Machine Learning-Driven Decision Support Framework to Improve Musculoskeletal Health in Electrical Workers Using Wearable Devices

Abstract

The construction industry is currently facing a significant shortage of skilled labour, limiting its ability to meet societal demands and sustain economic contributions. A key contributor is the high prevalence of musculoskeletal disorders (MSDs), which are often caused by repetitive motions, awkward postures, and sustained physical exertion inherent to various construction trade activities. This research focuses on electrical workers, a subgroup within the construction workforce that has been shown to experience high rates of MSDs. Recently, wearable support devices have emerged as a promising intervention for reducing fatigue and MSD risk. Although the construction sector has historically been slow to adopt new technologies due to fiscal and practical constraints, the growing availability of cost-effective, lightweight, and task-specific wearables has made their application increasingly feasible. Such innovations hold significant potential to enhance worker well-being, decrease the risk of work-related musculoskeletal injuries, and ultimately extend career longevity and earning potential within the industry.

This research examines the effectiveness of wearable support devices in reducing the risk of MSD injuries among electrical construction workers. To facilitate evidence-based implementation, a decision-support framework is developed to recommend cost-effective wearable devices tailored to individual workers’ needs while accounting for the financial constraints faced by both employees and organizations. The framework comprises three principal components: (1) analysis of injury databases and task-specific ergonomic assessments to identify high-risk body regions and exposure patterns; (2) evaluation and classification of commercially available wearable devices according to their provided biomechanical support levels; and (3) development of a decision support system (DSS) that integrates machine learning and optimization techniques to generate cost-effective, personalized recommendations.

First, publicly available injury data for electrical construction workers from WorkSafeBC and the Workplace Safety and Insurance Board (WSIB) were analyzed to identify the most frequently injured body parts, highlighting the three main areas most in need of support, namely the back, leg, and shoulder. These findings were validated through detailed task analysis and ergonomic assessment of common electrical construction activities. Next, a representative subset of commercially available wearable support devices targeting these body parts was experimentally evaluated. A diverse group of participants, including professional electricians and novice users, performed experimental tasks while performance metrics were recorded. Clustering analysis classified the devices into three support categories (Low, Medium, and High). Finally, a DSS was developed, integrating: (1) a machine learning model to predict support requirements, and (2) an optimization model that recommends specific wearable configurations aligned with worker needs and budget constraints. The DSS was validated through field testing with electrical workers who provided positive feedback on the wearables’ real-world effectiveness and usability.

The proposed framework provides a novel decision-support approach for recommending cost-effective wearable devices from a repository of low-cost, commercially available options. It integrates objective and subjective inputs, performance metrics, machine learning tools, and optimized decision logic to address a critical occupational health and safety concern prevalent in construction. Although this study focused on electrical workers and a specific subset of support wearables, the framework is inherently scalable. It can be adapted to other trades with different ergonomic demands, as well as to a broader and continuously expanding range of wearable technologies entering the market. By supporting the selection of effective and affordable devices, the framework has the potential to reduce MSD occurrences, enhance worker well-being, and improve overall productivity within the construction sector.