Unveiling Predictors of Parkinson's Disease Progression: A Machine Learning Approach with Questionnaire and Wearable...

Parkinson's disease is a progressive neurological disorder characterized by motor and non-motor symptoms that significantly impact patients' quality of life. In this longitudinal study, machine learning methods are employed to identify potential prognostic markers associated with the disease's progression. The innovation stems from the integration of diverse data sources such as questionnaire and monitoring device data, and the use of feature engineering to pinpoint statistically significant features that can only be obtained through wearable devices. The analysis revealed that the Unified Parkinson's Disease Rating Scale Part III questionnaire emerged as the most statistically important factor correlated with disease progression, with a ~10% increase in importance score from the second best. Additionally, engineered features were identified as statistically important, with their average importance score measuring ~10% higher than the average score of the rest of the features.Clinical RelevanceThe study demonstrates how combining machine learning with data from patient-reported outcomes and wearable monitoring devices can provide valuable insights into the progression of Parkinson's disease. These findings can inform personalized therapeutic interventions and contribute to more effective monitoring of disease progression, ultimately improving patient outcomes in clinical settings.