Exploratory analysis using machine learning of predictive factors for falls in older adults with and without type 2 diabetes

Falls remain a pressing public health concern among older adults, often resulting in significant physical, psychological, and financial consequences. The risk is notably higher among individuals with type 2 diabetes mellitus (T2DM) due to complications such as peripheral neuropathy, musculoskeletal impairments, and psychological distress. This study aimed to identify predictive factors for falls among communitydwelling older adults in Sarawak, with and without T2DM, using traditional statistical analysis and a machine learning approach specifically the Multilayer Perceptron (MLP) model. A cross-sectional dataset was collected from adults aged 60 and above, incorporating both intrinsic (e.g., balance, strength, fear of falling) and extrinsic (e.g., environmental hazards) risk factors. Data were pre-processed using normalization, outlier removal, and imputation techniques. Logistic regression and 10-fold crossvalidated MLP models were applied to explore predictive patterns. Key results showed that the MLP model achieved higher prediction accuracy compared to traditional methods. Physically, lower extremity muscle weakness and impaired mobility (measured by TUG and HGS) were significant predictors. Psychologically, high fear of falling (FES-I) and low balance confidence (ABC Scale) were associated with increased fall risk. Older adults with T2DM displayed a distinct fall risk profile, marked by reduced proprioception and greater psychological concern. These findings highlight the need for tailored fall prevention strategies that consider both physical and psychological components, particularly in older adults with T2DM. Integrating machine learning into clinical assessments could enhance early identification and guide personalized interventions. This study contributes to the growing evidence supporting the use of AIdriven tools in geriatric care and establishes a foundation for future public health initiatives targeting fall prevention among high-risk populations.