Factors contributing to chronic ankle instability in parcel delivery workers based on machine learning techniques

Area of Science:

• Biomechanics
• Machine Learning
• Occupational Health

Background:

• Ankle injuries, particularly sprains, are common in parcel delivery workers (PDWs) and often lead to chronic ankle instability (CAI).
• High recurrence rates of ankle sprains underscore the need for effective identification of PDWs at risk for CAI.

Purpose of the Study:

• To develop and compare the predictive performance of machine learning models for classifying PDWs with and without CAI.
• To identify key variables predictive of CAI in PDWs using postural control, range of motion, muscle strength, and anatomical deformity.

Main Methods:

• 244 PDWs were assessed using 13 predictors including balance tests, range of motion, muscle strength, and anatomical measures.
• Five machine learning algorithms (LASSO logistic regression, Extreme Gradient boosting, support vector machine, Naïve Bayes, random forest) were trained.
• Predictive performance was evaluated on training and test datasets.

Main Results:

• Support vector machine and random forest models demonstrated good predictive performance for classifying CAI in PDWs.
• Key predictive variables identified across models included limited ankle dorsiflexion, reduced lunge angle, higher BMI, older age, increased balance retrials, and lower evertor strength ratio.

Conclusions:

• Machine learning models can effectively classify PDWs with and without CAI, showing good predictive accuracy.
• The study identified significant predictors of CAI, offering potential for targeted interventions and injury prevention in this occupational group.