Video-derived balance score tracks age and fall risk

In a preprint evaluating 180 human participants, researchers developed and validated the Stanford Balance Level Evaluation (StaBLE) using video-derived digital metrics. The protocol assessed participants across an 18-task battery designed to measure a broad spectrum of balance performance.

The resulting StaBLE score correlated negatively with age (ρ = -0.76, p<0.001). It also correlated with the Activities-Specific Balance Confidence Scale (ρ = 0.60, p<0.001), the Mini-Balance Evaluation Systems Test task (τ = 0.55, p<0.001), the 4-Stage Balance Test (τ = 0.55, p<0.001), and the Short Physical Performance Battery (τ = 0.57, p<0.001). The digital score avoided ceiling effects present in conventional clinical scales and exhibited different distributions for participants identified at risk of falling versus those not at risk (KS: 0.67, p<0.001). The authors also identified task subsets to shorten testing time.