Extreme huge value of hazard ration from PHREG

Nancych — Mon, 12 Mar 2018 23:03:04 GMT

Hi,

I am running into an issue for a subgroup analysis (by sex). The model converged, but the Hazard ratio from PHREG is extremely huge. you might see from below, there is no event for category of female. Is there anyway we could deal with it, such as correction? how about the HR and CIs for male, is that still valid? I appreciate any suggestions in such situation.

	Treatment	Placebo
Female
Number assessed	20	14
Censored	9	14
Events	10	0
Hazard Ratio (CI)	1267890 (0.000,-)
Male
Number assessed	20	19
Censored	14	14
Events	6	5
Hazard Ratio (CI)	0.69 ( 0.23; 2.06)

proc phreg data = XXX;

class trt sex;

model aval*cnsr(1) = trt sex trt*sex/ties = efron rl alpha=0.05 ;

HAZARDRATIO trt/diff=ref;

run;

Thank you,

Lily

Re: Extreme huge value of hazard ration from PHREG

Nancych — Thu, 15 Mar 2018 22:07:12 GMT

anyone can help?

topic Re: Extreme huge value of hazard ration from PHREG in Statistical Procedures

Extreme huge value of hazard ration from PHREG

Re: Extreme huge value of hazard ration from PHREG