Finding studies
Finding studies
Take this into the appointment.
Saves the questions and what to expect into your notes, next to the visit they belong to.
Lead
University Hospital, Montpellier
Statistical analysis plan The number of patients solicited for the study and the number of subjects included and randomized in each group will be reported on the flow chart. Premature stops (lost to follow-up, death, study withdrawals) and their reasons will be reported as well as any deviations from the protocol. An initial descriptive analysis of each of the groups will be carried out. For qualitative variables, this description will include the number as well as the frequency of the different modalities. Concerning the quantitative variables, the description will include the number, the mean, the standard deviation, the median as well as the extreme values according to the distribution. The comparability of the 2 groups will be verified for all the initial characteristics likely to influence the results. In the event of non-comparability on one or more parameters, an adjustment will be made on this or these parameters for the comparisons between groups of judgment criteria. Analysis of the primary judgment criterion The variation in the evolution of the various main criteria will be compared between the 2 groups using a nonparametric means comparison test (Mann-Whitney test) or if the distribution turns out to be Gaussian and the conditions application are verified, a parametric test (Student's T). Analysis of secondary endpoints The qualitative variables will be compared by a Chi-square test. Otherwise, if the conditions for performing this test are not met, Fisher's exact test will be used. The quantitative variables will be compared between the two groups: * in the case of Gaussian variables, by a student's t test or an analysis of variance * in the case of non-Gaussian variables, by a Wilcoxon-Mann-Withney test Longitudinal data (quality of life scores, physical activities) will be modeled by a mixed regression with the patient effect as a random variable. A linear regression model modeling the variation of the longitudinal endpoints will be fitted to the covariates having at least one trend (p \<0.15) to the difference between the two arms of the study. A forward type procedure will be implemented to obtain a final multivariate model. Management of missing data: The IPTW method will be used. The lost to follow-up will be considered using a weighting on the inverse of the probability of being censored
Age
30–75
Sex
ALL
Healthy volunteers
Not accepted
