INTEGRATION

Charles Trinh — FHIR Facade for Legacy Systems | Las Vegas, NV · Las Vegas, Nevada · Henderson · Clark County, Nevada · North Las Vegas, Nevada

Diagram of a FHIR facade in front of a legacy database and message queue
Charles Trinh — FHIR Facade for Legacy Systems

A FHIR facade puts a current-appointment interface in front of a system that was never designed to answer one. I build facades that translate at the edge, isolate the legacy system's fragility and give the customer a documented exit path.

What a facade is and what it is not

A FHIR facade is an application that answers FHIR requests by translating them into calls against a legacy system. The legacy system stays as it is; the facade presents a US Core-conformant surface to the consumer. The facade is not a fork of the legacy system and not a re-platforming programme. It is a translation layer with its own lifecycle, its own owner and its own plan for retirement.

The facade's job is to make the legacy system's limitations visible rather than hidden. Every facade I build answers a CapabilityStatement that states exactly which resources, which search parameters and which operations it supports. I do not promise operations the legacy system cannot perform in a reasonable time, and I do not invent resource support that would prove unmappable in a real test. The statement is the contract; the facade's code is the implementation. The two are compared by a conformance test that runs against both and produces a diff.

The facade's second job is isolation. The legacy system's fragility is absorbed at the facade. A legacy database with a nightly maintenance window is handled by a facade that pauses writes and serves reads from a cached view during the window. A legacy queue that fails on malformed messages is handled by a facade that validates the message before forwarding. The facade absorbs the legacy's quirks and exposes a consistent surface above them.

Query translation and the search parameters you support

The first piece of the facade is query translation. A FHIR search such as GET /Patient?family=Smith&birthdate=ge1980-01-01 has to become a legacy query that the underlying system can execute efficiently. The translation is not mechanical: the legacy system may not index the field, or may have a different representation of the same value. I start each facade with a parameter matrix that lists the search parameters the consumer uses and whether the legacy system can execute each one at the consumer's expected volume.

A parameter that the legacy system cannot execute efficiently is a design decision, not a technical one. Options include rejecting the parameter (the CapabilityStatement declares it unsupported), accepting it with a documented limit (the legacy query is slow but works at small volumes), or accepting it and translating the query in the facade (a join or an in-memory filter). Each choice has a cost. The first choice moves the problem to the consumer. The second keeps the problem at the facade. The third hides the problem in the facade's performance. I put the choice to the customer's operations owner with the volume estimate.

The second trap is pagination. FHIR pagination uses _count and a continuation link. The legacy system may paginate differently, or not at all. I make the facade own the pagination, fetching from the legacy system in pages sized for the legacy system and returning FHIR pages sized for the FHIR consumer. This decouples the two paginations and gives the facade room to correct the legacy system's performance. It also means the facade's cache is in the facade's paging buffer, and the buffer is sized to the customer's peak query volume plus a safety margin.

The CapabilityStatement as an honest promise

The CapabilityStatement is the only place where the consumer can read what the facade supports, and a facade whose statement is aspirational is worse than no facade. I run the conformance test described in the ONC (g)(10) criteria against the facade's statement, and the test's output becomes the facade's release gate. A resource outside the statement is not served. An operation claimed and not implemented is a release-blocking defect, not a defect for the customer to discover during integration. One facade claimed $everything and answered 500. I have seen one facade where the statement claimed $everything and the implementation raised a 500,

The statement also declares the versions the facade supports. FHIR R4 is the current default, and the customer may need STU3 for one or more legacy consumers. Supporting multiple versions in a facade doubles the translation surface, so I support multiple versions only when the customer name one named consumer that requires it and the consumer is in the customer's migration plan. The (g)(10) criteria are written against R4; additional versions are the consumer's problem, not the facade's, and I write that in the statement and in the contract.

The statement's documentation element is often treated as a field to fill in. I use it as the facade's user guide. The narrative section describes each resource, each search parameter, the legacy system's known limitations (the field that is always empty, the code set that is local rather than standard) and the typical latency. It is the first thing the consumer's developer reads, and it saves a week of debugging per supported resource.

Deferred operations and the long-running jobs the facade owns

Some FHIR operations are long-running by nature: $export, $everything on a patient with a long chart, and bulk operations on large Group resources. The legacy system is unlikely to support these natively, so the facade has to. The facade implements them as asynchronous jobs with a tracking URL, a job queue and a completion webhook. The job queue has its own storage (a database or a durable queue), its own worker pool and its own monitoring; it is not the legacy system's async engine.

The long-running operation's cost is the facade's, not the legacy system's, and it shows up in the facade's budget. A bulk export of a ten-thousand-patient group is a multi-hour run and a substantial load on the legacy database. I schedule it in the customer's maintenance window, size the export's chunked reads to the legacy database's capacity and monitor the legacy database's latency during the export. The facade that exports without watching the legacy database is a facade that gets blamed for the legacy system's slowdown.

The facade also owns the long-running operation's permission model. The consumer requesting the export may not have access to every patient in the group. The facade applies the customer's consent rules before the export begins, and the export's summary records how many patients were excluded and why. This is the same model as the gateway's consent filter, but applied to the batch. Without it, the export becomes a bypass of the customer's consent controls. See the bulk FHIR export work for the export design and the patient identity and consent work for the filter rules.

Write operations and the risk they carry

Most facades begin as read-only because the legacy system's write semantics are not safe to expose. A read facade can always be corrected by changing what it returns; a write facade can produce data the customer cannot undo. I implement reads first, prove them with a conformance test and a clinical sample review, and then add writes resource by resource with a named customer owner for each resource. The owner is the clinical or operations lead whose data it is, not the IT director.

The write operations require a translation from FHIR to the legacy system's storage model, and that translation is where I spend most of the validation time. A FHIR Observation with a LOINC code has to land in the legacy lab table with the legacy system's local test code and a reference to the patient in the legacy system's identifier. A round-trip test for each write resource is the only proof that the translation is lossless.

The second write risk is concurrency. Two consumers can write the same resource in quick succession, and the legacy system may overwrite silently. I implement optimistic concurrency with the resource's meta.versionId and return a 409 on a stale update. The legacy system may not carry a version identifier, in which case the facade keeps its own version table alongside the legacy data and rejects updates that do not match. This adds a facade-owned table but is the only way to give the consumer the semantics the FHIR specification promises.

Retiring the facade and the owner who keeps it honest

A facade has a lifespan. The legacy system may be retired, or replaced, or upgraded to a version that supports FHIR natively. The facade's retirement is planned from the day it is built, and the plan is in the customer's documentation. The plan names the trigger (the legacy system's retirement date, or its native FHIR support), the migration path for each consumer and the owner of the transition. A facade without a retirement plan is a permanent piece of debt that nobody wants to own.

The facade's owner is a named person on the customer's team, usually an integration engineer or an architect. The owner reviews the facade's statement against the legacy system's current state every quarter, checks the mapping document against the legacy system's current schema and signs the facade's release notes. The owner's name is in the facade's README and in the customer's interface register. Facades without owners rotate through support teams; facades with owners are maintained.

The facade's operating cost is modest but real: the facade's compute, the job queue's storage, the monitoring, the quarterly review and the release process. I quote these as a running cost so the customer can compare the facade's total cost against the alternative (a full re-platforming programme) over three years. The facade almost always wins on cost. See the US Core conformance work for the validation detail and the HL7 interface engine work if the legacy system also speaks HL7v2 and the facade has to source data from a message flow.

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 — fhir facade for legacy systems — 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