October 12, 2023
Keio University
Nara Institute of Science and Technology
A research group led by Satoshi Nishioka, a third-year student in the Doctoral Programs at the Graduate School of Pharmaceutical Sciences, Keio University, and Professor Satoko Hori of the Faculty of Pharmacy at the same university, in collaboration with Professor Eiji Aramaki of the Nara Institute of Science and Technology, has developed a method for extracting adverse event signals from text posted by patients on internet blogs (provided by Medi-Aid Co., Ltd.), using deep learning to focus on severe events that disrupt daily life. This method is expected to contribute to improving the monitoring of side effects in cancer patients. The results of this research were published in the online edition of the international academic journal "Scientific Reports" on September 19, 2023.
Highlights of this research
There have been no previous attempts to utilize information from text generated by patients (patient text) for the side effect management of individual patients by focusing on the severity of adverse events.
In this study, we developed a method for extracting adverse event signals based on the degree of disruption to patients' daily lives using three deep learning-based natural language processing models (BERT, ELECTRA, and T5).
By secondarily utilizing the raw voices of patients expressed outside of clinical consultations to automatically extract adverse event signals focused on high-priority events and connecting patients to healthcare professionals, it may be possible to contribute to improving the quality of anticancer drug side effect management.
For the full press release, please see below.