Predicting future fallers in Parkinson's disease using kinematic data over a period of 5 years

Abstract

Parkinson's disease (PD) increases fall risk, leading to injuries and reduced quality of life. Accurate fall risk assessment is crucial for effective care planning. Traditional assessments are subjective and time-consuming, while recent assessment methods based on wearable sensors have been limited to 1-year follow-ups. This study investigated whether a short sensor-based assessment could predict falls over up to 5 years. Data from 104 people with PD without prior falls were collected using six wearable sensors during a 2-min walk and a 30-s postural sway task. Five machine learning classifiers analysed the data. The Random Forest classifier performed best, achieving 78% accuracy (AUC = 0.85) at 60 months. Most models showed excellent performance at 24 months (AUC > 0.90, accuracy 84-92%). Walking and postural variability measures were key predictors. Adding clinicodemographic data, particularly age, improved model performance. Wearable sensors combined with machine learning can effectively predict fall risk, enhancing PD management and prevention strategies.

Fig. 1. Receiver operating characteristic (ROC) curves.

Five classifiers were evaluated in the study: Logistic Regression (log_reg), Random Forest (rf), Support Vector Machine (svm), Elastic Net and XGBoost at each of the a 24, and b 60 months. The performances of the best models are illustrated and trained on the selected set of features. AUC area under the curve.

Fig. 2. Comparison of clinically assessed risk factors of falls in the studied population.

a EQ-5D-5L (EuroQol—5 Dimensions—5 Levels) index, b Movement disorders society-Unified Parkinson’s disease rating scale-part 3 (MDS-UPDRS-III), c PDQ39mobility which is the sum of items 4–10 of PDQ39 questionnaire (self-rated walking, fear of falling, confidence in getting around), d number of participants in both groups, per Hoehn and Yahr (H&Y) stage, e disease duration and f age.

Fig. 3. Follow-up flowchart.

Data were analysed at two time points: 24 and 60 months. Initially, 109 participants with idiopathic PD were recruited. Participants were lost to follow up leaving 98 participants for the 24-month time period (13 fallers and 85 non-fallers) and 97 participants for the 60 months (23 fallers and 74 non-fallers).

Fig. 4. Placement of Inertial Measurement Units (IMUs).

Six IMUs were placed on the participants wrists, feet, sternum, and lumbar region.