Defining responders to altered running footwear based on measures of running economy

This study compared three different ways of defining a responder to a shoe intervention when assessing running economy (RE). Furthermore, we investigated the influence of multiple trials on the classification of responders. This was done using indirect calorimetry to measure RE in thirty-seven runners, in three different running shoes. Over two visits, each participant ran a total of four times in each shoe. Three responder definitions were used to classify participants into responders or non-responders. The strength of each definition was evaluated using a logistic regression analysis with a pseudo r². The effect of multiple trials was analysed using number of trials defined as responders. Two definitions had similar high McFadden’s Pseudo r-squared (r² = 0.89, r² =0.86), being definitions that i) included a coefficient of variation (CV) derived from the literature and ii) a CV calculated by averaging all participants day-to-day CV from the present study. Inconsistent responders were present in 57% of the incidences, meaning that participant showed both responding and non-responding trials for the same shoe condition. For defining responders to a shoe intervention (RE specifically), it is recommended to use a cut-off value based on the literature which is easily available. Furthermore, conducting multiple trials is beneficial as this allows a separation between consistent- and inconsistent responders.