# Thread: Univariate and multivariate cox regression analysis

1. ## Univariate and multivariate cox regression analysis

I have just started with statistics but please can someone explain me ho to make a table like this (see the attachment)? Can I do that with SPSS? Any help would be very appreciated.

2. ## Re: Univariate and multivariate cox regression analysis

Well you make the table in a basic word formatting program like Microsoft Word.

The analyses were based on an analytical procedure typically called one of the following:

Cox Regression
Proportional Hazards regression
Survival Regression

So you would have to know how to run the procedure, its assumptions, and interpret the results. I haven't really used SPSS too much, but it would be hard for me to imagine it didn't include that procedure, which is pretty common in programs. The univariate approach is just one predictor in the model and I am guessing the multivariate should have been called multiple Cox regression, if so, it is the same thing as univariate but the model has more than 1 predictor.

Have you conducted any regression models? Have you conducted any binary outcome regression models? And do you have time to event data?

3. ## The Following User Says Thank You to hlsmith For This Useful Post:

ondansetron (05-30-2017)

4. ## Re: Univariate and multivariate cox regression analysis

here's where I got that table: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5135408/

I understand that , with SPSS, If I want to generate a table like that I need to test each line of that table separatedly. There's no way to get a table like that automatically.
It is a long thing to do, I can do that for the univariate side of the table.
It is less obvious to me the meaning of the "multivariate part". Can you explain me the meaning and how to get that part of the table?

5. ## Re: Univariate and multivariate cox regression analysis

Originally Posted by attorianzo

here's where I got that table: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5135408/

I understand that , with SPSS, If I want to generate a table like that I need to test each line of that table separatedly. There's no way to get a table like that automatically.
It is a long thing to do, I can do that for the univariate side of the table.
It is less obvious to me the meaning of the "multivariate part". Can you explain me the meaning and how to get that part of the table?
The author is mistaken calling it "multivariate" in this case. It's a mistake common in medical/public health/epidemiology/nursing and others. Multivariate tends to refer to multiple dependent variables being analyzed as once due to an underlying covariance structure. Refer to a MANOVA for an example. The paper is actually using a "multivariable" cox regression, which means there is one dependent/outcome variable, but more than one independent variable. As hlsmith said, this could also be called a "multiple cox regression", but the author's use of multivariate is incorrect in this case. Further, the paper should tell you how each of those multivariable parameter estimates were calculated. In other words, it should specify what other variables and terms were "adjusted for" in that model.

In SPSS, you would only need to include the other independent variables in the cox regression when the interface box asks for "covariates" (I believe this is how it's called in SPSS). You'll get the estimates as an output and can make the right side of the table. You may need to specify an option to generate a confidence interval for each parameter estimate.

6. ## Re: Univariate and multivariate cox regression analysis

Ditto, yeah, it is not too hard are all to make that table. Run a bunch of univariates and record measures then run a single model with all of them included at once.

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts