Relative risk for a table 2xK with a reference group

[antor82]

Hi All!!!

I need to compute the relative risk for a 2x5 table, having the lowest group as the reference value.

Can anybody help me?

Thank you in advance

Antonio

[Ksharp · Accepted solution]

/*
Relative Rick is only calculated for 2x2 tables.
2xK table I think you mean multiple 2x2 tables,
*/
data have;
call streaminit(123);
do page=1 to 20;
 do obs=1 to 100;
  row=rand('bern',0.4);
  col=rand('bern',0.6);
  output;
 end;
end;
run;


proc freq data=have;
table page*row*col/relrisk(column=1)  ;
run;

[antor82]

Thank You very much Ksharp for Your prompt and clear reply!

Actually I have a binary outcome (0/1) and a categorical predictor with four mutually exclusive values (0/1/2/3).

When a run a proc freq, I obtain a 2x4 table.

Should I create different columns for each of the four values of the predictor variable?

Then should I run a proc freq as in Your example?

eg:

table outcome*pred1*pred2*pred3*pred4

?

Thank You in advance

[antor82]

trying to be clearer...

Binary outcome 0/1

Four class of renal damage A/B/C/D

I need to compute the RR for the outcome "1" in B vs A; C vs A; D vs A.

Hope is clearer now.

Similar to a proc logistic with A as the reference value. Only needing to compute RR instead of OR

[Ksharp]

"I need to compute the RR for the outcome "1" in B vs A; C vs A; D vs A."
Then you need make three sas datasets(2x2 table) for RR separatedly .
And @StatDave posted a very useful URL, check it out:
https://support.sas.com/kb/23/003.html