Validation of a Commercial and Custom Made Accelerometer-Based Software for Step Count and Frequency during Walking and Running

Background: Walking and running are main human locomotor activities during daily living, and well known to strongly predict health impairment and mortality. Hence, the main aim of this study was to assess the ability of a commercial and a custom made software for determining number of steps and step frequency during walking and running with an accelerometer in a semi-standardized setting. Methods: 20 subjects (6 males and 14 females) equipped with the Actigraph GT3X+ tri-axial accelerometer at the thigh and the hip carried out a protocol of three walking speeds and three running speeds. The validity of the accelerometer ability to count steps and estimate step frequency was determined by comparing data from ActiLife 5 and custom made software (Acti4) with observations from video recordings from the different activity speeds. Results: No significant differences in number of steps or step frequencies were found between the video observations and Acti4 measures in any walking and running speeds. The ActiLife 5 software recorded a significantly lower number of steps and step frequencies compared to the video observations in the three walking speeds and in the fastest running speed. Pearson’s correlations and Bland-Altman plots indicated large to very large correlations and a high degree of agreement between the video observations and both the custom made Acti4 software and commercially available ActiLife software at all speeds of walking and running. Conclusion: The custom made Acti4 software showed valid for estimating steps and step frequency at slow, moderate and fast speeds of walking and running. Combined with the ability to detect activity type, the Acti4 software provides a valid objective method for measurements of number of steps and step frequencies.