STRP

To begin feeding the pipeline of projects moving through the research translation process, TRAC started incubating this project in 2024. Professor Bei Xiao’s project couples machine learning for natural language processing and human predictions to determine the success of new treatments in clinical drug trials based on trial descriptions. This work will translate NSF-funded research by Dr. Xiao and others (awards #2050727 and #2030015) that developed advanced algorithms for investigating factors that are linked with failures in clinical testing of new drugs. In this work, the researchers found that the success of clinical testing can be predicted with an average accuracy of 80% using their models (Feijoo et al. 2020). The researchers also found common protocol characteristics across therapeutic areas that are linked to clinical trial success, and identified key features of stakeholder collaboration that influence the success or failure of a trial (Lin et al. 2021).