Using R for the analysis
Risk factors are composed of two continuous variables and 5 or so factor variables with varying levels (mostly 0,1 but a couple with 0,1,2,3)
I am doing survival analysis for a cancer project. Basically my dataset is a series of risk variables ( t stage, tumor size, age at dx, etc, the event codes (local, regional, and distant recurrences), and then the time variable (followup time). I'm using R to do the analysis (i'm a beginner) and the problem i'm having is with the univariate cox regressions. Right now i'm just entering the line of code for each variable i.e. coxph(survobj~[risk factor],data), This obviously gets very tedious doing this for all the covariates and the different types of recurrences so I figure there has to be a way to apply the function to each variable in one line of code... something like what the apply function could do but i'm not sure how to do it. The idea would be to have something like the general function coxph(survobj~x) where x is the different risk factors, and then have it spit out an object with the HR's etc. Anyone know how I would go about coding this? I tried the following code to create a formula but it spit out an error that i'm not sure how to deal with...
f
You might have more luck asking this question on Stack Overflow as it is a coding question rather than a pure science question. Stack Overflow deals with lots of R queries.
http://stackoverflow.com/search?q=survival+analysis+coxph
That is actually what I ended up doing! In case anyone stumbles on this question and is wondering something similar here this is the answer they gave which I found very helpful.
If you want to go the formula-route (which in your case with multiple outcomes and multiple variables might be the most practical way to go about it) you need to create a formula per model you want to fit. I've split the steps here a bit (making formulas, making models and extracting data), they can off course be combined this allows you to inspect all your models.
#example using transplant data from survival package
#make new event-variable: death or no death
#to have dichot outcome
transplant$death
Ah excellent suggestions veit! Thank you! Do you by any chance know how to add extracting the wald p-value to the results return part?
For the cancer papers in my area they generally report both it seems, at least in the table summarizing the salient univariate and multivariate data. Probably excessive but I know personally I often like to see both but maybe I'm just weird ha. Thanks for the help, I'll have that a shot!
- Badges
- Science topic
- Similar topics
- Chemistry