ehr · adt · orders · results
Charles Trinh — Health System Integration, Las Vegas NV

Integrating with a hospital or health system: interface agreements, ADT and results flows, downtime behaviour, and the testing that proves it before go-live.
Start with the interface agreement, not the code
Who sends what, on which trigger, with what acknowledgement, and who is called when it stops.
Write it down and have both sides sign it, because most integration failures are disputes about expectations rather than defects in code.
Name the message owner on each side, with a deputy, and put both in the document.
ADT as the spine
Admissions, discharges and transfers drive every downstream system; getting ADT right makes the rest tractable and getting it wrong makes everything else unreliable.
Handle out-of-order delivery explicitly with a monotonic marker and a dead-letter path, because replays after downtime are normal.
Treat a missing discharge as an incident, because a patient who never leaves the census distorts everything downstream.
Orders and results
Order messages carry intent; result messages carry findings, and the two need distinct handling and distinct acknowledgement.
Map to a model you own rather than field to field, so a second consumer does not become a second full project.
Log the acknowledgement per message, at application level, because transport success with application failure is the classic silent break.
Downtime and recovery behaviour
Decide what happens during a planned outage and an unplanned one, and test both before go-live.
A queue that survives restart plus a documented catch-up procedure is the difference between a blip and a data gap.
Record the last successfully processed message on every interface, so recovery starts from a known point.
Testing that proves it
Volume test at peak, not average; order test with replayed and out-of-sequence messages; failure test with the receiving system down.
Test the clinical workflow, not the message, because a message that parses and a workflow that works are different outcomes.
Keep the test evidence, since it is the artefact you will be asked for after the first real incident.
Operational handover
The runbook names each interface, its owner, its expected volume, its last success and its first diagnostic step.
Monitor from outside the interface engine, because the engine's own health does not tell you whether the hospital received anything.
Review the interface inventory quarterly; interfaces accumulate and nobody retires them.
How I do it
One trace of a real case with timestamps at every hop, before any document is written.
A findings list naming the defect, the system, the owner and the date, reviewed monthly.
No attestation I did not test, and no template sold as an assessment.
Interface inventory
Before designing anything, list every interface the organisation runs, with its owner, its volume, its last success and its business purpose.
The inventory usually reveals three interfaces nobody can explain and two that fail silently every week.
Retire what nobody uses before adding anything, because most integration estates grow without ever shrinking.
Mapping to a model you own
Do not map the sender's fields onto your database columns; map them onto your own domain model and keep the translation explicit.
The extra layer is what lets a second sender arrive without a second project, and it makes the loss visible when a field is dropped.
Record unmapped fields, because a dropped field is a clinical question waiting to be asked.
Security and consent on the wire
Every interface carries a legal basis, and a message that arrives without one should be quarantined rather than processed.
Encrypt in transit, authenticate both ends, and log the identity of the sender rather than only the network address.
For anything containing sensitive categories, enforce segment-level rules before the data reaches the store.
Change management
A change on either side is an integration change, and the only way to survive it is notification before deployment.
Exchange a written change notice with each counterparty and keep the contact list current, because the names in it change every year.
Test the change in a shared environment where one exists, and where it does not, test against captured production messages.
The people half of integration
Every interface has a person on each side who is called when it breaks, and both names should be current and in the runbook with a deputy. The list ages faster than the technical configuration, because people change roles and the interface does not. Reviewing that list quarterly, and testing that the named person answers, is a more valuable exercise than most of the technical testing because the technical failures are usually resolved by the conversation that follows.
Interface work also crosses clinical and technical audiences, and the successful projects have one person who can translate. Without that translator, requirements arrive as message formats and clinical needs arrive as complaints, and the gap between them is filled by assumptions in code. Naming the translator in the project charter is unusual and effective.
What go-live actually requires
Go-live is not the first successful message; it is the first week of clinical use with someone watching the queues. Plan for a person on both sides during that week, a defined check several times a day, and a rollback that does not require a project meeting. The interfaces that fail in production are almost never the ones that failed in test, because test traffic is orderly and production traffic is not.
Keep the test evidence and the go-live checklist together, because the artefacts are what you will be asked for when something goes wrong. The checklist should name the checks that were performed, by whom and when; a checklist with boxes ticked and no names is a document that records enthusiasm. This is cheap to do properly at the time and impossible to reconstruct later.
Volume, throughput and back-pressure
Peak clinical volume is concentrated and repetitive, and an interface that keeps up at average load can still fall behind during a morning clinic. Measure throughput per interface, define the maximum acceptable backlog, and alert on backlog rather than on errors, because a slow interface produces no errors while it accumulates work. Backlog is the number that tells you whether the clinical workflow is being affected.
When the backlog grows, the response should be defined in advance: shed non-critical message types, increase workers, or notify the receiving side. An interface with no shedding policy saturates and then behaves unpredictably. Writing the priorities down before the incident turns a negotiation into a lookup.
Reconciliation after an outage
After any outage, both sides believe different things about what was exchanged, and the reconciliation compares the sender's record of what it transmitted against the receiver's record of what it processed. That comparison needs a stable message identifier on both sides and a time window to compare over, and it should produce a list of specific messages rather than a count. Counts hide the individual patient record that never arrived.
The report should be produced automatically on request rather than reconstructed during an incident, which means the query has to exist before it is needed. Building it once, testing it against a deliberate gap, and keeping it in the runbook turns a multi-hour reconciliation into a ten-minute check. The organisations that do this find gaps every quarter; the ones that do not find them in a chart review months later.
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 — health system integration, las vegas nv — 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
- Phone: 702-559-3489
- Email: charles@trinh.io
- Based in Las Vegas, Nevada · working remotely across the United States
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.