JMIR Publications

ABSTRACT

Oral disease and malnutrition are common, intertwined problems in long term care. Routine monitoring—monthly weights, occasional diet reviews, and infrequent dental assessments—often misses slow decline. Recent tools, including computer vision intake tracking and smartphone based gingival screening, make continuous, objective surveillance feasible. This viewpoint synthesizes emerging evidence and proposes a nursing led policy framework to operationalize AI early warning systems in nursing homes. The framework specifies governance with oral health and nutrition champions, a pragmatic 48 hour bedside evaluation standard for AI flagged risks, standards based electronic health record integration, and escalation pathways aligned with prevention oriented reimbursement. We address explainability, consent for vulnerable residents, data minimization, and equity, and we emphasize threshold calibration to limit alert fatigue and subgroup monitoring to ensure fairness. An implementation playbook outlines staged rollout and key performance indicators across clinical outcomes, workflow burden, and cost. Our distinctive contribution is a nursing led design with protected champion roles, a time bound response rule, auditable KPIs with scheduled threshold tuning, and concrete Electronic Health Record (EHR) write back steps. Embedded in nursing workflows and coupled to accountable policy, AI enabled surveillance can shift long term care from reactive to preventive practice while protecting resident dignity. We conclude with a call for co designed implementation studies to move from pilots to reliable, equitable practice.