CMS-9115 · FHIR R4 · app ecosystem

Charles Trinh — Patient Access API | Las Vegas, NV · Las Vegas, Nevada · Henderson · Clark County, Nevada · North Las Vegas, Nevada

A service card reading PATIENT ACCESS API — member-authorised apps against a conformance-tested surface
Charles Trinh — Patient Access API

The member is the client

A patient access API exists to be consumed by an application the patient chose. That single sentence changes the design. A claims export behind a login is not a patient access API; it is a report with an OAuth wrapper, and the member experiences it as a portal rather than as access.

The obligation under CMS-0057-F is a member-facing data surface using FHIR and US Core, with coverage information conveyed per the CARIN digital insurance card guide, consumable by a third party without a bespoke integration with the plan. The operative test is not whether the endpoint returns 200. It is whether a member can connect an app they picked, and whether that app can still read their data six months later.

What has to be in the payload

Claims and encounters, because that is what the member recognises and what most plans already hold. Clinical data — conditions, medications, observations, immunisations, allergies — because a response that describes what was billed and nothing about the patient's health is a billing history wearing a clinical label.

Coverage, per CARIN, including the member identifier, the plan and product, the group, the network, and the contact information the member would otherwise have to call to obtain. A digital insurance card that omits the phone number on the back of the physical card has not replaced the card.

Provenance matters here in a way it does not in provider exchange: a member looking at their own record needs to know whether a value came from a lab system or from something they typed into an app. Collapsing that distinction produces a record the member cannot evaluate.

Authorisation, and the grant as an artefact

SMART on FHIR with the patient-access scopes, and the member's grant recorded as a first-class object: which application, which scopes, which patient, granted when, expires when, revoked when. That record is what you produce when a member asks who has been reading their data, and it is what you cannot reconstruct after the fact if you did not write it down.

Present the grant clearly. A consent screen that says the app will have access to health information is not a disclosure a member can evaluate. Name the scopes in plain language, name the application and its publisher, and make refusal the easy path rather than the buried one.

Revocation has to be immediate

Revocation that takes effect at the next token refresh is the single most common defect in patient access implementations. From the member's side, revoking access and continuing to see data flow is indistinguishable from the platform ignoring them, and under information-blocking scrutiny it is a withholding question dressed as a token lifetime.

The mechanism is a grant-version stamp checked on the resource path, not a shortened token lifetime. Cheap, fails safe, and testable in an afternoon.

Application registration is a control, not a form

Third-party applications need a registration process with identity verification, redirect URI validation, a named contact, a published privacy policy, and a way to suspend an application that misbehaves. An open registration endpoint with no verification is an incident waiting for a name.

Equally, registration cannot be a months-long gate that no small developer can pass, because the point of the rule is that the member's chosen application can connect. The balance is verification that is fast and real, not diligence that is slow and theatrical.

Rate limits, published

Publish them: per client, per member, with a documented retry contract and a contact path when a legitimate application needs more. An app that hits an undocumented limit and receives a 429 with no guidance has no behaviour available to it, and the member blames the plan.

Reliability is the feature

Members do not care that the endpoint exists; they care that their app works on a Sunday. That means monitoring from outside the estate — synthetic requests against the production surface, from a real client, on a schedule — rather than dashboards built from your own logs. Your logs will show success while a certificate expiry takes the surface down for everyone.

Testing from the outside

Register a test application exactly as a third party would, complete the flow, refresh a token, revoke consent, confirm the grant dies, and confirm the member's data disappears from the app. Quarterly, and after every release. Everything else you measure about this API is a proxy for that one journey.

What I do

I have implemented and operated the patient access surface as part of a platform that passed the ONC (g)(10) Standardized API suite at 317 of 317 subtests, covering US Core, SMART App Launch and Bulk Data. I know which failures in that suite are cosmetic and which are the ones a real application will hit on its first day.

If your patient access endpoint passes conformance and no member has ever connected an app, the gap is not in the specification. That is the conversation I have.

What breaks in month six

Almost nothing fails on launch day. The failures that matter arrive once the application is connected and the novelty has passed.

Certificate expiry. The TLS certificate on the authorisation server rotates on a schedule nobody wrote down, and the app that has worked since March stops on a Saturday. Monitor from outside the estate, not from your own logs — your logs will show success while the surface is unreachable.

Scope drift. A release changes the scopes a client is granted without any change to the client. The app did not break; the contract did. Version your client registrations and diff the granted scopes on every release.

Refresh token expiry that surprises. A refresh window shorter than the member's usage cadence forces re-authorisation, which members experience as the plan revoking access. Decide the window deliberately against real usage and write the decision down.

Silent data gaps. A source system stops feeding one resource type and nobody notices, because the response is still a valid bundle with fewer resources in it. Monitor counts per resource type, not only HTTP status.

The reporting a plan actually needs

Four numbers, weekly, and they should be visible to the compliance owner rather than buried in an engineering dashboard.

Registered applications and the count that have made a successful request in the last thirty days. A large registration number with a small active number means the surface is technically open and practically unused, which is a different problem from an outage and needs a different remedy.

Members with an active grant. The proportion of the population that has connected anything tells you whether the member-facing story is working.

Requests per member per week, distribution rather than mean. A mean hides the single heavy integrations and the long tail that is barely functioning.

Errors by class — authorisation, scope, validation, availability. A rising authorisation error count is either a broken app or a broken server, and the distinction is decidable from the error code if you are logging it properly.

Vendor questions that separate real from demonstrated

Ask for the production endpoint and a test account, and complete the flow yourself. Ask which scopes are enforced and where in the data path. Ask how revocation propagates and how quickly. Ask for the last ninety days of availability measured externally. Ask what happens when a member's coverage ends and their data should no longer be served.

A vendor that cannot answer the revocation question has not implemented revocation. That single question is the most efficient discriminator I know.

What I do

I have built and operated the patient access surface inside a platform that passed the (g)(10) suite at 317 of 317. I fix the enforcement layer and the observability around it, because an endpoint that passes conformance and has no external monitor is an endpoint that will fail without anyone knowing.

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.

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

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 — patient access api — 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.

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