Purpose
The purpose of this thesis was to develop a new test of dynamic postural control to target these shortcomings and to establish 1) validity; 2) reliability; 3) the influence of potential covariates such as anthropometry, age, sex and level of physical activity; and 4) the influence on overhead athletic performance.
Project description
Standardized testing procedures were developed by a group of experts, based on: 1) starting position; 2) task; 3) measurement; and 4) ending position, which served as content validity. In study I, criterion-related and construct validity were explored. Specifically, the magnitudes of joint movements used to assume maximum HSEBT reach positions were quantified using motion capture (Qualisys Oqus 400 cameras, Qualisys AB, Gothenburg, Sweden) and compared to joint movements in the comparable star excursion balance test (SEBT) and normative ROM values. Criterion-related (concurrent) validity was established by comparing reach measurements calculated from motion capture data to those visually obtained using Bland Altman and correlational analysis. Construct validity was assessed by correlating outcome measurements (reach, composite scores and area calculations) from the HSEBT with the comparable SEBT. In study II, inter-rater and test-retest reliability was assessed from the outcome measurements of three experienced testers using interclass correlation coefficients (ICC), with the calculation of stability measurements (standard error or measurement and coefficient of variation) and minimal detectable change. The influence of anthropometry, age, sex and level of physical activity was explored in study III. Stepwise linear regression was used to determine the influence of these factors on reach measurements. Independent samples t-tests were used to determine between group (age, sex and level of physical activity) differences with calculation of effect sizes and group difference comparisons to minimal detectable change values (study II). The influence of HSEBT reach measurements on athletic performance (overhead team handball throwing) in an elite female population was explored for both throwing velocity, calculated from motion capture data, and accuracy, via mean radial error calculated from video records using Pearson correlational analysis.
Results
The HSEBT elicited significantly greater joint movements than the SEBT in 18 out of 22 joint movement comparisons. The magnitude of these joint movements was comparable to the ranges of normative ROM values for 8 out of 22 joint movements. Excellent correlations were observed between visually observed and calculated reach measurements from motion
capture data for 18 out of 20 tests (r ≥ 0.90) with a shared variance that ranged from 81 to 97%. For the remaining two tests good correlations were observed (r = .79 and .89). The fixed biases observed (range = 2.2 to 12.8 cm, −6.0 to 11.2º and 23.7%) can be partially explained by the methods used to calculate reach measurements. Different composite and area scores for the HSEBT and SEBT had variable correlations (range r = .269 to .823), with a wider range of observed values for the individual reaches (range r = -.182 to .822). The strongest correlations were observed for the anterior composite, area and reach measurement comparisons (range r = .515 to .823). In Study II moderate to high test-retest reliability was observed for 19 out of 20 reaches (range ICC = 0.80 to 0.96). The inter-rater reliability was high for all reaches (range ICC = 0.90 to 0.98). Minimal detectable change values ranged from 0.9–7.9 cm and 4.7–7.2º for all reaches. Wingspan (study III) explained 34.6 and 11.7% of the variance of two HSEBT reaches. When normalized (% of wingspan) the
same reaches were influenced by age, sex and level of physical activity with significant between-group differences, and moderate effect sizes (range d = .50 to .72). In addition, one non-normalized reach was influenced by age and level of physical activity (range d = .55 to
.75). HSEBT reach measurements are not correlated with throwing velocity (range r = -.530 to .395), but with mean radial error for some reaches (range r = .149 to .666) (study IV).