Objective: To evaluate how accurately a new device can discriminate different clinical severities of dyskinesia from non-dyskinetic movements in patients with Parkinson’s disease (PD).

Background: PD dyskinesia is a leading cause of falls and unplanned hospital admissions and leads to reduced quality of life (QOL). It may occur unpredictably and frequently throughout the course of a day, making it difficult for patients to report their symptoms in detail. Furthermore, not all patients are aware of their own dyskinesia. New methods for objectively monitoring dyskinesia over 24 hours at home would enable clinicians and patients to make informed decisions on drug management.

Methods: 23 PD patients wore small electromagnetic movement sensors on their limbs, head and trunk so their movement data could be continuously recorded onto a mobile phone. They were video-recorded and clinically assessed every hour using the UPDRS and UDysRS. The first 6 patients (TRAIN) had 7 assessments and the next 17 patients (TEST) had 3 assessments. The TRAIN movement sensor data and clinical ratings were used to develop a computer program called as ‘classifier’ that discriminates different severities of dyskinesia. The classifier was developed using purpose written computer evolutionary algorithms. The accuracy of the classifier was then evaluated on the previously unseen TEST movement sensor data.

Results: Table 1 shows that the patients in the TRAIN and TEST data sets were broadly similar, although the TRAIN patients were slightly older and their motor and dyskinesia scores more severe. The classifier trained on the first data set generalised well when tested on the second data set, achieving useful levels of sensitivity/specificity when distinguishing samples with UDysRS levels 3 (0.85/0.80) and 4 (0.93/0.90) from samples with no dyskinesia.

Conclusions: This technology shows promise for development into a useful home-monitoring device that can objectively measure dyskinesia. It has the potential to enable better management of dyskinesia and hence improve QOL, reduce unplanned hospital admissions and reduce medical costs.

The TRAIN data was previously presented at Parkinson’s UK research conference in York in 2014.

« Back to 2016 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/developing-a-new-home-monitoring-device-for-dyskinesia-in-parkinsons-disease/