Quantification of Outdoor Mobility by Use of Accelerometer-Measured Physical Behaviour

Abstract

Hip fractures in older persons are associated with both low levels of daily physical activity and loss of outdoor mobility. The aim was to investigate if accelerometer-based measures of physical behaviour can be used to determine if people undertake outdoor walking and to provide reference values for physical behaviour outcomes related to outdoor mobility. Older persons (n = 245), ≥70 years, one year after hip fracture, participated. Six objective measures of physical behaviour collected by an activity monitor were compared with self-reported outdoor mobility assessed with the Nottingham Extended ADL scale. All measures of time and length in upright periods were significantly lower in participants who reported not walking outdoors (p < 0.001). A set of cut-off points for the different physical behaviour variables was generated. Maximum length of upright events discriminated best between groups, with 31 minutes as a threshold to determine if a person is more likely to report that they walk outdoors (sensitivity: 0.805, specificity: 0.704, and AUC: 0.871) or 41 minutes or more to determine if a person is more likely to report outdoor walking on their own (AUC: 0.891). Physical behaviour variables from activity monitoring can provide information about patterns of physical behaviour related to outdoor activity performance.