Analyse longitudinale de l'équilibre statique et de la marche pour prévenir les chutes chez les sujets âgés vivant en EHPAD

Falls among the elderly remain a major public health issue, particularly in nursing homes, despite growing awareness in healthcare settings. Current prevention strategies have not significantly reduced fall rates or the number of fallers. Falls thus remain the leading cause of fatal accidents in older adults. While technological solutions¿such as fall detection systems¿have improved incident response time, no current approach provides longitudinal monitoring capable of identifying evolving fall risk with sufficient sensitivity and reliability. This thesis investigates predictive factors of falls in frail, dependent older adults living in nursing homes. It focuses on functional activities¿gait and balance¿measured through wearable devices that capture physiological data. Particular attention is given to sensory integration, given its crucial role in posture and movement control. Despite extensive literature on gait, balance, and falls, studies involving highly dependent individuals remain rare. This work addresses that gap through two complementary components: (1) a systematic literature review with meta-analyses on the effects of sensory disturbances on static balance, and (2) a six-month observational study of 36 nursing home residents, collecting monthly data on gait (IMUs), balance (force platforms), and muscle strength, alongside daily fall reports and medical records. An adapted physical activity program introduced mid-study enabled the exploration of changes in fall risk indicators over time. Results highlight the feasibility of integrating non-invasive, time-efficient assessments into routine care, and suggest potential markers for long-term fall risk monitoring. Sensory input emerged as a key factor, with even simple eyes-closed tests revealing significant challenges for many participants. Ultimately, this thesis provides new insights into fall risk among a clinically complex and often overlooked population, and emphasises the value of longitudinal, instrumented monitoring in improving fall prevention strategies.