Accuracy of the FitBit at walking speeds and cadences relevant to clinical rehabilitation populations

Background Activity monitoring is used for motivation and assessing patients in the rehabilitation setting. Monitors available on the market have not been assessed for accuracy at walking patterns seen in the clinical setting. Objectives To assess the accuracy of the Fitbit Zip device at different anatomical locations at walking speeds, step length, and cadence similar to those seen in clinical rehabilitation populations. Methods Ten healthy participants completed 3 sets of 12 combinations of speed and cadence ranging from 0.2 m/s at 48 steps per minute to 1.0 m/s at 120 steps per minute along a 20 m walkway. Step length was controlled by ground markings and cadence was controlled by a metronome. Five Fitbit Zip devices were concurrently worn at the chest, hip, shin, ankle and forefoot. Percent accuracy for each location and walking condition combination were calculated. Results At the chest and hip the Fitbit has poor accuracy below 0.7 m/s, however when worn distally, can be accurate to 5.5% at speeds as slow as 0.5 m/s. Conclusions For patient populations with slow walking speed, activity monitoring with the Fitbit can be achieved provided the device is located distally on the lower limb.