Spatial access to sport facilities from the multiple places visited and sport practice: Assessing and correcting biases related to selective daily mobility

Abstract Background Studies investigating the association between spatial accessibility to environmental resources from the various places a person visits during daily activities and use of corresponding resources often do not account for potential biases related to selective daily mobility. This bias occurs when accessibility is also measured from places intentionally visited to access the resources of interest. The aim of this study was to examine associations between spatial accessibility to sports facilities from multiple places and sport practice while addressing the selective daily mobility bias. Methods The second wave of the RECORD Cohort was used to examine the relationship between the spatial accessibility to sport facilities and the practice of three sport categories (swimming, racket, and team sports), using multilevel linear probability models (n = 5327 participants) adjusted for individual and contextual characteristics. Street network distance to the nearest sport facility was considered as a measure of spatial accessibility [from the residence; from the residence and workplace; from all visited locations (full activity space), biased; and from all locations excluded those visited for sports (truncated activity space), corrected]. Results The residential and residential-workplace accessibility to facilities was not associated with sport practice. The spatial accessibility to facilities from all places visited (full activity space) was associated with the practice of the three categories of sports (biased relationships). After correcting the bias (truncated activity space), the strength of the relationships was markedly reduced. An association remained only for swimming sports. Conclusion This study underlines the need to account for selective daily mobility bias when determining spatial accessibility to resources from the various places visited. Such bias, if not addressed, may result in overestimated associations between spatial accessibility and use, leading to potentially erroneous conclusions in terms of planning.