The McMaster Health Information Research Unit: Over a Quarter-Century of Health Informatics Supporting Evidence-Based...

Area of Science:

• Health Informatics
• Evidence-Based Medicine
• Artificial Intelligence in Healthcare

Background:

• Evidence-based medicine (EBM) principles were established at McMaster University, emphasizing research evidence, clinical expertise, and patient values.
• The Health Information Research Unit (HiRU) was founded in 1985 to advance EBM through health informatics.
• Early digital health informatics focused on teaching clinicians database search skills, with PubMed becoming a key resource.

Purpose of the Study:

• To explore the 25+ year evolution of health informatics at HiRU in supporting evidence search and retrieval.
• To highlight the increasing role of machine learning and large language models (LLMs) in managing vast biomedical literature.
• To discuss the integration of responsible artificial intelligence in clinical knowledge dissemination.

Main Methods:

• Review of the historical development of search strategies and tools at HiRU.
• Application of classical machine learning, deep learning, and LLMs for filtering biomedical literature.
• Utilizing gold-standard annotated datasets and human-in-the-loop active machine learning.

Main Results:

• The development of validated search filters (Clinical Queries) enhanced search precision for clinicians.
• Significant advancements in digital health informatics have transformed access to clinical studies, systematic reviews, and guidelines.
• Current research leverages sophisticated AI, including LLMs, to navigate the rapidly growing volume of PubMed publications.

Conclusions:

• Health informatics has continuously adapted to improve the accessibility and relevance of medical research.
• LLMs and responsible AI represent the next frontier in facilitating the integration of best evidence into clinical practice.
• The ongoing work at HiRU aims to empower clinicians with efficient tools for evidence retrieval and application.