Statistical considerations
Quantitative data collected at Y-1, Y0, Y+1 and Y+2 were summarized
using mean and standard deviation and were compared using Student’s
t-test. Descriptive data were expressed in percent and compared using
Fisher’s Test.
Figures 1 to 3 show the distributions of the variables separately for
groups A (0), A (1) and A (2). In order to compare these distributions,
we use Wilcoxon’s signed rank test. This test is used for paired data,
the null hypothesis being that the distributions of the two samples are
identical. The p-values of the tests comparing A (0) and A (1)
respectively, and A (1) and A (2) are indicated in green. The only test
rejected at the 10% threshold is for the number of consultations
between A (0) and A (1) (p = 0.059). Figures 4 and 5 show the
distribution of the differences between the groups (from left to right)
A (0) and A (-1), A (1) - A (0) and A (2) - A (1) for the different
variables. We use the same Wilcoxon test as before to compare them. This
time the two tests are rejected at the 10% threshold for VEMS but
neither is rejected for BMI. Analyses were executed with Python 2.7,
scipy 1.1.0.