Extract medical documents and map standard codes

Goal

Turn unstructured clinical documents into reviewable coded records rather than a free-text summary.

Steps

1. Submit an authorized clinical PDF/image, OCR its text, and retain source pages. [src10-strands-normalize-medical-documents-to-standard-codes]
2. Extract diagnoses, medications, treatments, patient details, and vitals into structured JSON fields. [src10-strands-normalize-medical-documents-to-standard-codes]
3. Map ICD-10, RxNorm, and SNOMED CT separately; flag low-confidence or unmapped items for human review. [src10-strands-normalize-medical-documents-to-standard-codes]

How to verify

Check each code against its source page and term; unknown terms should remain unmapped rather than receive invented codes.

Troubleshooting

For OCR errors in dose or units, inspect the source page and flag review; for code-version mismatches, update the terminology source and retry that field.