Digital Bioimpedance for Physical Activity Detection in Type-2 Diabetes: Quasi-Experimental Validation Study

This pragmatic implementation study demonstrates that automated BIA is feasible for routine diabetes care and suggests potential as a digital biomarker of activity-related glycemic control. While temporal allocation precludes definitive causal inference, and findings should be interpreted as associational, the observed Walk Score moderation and bioimpedance-HbA1c dose-response patterns are consistent with behavioral mechanisms rather than pure confounding. Left-arm reactance warrants randomized validation as a scalable, passive surveillance tool for precision diabetes management.