I have two binary variables: treatment arm (arm 1, arm 2) and event (Yes, No) for each participant. I transformed my participant-level data into a long format summary data (so now I have 3 variables: count, arm and event). My goal is to compute the relative risk of event in arm 1 comparing to arm 2, and the exact confidence interval. The problem is the frequency of the event in arm 2 is 0 (no one had event in arm 2) so I always get infinite RR. And upper bound of my confidence interal goes to infinite as well. Is there a way to avoid this problem? I learned that proc freq would add 0.5 to each frequency to ensure that the statistics is defined when there is 0 frequency table cells but I couldn't figure out the code.
Here is my code:
proc freq data=data1;
weight Count;
tables arm*event ;
exact RELRISK;
run;
Can you share the cell counts you are getting in the 2x2 table?
my cell count frequencies are:
arm 1, event = yes: 5
arm 1, event = no : 70
arm 2, event = Yes: 0
arm 2, event = No: 26
proc freq data=data1;
weight Count/ZERO;
tables arm*event ;
exact RELRISK;
run;
I'm still getting infinite relative risk with the ZERO option.
Are you ready for the spotlight? We're accepting content ideas for SAS Innovate 2025 to be held May 6-9 in Orlando, FL. The call is open until September 25. Read more here about why you should contribute and what is in it for you!
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.