Likelihood ratio test confidence interval

[keldel]

I would like to obtain the confidence interval from the likelihood ratio test (G^2) for a binomial function with n=10000 and x=50 (the number of yes responses). 

options ls=72;
data conditionprevalence;
input condition count; 
cards;
1 3
2 997
;
run;
proc freq;
weight count;
tables condition/cmh chisq;
run;

I know this isn't right since I am not getting a confidence interval and this doesn't seem to be the LRT. Thank you for any insight.

@jimbarbour 

[keldel]

Sorry the sample code I provided had the wrong values. I had cards

1 50

2 9950 

[Watts · Accepted solution]

See the BINOMIAL(CL=LIKELIHOODRATIO) option in the TABLES statement (and details in the documentation section Likelihood Ratio Confidence Limits under Binomial Proportion).