[R] log rank test p value

[Marc Schwartz]

On Wed, 2007-05-16 at 11:54 +1000, Murray Pung wrote:
> How can I get the Log - Rank p value to be output?
> 
> The chi square value can be output, so I was thinking if I can also have the
> degrees of freedom output I could generate the p value, but can't see how to
> find df either.
> 
> > (survtest <- survdiff(Surv(time, cens) ~ group, data = surv,rho=0))
> Call:
> survdiff(formula = Surv(time, cens) ~ group, data = surv, rho = 0)
> 
>          N Observed Expected (O-E)^2/E (O-E)^2/V
> group=1 20       16     11.0      2.23      4.64
> group=2 20       12     17.0      1.45      4.64
> 
>  Chisq= 4.6  on 1 degrees of freedom, p= 0.0312
> 
> > survtest$chisq
> [1] 4.641028
> 
> > survtest$df
> NULL

That part of the output is created in print.survdiff(), which you can
review by using:

  survival:::print.survdiff

The degrees of freedom and the p value are calculated in that function.


Using the example in ?survdiff:

survtest <- survdiff(Surv(futime, fustat) ~ rx, data = ovarian)

> survtest
Call:
survdiff(formula = Surv(futime, fustat) ~ rx, data = ovarian)

      N Observed Expected (O-E)^2/E (O-E)^2/V
rx=1 13        7     5.23     0.596      1.06
rx=2 13        5     6.77     0.461      1.06

 Chisq= 1.1  on 1 degrees of freedom, p= 0.303 



> str(survtest)
List of 6
 $ n    : 'table' int [, 1:2] 13 13
  ..- attr(*, "dimnames")=List of 1
  .. ..$ groups: chr [1:2] "rx=1" "rx=2"
 $ obs  : num [1:2] 7 5
 $ exp  : num [1:2] 5.23 6.77
 $ var  : num [1:2, 1:2]  2.94 -2.94 -2.94  2.94
 $ chisq: num 1.06
 $ call : language survdiff(formula = Surv(futime, fustat) ~ rx, data = ovarian)
 - attr(*, "class")= chr "survdiff"



We can then do the following to secure the p value:

> 1 - pchisq(survtest$chisq, 1)
[1] 0.3025911


See ?str and ?pchisq for more information.

HTH,

Marc Schwartz