Please login to view this media
- Talk
- 15/09/2021
- Canada
Can Wearable Sensors and Machine Learning Predict Functional Recovery Following Total Knee Arthroplasty?
Description
In this presentation led by Riley Bloomfield, the speaker addresses the critical issue of patient dissatisfaction following total knee arthroplasty (TKA) despite advances in surgical techniques and implant designs. With knee osteoarthritis becoming more common among active individuals, TKA remains the primary treatment for severe cases, though up to 20% of patients express disappointment post-surgery. The presentation outlines the importance of managing patient expectations, particularly relating to functional recovery. Citing research by Bourne et al, Bloomfield emphasizes that dissatisfaction is often tied to unmet preoperative expectations, especially regarding functionality.
To investigate this further, Bloomfield's research employs wearable sensors to collect data on functional movements from 82 TKA patients during preoperative and three-month postoperative assessments. Using this data, the study aims to predict which patients are likely to see significant functional improvement, thus providing a basis for setting realistic expectations.
The methodology includes machine learning models—specifically support vector machine, naive Bayes, and random forest classifiers—to analyze patient recovery patterns based on their timed performance in a specific mobility test. The findings indicate that the random forest model outperformed others, demonstrating high accuracy in predicting functional recovery, while highlighting the pitfalls of the naive Bayes classifier in accurately identifying recovery responders.
The presentation concludes with a call for further analysis involving larger patient populations to enhance predictive reliability and considers the implications for patient communication in clinical settings. This foundational work lays the groundwork for ongoing research aimed at improving patient satisfaction and outcomes in total knee replacement surgeries.