Objective:

Evaluate mobile application and wearable device for detecting clinically significant changes in Parkinson’s motor symptoms.

Background:

Monitoring motor and non-motor symptoms in Parkinson’s disease (PD) poses a significant challenge for clinicians and clinical trials, with wearable technologies providing a promising objective assessment tool. This study assesses algorithms for detecting changes in motor symptoms using a wearable device and mobile app, with emphasis on discerning clinically significant changes from random variation. In contrast to commercially available Parkinson’s disease (PD) monitoring solutions, which focus on changes within brief trial periods, our study evaluated algorithms capable of analyzing data across any testing duration.

Method:

The study analyzed data from 600 patients who had recorded at least seven days of data with over six hours of data per day. The analysis included several data streams, such as tremor and dyskinesia measurements from the wearable device. In addition the mobile application captured walking speed, step length, double support time and walking asymmetry. Patients who exhibited detected changes were contacted to determine the degree of agreement between their personal experience and the algorithmic results.

Results:

Our analytical approach evaluated several statistical algorithms, such as cumulative sum and robust statistical detectors, that assess changes in mean and variance or deviations and are less sensitive to outliers. We also employed a Bayesian online change point detection method, suitable for real-time applications which incorporates prior knowledge into analysis. We successfully detected statistically significant changes in objectively monitored motor symptoms, which were validated by patients.

Conclusion:

This investigation aims to evaluate the feasibility of using statistical algorithms to detect clinically meaningful changes in long-term, wearable-monitored Parkinson’s symptoms.  By creating dependable and precise algorithms capable of robustly detecting deviations from a specific clinical state, this strategy may facilitate screening of large cohorts of patients in clinical trials. Robust detection and monitoring of treatment response over time in a well-defined population has the potential to inform disease progression, aid in treatment decision-making, and facilitate the development of more efficacious therapies for Parkinson’s patients.

« Back to 2023 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/detecting-changes-in-parkinsons-motor-symptom-patterns-using-wearable-devices/