Biomedical Engineering Reference
In-Depth Information
above, the two treatments do not have the same survival distribution. In fact,
patients in treatment 1 survived significantly longer than those in treatment
2. To conduct a three-sample test, we artificially create another group based
on the data
cosmesis
. Here are the results:
>gLRT(cosmesis,k=3,method="glrt3",rho=0,gamma=0,inf=100)
$method
[1]"Generalizedlogranktest(Zhaoetal.,2008)"
$u
[1]-20.32072224.3319828.786514
$var
[,1][,2][,3]
[1,]0.6033566-0.5782167-0.5782167
[2,]-0.57821671.0688249-0.5782167
[3,]-0.5782167-0.57821673.0452748
$chisq
[1]7.768535
$df
[1]2
$p
[1]0.02056289
Steno Memorial Hospital Diabetic Data from Denmark: This
study involved a total of 731 patients who were younger than 31, diagnosed
as Type I diabetics between 1933 and 1972, and followed until death, emi-
gration, or 31 December 1984. The survival time of interest is the time from
the onset of diabetes to the onset of diabetic nephropathy (DN), a major
complication of Type I diabetes and at least four samples of 24-hour urine at
time intervals of at least 1 month contain more than 0.5 g protein. All 731
patients considered here had developed DN at the time of admission or by the
end of the study, meaning that there is no right-censoring in the data. There
were 595 exactly observed and 136 interval-censored observations. Among the
Search WWH ::
Custom Search