Automated medical literature retrieval

Area of Science:

• Medical Informatics
• Biomedical Research

Background:

• Increasing publication rates challenge medical specialists' awareness of current research.
• Existing literature retrieval systems are often repetitive and time-consuming.

Purpose of the Study:

• To describe a system that automatically generates literature search queries from electronic patient records.
• To enable clinicians to focus on patient care supported by current research.

Main Methods:

• A prototype system generates weighted query terms from patient data (age, blood tests).
• Utilizes fuzzy logic for age-based weighting and conditional homonyms for query expansion.
• Retrieves and ranks PubMed publications based on patient relevance, highlighting recent articles.

Main Results:

• Nine medical specialists evaluated the system, providing positive feedback.
• Participants found the prototype supportive and desired its integration into daily practice.

Conclusions:

• Automated query formulation and patient-specific relevance ranking integrate literature retrieval into clinical workflows.
• The system supports busy professionals by reducing time spent on literature searching and query modification.