DATA
Charles Trinh — Healthcare Data Quality Programmes | Las Vegas, NV · Las Vegas, Nevada · Henderson · Clark County, Nevada · North Las Vegas, Nevada

Data quality is a programme, not a project. It needs a definition of correct for each field, a monthly measurement, a named owner, and a fix when the measurement moves. I set up the four together.
Why quality programmes stall
Most quality programmes stall because they begin with a dashboard and never agree on what correct means. A completeness figure of 94 per cent is meaningless if nobody has decided which of the missing records are legitimately absent, which are ingest failures, and which are data entry gaps. Without that definition, the number moves when the denominator changes and cannot tell anyone whether the system improved or the population changed.
The first deliverable is a definition catalogue: one line per field per source naming the domain, the validity rule, the allowed missingness, the expected latency and the owner. A catalogue of a few hundred fields for a mid-sized hospital is normal. The catalogue is the artefact everyone argues about instead of the score. Once it exists, the argument has a reference point and it becomes a bounded conversation.
I write the catalogue with the source system owner next to me. When the owner says the field is 98 per cent complete and the extract says 71 per cent, that gap is itself the finding and it is usually something upstream of the warehouse. Quality work moves fastest when the first month is spent finding disagreements between owners and extracts rather than publishing dashboards.
The dimensions that matter in healthcare
I measure completeness, conformity, timeliness, plausibility and uniqueness. Conformity covers code systems and formats: a diagnosis that is not in the agreed ICD-10-CM release, a race field outside the census categories, a date of birth in the future. Plausibility covers clinical ranges: a systolic pressure of 410, a weight of 4 kilograms on an adult, an allergy coded to a drug the patient has received for years. These are the fields where errors are silently expensive.
Timeliness is the dimension most often skipped. A lab result that arrives three days after the draw breaks a clinical quality measure as effectively as one that never arrives. I measure the distribution of arrival lag per source and flag when the 95th percentile slips past the contractual window, rather than only counting late records. The percentile view catches slow drift that averages hide.
Uniqueness matters at the patient level and at the event level. Duplicate encounters inflate denominator counts, duplicate problems create false precision in cohorts, and duplicate observations make a trend look like a spike. The measurement ties into the patient matching work, because many duplicates originate at registration rather than in the warehouse., which is where most of the reconciliation work actually sits.
Scoring without gaming
Any score becomes a target and targets get managed. I keep the detailed per-field measurements in analysts' hands and publish only a small set of composite measures to leadership, each with a definition that was agreed before the first score was computed. A composite that can be improved by widening the denominator is a defect, and I test that by running the measure against the prior quarter's population before anyone sees the number.
Owners see their own score daily, not monthly, with the ability to drill into failing records. Monthly review is where the programme meets; daily feedback is how the owner fixes issues. A quarterly-only cycle to a team that can fix the issue in an afternoon wastes most of the available improvement. The exposure is the point: an owner who can see failing rows fixes them far faster than one reading a percentage.
Scorecards carry the date of last definition change. When a score jumps, the reader can immediately see whether it was the data or the rule. Without that line, every improvement conversation starts with a dispute about whether the definition moved and the programme spends most of its meeting on archaeology instead of action., and the scorecard shows drift before it becomes a complaint from a clinician reading the chart.
Root cause and the fix process
Every failing cell in the scorecard links to sample records and to the upstream step where the failure entered. I do not accept a fix that cleans the warehouse but leaves the source. When a completeness issue traces to a registration screen that permits empty fields, the fix is the screen, and the fix is verified by the score returning to threshold after the change. Warehouse cleaning papers over the source and the failure returns next week.
Fixes go through the same change process as any production change: a ticket, a review, a test, a rollback plan. Quality fixes that bypass change control are the ones that hurt most when they go wrong, because the person fixing usually knows the least about the downstream consumers of the field. I have seen a well-intentioned cleaning job overwrite reference values used in an external submission, and the recovery took days.
Remediation is tracked with a due date and a named owner. When the due date passes, the cell escalates to the data governance body. The escalation ladder is the only part of a quality programme that consistently gets skipped, and it is the part that turns measurement into improvement rather than a conversation that repeats every quarter., and the same rule caps the length of a fix so one stalled field does not block others.
Contracts between producers and consumers
Each source system and each downstream consumer agree a contract: the fields, the definitions, the refresh frequency, and the tolerances. The contract is tested nightly and a breach opens a ticket against the producer, not against the consumer. In practice this single change moves ownership disputes from a standing argument to a queue with a clear owner in the producer's team.
I extend the contract to systems outside the hospital where the data originates there. Payer files, external laboratories and registries each have their own conventions and the buyer usually learns about a change when totals stop matching. A contract and a nightly conformance check turn that discovery into a scheduled conversation rather than an emergency when the numbers are reconciled at quarter end.
Contracts are reviewed annually and whenever a source system is upgraded. A vendor upgrade that renames a field or narrows a code set breaks silently. The version on the contract names the source release it was tested against so the upgrade review has a reference point. Systems that pass the tolerance check on dev fixtures can fail in production when the vendor changes an internal format.
Instrumentation and the scorecard
The instrumentation itself is small: one table of definitions, one of measurements by date and source, and one of active issues with ownership and history. Reports are views on top. I keep the store in a clinical data warehouse alongside the data being measured, so every score is queryable against the exact records behind it and a challenged score can be traced to a specific row.
Alerts fire when a score moves beyond its control band, and the control band is computed from the field's own history rather than a fixed threshold. A field that normally varies by ten per cent will not page the team for that. The alerting builds on the same platform as other system health work covered in observability and synthetic monitoring, because a source system that has stopped sending data is usually visible as an outage first.
The programme has a monthly report, a written account of what changed, what was fixed, what remains open, and what definition changed. The monthly report is short and it is the artefact I hand to the executive sponsor and to the audit committee. A programme without that document looks the same to a reviewer as one that has stopped working, regardless of what the scorecard shows.
Building the team to run it
The programme is run by the hospital after handover. I work with the data steward community through the first two monthly cycles, writing the review procedure with them rather than setting it and leaving. The procedure names who reviews which score, what counts as an exception, and how a new field enters the catalogue. Without that, the second cycle becomes a copy of the first without improvement and the scorecard stops being read.
Stewards are given protected time. A steward who is expected to review data quality on top of a full clinical schedule makes choices that favour the clinical work every time, and that is correct from their perspective. The programme budget has to include the stewards' hours or the programme is a fiction in the first month and a recognised fiction in the second.
The measure of a successful programme is not the score. It is that the monthly review discusses specific exceptions with named owners and due dates, and the score moves as fixes land. When a quarter passes with the same failing cells and the same priorities, the programme has quietly become a reporting exercise and the readiness assessment is the right next step.
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 — healthcare data quality programmes — the scope, the failure modes it prevents and how the work is measured. It sits in the X12 and EDI 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.