Feature-level gait data and frailty-related variables derived from three-dimensional skeleton analysis in older outpatients

This dataset contains de-identified, feature-level gait data and frailty-related variables used in a cross-sectional study of older outpatients. Gait features were extracted from walking videos using a three-dimensional skeleton detection system and include parameters such as stride length, gait cycle, and upper-limb and trunk movement variability. In addition to gait features, the dataset includes demographic and clinical variables required to construct the frailty prediction model, including age, sex, Kihon Checklist (KCL) scores, and Japanese version of the Cardiovascular Health Study (J-CHS) frailty status.