Abstract
BACKGROUND: Identifying older adults at high risk of progressive oral functional decline is a critical public-health priority, as early decline often manifests initially as subtle swallowing difficulties. Standard diagnostic tools (e.g., VFSS, FEES) are invasive, whereas most existing screening methods assess only a single functional dimension. Therefore, this study aimed to develop and validate a fully non-invasive, multidimensional screening approach that integrates high-resolution video-based jaw-movement analysis with diadochokinetic (DDK) performance to enable the early detection of older adults at risk. METHOD: Eighty-nine participants aged 65-95 years were recruited from community care stations in Taiwan. Participants were classified as high risk for oral functional decline if they performed fewer than three swallows within 30 s on the repetitive saliva swallowing test (RSST). The assessment included DDK performance and jaw movement analysis using video tracking. Logistic regression was used to identify significant predictors, followed by receiver operating characteristic (ROC) curve analysis to evaluate diagnostic performance. RESULTS: Among the 89 participants (mean age = 77.6 years; 22 men), 41 (46%) met the RSST high-risk criterion. Female participants exhibited significantly faster jaw-movement velocities compared to male participants (p < 0.05). Multivariate logistic regression confirmed age, DDK performance, and jaw-kinematic variables as independent predictors of high-risk status. A combined model of key DDK measures and average jaw-closing angular velocity achieved optimal discrimination (AUC = 0.83) and high sensitivity (0.95), outperforming individual predictors. CONCLUSION: This study presents a practical, cost-effective, and non-contact screening method for early identification of oral function high risk in older adults. Integrating jaw movement with DDK performance offers high sensitivity, supporting potential use in early clinical interventions and community-based screening programs. TRIAL REGISTRATION: ISRCTN11010971 (https//www.isrctn.com/ISRCTN11010971). Registered on 17 April 2025.