Wearable sensing technology for knee joint monitoring and functional assessment: a systematic review.
systematic_review · Level I
Where this comes from
- Record sourced from PubMed, PMID 41915525.
- Also identified by DOI 10.1109/JBHI.2026.3679290.
- No licence information is recorded for this record.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
Traditional knee joint monitoring and assessment methods rely heavily on subjective evaluations, leading to an increasing interest in wearable sensing technology for objectively monitoring knee joint function. This review evaluates current research on wearable devices for knee monitoring and functional assessment, highlighting current approaches, limitations, technological advancements and identifying knowledge gaps to guide future research. Following PRISMA guidelines, a systematic search of PubMed, Scopus, IEEE Xplore, Web of Science, Cochrane Library, and Google Scholar identified 77 eligible studies, with methodological quality assessed using the COSMIN Risk of Bias tool. Findings were categorized into four domains: kinetics, kinematics, muscle activity, and integrated domains based on the measured biomechanical parameter. Knee angles were the most frequently measured parameter; however, significant methodological variability was identified, particularly inconsistent sensor placement protocols, calibration methods, and limited real-life applicability. The COSMIN assessment revealed few studies achieved high overall reliability ratings, with kinematics showing the highest consistency and muscle activity the lowest. Kinetic assessments primarily measured knee adduction moments with generally acceptable accuracy but were constrained by indirect measurement methods. Muscle activity assessments predominantly employed surface electromyography, yet exhibited substantial inconsistencies, particularly in electrode placement and analysis techniques. Integrated biomechanical assessments showed significant potential but were constrained by the lack of simultaneous multi-parameter measurement capabilities. Notably, ML/DL integration demonstrated promise, though inconsistencies in performance metrics and sensor configurations complicate direct comparisons, emphasizing the need for standardized evaluation criteria for broader clinical applicability.