topic proc freq riskdiff how to structure data in SAS Programming

proc freq riskdiff how to structure data

kalbo — Wed, 11 Dec 2024 16:37:50 GMT

Hi all I have AE data which is structured like:

Arm Preferred_term

A Headache

A Zoster

B Fatigue

B Headache

I want to use proc freq to create a data set containing risk difference (comparing Arm A and Arm B) like this:

Arm Preferred_term Risk_diff

A Headache 0.45

A Zoster 0.5

B Fatigue 0.1

B Headache 0.2

How do I structure the data prior to using proc freq?

Thanks.

Re: proc freq riskdiff how to structure data

ballardw — Wed, 11 Dec 2024 18:26:43 GMT

Relative risk and risk difference in Proc freq data structure would be basically one of two forms.

Summarized data would have a risk variable with two levels (had/did not have exposure or risk) outcome (did/did not have outcome) and a count of cases. The count variable would be used as WEIGHT variable in Proc Freq.

An example of this data appears in the Example section of Proc Freq in "Analysis of a 2x2 Contigency Table".

as:

data FatComp;
   input Exposure Response Count;
   label Response='Heart Disease';
   datalines;
0 0  6
0 1  2
1 0  4
1 1 11
;

The other structure would be person/case level data where each observation is one case and just has the Exposure and Response.

I might guess that your outcome or response is "have headache"/"do not have headache" "have zoster/do not have zoster".

I really hesitate to guess what your exposure or actual risk might be though. "Arm" sounds very unlikely to be a risk. So I suspect we are missing some details.

Re: proc freq riskdiff how to structure data

Ksharp — Thu, 12 Dec 2024 01:50:33 GMT

If you want to calculate risk difference, you need 2x2 contingency table.
Assuming row variable is ARM which have two levels (A and B),
But the column variable ( Preferred_term ) have four levels,
Therefore you are NOT able to get risk diff.

Better give an example to illustrate how to calculate risk diff.