The analysis of primary and secondary outcome measurements
will include all randomized patients. To be
able to use all patient data, the mixed models approach
will be adopted with individuals at level one and center
at level two. The analysis will demonstrate any improvements
at follow-up, comparing the intervention and
control group. In linear mixed models the variable group
is considered main, controlling for baseline value, center
and gender. To judge the quality of the model we will
analyze the residuals. A difference at the P