testing incidence rate ratios

testing incidence rate ratios

by Rena -

Hi, I have two questions regarding an RCT I am participating in and my outcome of incident pregnancy at 12 months of follow-up.

1) Should I be calculating incidence proportions (i.e. number of women pregnant during last year/total number of women eligible to be pregnant) or incidence rate (i.e. number of pregnancies during last year per person-months--remember that everyone is being followed for one year)?

2) I had wanted to calculate incidence rate ratio (intervention/control arm).  If so, what is the test of significance?  I have concluded ANOVA, is that correct?

Thanks,

Rena

Re: testing incidence rate ratios

by Mark Pletcher -

This question is getting a bit beyond the scope of Biostat 212!  A good one for Epi Methods...but happy to chat about it sometime in person if you see me.  If I understand your question, ANOVA is almost certainly NOT correct.  Probably just use proportions?