Stair fall risk profiling using a novel multivariate approach

Stair falls are a major problem for older people. The present study: a) applied a novel multivariate approach to characterize the overall stair behaviour while ascending and descending stairs and b) investigated whether the selected stepping behaviour was maintained irrespective of step dimensions. K- means clustering was used to characterize the overall behaviour of 70 older (>65 y) and 25 younger adults based on biomechanically risky and conservative strategies. Age and fall history were not unique factors of the clusters, highlighting the limitations of conventional comparisons. Moreover, changing the staircase to ‘easier’ step dimensions tended to have no effect on the selected stepping behaviour. Further research should implement this multivariate method using a longitudinal approach to identify the behaviours that can differentiate those who will experience a stair fall from those who will not.