I had 220 cases that were all treated using the same method but with slightly different baseline values (category of disease at start), age groups, gender.... The number of treatments were also different. Success was evaluated as change from a higher category (extreme severity) to a lower (mild disease) and also by a drop in breathing rate. I used the paired samples T-test to see the change in ordinal or scale data from baseline (before test started) to end of treatment.

To see differences in baseline and treatment effect variables for the cases that had had environmental changes 1. before the trial, 2. during the trial or 3. not at all, I used ANOVA and LSD as a post hoc test.

20 variables (such as sex, age group, baseline severity group, different environmental factors, numer of treatment times, number of relapses, reactions to treatments etc) were compared to the 2 main outcome varaiables as well as to the degree of treatment success (how many categories the treated either had got better, worse - or stayed the same) using an exact fisher's test.

Lastly change in breathing rate (scale) was correlated (spearman) to change in categories (ordinal). Would you suggest anything else and, especially, are these test suitable?