DATA

Charles Trinh — Terminology Services: SNOMED CT and ICD | Las Vegas, NV · Las Vegas, Nevada · Henderson · Clark County, Nevada · North Las Vegas, Nevada

Terminology service interface showing a SNOMED CT concept with ICD-10-CM map targets and version history
Charles Trinh — Terminology Services: SNOMED CT and ICD

Two systems that code the same clinical concept differently will disagree about the count. Terminology services map the disagreement to a rule, a version and a published mapping rather than to a meeting.

Why terminology work matters more than it looks

The difference between two code systems is not a translation exercise. A SNOMED CT concept is a clinical term with a logical definition and a hierarchy; an ICD-10-CM code is a classification designed for reporting and payment. The concept Acute pharyngitis exists in SNOMED CT with diagnostic criteria and synonyms. In ICD-10-CM the equivalent code exists as J02.9 within a chapter and a billable-code set. When a clinical decision support rule fires on one and a payment claim is coded to the other, the mapping is a decision about usability, not a translation.

Most terminology problems are not the code systems; they are the mappings between them. A hospital that maps SNOMED CT to ICD-10-CM locally without published rules produces a mapping that is unmaintainable, undocumented and quietly wrong at the edges. The mapping needs to be a first-class artefact with a published policy, a version, and a test. The test is the difference between a mapping that someone believes and a mapping that someone can audit.

Terminology services sit underneath every meaningful analytic and every decision support rule. A bad value set makes a quality measure wrong. A bad map makes a claim wrong. The cost is not in the server; it is in the downstream reporting that nobody inspects until the numbers disagree and the root cause traces back to a local mapping. I treat terminology as engineering infrastructure with a contract, not as a librarian's spreadsheet, and I hold it to the same tests, the same versioning and the same review as the data pipelines that feed the warehouse.

Value sets as first-class assets

Every downstream analytic that uses coded data is using a value set, whether the team knows it or not. The list of ICD-10 codes that counts as diabetes, or the SNOMED concepts that count as a medication allergy, is a value set. I require that every value set used in a report, a quality measure or a prior-authorisation rule be stored in the terminology service with an owner, a version, and a description. Inline code lists in SQL are the defect.

Value sets carry intension and extension. Intension is the rule: an ECL expression or a query. Extension is the list of codes resolved at a specific version. I store both, with the resolved extension frozen per version so the set used in a report is reproducible. When an SCT release lands, the new version produces a new extension, and the reports that referenced the old version still reproduce. Without this, a report from March reads differently in June when the release updates.

Value set governance names who can change which set, and how. Clinical teams own clinical value sets; analytics teams own measure definitions; finance owns billing subsets. Ownership is visible in the service. When a change is proposed to a set used in a reported measure, the change routes to the measure owner for approval before it takes effect., and the reviewer records the date on which the ownership was last exercised.

The mapping policy

Local maps are sometimes unavoidable, but they are never a substitute for a published policy. I write the mapping policy before the first local map: who produces maps, which direction (SNOMED to ICD, ICD to SNOMED), the source of truth, the process for handling ambiguity, the approval route, the version controls, and the deprecation path. The policy document is the artefact the mapping team operates from and the auditor reads.

Direction of mapping changes the semantics. A SNOMED-to-ICD map takes a clinical concept and produces the best equivalent in the classification for reporting; unmapped concepts usually fall to an "other" code and that is acceptable. An ICD-to-SNOMED map takes a billable code and produces clinical concepts; that is often one-to-many and the policy has to state whether the mapping produces all candidates for picking, or a single default. Conflating the two directions is where most local maps become unmaintainable.

For the US, the published mappings from the terminology authorities are the starting point. I use the national release as the base, with a local overlay only for clinical concepts that national does not cover. Overlays are maintained with a journal: each addition, its rationale, who approved it, and which concept it maps from. An overlay with no journal becomes a historical accumulation nobody can explain within a year.

Version management and the release cycle

Terminology releases are the main source of quiet change. SNOMED CT International, the US extension, ICD-10-CM annual and intermediate updates, LOINC updates, and every source system's own reference release all move. I tie every value set, every mapping and every report to a specific release. The terminology service stores which release each artefact depends on and which releases are currently installed.

When a new release lands, the service runs a difference report against the previously installed release: concepts added, concepts deprecated, hierarchy changes, and every value set whose extension changed as a consequence. The difference report is routing, not reading. It names which value sets and mappings need review, and the review queues to the owning team. Most releases bring many changes and only a handful that touch a reported set; the report makes the handful visible and lets the team spend review time on the changes a payer or regulator will see.

In-place upgrades are not acceptable. The service runs two versions side by side during the transition, with reports pinned to the version they were signed off on. After the review of the difference report, the value sets move to the new version and the reports are re-run. A cutover where the old version disappears the same day is a source of audits, because reports signed off on Friday read differently on Monday with no record of the change.

The terminology service itself

In FHIR terms the service exposes a CodeSystem, a ValueSet and a ConceptMap resource per artefact, with the standard operations for lookup, validate, subsumes, expand and translate. The US Core conformance work covers the profile details. The important part is that the operations are the only way to resolve a code; teams do not read the release files directly. When teams read the raw release, the service cannot version the read.

I build the service to be tolerant of malformed input from upstream systems. Real coded data carries unrecognised codes, deprecated codes, local extensions and codes from the wrong release. The service records what it saw and flags it, rather than rejecting the row. The rejection rate for a new source is usually the first clue that a source is using a different release of a code system than the service was configured with, and rejecting rows loses that signal.

Performance matters because value set expansion against large ICD-10 sets can be slow if done naively. I cache expansions by value set version, and I warm the cache for the value sets used in every scheduled report. The report jobs hit warm cache at 02:00 rather than cold cache. The difference is usually between a minute and an hour for a measure run.

