Real-world performance evaluation of a telemonitoring system for Parkinson’s disease symptom assessment

Abstract

Background

Parkinson’s disease (PD) nurses constitute a reliable modality for appraising the motor status of patients, second only to expert neurologists. Consequently, employing PD nurses for the assessment of medical device performance is appreciated to be a fundamental benchmark in the field.

Objective

To evaluate the precision of a telemonitoring device in measuring motor symptoms when utilized at patient’s home during everyday life conditions.

Methods

To ascertain the extent of correlation between device-reported outcomes and PD nurse assessments and the precision with which the device detects symptoms such as gait impairment, rest tremor, bradykinesia, and the presence of OFF, a total of 34 evaluations were conducted by a PD nurse at the patients’ home, each lasting 3 h.

Results

A robust correlation was noted for all three symptoms when analyzed at both 30-minute intervals and cumulatively over the entire 3-hour duration with Pearson’s r range 0.68–0.76 and 0.83–0.90 respectively. Furthermore, substantial agreement was observed for the identification of the presence of all three symptoms with Cohen’s kappa range from 0.61 to 0.71 across all 30-minute evaluations. Finally, the device demonstrates high accuracy in detecting rest tremor (90%), bradykinesia (89%), gait impairment (95%) and OFF state (94%).

Conclusion

This study evaluates for the first time the performance of a remote monitoring device, unobtrusively, at patients familiar environment, with a PD nurse present in non-scripted scenarios. The device exhibited commendable efficacy in the accurate detection of PD symptoms and state.