Accepted for/Published in: Journal of Medical Internet Research
Date Submitted: Sep 10, 2023
Date Accepted: Dec 19, 2023
(closed for review but you can still tweet)
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
The Performance of Wearable Artificial Intelligence in Detecting Stress Among Students: Systematic Review and Meta-Analysis
ABSTRACT
Background:
Students usually encounter stress throughout their academic path. Ongoing stressors may lead to chronic stress, adversely affecting their physical and mental well-being. Thus, early detection and monitoring of stress among students is crucial. Wearable artificial intelligence (AI) has emerged as a valuable method for this purpose. It offers an objective, non-invasive, non-obtrusive, automated approach to continuously monitor biomarkers in real time, thereby, addressing the limitations of traditional approaches such as self-reported questionnaires.
Objective:
This systematic review and meta-analysis is aimed at assessing the performance of wearable AI in detecting and predicting stress among students.
Methods:
Search sources in this review included searching 7 electronic databases, checking the reference list in the included studies, and checking studies cited that included studies. This review included research articles centered around the creation or application of AI algorithms for the detection or prediction of stress among students using data from wearable devices. Two independent reviewers performed study selection, data extraction, and risk of bias assessment. Quality Assessment of Studies of Diagnostic Accuracy-Revised (QUADAS-2) tool was adapted and used to examine the risk of bias in the included studies. Evidence synthesis was performed using narrative and statistical techniques.
Results:
This review included 19 studies out of the 327 studies retrieved from the search sources. A meta-analysis of 37 accuracy estimates derived from 6 studies revealed a pooled mean accuracy of 0.856 (95% confidence interval (CI) 0.70 to 0.93). Subgroup analyses demonstrated that the accuracy of wearable AI was moderated by the number of classes (P=0.022), type of wearable device (P=0.049), location of wearable devices (P=0.022), dataset size (P=0.009), and ground truth (P=0.001). The average of estimates of sensitivity, specificity, and F1 score was 0.755 (Standard Deviation (SD) 0.181), 0.744 (SD 0.147), and 0.759 (SD 0.139), respectively.
Conclusions:
Wearable AI shows promise in detecting student stress but currently has suboptimal performance. It should be used alongside other assessments (e.g., clinical questionnaires) until further evidence is available. Future research should explore the ability of wearable AI to differentiate types of stress, distinguish stress from other mental health issues, predict future occurrences of stress, consider factors like the placement of the wearable device and the methods used to assess the ground truth, and report detailed results to facilitate the conduction of meta-analyses.
Citation
