Abstract
BACKGROUND: Early diagnosis is crucial for ensuring that patients with Parkinson disease (PD) receive timely treatment, which can improve their quality of life and prolong lifespan. Wearable sensors have emerged as promising tools for early PD diagnosis, offering noninvasive, continuous symptom monitoring. OBJECTIVE: This review aimed to evaluate how wearable sensors have been applied in early diagnosis of PD over the past decade, focusing on sensor types, methods, findings, and limitations. METHODS: The systematic review was conducted following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Studies were sourced from PubMed, IEEE Xplore, Scopus, and Web of Science and screened based on predefined criteria. The inclusion criteria were as follows: (1) the study was observational or experimental, (2) wearable sensors were applied for the early diagnosis of PD, (3) participants were diagnosed with early-stage or prodromal PD, (4) the study included at least 10 participants with PD, and (5) the article was published between 2013 and 2023. Studies were excluded if they focused solely on treatment, rehabilitation, symptom monitoring, or nonwearable devices; lacked diagnostic clarity; were not published in English; or were not primary research articles. All the selected studies were assessed for quality using the Quality Assessment of Diagnostic Accuracy Studies-2, the quality assessment tool recommended by the Cochrane Collaboration. RESULTS: Overall, 1888 records were retrieved from the selected databases, with 1044 records remaining after duplicate removal. Following the screening of titles and abstracts, 949 ineligible records were excluded, leaving 95 articles for eligibility. Eventually, of the 1044 studies, 12 (1.12%) met the inclusion criteria, validating the feasibility of wearable sensors in the early diagnosis of PD. Most (10/12, 83%) were cross-sectional studies, with 1 longitudinal and 1 mixed-design study. Of the 12 studies, 4 (33%) focused on identification diagnosis, 2 (17%) addressed the staged diagnosis of PD, and 1 (8%) focused on the identification of specific symptoms. Of the 12 studies, 5 (42%) assessed the overall feasibility and performance of wearable sensors in early PD detection without targeting specific classification purposes. The main wearable sensors used were inertial measurement units (8/12, 67%) and accelerometers (4/12, 33%), which primarily captured motion-related data. While initial findings suggest that wearable sensors are feasible for early PD diagnosis, the evidence is still limited by small sample sizes and short study durations. CONCLUSIONS: Wearable sensors show promise in supporting the early diagnosis of PD, particularly for motor symptoms monitoring. However, several limitations remain in validating and applying wearable sensors in clinical contexts, including cross-sectional designs and limited diagnostic standardization. More diverse studies are needed to further validate these findings and address existing shortcomings to better advance the use of wearable sensors in the early diagnosis of PD. TRIAL REGISTRATION: PROSPERO CRD42024544198; https://www.crd.york.ac.uk/PROSPERO/view/CRD42024544198.