Prospective study. What test would be appropriate?

#1
Hey there!

I am practicing on these data I have and I want to apply a statistical test.

I want to compare 2 treatments: evolution and incidence of a certain disease after treatment.

There were 3 groups taken into study: treated with A (n=78), treated with B (n=65) and C=control group, untreated (n=72).

I evaluated them initially, 6 months and 12 months after the initiation of treatment. Each subject received score 0, 1 or 2 at these time points.

Since the data are categorical, I am thinking of applying chi-square test, but I am not quite sure about the steps I have to follow.

Should I calculate expected value (and how?), the distribution of the data (?). Could you suggest some steps (mean, standard deviation)? Thank you!
 

hlsmith

Omega Contributor
#2
Do we assume underlying characteristics of data are comparable or randomization occurred. Otherwise, there could be confounders.


So you have repeat measures with no baseline value? And the outcome is categorical or is it ordinal?
 
#3
Do we assume underlying characteristics of data are comparable or randomization occurred. Otherwise, there could be confounders.


So you have repeat measures with no baseline value? And the outcome is categorical or is it ordinal?
As far as I know, subjects were randomly assigned to treatment A, B or control group. the outcome can take values between 0 and 16. I guess it is ordinal, isn't it?
 

hlsmith

Omega Contributor
#4
Do you know the context of the example? So subjects assigned to one of three groups and you collect a variable at six and 12 months that is on the scale 1-16, correct?


What does the dispersion of the outcomes look like, and once again, we don't know what they represent.
 
#5
Do you know the context of the example? So subjects assigned to one of three groups and you collect a variable at six and 12 months that is on the scale 1-16, correct?


What does the dispersion of the outcomes look like, and once again, we don't know what they represent.
It's a score that classifies incontinence on a scale 0-16 based on 5 questions. In order to answer to your question about dispersion should I make a graph?
 

hlsmith

Omega Contributor
#6
Yes, a histogram would be nice. I have not seen it, but there is some literature stating an ordinal range with 7 or more values might be modeled as a continuous variable. I was just curious in your sample what values were seen.


Once again, do you have a baseline incontinence score?
 
#7
Yes, a histogram would be nice. I have not seen it, but there is some literature stating an ordinal range with 7 or more values might be modeled as a continuous variable. I was just curious in your sample what values were seen.


Once again, do you have a baseline incontinence score?
Well, all the data I have are in the attached document: no baseline incontinence score. The values obtained by the authors vary from 0 to 10.