Rethinking information delivery: using a natural language processing application for point-of-care data discovery

Area of Science:

• Medical Informatics
• Natural Language Processing
• Clinical Decision Support

Background:

• Clinicians require efficient access to relevant medical information for effective patient care.
• Existing methods for retrieving information from biomedical literature can be time-consuming.
• Decision support tools aim to integrate evidence-based information into clinical workflows.

Purpose of the Study:

• To evaluate Semantic MEDLINE, a natural language processing application, as a clinical decision support tool.
• To assess the efficacy of the Combo algorithm in enhancing information retrieval from PubMed citations.
• To determine the system's capability in delivering tailored information based on user needs.

Main Methods:

• Selected three disease topics and formulated PubMed search queries for prevention literature.
• Processed retrieved PubMed citations using Semantic MEDLINE with the Combo algorithm.
• Established a reference standard using a commercial decision support tool for comparison.

Main Results:

• Semantic MEDLINE with Combo achieved an average recall of 79%.
• The system demonstrated an average precision of 45%.
• The overall average F-score for the system was 0.57.

Conclusions:

• Semantic MEDLINE, augmented by the Combo algorithm, presents a viable approach for point-of-care information delivery.
• This technology has the potential to serve as an effective decision support tool for clinicians.
• Health sciences libraries can leverage such tools to provide customized information resources to their users.