BACKGROUND:
Falling is a common but devastating and costly problem of aging. There is no consensus in the literature on whether the spatial and temporal gait parameters could identify elderly people at risk of recurrent falls.
OBJECTIVE:
To determine whether spatiotemporal gait parameters could predict recurrent falls in elderly women.
METHOD:
One hundred and forty-eight elderly women (65-85 years) participated in this study. Seven spatiotemporal gait parameters were collected with the GAITRite(r) system. Falls were recorded prospectively during 12 months through biweekly phone contacts. Elderly women who reported two or more falls throughout the follow-up period were considered as recurrent fallers. Principal component analysis (PCA) and discriminant analysis followed by biplot graph interpretation were applied to the gait parameters.
RESULTS:
After 12 months, 23 elderly women fell twice or more and comprised the recurrent fallers group and 110 with one or no falls comprised the non-recurrent fallers group. PCA resulted in three components that explained 88.3% of data variance. Discriminant analysis showed that none of the components could significantly discriminate the groups. However, visual inspection of the biplot showed a trend towards group separation in relation to gait velocity and stance time. PC1 represented gait rhythm and showed that recurrent fallers tend to walk with lower velocity and cadence and increased stance time in relation to non-recurrent fallers.
CONCLUSIONS:
The analyzed spatiotemporal gait parameters failed to predict recurrent falls in this sample. The PCA-biplot technique highlighted important trends or red flags that should be considered when evaluating recurrent falls in elderly females.
falls; elderly; gait; principal component analysis; biplot; physical therapy
Thumbnail
