[Video] Enhancing SDC Workflows with AI: PDF Extraction and Multilingual Support
[Video] Enhancing SDC Workflows with AI – Maria Ryzhikova at
Maria Ryzhikova presented on enhancing SDC workflows using AI at FHIR DevDays 2025.
Summary
Maria Ryzhikova of Aidbox demonstrates how AI can significantly improve FHIR Structured Data Capture (SDC) workflows. She highlights the ability to convert clinical PDFs into structured Questionnaires and achieve multilingual support from a single definition.
Key Players
Details
This article summarizes session content from HL7 FHIR DevDays 2025, detailing the potential of AI in FHIR SDC (Structured Data Capture). Traditionally, designing SDC required significant manual effort for defining complex branching logic and expressions. However, utilizing AI dramatically simplifies this process. A key technical advancement demonstrated is the ability to upload a PDF document (e.g., GAD-7 anxiety screening) and have OCR and an AI model automatically generate a Questionnaire structure proposal, which can then be stored on a FHIR server. This drastically reduces manual form building time and enables structured data capture from the initial development stage. Furthermore, multilingual support is streamlined: instead of duplicating Questionnaires, AI uses translation extensions to generate and maintain item text and answer labels for a single Questionnaire definition, simplifying multi-language version management. Future applications include generating calculated expressions, prefill rules, and extracting data via conversational interfaces. These features are expected to enhance non-technical user accessibility and cut design time dramatically. These capabilities hold high relevance for the digitalization and standardization of diverse clinical processes and questionnaires within Japanese electronic health record systems.
Technology Note
FHIR（Fast Healthcare Interoperability Resources）は医療データ交換の国際標準。このエントリの関連技術: FHIR
Original content copyright by respective publishers