Empowering Direct Care Workers with AI-Integrated Wearable Technology to Prevent Back Injuries

Care4qol Llc, Albany NY

Investigators

Abstract

Abstract Work-related injuries (WRIs) among direct care workers (DCWs), including home health aides and nursing assistants, are among the highest of all professions, with nursing assistants experiencing WRIs at 5.5X higher than the average workers. The injuries contribute to ~ 20 billion annually in healthcare expenses, lost productivity, and workers’ compensation claims. Unlike other industries with manual lifting limits of ≤50 pounds, DCWs routinely transfer patients weighing 100-250 pounds. Current prevention solutions, such as routine ergonomic training or mechanical lifting devices, are often ineffective due to logistical challenges, limited real-world applicability, and high cost in diverse care settings. To address this critical issue, Care4QoL will develop the Wearable Technology for Injury Prevention (WeTIP), which integrates a smart back brace and shoe inserts equipped with advanced sensors. These sensors monitor key ergonomic metrics, including spine twist, back bend, foot pressure, balance, and leg squat depth. A machine learning algorithm, trained on real- world data, continuously improve risk detection and personalized feedback. Building on our feasibility study results, Phase I has two specific aims: (1) Develop and validate the Gen 2 Prototype: this version will reduce sensors from 17 to 6 or fewer, with enhanced capabilities for tracking and analyzing body mechanics. The Gen 2 prototype will include both the smart back brace and shoes inserts for monitoring and feedback. Prototype validation will compare the system’s performance against physical therapist assessments (gold standard) using metrics such as classification accuracy and response time. (2) Evaluate the WeTIP system in clinical settings: testing with 30 DCWs will evaluate lifting and transferring techniques against gold-standard ergonomic practices. Performance metrics including accuracy, false positive rates, and feedback response time, will guide refinement of the machine learning model. These aims mitigate technical risks and lay the foundation for Phase II study, which will integrate visual and auditory biofeedback and evaluate the clinical effectiveness of WeTIP. The goal is to reduce WRIs among DCWs, improving their quality of life, lowering healthcare cost, and enhancing patient care. With a cost-effective, portable, and scalable design, WeTIP aligns with public health priorities and holds strong commercialization potential across high-risk occupations.