Awardee Organization(s): University of Texas Health Science Center at Houston
Principal Investigator(s): Min Ji Kwak, MD, MS, DrPH
Official Project Title: Detection of adverse drug event using NLP among older adults with heart failure
AITC Partner: PennAITech
Website(s): https://med.uth.edu/internalmedicine/2022/11/17/min-ji-kwak-md-ms-drph/
Adverse drug events (ADEs) in older adults with heart failure are a serious public health concern. These drug-related complications can be life-threatening and significantly reduce quality of life. To capture ADEs correctly, doctors need to carefully review a patient’s symptoms, medical history, prescription changes, and past records. However, this can be difficult to do thoroughly during a busy clinic visit.
An automatic tool using an Artificial Intelligence tool can help by scanning a patient’s past medical records for signs of an ADE. This tool can be built into electronic health records to provide real-time assessments. While current AI models typically handle only one task at a time—like identifying medications—ADE detection requires a more advanced system that can process multiple tasks and make complex decisions.
To address this, we are developing a specialized AI framework called AIDE4HF. This system will leverage existing tools to detect ADEs in older adults taking heart failure medications. Our project has two main goals:
1.Create a high-quality dataset of ADEs related to heart failure medications in older adults.
2.Develop and test a powerful AI system that can accurately detect these ADEs.
This research is a collaboration between UTHealth McGovern Medical School and UTHealth McWilliams School of Biomedical Informatics. By combining expertise from multiple fields, we aim to create a model that closely mimics real clinical decision-making and has the potential to be widely used in medical practice.