The interplay of risk factors for future falls and fall-related injuries among community-dwelling older adults: a Bayesian network analysis

Falls are a leading global cause of injury and mortality in older adults, necessitating understanding risk factor interplay for prevention. This study aims to investigate complex interactions among risk factors associated with future falls and fall-related injuries in older adults. A prospective cohort study used China Health and Retirement Longitudinal Study (2015 and 2018) data from 8,316 community-dwelling older adults. Falls and fall-related injuries were self-reported. The least absolute shrinkage and selection operator (LASSO) regression was employed for variable selection. A Bayesian network analysis utilizing the Max-Min Hill Climbing algorithm and Bayesian estimation method was subsequently implemented to identify interactions among risk factors related to falls and fall-related injuries. Fall incidence was 23.29%, and the incidence of severe injuries after falls was 10%. According to variables identified by LASSO regression, two Bayesian networks were constructed for falls (21 nodes and 48 directed edges) and fall-related injuries (20 nodes and 48 directed edges). Results revealed complex associations among the risk factors for falls and fall-related injuries, highlighting direct factors such as fall history, hip fracture history and balance, as well as the indirect effects of factors such as age, vision and cognitive function. Some specific combinations of direct risk factors contributing to a high probability of falls and fall-related injuries were also identified. Falls and related injuries are frequent in community-dwelling older adults, associated with by multiple interconnected factors. Bayesian network analysis helps uncover the relationships between these variables and quantify the associations of major factor. Clinical practice may benefit from prioritizing multifactorial exercise programs targeting grip strength, lower limb strength, and balance. Consideration should be given to fall risk assessments and tailored interventions for older adults with prior falls, hip fracture history, or those using walking aids to effectively reduce falls and related injuries.