Biomedical Engineering Reference
In-Depth Information
method=\wsr.mc" gives the multiple imputation method where the infer-
ences are by Monte Carlo within each imputation (Section 13.3.6).
Consider the breast cosmesis data given in Finkelstein and Wolfe (1985),
which is included in theintervalpackage. Here is the R code to load the package
and data set, and print the first few observations from the data set, together
with those observations:
>library(interval)
>data(bcos)
>head(bcos)
leftrighttreatment
1 45 Inf Rad
2 6 10 Rad
3 0 7 Rad
4 46 Inf Rad
5 46 Inf Rad
6 7 16 Rad
Here is the code and results for the logrank test of Finkelstein (1986) test
using the normal approximation to the permutation test:
>ictest(Surv(left,right,type="interval2")~treatment,
+scores="logrank2",method="pclt",data=bcos)
AsymptoticLogranktwo-sampletest(permutationform),
Finkelstein
'
sscores
data:Surv(left,right,type="interval2")bytreatment
Z=-2.6839,p-value=0.007277
alternativehypothesis:survivaldistributionsnotequal
nScoreStatistic*
treatment=Rad 46 -9.944182
treatment=RadChem48 9.944182
*likeObs-Exp,positiveimpliesearlierfailuresthanexpected
Search WWH ::
Custom Search