Significance test for differences in baseline and re measurement rates

#1
Hi,

I need help in knowing which test to perform with this analysis.
I have two groups (lets say group A and B). A study was performed with initial rates of heart failures in group A and B (i.e, baseline rates) were measured. After 6 months of follow up the heart failure rates in group A and B (i.e, re measurement rates) were measured again.
By simple mathematics, I know that the improvement rate in better in group A as compared to group B.
I have to prove that the differences in rates in both groups are statistically significant.
I am using t-test for this but not sure how to go about it.

I would appreciate your suggestions.

Thanks a lot.
 

bukharin

RoboStataRaptor
#2
It sounds like you're measuring heart failure as yes/no. In this case you're comparing the change in proportion with heart failure in group A, with the change in proportion with heart failure in group B. For this I'd recommend McNemar's test.
 
#3
Thanks for your reply. Isnt Mcnemar's test for paired data? the two groups are not paired.
To be more explicit. I have the baseline and remeasurement rates for both groups (each group has 10 hospitals in it). I calculated relative improvement rate, RIR (which is [baseline - remeasurement] / baseline) for each hospital in both groups.
With t-test, I analysed if the mean RIR for both groups are statistically significant? Is this a right way of doing it?
to confirm the results, I also categorized each hospital into improvement/no-improvement based on RIR and did a chisq.
 

trinker

ggplot2orBust
#4
Isnt Mcnemar's test for paired data?
Yes it is.

Why do you say your data's not paired? It sounds like you used the same hospitals though with count data. This seems paired or matched to me. The count data leads me to think repeated measures with poisson or negative binomial distribution. I can't comment on a test though because I have not encountered a need for such a test in my field. I'd say Poisson or negative binomial regression may be options but this requires more exploration on your part. And I'm still not sure if the McNemar's test is out either because you're hospitals seem paired (not the patients).

Caveat: this is not meant to give you a definitive answer but guide to give more options/information