To complement subjective patient-reported outcome measures, objective assessments are needed. Activity is an objective clinical outcome which can be measured with wearable activity monitors (AM). AM's have been validated and used in joint arthroplasty patients to count postures, walking or transfers. However, for demanding patients such as after sports injury, running is an important activity to quantify. A new AM algorithm to distinguish walking from running is trialed in this validation study. Test subjects (n=9) performed walking and running bouts of 30s duration on a treadmill at fixed speeds (walking: 3, 4, 5, 7km/h, running: 5, 7, 9, 12, 15km/h) and individually preferred speeds (slow, normal, fast, maximum, walk/run transition). Flat and inclined surfaces (8%, 16%), different footwear (soft, hard, barefoot) and running styles (hind/fore-foot) were tested. An AM (3D accelerometer) was worn on the lateral thigh. Previously validated algorithms to classify all gait as walking were adapted to differentiate running from walking, the main criterium being vertical acceleration peaks exceeding 2g within each subsequent 2s-interval. Independently annotated video observation served as reference.Background
Methods
