Biomedical Engineering Reference
In-Depth Information
baseline - follow-up
70
80
60
70
subj1
subj2
subj3
subj4
subj5
subj6
subj7
subj8
subj9
subj10
subj11
50
60
40
50
30
40
walker
crutches
one crutch
none
30
Fig. 6. Symmetry results according to walking
aid. Box-plot representations of the distribu-
tions. The whiskers represent the smallest and
largest observations, the edges of the box cor-
respond to the lower and upper quartiles, the
horizontal line indicates the median.
20
baseline
follow-up
Fig. 5. Symmetry results at baseline and
follow-up. The two distributions are not sta-
tistically significantly different.
baseline - follow-up
100
100
95
subj1
subj2
subj3
subj4
subj5
subj6
subj7
subj8
subj9
subj10
subj11
95
90
90
85
walker
crutches
one crutch
none
baseline
follow-up
Fig. 8. Normality results according to walking
aid. A Kruskal-Wallis test indicates that the
distribution of no walking aid is significantly
different from distributions of two crutches
and walker.
Fig. 7. Normality results at baseline and at
follow-up. The two distributions are statisti-
cally different according to a Wilcoxon rank
sum test, p < 0.0001.
represented by 1 and no-aid was represented by 4. The resulting Spearman's rank cor-
relation coefficient was r=-0.78, p < 0.0001.
The normality index also correlates well with both average speed, r=-0.79 p < 0.0001,
and normalized average step length, r=-0.76 p < 0.0001. Normality values for each indi-
vidual trial are shown against average speed values in Figure 9, and against normalized
step length in Figure 10. On both plots the linear model approximation is shown as a
solid line, and the 95% confidence interval (CI) for predicted observations is shown
as dotted lines. The mean average speed at baseline, 0.46
±
0.16 m/s, was significantly
different from the speed at follow-up, 1.06
0.22 m/s, p < 0.0001.
Normality results were also compared to the EQ-5D TM answers that varied the
most between before the surgery and after follow-up, namely mobility (Figure 11),
usual activities (Figure 12), and pain/discomfort (Figure 11). In all cases, there is a
±
Search WWH ::




Custom Search