Please login to view this media
- Talk
- 25/09/2023
- UK
A Precision Health Approach For Knee Osteoarthritis Prediction of Rapid Progression Using Automated Machine Learning
Description
In this presentation, Simone Castagno discusses the development of a precision health tool aimed at predicting rapid progression of knee osteoarthritis (OA). The field of precision health seeks to enhance disease prevention and management at both individual and population levels, with a focus on using machine learning to predict disease trajectories. Castagno begins by outlining the challenges posed by osteoarthritis, a complex condition currently lacking effective preventative or curative treatment options.
Through machine learning, the project aims to identify patients at risk for rapid OA progression, leveraging a public dataset that includes clinical, radiographic, and biomarker data from 600 new OA patients over a four-year period. The analysis involved categorizing outcomes into specific classes based on pain and radiographic findings and employing automated algorithms for model development.
The results highlight the tool's high accuracy rates and demonstrate the predictive value of certain clinical assessments, such as the WOMAC pain score. Castagno emphasizes the identification of crucial predictors of OA progression and notes the creation of a robust prototype application. This prototype allows for quick predictions and highlights the relative importance of various features per individual patient, aiming to bridge the treatment gap for younger patients who are often overlooked in standard management practices.