Improving Patient Access and Comprehension of Clinical
Notes: Leveraging Large Language Models to Enhance
Readability and Understanding
Name
mannhardt-niklasm-meng-eecs-2023-thesis.pdf
Description
Thesis PDF
Size
1.49 MB
Format
Adobe PDF
Checksum (MD5)
8a8bf3c0b5f190930448672af270b1ae
Author(s)
Mannhardt, Niklas
Advisor(s)
Sontag, David A.
Date Issued
September 2023
Publisher
Massachusetts Institute of Technology
Abstract
Patient access to clinical notes has demonstrated numerous benefits, including an increased sense of control over their condition, enhanced engagement, improved medication adherence, and greater clinician accountability. However, the presence of medical jargon, abbreviations, and complex medical concepts within clinical notes hinders patient comprehension, thus diminishing the positive effects of note accessibility. These notes, primarily intended for clinicians, often appear disorganized and contain an abundance of technical terms. Breast cancer patients, in particular, face information overload and experience taxing symptoms related to their treatment, exacerbating this issue. Although some clinicians are adapting their writing style to meet patients’ needs, time constraints limit the feasibility of comprehensive note-taking. We propose the development of a patient-facing tool, in the form of a web application, to make information contained in clinical notes more accessible by leveraging machine learning models to simplify, summarize, extract information from, and add context to clinical notes. Through a series of user studies, we demonstrate that our proposed augmentations to clinical notes significantly improve comprehension and enhance patients’ reading experience.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
In Copyright - Educational Use Permitted
Copyright retained by author(s)
Persistent DSpace Link
