public download:
- The repository can validate that a supplied file matches the expected family and version.
- The repository should not bundle the licensed asset if redistribution is not allowed.
- A local path, command, service endpoint, or file-exchange location can point to the user-owned artifact.
- Compatibility checks should fail closed when the expected local asset is missing or the version does not match the selected pricing year.
- Precomputed AR-DRG outputs can also be treated as local inputs when the provenance is clear and the study does not require regrouping.
This lets the docs explain how the software behaves without implying that the licensed asset is shipped in the repository.
Relationship to stream input validation
Section titled “Relationship to stream input validation”The coding-set registry is the canonical version source. Stream-specific input validation is the consumer that rejects incompatible versions or missing local assets before a calculation runs.
The current validation work is tracked in the stream input validation docs:
- Classification input validation track
- Classification input validation spec
- Classification input validation plan
Those docs describe the preflight rules for stream-specific required fields, pricing-year compatibility, and the boundary between shared validation and local licensed inputs.