One patient record across organizations
Records from every partner resolve to the same person and the same encounter history.
For statewide & multi-organization health programs
Program outcomes you can defend, from data your partners already send.
Hospitals, clinics, pharmacies and EMS agencies each record the same care differently. CareCompile builds the shared data layer underneath your program, so every partner’s records resolve to one patient and one set of codes, and your outcome measures hold up under review.
LOINC codes loaded
the full release, 2.83
Source spellings mapped
resolved to 276 shared concepts
Synthetic lab results resolved
72,743 of 79,298
Measured on synthetic data in a development environment, no live facility. Synthetic data generated by MediFlow, our clinical simulator (current release v14), updated 22 Oct 2026, 3:00 AM ET. No real patient data.
01 — THE PROBLEM
Programs rarely fail on ambition. They fail on the data.
When thirty organizations report the same lab result thirty different ways, a care team gets a pile of documents instead of a timeline, and the program office can’t compute a measure anyone trusts.
Fixing that by hand doesn’t scale, and guessing at codes produces numbers that fall apart in an audit.
02 — WHAT YOU GET
What your program gets.
Six things a program office can point to when someone asks how the numbers were made.
01
02
One set of codes
Lab results are checked against the full LOINC release. A code is accepted only if it exists and is active, never by guesswork.
03
A person approves every uncertain match
Software proposes, a named reviewer decides, and nothing unresolved is silently dropped.
04
Standards-based exchange
FHIR R4 endpoints, designed to run in your cloud environment or on infrastructure your program already operates.
05
Documentation an auditor can use
System architecture, data dictionary and mapping guide, delivered as program artifacts.
06
Measures you can defend
Impact reporting computed from coded, normalized results, not free text.
03 — HOW IT WORKS
Three systems, three labels. One clinical meaning.
Each partner names the same test its own way. We resolve those names to a shared concept and hold back anything we can’t verify instead of guessing.
As received — three charts
2160-0 (lab A, analyzer)1.1 mg/dL
CREAT-S (lab B, vendor default)1.4 mg/dL
CREATININE, SERUM (lab C, local name)168 µmol/L
Counted naively: three different tests, one result each, no trend.
As resolved — one chart
Creatinine [Mass/volume]n=3 · mg/dL
1.1 → 1.4 → 1.9rising
LOINC 2160-0, release 2.83sources: A, B, C
One concept, three sources, one rising trend. The local label each partner sent stays attached to every point. Illustrative example on synthetic values.
Technical detail: the five steps and what happens when a result can’t resolve
01
Receive
HL7 v2.x and FHIR R4 arrive with each system’s own codes, local mnemonics and units intact.
02
Normalize
Structure, specimen and units are read as sent. Nothing is coerced to fit; conflicts are kept, not smoothed.
03
Resolve
Each label is matched against the canonical dictionary and the loaded LOINC release. Only a release-verified exact match commits on its own.
04
Review
Everything else becomes a proposal and waits for a person. Approval is a named, timestamped action.
05
Publish
Unverified bindings ship with the code omitted and a flag in its place, so a guess cannot be mistaken for a confirmation.
Specimen
Code bound to the wrong specimen. Held and not counted.
Units
Units disagree with the concept’s canonical units. A conversion is offered to a reviewer, never applied silently.
Unknown
Local label matches nothing in the release. Quarantined on arrival and excluded from every measure until a person binds it.
Export
Published FHIR omits an unverified code. It carries
has-unverified-loinc-binding in its place.
04 — AN ENGAGEMENT
How an engagement runs.
01
Assess
Inventory partner systems and data feeds, and agree on the measures the program has to report.
02
Build
Stand up the shared data layer and terminology services, tested first on synthetic data.
03
Onboard partners
Connect organizations one at a time, with mapping review and quality checks for each.
04
Report and hand over
Deliver the measures, the documentation, and training for your team.
We work as the technical layer behind your program, not its front door. Community, workforce and governance work stays with the people who already do it well.
05 — PRIVACY & OVERSIGHT
Evidence for your privacy officer, without handing over the console.
Every release of information is logged, including refusals, and can be exported per person as an accounting of disclosures.
06 — RUNNING TODAY
What is built and exercised today.
Each of these can be demonstrated end to end in a live session.
Running
Checked against the full LOINC release
Never by check digit
A code is accepted only if it exists and is ACTIVE in the loaded release (98,101 codes, release 2.83). A correct check digit alone proves nothing, so arithmetic is never trusted.
Running
Machines propose. People decide.
Nothing is silently dropped
Only an exact, verified match commits on its own. Everything else waits for a person, with 87 proposals pending and 20 approved by a named reviewer. Unresolved results are held, never discarded.
Running
FHIR terminology any server can load
Vendor-neutral
The dictionary publishes as FHIR R4 CodeSystem, ValueSet and ConceptMap, and loads into stock open-source HAPI FHIR with zero validation errors.
07 — NOT YET PROVEN
Designed, not yet run with a real site.
Kept separate from the section above, on purpose.
Would be added
Pharmacy, EMS and devices
The other record traditions, mapped onto the same patient.
Would be added
Per-facility reporting
Needs each organization identified in the inbound stream.
Would be added
Onboarding at program scale
Designed, not yet exercised end to end.
What we will not claim
We hold no SOC 2, HITRUST or FedRAMP certification, and have no live production cloud deployment. If those are prerequisites, we would rather you learn it here than in procurement.
08 — WHO YOU WORK WITH
Two founders, no layers in between.
Bill Sackewitz
Founder and CEO
Public-sector health and program leadership. Sits with your leadership.
Mario Casamalhuapa
Founder and CTO
Platform and integration engineering. Works on your systems.
See it on synthetic data in thirty minutes.
No patient data and no data agreement needed. We'll walk your team through the full flow, from a partner's raw message to a reportable measure.
Programs notice — CareCompile provides data-integration and terminology services. Figures on this page are measured on synthetic data in a development environment, as of 22 Oct 2026, 3:00 AM ET.