Solution · Healthcare

Clinical & claims data mapped to FHIR.

Hospitals, labs and payers receive patient, lab and claims data in every possible shape. AdaptivMapr maps it to FHIR R4 with medical-code validators — without shipping raw records to an LLM.

Get an API key Read the docs

The problem

Why this is hard

Hospitals, labs and payers receive patient demographics, lab results and claims in every possible CSV, Excel and HL7-adjacent shape — one export per source system, none of them alike. Mapping each to a canonical model like FHIR R4 is slow, error-prone and manual, and it is exactly the data you least want to hand to a generic AI. The AdaptivMapr healthcare pack ships 10 FHIR-mapped templates (patient_demographics, lab_results, claims_line_items, drug_formulary and more) with medical-code validators and direct FHIR R4 emit. Schema-only mapping never lets a record leave your system; when you need full-data routing, it runs in-region under a BAA.

How it works

The pipeline, end to end

1
Upload a clinical or claims file
Send a CSV or Excel export from your EHR, LIS or claims system. In schema-only mode only the headers and up to 3 short sample rows leave your infrastructure — no patient record is transmitted.
Schema-onlyCSV / ExcelNo PHI leaves
2
The cascade maps headers to a FHIR template
Five deterministic-first layers (statistics, heuristic, fuzzy, semantic, LLM) resolve each column against the healthcare pack template. Multilingual DE/FR/IT/EN/ES hints catch vocabulary drift; most columns never reach the metered LLM layer.
Cascade engineDE/FR/IT hintspatient_demographics
3
Validators check the medical codes
Field-level validators verify LOINC (lab codes), ICD-10 (diagnoses), ATC (drugs), CPT (procedures) and NPI (providers) before anything is accepted — so a mismapped or malformed code is caught at map time, not downstream.
LOINCICD-10ATCCPTNPI
4
Emit FHIR R4 or deliver to your warehouse
Confirmed mappings emit FHIR R4 resources (Patient, Observation, Claim) or write straight to your warehouse or webhook. Medium/high-risk templates auto-flag for human review before commit.
FHIR R4Patient / Observation / ClaimHITL flag

Why AdaptivMapr

What makes it fit

FHIR-native

Healthcare templates carry a fhir_resource mapping and emit FHIR R4 (Patient, Observation, Claim) directly — the canonical model is built in, not bolted on.

Medical-code validators

LOINC, ICD-10, ATC, CPT and NPI validators verify codes at map time, so malformed or mismatched values never reach your canonical store.

PHI via BAA

Schema-only mapping ships no records at all. Full-data routing goes through our BAA-covered infrastructure, pinned in-region — PHI never touches an unverified model.

Multilingual hints

Every template carries DE / FR / IT / EN / ES hints, so a German lab export and a French claims file map to the same FHIR fields. Localized pack pages exist per language.

In code

A representative call

Map a lab-results export to the FHIR lab_results template and commit — schema-only, so only headers and a few sample rows leave your system.

curl

# 1. upload (schema-only: headers + <=3 short sample rows)
curl https://api.adaptivmapr.com/v1/uploads \
  -H "Authorization: Bearer $MAPR_KEY" \
  -F file=@lab_results.csv

# 2. map headers to the FHIR healthcare template
curl https://api.adaptivmapr.com/v1/uploads/$ID/mappings \
  -H "Authorization: Bearer $MAPR_KEY" \
  -X PATCH -d '{"template":"lab_results"}'

# 3. commit the confirmed mapping (emits FHIR R4)
curl https://api.adaptivmapr.com/v1/uploads/$ID/commit \
  -H "Authorization: Bearer $MAPR_KEY" -X POST

response

{
  "template": "lab_results",
  "risk": "medium",
  "requires_hitl": true,
  "hitl_status": "pending_review",
  "fields": {
    "test_code":  { "source": "LOINC", "validator": "loinc", "ok": true },
    "value":      { "source": "Result", "ok": true },
    "patient_id": { "source": "MRN", "fhir_resource": "Patient" }
  }
}

FAQ

Common questions

Not in schema-only mode — the default. Only column headers and up to 3 sample rows (each clamped to 80 characters) are sent; the cascade runs on those. Full-data mapping is opt-in and, when enabled, routes through our BAA-covered infrastructure in your region.

The healthcare pack maps to FHIR R4 Patient, Observation and Claim, among others — each template carries a fhir_resource mapping where one applies. You can also deliver the mapped rows to your own warehouse or webhook instead of emitting FHIR.

Field-level validators run at map time: LOINC for lab codes, ICD-10 for diagnoses, ATC for drugs, CPT for procedures and NPI for providers. A value that fails its validator is not silently accepted. (Note: SNOMED is intentionally not validated — it requires an Affiliate License and no template uses it.)

patient_demographics, lab_results, claims_line_items and drug_formulary are medium-risk, so the mappings response flags requires_hitl: true / hitl_status: "pending_review". That lets your workflow gate the commit — the flag is derived from the template risk level, nothing to configure.

FHIR emit Field validators Template packs

Ready when you are

Map healthcare data without shipping raw records.

Test on schema-only for a few tokens per map. Full-data routing under a BAA when you go live.

Get an API key See pricing →

$10 minimum to start · pay only for what you map · PHI under BAA coverage