for testing difference in pre-post change between males and females do unpaired t test of mean difference, males v females

For relation between outcomes A and B, do correlation coefficient between A and B . If there is a difference between males and females do seperately or a do a multiple regression of A on B or B on A with sex as a binary variable in the regression.

If data severely non normal do appropriate non parametric tests.

Hello Nasser,

Regarding your wish to evaluate whether changes in outcome "A" are related to changes in outcome "B:"

If you assume that both outcomes are measured without error on both occasions, you could compute differences in each (e.g., post A - pre A; post B - pre B), then correlate the scores (and, of course, evaluate a scatter plot for the two difference scores.

If the outcomes are not measured without error, you'd likely be better off estimating true scores for each measure on each occasion, and working with those instead. The danger of working with fallible scores when it comes to differences is that the differences are less reliable than either of the two scores used to compute them.

Good luck with your work.

Hello Nasser,

One of the scenarios begins by comparing the outcome in pre-and post-intervention and then adjusting the potential confounders in your prediction model for each survey phase. You can use logistic regression if your outcome was categorical or linear logistic mode if your outcome was continuous.

I assume the outcome measures is in numerical variable. therefore, you can use repeated measures ANOVA to identify several hypothesis:

1. is there any significant differences of mean outcome (e.g., HbA1c) between pre-treatment & post-treatment within the male or female group itself?

2. is there any significant differences of mean outcome between the male and female group (regardless of time)?

3. is there any significant differences of mean outcome among group (male and female) with regard of time (time*group interaction)?

Thank you all for your answers. I appreciate the information you have provided.