cif {surv2sample} | R Documentation |
Compute estimates of cumulative incidence functions for one or more samples of censored data with several competing risks (types of failure).
cif(x, group)
x |
an object of class "Survcomp" . |
group |
a vector giving for each observation the group
to which the observation belongs. If there are K samples,
group may only contain values 1,...,K.
If group is missing, observations are assumed to be
one sample. |
The cumulative incidence function F(t,j) is defined as the probability of failure by time t from a particular cause j in the presence of other competing risks 1,...,j-1,j+1,...,J, that is, F(t,j)=P(T<=t, e=j), where T is the failure time and e is the failure cause.
The function cif
estimates cumulative incidence functions for
all causes and all groups.
cif
returns a list of class "cif"
with the following components:
[[k]] |
a list containing estimates for the kth
sample. It has components time (sorted times in the
group k), surv (the Kaplan–Meier estimate of the
overall survival in this group), and f , which is
a matrix containing estimates of cumulative incidence functions
for the kth sample. The jth column of f is
the cumulative incidence function for the cause j. |
time |
the vector of sorted times. |
event |
the vector of corresponding event types. |
group |
the vector of corresponding group indicators. |
ncauses |
the number of causes present in the data. |
ngroups |
the number of samples. |
David Kraus (http://www.davidkraus.net/)
Klein, J. P. and Moeschberger, M. L. (2003) Survival Analysis. Techniques for Censored and Truncated Data. Springer, New York. (Section 2.7)
## bone marrow transplant data data(bmt1) print(a <- cif(Survcomp(bmt1$time, bmt1$event), bmt1$donor)) str(a) ## several first times and cifs for group 1 (HLA-identical ## sibling donors) head(cbind(a[[1]]$time, a[[1]]$f)) ## several last times and cifs for group 2 (HLA-matched ## unrelated donors) tail(cbind(a[[2]]$time, a[[2]]$f)) ## plot aggregate cumulative incidence functions for each ## donor type to see probabilities within groups plot(a)