Feasibility of a wearable inertial sensor to assess motor complications and treatment in Parkinson’s disease

Abstract

Background Wearable sensors-based systems have emerged as a potential tool to continuously monitor Parkinson’s Disease (PD) motor features in free-living environments. Objectives To analyse the responsivity of wearable inertial sensor (WIS) measures (On/Off-Time, dyskinesia, freezing of gait (FoG) and gait parameters) after treatment adjustments. We also aim to study the ability of the sensor in the detection of MF, dyskinesia, FoG and the percentage of Off-Time, under ambulatory conditions of use. Methods We conducted an observational, open-label study. PD patients wore a validated WIS (STAT-ONTM) for one week (before treatment), and one week, three months after therapeutic changes. The patients were analyzed into two groups according to whether treatment changes had been indicated or not. Results Thirty-nine PD patients were included in the study (PD duration 8 ± 3.5 years). Treatment changes were made in 29 patients (85%). When comparing the two groups (treatment intervention vs no intervention), the WIS detected significant changes in the mean percentage of Off-Time (p = 0.007), the mean percentage of On-Time (p = 0.002), the number of steps (p = 0.008) and the gait fluidity (p = 0.004). The mean percentage of Off-Time among the patients who decreased their Off-Time (79% of patients) was -7.54 ± 5.26. The mean percentage of On-Time among the patients that increased their On-Time (59% of patients) was 8.9 ± 6.46. The Spearman correlation between the mean fluidity of the stride and the UPDRS-III- Factor I was 0.6 (p = <0.001). The system detected motor fluctuations (MF) in thirty-seven patients (95%), whilst dyskinesia and FoG were detected in fifteen (41%), and nine PD patients (23%), respectively. However, the kappa agreement analysis between the UPDRS-IV/clinical interview and the sensor was 0.089 for MF, 0.318 for dyskinesia and 0.481 for FoG. Conclusions It’s feasible to use this sensor for monitoring PD treatment under ambulatory conditions. This system could serve as a complementary tool to assess PD motor complications and treatment adjustments, although more studies are required.

Holter STAT-ON™ against other tools for detecting MF in advanced Parkinson’s disease: an observational study

“…The Holter registers were found to be more substantial than diary monitors, which supported the use of the STAT-ON™ to guide therapeutic decisions in clinical practice, especially in APD patients needing SLT. This Holter device has also been validated previously in further studies, demonstrating a better detection of ON/OFF MF, dyskinesia, and falls against patients' diaries (23), as well as supporting its use as a valuable complementary tool to assess PD motor complications and the need for treatment adjustments (24). Furthermore, the ongoing multicenter, randomized clinical trial MoMoPa-EC will investigate this issue by establishing three study arms (STAT-ON™ data vs. Hauser diaries vs. clinical data only collected during the visit) (25,26).…”

Wearable sensor devices can automatically identify the ON-OFF status of patients with Parkinson's disease through an interpretable machine learning model

“…The WSD refers to portable and mobile devices worn on the body or embedded in clothing, such as smart glasses, watches, clothes, and pressure shoes, which contain hardware and software technologies and have special functions for collecting spatiotemporal kinematic parameters, data processing, transmission, and storage. At present, the WSD used in the field of PD can realize quantitative evaluation of PD through specific motor tasks and establish data models so that doctors can accurately analyze the movement status of patients ( 5 , 14 , 34 ). Compared with the results of MDS-UPDRS-Part III evaluated by doctors, WSDs were reported as a method with higher objectivity, accuracy, and sensitivity for evaluating the status of patients' motor abilities, such as bradykinesia, dyskinesia, tremor, and freezing of gait ( 5 , 15 ).…”

“…In recent years, the use of different sensor devices for the quantitative evaluation of motor ability in patients with PD has been increasingly explored. Many studies supported the reliability of motor data collected by wearable sensor devices (WSD) to assist in the diagnosis of patients with PD and the assessment of disease progression and to guide clinical practice (9)(10)(11)(12)(13)(14). Some studies have combined machine learning (ML) or deep learning (DL) algorithms with WSD to estimate MDS-UPDRS-Part III and assist in the diagnosis of PD (15-17).…”

Need for personalized monitoring of Parkinson’s disease: the perspectives of patients and specialized healthcare providers

“…These episodic assessments do not always provide a representative and complete picture of the patient’s actual functioning in daily life, for example due to recall bias ( 4 ) and observer effects ( 5 ). Remote monitoring tools such as wearable sensors may partly fill this gap and provide opportunities for personalized care, telemedicine and improved self-management ( 6 , 7 ).…”