Aging effects on the structure underlying balance abilities tests

Abstract

Balance impairment is one of the biggest risk factors for falls reducing inactivity, resulting in nursing care. Therefore, balance ability is crucial to maintain the activities of independent daily living of older adults. Many tests to assess balance ability have been developed. However, few reports reveal the structure underlying results of balance performance tests comparing young and older adults. Covariance structure analysis is a tool that is used to test statistically whether factorial structure fits data. This study examined aging effects on the factorial structure underlying balance performance tests. Participants comprised 60 healthy young women aged 22 ± 3 years (young group) and 60 community-dwelling older women aged 69 ± 5 years (older group). Six balance tests: postural sway, one-leg standing, functional reach, timed up and go (TUG), gait, and the EquiTest were employed. Exploratory factor analysis revealed that three clearly interpretable factors were extracted in the young group. The first factor had high loadings on the EquiTest, and was interpreted as 'Reactive'. The second factor had high loadings on the postural sway test, and was interpreted as 'Static'. The third factor had high loadings on TUG and gait test, and was interpreted as 'Dynamic'. Similarly, three interpretable factors were extracted in the older group. The first factor had high loadings on the postural sway test and the EquiTest and therefore was interpreted as 'Static and Reactive'. The second factor, which had high loadings on the EquiTest, was interpreted as 'Reactive'. The third factor, which had high loadings on TUG and the gait test, was interpreted as 'Dynamic'. A covariance structure model was applied to the test data: the second-order factor was balance ability, and the first-order factors were static, dynamic and reactive factors which were assumed to be measured based on the six balance tests. Goodness-of-fit index (GFI) of the models were acceptable (young group, GFI=0.931; older group, GFI=0.923). Static, dynamic and reactive factors relating to balance ability had loadings 0.21, 0.24, and 0.76 in the young group and 0.71, 0.28, and 0.43 in the older group, respectively. It is suggested that the common factorial structure of balance abilities were static, dynamic and reactive, and that for young people reactive balance ability was characterized and explained by balance ability, whereas for older people it was static balance ability.

Keywords: Older people; covariance structure analysis; reactive balance ability.

Fig. 1.

The structure of balance ability model; AREA-EO: COP area with eyes open; AREA- EC: COP area with eyes closed; C2: score under Condition 2 of SOT; Velocity-PW: velocity of preferred walking; Velocity-MW: velocity of maximum walking; TUG: Timed Up & Go; MAP: medium anterior perturbation of MCT; LAP: large anterior perturbation of MCT; LPP: large posterior perturbation of MCT.

Fig. 2.

The structure of balance ability model for the young group; AREA-EO: COP area with eyes open; AREA- EC: COP area with eyes closed; C2: score under Condition 2 of SOT; Velocity-PW: velocity of preferred walking; Velocity-MW: velocity of maximum walking; TUG: Timed Up & Go; MAP: medium anterior perturbation of MCT; LAP: large anterior perturbation of MCT; LPP: large posterior perturbation of MCT. Sixty subjects were observed, and nine measuring errors were left out. All coefficients were significant statistically.

Fig. 3.

The structure of balance ability model for the older group; AREA-EO: COP area with eyes open; AREA- EC: COP area with eyes closed; C2: score under Condition 2 of SOT; MAP: medium anterior perturbation of MCT; LAP: large anterior perturbation of MCT; LPP: large posterior perturbation of MCT; Velocity-PW: velocity of preferred walking; Velocity-MW: velocity of maximum walking; TUG: Timed Up & Go. Sixty subjects were observed, and nine measuring errors were left out. All coefficients were significant statistically.