Medical document extraction: from scans to structured JSON.
YIRA MedSense reads medical histories (including handwritten notes), lab reports, discharge and billing documents and hospital tariffs, and returns clean structured data: diagnoses, ICD-10 codes, drug names, lab values with flags, patient details and billing lines, ready for your claims or hospital system.
Medically reviewed by Dr. N. S. Santosh Varma, MBBS, FIIM, FIDM, FICCM · Medical Officer, YIRA MedSense · Updated 12 October 2026
{
"document": "medical_history_handwritten.jpg",
"patient": { "age": "…", "gender": "…" },
"answers": [
{ "q": "Known diabetes?", "a": "…" },
{ "q": "Current medication", "a": "…" }
],
"codes": { "icd10": ["…"], "drugs": ["…"] },
"status": "ready for review"
}Illustrative structure. Field values depend on the document.
Every document type in a claim file.
| Document | What MedSense extracts |
|---|---|
| Medical histories (typed or handwritten) | Structured question-and-answer fields, past conditions, current medication |
| Lab reports | Test names, values, units and reference ranges, with values flagged High or Low against the lab's own range |
| Discharge summaries and doctor notes | Diagnoses, ICD-10 codes, procedures, drug names (including NLEM names) |
| Hospital bills | Billing line items, quantities and amounts |
| Hospital tariffs | Procedure-level rate tables by room category; missing rates kept as NA |
| Patient details | Demographics needed for matching and identity checks |
Four stages from upload to integration.
Ingest
API, mobile app or bulk scanner portal: scanned PDFs, digital PDFs, handwritten notes and mobile photos.
Understand
Medical-grade OCR and clinical NLP identify entities: ICD-10 codes, drug names, demographics and billing lines.
Validate
Review checks and standardisation so every hospital's and lab's format comes out the same.
Integrate
Structured JSON pushed to TPA portals, insurance systems, hospital EMRs and analytics databases.
Generic OCR reads text. MedSense reads medicine.
| Need | Generic OCR / IDP | YIRA MedSense |
|---|---|---|
| Handwritten doctor notes | Text fragments at best | Structured question-and-answer fields |
| Lab values | Numbers without meaning | Value, unit, range and High/Low flag |
| Diagnoses and drugs | Free text | ICD-10 codes and drug names identified |
| Hospital tariffs | Tables as text | Procedure-level rates with contract rules and visible gaps |
| Claims use | Needs custom work | Built for insurers, TPAs and hospitals, with identity checks and prior-record matching |
Built for claims and hospital teams.
Insurers and TPAs
Faster pre-auth and claim review from structured files.
Hospitals
Claim and pre-auth documents ready for payers.
Medical officers
Clean clinical data instead of scans.
Software vendors
Embed MedSense through partner API or white label.
In production at Health Assist Insurance TPA.
MedSense has parsed more than 1 lakh (100,000+) medical documents, flagged over 5,000 fraud cases for review and supported 100+ insurance operations.
Sandbox
Test MedSense on your own sample documents.
Provision
Set up document types, fields and rules.
Integrate
Connect by API to your claims, TPA or hospital system.
Go live
Run in parallel, then switch over.
Timing depends on scope, data access and integration readiness. Client-managed deployment is available, so MedSense can run inside your own infrastructure.
Questions from claims teams
What is medical document extraction?
Turning unstructured medical documents, such as scanned reports, handwritten notes, lab reports and bills, into structured data fields that software can use. YIRA MedSense does this for insurers, TPAs and hospitals, with clinical understanding of diagnoses, drugs and lab values.
Can MedSense extract data from handwritten medical records?
Yes. Handwritten medical histories are converted into structured question-and-answer fields that teams can review and search.
What output format does MedSense return?
Structured JSON, delivered by API to your claims system, TPA portal, hospital EMR or analytics database.
Does MedSense extract ICD-10 codes?
Yes. Clinical NLP identifies diagnoses and ICD-10 codes, along with drug names, patient demographics and billing line items.
How is MedSense different from generic OCR?
Generic OCR returns text. MedSense understands medical content: it flags lab values against reference ranges, identifies ICD-10 codes and drug names, structures handwritten histories and hospital tariffs, and adds identity checks and prior-record matching for claims.
Can software companies embed MedSense?
Yes. Partner API licensing and white label are available for HIMS, claims and health-tech software vendors.
Where is the data processed?
MedSense supports client-managed deployment, so it can run inside your own infrastructure. Hosting and data residency are agreed per contract.
Book your demo today.
See MedSense on the documents and fields that matter to your claims team.