Assessing Translation Science in Exercise Using Text Analytics

Author: ORCID icon
Morrow, Kristin, Systems Engineering - School of Engineering and Applied Science, University of Virginia
Brown, Don, DS-Research, University of Virginia

A critical challenge facing biomedical investigators and frontline clinicians is the significant delay between clinical research and its adoption in real-world medical settings. Through an investigation of the gap between clinical research related to exercise and its appearance in electronic health records (EHRs), this study aims to connect advances in research to EHRs and back to medical decision-making. Current medical knowledge states that cardiorespiratory fitness (CRF) is an essential component of healthy lifestyles and can play a beneficial role in a variety of specific disease states and conditions. Consequently, care providers should recommend regular exercise whenever appropriate in patient interactions. However, our findings at the University of California, Irvine Medical Center show that the discussion of established guidelines regarding CRF recommendations to patients rarely occurs in the context of EHR in the emergency medicine department. For coronary artery disease, the most common type of heart disease, only 0.34% of EHRs mention exercise-related treatment plans. Nonetheless, emergency department visits could provide a valuable opportunity for physicians to influence patient well-being through health messaging. While the time between patients and physicians in this setting will not change, direct querying from PubMed, patient charts, and social determinants can be leveraged to find important information and deliver targeted messages. In this way, the emergency department can serve as a powerful tool for physicians to effect positive change in patients' lives, despite the challenges associated with emergency care. This study presents a robust, self-supervised framework that quantifies the relationship between clinical research and EHRs through semantic similarity analysis of mentions within EHRs, thereby enhancing our understanding of the connection between the two.

MS (Master of Science)
Natural Language Processing (NLP), Natural Language Understanding (NLU), Medical Decision-Making, Bioinformatics, Electronic Health Records, Machine Learning, Data Analytics, Sentence Transformers, Topic Modeling , Similarity Search, Semantic Search, Summarization, Word Embeddings, Translation Science
All rights reserved (no additional license for public reuse)
Issued Date: