Read more

June 07, 2021
1 min read
Save

Machine learning algorithm may predict improvements in athletes after hip arthroscopy

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Use of a machine learning algorithm may predict clinically relevant sports-specific improvements in athletes with femoroacetabular impingement undergoing hip arthroscopy, according to published results.

Shane J. Nho, MD, MS, and colleagues identified 1,118 patients who participated in a formal sports program or athletic activity prior to primary hip arthroscopy for femoroacetabular impingement syndrome between February 2018 and January 2021. Researchers used recursive feature selection to identify a combination of variables that optimized model performance from an initial pool of 26 features and used 10-fold cross-validation with three iterations of training to develop six machine learning algorithms, which were applied to an independent testing set of patients. Researchers evaluated the models using discrimination, decision-curve analysis, calibration and the Brier score. Researchers considered achieving the minimal clinically important difference in the hip outcome score-sports subscale (HOS-SS) at a minimum of 2 years postoperatively as the primary outcome measure.

Results showed 76.9% of athletes achieved the minimal clinically important difference for the HOS-SS. Researchers found a preoperative HOS-SS of 58.3 or greater, Tönnis grade of 1, alpha angle of 67.1° or greater, BMI of greater than 26.6 kg/m2, Tönnis angle of greater than 9.7° and age older than 40 years optimized algorithm performance and decreased the likelihood of achieving minimal clinically important difference. Of the machine learning algorithms used, researchers noted the elastic-net penalized logistic regression model demonstrated the best performance and was transformed into an online application as an educational tool to demonstrate machine learning capabilities.

Shane J. Nho
Shane J. Nho

“With machine learning algorithms, we are now able to talk to a patient in the office and tell them what will be their anticipated surgical outcome. Using the machine learning algorithm, we can enter in a patients age, weight and level of activity, and the computer program will tell the patient what they can expect from surgery,” Nho told Healio Orthopedics. “Rather than relying on a surgeon to describe in general terms how a patient might perform after surgery, it is possible to individualize the treatment to the patient in real time while talking to them in the exam room. Patients will be equipped to make shared decisions about their own health care. The future is now.”