Arcametric
Researcher Due Diligence
draft · 18 SEP 2026
Researcher Due Diligence | Terminology Governance

Arcametric Terminology and Mapping Manifest

This manifest defines how terminology mappings should be disclosed to researchers. It prevents three common errors: presenting a near match as an exact code, presenting a validated instrument as though it automatically has an external terminology identifier, and presenting a local concept used by Arcametric as an industry standard.

Status: DraftUpdated: 18 SEP 2026Version: 1.0

Purpose

This manifest defines how terminology mappings should be disclosed to researchers. It prevents three common errors: presenting a near match as an exact code, presenting a validated instrument as though it automatically has an external terminology identifier, and presenting a local concept used by Arcametric as an industry standard.

Arcametric governs its own canonical concept identity. External terminologies, research frameworks, registries, and consensus sets relate to that identity through governed mappings; they do not automatically replace it.

The Durable Grammar Framework is used as a governed architectural lens for semantic durability. It asks whether meaning, applicability, timing, provenance, and historical interpretation remain intact. It does not prescribe physical table names, generic database primitives, or workflow structure.

A mapping is a governed assertion about the relationship between two concepts for a stated use. Exact equivalence must be demonstrated rather than inferred from similar labels. Where no verified external concept preserves the intended meaning, a governed local concept is valid and should remain explicit.

Mapping classes

Class Meaning Research treatment
Exact external mapping Arcametric concept and external concept are verified as equivalent for the stated use Export external system, identifier, display, version, and verification date where permitted
Broader external concept External concept is less specific than the Arcametric concept Preserve Arcametric meaning and label relationship as broader
Narrower external concept External concept is more specific than the Arcametric concept Do not substitute it as equivalent
Related external concept Meaning overlaps but is not equivalent Preserve relationship explicitly as related
Validated instrument Field belongs to a recognized assessment instrument Identify the instrument separately from terminology mapping status
Local concept Needed to preserve meaning but no exact verified external mapping is established Export local code and definition, clearly labeled local
Safety overlay A separate safety classification applies because harm or clinical concern occurred Preserve both the experiential record and safety record when appropriate

Terminology families relevant to the current research model

System or family Primary role Boundary
RxNorm Medication or substance identity where verified Use only verified mappings and appropriate concept level
MedDRA Adverse-event terminology where verified and licensed for the use Do not use an adverse-event code merely to represent a non-harmful subjective experience
CTCAE Adverse-event severity grading where applicable Severity grade is separate from event identity
LOINC Observations and assessment-related identifiers where an exact verified concept exists Do not invent instrument mappings
SNOMED CT Clinical concepts where a verified relationship exists Preserve exact versus broader, narrower, or related status
UCUM Units of measure Keep units explicit and document conversions
Governed local experiential concepts Practitioner experiential concepts not adequately represented by verified external terminology Use current governed reference identity. This manifest does not prescribe an EXP_* prefix namespace; local identity is not an external standard.

Dataset-level manifest fields

For each coded field included in an approved research dataset, document:

Manifest field Requirement
dataset_field Required
local_code Required when a local concept exists
local_display Required when a local concept exists
code_system Required when an external mapping is supplied
external_code Required when an external mapping is supplied
external_display Recommended
relationship Required: exact, broader, narrower, related, instrument-only, local, or safety-overlay
terminology_version Required when material and available
verified_at Required for governed mappings
verification_source Required for governed mappings
licensing_or_distribution_note Required when redistribution may be constrained
status Current, under review, superseded, or not mapped
applicability_context Required when modality, respondent, instrument/form, workflow phase, value scale, or other context changes mapping meaning
effective_or_supersession_note Required when a mapping revision would otherwise make historical interpretation ambiguous

Experiential-layer rule

The practitioner experiential layer may use governed local concepts when no exact verified external terminology preserves the intended meaning. Local codes must come from current governed reference data. This manifest does not prescribe a canonical prefix namespace.

An experiential observation and a safety event are different records with different meanings. An unusual, difficult, meaningful, or intense experience should not be coded as harm merely because it lacks a conventional terminology code.

When harm or another safety condition occurs, the appropriate safety terminology can be recorded separately without erasing the experiential meaning.

Validated-instrument rule

A validated instrument can be identified by its recognized instrument name even when an exact LOINC or SNOMED CT identifier has not been verified.

Do not manufacture an external terminology mapping to make an export appear more standardized.

Applicability and context rule

Mapping identity alone is not enough when the relationship depends on context. Preserve the material applicability needed to interpret the mapping, such as modality, respondent or observation source, assessment instrument and form/version, value scale or unit, workflow phase, or safety-versus-experiential meaning.

Do not create a new physical field solely because this manifest names a semantic requirement. The requirement is that the approved dataset or governed mapping record can reconstruct the intended meaning.

Versioning rule

Terminology systems and local vocabularies evolve. Research packages should preserve the version or verification date that applied when the dataset was prepared when that information could affect reproducibility.

If a mapping changes after a dataset is released, do not silently rewrite the historical dataset. Preserve the mapping revision, effective or verification context, and supersession history needed to reproduce the prior interpretation.

Implementation evidence boundary

The presence of a concept, mapping, identifier, or field in this manifest does not by itself prove that the current production application captures it, that a current export serializes it, or that a live reference row is populated correctly. Implementation and live mapping state require their own current evidence.

Historical proposals and generic architecture examples are not production truth. Current schema, repository source, runtime, and live mapping evidence should be used only for the claims each source can actually prove.

Licensing boundary

Some external terminology systems have licensing or redistribution conditions. The existence of a mapping in Arcametric does not automatically authorize unrestricted redistribution of an external terminology table.

Dataset delivery should include only the terminology content permitted for the approved use and should identify any external licensing dependency.

Researcher review checklist

Before relying on coded Arcametric data, confirm:

  1. Which mappings are exact?
  2. Which are broader, narrower, or related?
  3. Which fields identify instruments rather than terminology concepts?
  4. Which concepts remain local?
  5. Which terminology versions apply?
  6. Which mappings have verification evidence?
  7. Does the mapping depend on modality, respondent/source, instrument/form, scale/unit, workflow phase, or other applicability context?
  8. If a mapping changed, can the prior dataset interpretation still be reconstructed?
  9. Which safety codes represent actual safety events rather than experiential phenomena?
  10. Are there licensing restrictions on redistribution?