05-26-2018 10:12 PM
I am trying to see whether there is a significant upward or downward trend in certain infection over time. I calculated rates/95% confidence intervals using proc genmod. Can I use Cochran-armitage test (year vs. mean proportion) for trends? If not, what method would be appropriate?
05-27-2018 05:12 PM
I got yearly rates with below:
proc genmod data=cdi;
class year sex age_group / param=glm;
model cdiff= year sex age_group / type3 dist=poisson link=log offset=log_discharge;
estimate "rate: year=2003" intercept 1 year 1 0 0 0 0 /e
I would appreciate sas codes or references for trend analyses.
05-28-2018 09:07 AM
Since all your independent variables are CLASS variables, then your data indeed like a contingency table.
Check PROC FREQ + TEAND option. there is an example of it in PROC FREQ's documentation.