Quantifying Parkinson's disease severity using mobile wearable devices and machine learning: the ParkApp pilot study...

Area of Science:

• Neurology
• Biomedical Engineering
• Digital Health

Background:

• Clinical Parkinson's disease (PD) assessment faces reliability challenges and infrequent monitoring (6-month intervals).
• Frequent symptom variability in PD is difficult to capture with traditional clinical visits.
• Objective, remote monitoring using technology can overcome these limitations.

Purpose of the Study:

• To assess the feasibility, compliance, and user experience of remote Parkinson's disease symptom measurement.
• To investigate the use of smartphone and wearable sensors for passive and active data collection in PD patients' homes.
• To explore the potential of digital health tools for frequent and objective PD symptom assessment.

Main Methods:

• A feasibility study involving PD patients using the Mobistudy smartphone app for weekly activity tests (e.g., TUG, tremor, finger tapping) over 2 months.
• Continuous actigraphy data collection using a GENEActiv wrist device for 28 days.
• Integration of questionnaires (e.g., NMSQuest, PDQ-8, sleep scales) and clinical assessments (MDS-UPDRS) for comprehensive data.
• Application of signal processing and machine learning to analyze sensor data and correlate with clinical outcomes.

Main Results:

• Data collection is ongoing, with participants actively engaged in weekly remote assessments.
• User experience and technology acceptance are being evaluated through dedicated questionnaires.
• Correlation analysis between sensor-derived metrics and clinical PD severity scores (MDS-UPDRS) is planned.

Conclusions:

• Smartphone and wearable technology show promise for frequent, remote, and objective assessment of Parkinson's disease symptoms.
• This approach can potentially improve the management of PD by capturing disease fluctuations.
• Further analysis will determine the long-term feasibility and clinical utility of this digital health strategy.