HR and 95% CI for Univariate Survival Analyses, logrank,etc?

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

crusthov
Posts: 4
Joined: Sat Sep 28, 2013 6:02 pm

HR and 95% CI for Univariate Survival Analyses, logrank,etc?

Postby crusthov » Sat Sep 28, 2013 6:14 pm

Does anyone know how to generate the overall Hazard Ratio and 95% CI for log rank tests, breslow, and tarone ware? I only see default output for mean and median survival estimates, not HR. By convention of the medical literature, I need to report the HR and 95% CI.

Thanks!
statman
Administrator
Posts: 2757
Joined: Tue Jun 12, 2007 12:08 pm
Location: Florida, USA

Re: HR and 95% CI for Univariate Survival Analyses, logrank,

Postby statman » Sun Sep 29, 2013 12:20 pm

Beside explaining the test and how determined see the Note below
See the note below

NOTE: Please read the Posting Guidelines and always tell us your OS, the SPSS version and information about your study and data!

Statman
Statistical Services
crusthov
Posts: 4
Joined: Sat Sep 28, 2013 6:02 pm

Re: HR and 95% CI for Univariate Survival Analyses, logrank,

Postby crusthov » Sun Sep 29, 2013 5:17 pm

Thanks and sorry about that!

OS: Windows 7, 64-bit
SPSS: version 22

Study:
-Overall I am looking at the impact of radiation therapy on survival for a certain type of cancer
-So, one analysis includes performing a Kaplan-Meier to show the survival curves with and without radiation therapy
-I can do this just fine and get p-values, the mean-median estimates with 95% CI and curves.
-But what I want also is the Hazard Ration with 95% CI for the Logrank test, or breslow, or tarone ware.
-I know I can calculate a hazard ration using the cox model and just enter the one variable so it remains univariate - however, it is my understanding that this is a different test and has slightly different pvalues.
-By convention, the medical literature reports logrank tests with HR, 95%CI, and pvalue - so I want to know if this is possible with SPSS - If it is, it would make my life a lot easier. As it is now, I am copy-pasting my data into graphpad to get the log-rank HR and CI


Any help would be greatly appreciated! (even if its just letting me know if SPSS cant generate HR with 95% for univariate survival tests.) Thanks!
statman
Administrator
Posts: 2757
Joined: Tue Jun 12, 2007 12:08 pm
Location: Florida, USA

Re: HR and 95% CI for Univariate Survival Analyses, logrank,

Postby statman » Sun Sep 29, 2013 8:55 pm

I found info on the "Hazard Ratio", is this what you are talking about
crusthov
Posts: 4
Joined: Sat Sep 28, 2013 6:02 pm

Re: HR and 95% CI for Univariate Survival Analyses, logrank,

Postby crusthov » Sun Sep 29, 2013 10:11 pm

Yes! I specifically want to generate Hazard Ratios and 95% Confidence Intervals (of the Hazard Ratio) for univariate survival analyses (logrank, terone-ware, breslow) that come with the Kaplan Meier analyses under the "compare factor" button options in SPCSS

For example, I want to generate something that tells me this for a Logrank test:
Radiation Therapy = reference, HR=1
No Radiation therapy. HR 1.35, 05% CI 1.20-1.40, p=0.001

As it is now, I can only seem to generate the estimate for the mean and median survival with 95% CI of the mean and median estimates.

I know I can generate the HR, CI and p with a cox regression, but that is a different model than a logrank test. The logrank test with HR, 95% CI, and p-value and used by convention in the medical literature.

Do you know how I can generate this in SPSS?

Thanks!

Who is online

Users browsing this forum: No registered users and 1 guest

cron