Generating Form from PDF

AI-powered extraction of structured FHIR SDC Questionnaires from scanned medical forms

Overview

This API enables the automatic digitization of paper-based medical forms by converting scanned PDFs into structured FHIR SDC (Structured Data Capture) Questionnaire resources. Using either OpenAI or Gemini for intelligent inference, the system extracts relevant form fields from OCR-processed documents and maps them to a standardized FHIR structure.

Prerequisites

To use this API, you must set one of the following environment variables. These keys determine which AI model is used for processing and extracting structured data from scanned forms:

Environment Variables

• BOX_SDC_OPENAI_API_KEY – Uses GPT-4o Mini for inference.

• BOX_SDC_GEMINI_API_KEY – Uses Gemini Pro 2 for inference.

Endpoint

Request

POST /$ai-generate-questionnaire HTTP/1.1
Content-Type: multipart/form-data; boundary=------------------------boundary12345

--------------------------boundary12345
Content-Disposition: form-data; name="file"; filename="example.pdf"
Content-Type: application/pdf

<file-content>
--------------------------boundary12345--

Where <file-content> is the content of the PDF file being uploaded.

Response

{
  "resourceType": "Questionnaire",
  "title": "<questionnaire-title>",
  "item": [
    ...
  ]
}

The response body contains a FHIR Questionnaire resource with the extracted form fields. The title field contains the title of the questionnaire, and the item field contains an array of form items.

Error Responses

If the request is invalid or the file cannot be processed, the API may respond with an appropriate HTTP error code and a standard Operation Outcome resource.