Wearable-Derived Data for Patient Frailty: Extending Hospital Frailty Risk Score While Confronting Bias and Inequities.

Objective, continuous measurements from wearable devices present a unique opportunity to enhance traditional, episodic clinical assessments of frailty by passively capturing real-world data on mobility, activity, and physiological signals. Integrating wearable-derived data with validated risk stratification tools like the Hospital Frailty Risk Score (HFRS) allows for earlier detection of functional decline across adults of all ages and supports timely, targeted interventions to prevent progression to severe frailty. However, deploying AI-enabled, wearable-based remote patient monitoring at scale requires confronting both device-related measurement biases and algorithmic inequities, particularly those affecting racially minoritized groups. Without rigorous validation and equity-driven calibration, these biases risk amplifying disparities in triage and clinical outcomes. This Perspective reviews recent advances in national-scale frailty risk analysis, the practical application of wearables in patient monitoring, and the critical need for performance auditing and inclusive algorithm design. It proposes a path forward for extending frailty dashboards with wearable data, emphasizing transparent validation across demographic strata, continuous post-deployment monitoring, equitable device access, and governance centered on clinical relevance. This proposal integrates continuous wearable monitoring data with the HFRS framework while centering equity-by-design principles and governance recommendations as co-equal priorities alongside technical integration. By prioritizing health need and function in algorithmic targets and validating models with locally representative data, health systems can realize the promise of AI-augmented frailty care while advancing equity and precision for all populations.