Integration with reports, measures and claims

Reports and measures consume value sets by identifier and version. The report code names the value set ID and the version it was signed off against; the service resolves the member codes. When a new version is approved, the report moves forward only via a change that records the version bump. The report is reproducible either way. A measure that names "all diabetes codes" without a version produces a different rate every time the release changes.

Claims and billing subsets are downstream of the same service. A code that is valid clinically but not billable, or billable but not covered by a payer, is a value set question, and the set differs per payer. I keep the payer subsets as separate value sets under the same service, with their own owners and review cycles. The service becomes the only place these differences are described, rather than a set of Excel files passed between teams.

The finished service is used by clinical systems for decision support, by analytics for measure and report logic, and by revenue cycle for coding and claim rules. All three see the same value sets and the same releases. The healthcare data governance work names the owners and the escalation when a set is disputed. A terminology service without governance questions is a good database with no decisions being made about its content.

Proof

  • ONC (g)(10) self-test PASS, 317 of 317 subtests, April 2026 — public suite URL
  • 67 public hostnames, 51 serving, one AWS account
  • 104 clearinghouse enrolments across 53 jurisdictions
  • USCDI v7 proposal — Patient Digital Identity Credential — listed at isp.healthit.gov
  • SAM.gov UEI MVKQW9DKKGB8 · CAGE 20KM5 · NPI 1194645564
  • AWS Business Associate Addendum active since 18 August 2026

Where this work is delivered

Engagements run from Las Vegas and reach across the valley: clients in Las Vegas, NV; Las Vegas, Nevada; Henderson; Clark County, Nevada; and North Las Vegas, Nevada — plus remote delivery nationwide.

How an engagement runs

1. Trace one real case

Follow a single patient or claim end to end with a timestamp at every hop, before any document is written.

2. Written findings

A findings list naming the defect, the system, the owner and the date.

3. Monthly review

Review the list monthly and check whether the dates held.

Frequently asked

What is this service?

This page describes charles trinh — terminology services: snomed ct and icd — the scope, the failure modes it prevents and how the work is measured. It sits in the Interoperability category at trinh.io.

Where is the work delivered?

Engagements run from Las Vegas, Nevada across the valley — Las Vegas, Henderson, North Las Vegas, Spring Valley, Paradise, Winchester, Sunrise Manor, Enterprise, Summerlin, Clark County — and remotely anywhere in the United States. On-site work in the Las Vegas valley, remote delivery elsewhere.

What does it cost?

Standard rate $195 per hour, committed $165 per hour at twenty hours or more per week for three months or more; $1,600 standard day, $1,350 committed day. Scoped in writing before work starts, fixed fee where the deliverable is defined.

Who is Charles Trinh?

A technology architect in Las Vegas, Nevada, working in healthcare identity, consent and interoperability. Not a physician, not a medical practice, and not affiliated with any other Charles Trinh.

How does an engagement start?

With one trace of a real case and a written findings list naming the defect, the system, the owner and the date. The first call is twenty minutes and free: 702-559-3489.

Engagement

Scoped in writing before work starts; fixed fee where the deliverable is defined. First call is 20 minutes and free. 702-559-3489 · charles@trinh.io

Talk to Charles

Record

Profile

Healthcare technology architect working where federal regulation becomes production infrastructure. Built and operates a FHIR R4 clinical platform — SMART App Launch authorization server, patient and provider portals, SQLite — on a hardened single-host AWS estate of 67 public hostnames. The platform returned a self-test PASS on the ONC (g)(10) Standardized API test suite in April 2026, 317 of 317 subtests, publicly verifiable. Designed a consent architecture in which a permissioned ledger records patient grant, refusal and revocation as immutable events while all clinical data stays off-chain and encrypted per patient — no PHI on the ledger by design. Filed a proposed USCDI v7 data element with ONC on verifiable patient identity. Translates the Federal Register, HHS/CMS guidance and HITRUST requirements into architecture, controls and audit evidence.

Selected engineering scale

  • Self-managed estate of 67 public hostnames, 51 serving, on one AWS account.
  • FHIR R4 API with a SMART App Launch authorization server; patient and provider portals; an AI service layer; Hyperledger Fabric peer, orderer and CA; IPFS; a 337-table SQLite clinical store.
  • Two permissioned distributed ledgers in production, one operating a consent registry that records grant, refusal and revocation with no PHI on chain.
  • X12 005010 pipelines across eight transaction sets.
  • Designed, built, deployed and operated end to end by one person.

Credentials and standards work

  • Inferno (g)(10) self-test PASS, 317 of 317 subtests, April 2026 — the ONC Standardized API test suite. Standards validated: US Core 6.1.0, USCDI v3, SMART App Launch 2.0.0, Bulk Data 2.0.0. Public suite URL.
  • USCDI v7 public comment submission, April 2026 — proposed Patient Digital Identity Credential under the Provenance data class, listed publicly on the ONC Interoperability Standards Platform.
  • SAM.gov UEI MVKQW9DKKGB8 · CAGE 20KM5 · NPI 1194645564. AWS Business Associate Addendum active 18 August 2026.

Technical environment

AWS us-east-1 (compute, storage, messaging, SMS, email, voice) · Linux and systemd hardening · nginx · Fabric and EVM ledgers · FHIR R4 · X12 005010 · HL7 · SQLite · self-hosted mail, drive, search and monitoring · Git-based change control and CI release gates.

TRINH — Charles Trinh, technology architect
TRINH

Charles Trinh

Fractional CTO/CIO and Technology Architect — Healthcare Data, Identity & Regulated Infrastructure · Las Vegas, NV

Las Vegas, Nevada · US citizen · Fractional and project engagements

(702) 559-3489 Email Charles Trinh