Gait, gait variability and the risk of multiple incident falls in older people: a population-based study

Background: it is uncertain as to which measures of gait best predict those who are likely to fall. Our aim was to investigate the associations of gait and gait variability measures with incident falls risk. Methods: individuals aged 60-86 years (n = 412) were randomly selected from the Tasmanian electoral roll. Average gait and gait variability measures were collected on a computerised walkway. Falls were recorded prospectively over 12 months. Log multinomial regression was used to estimate the relative risk of single and multiple falls associated with gait measures. Covariates included age, sex, sensorimotor and cognitive measures, mood and medications. Results: in this population-based study greater intra-individual variability in step length and double-support phase were linearly associated with increased risk of multiple falls (P = 0.04). Non-linear associations with multiple falls were found for gait speed P = 0.002, cadence P = 0.004 and step time variability P = 0.03. None of the gait measures predicted risk of single falls. Conclusion: there is an increased risk of multiple falls, but not single falls, in older people with poorer gait. Specific measures of gait and gait variability seem to confer this risk and may be amenable to interventions designed to reduce the risk of multiple falls in older people.