[R] KM estimation for interval censoring?

[Robert Gentleman]

The Icens package implements a number of different estimation paradigms. 
Briefly, for interval censored data, the KM does not apply, but 
Turnbull's method does. Relevant citations for Icens are in it.

For a Cox model approach, there was at one time a package, not at CRAN, 
by Commenges (and a Biometrics paper, I believe), Google suggests that 
Dr. Commenges is pretty active in this area.


David Winsemius wrote:
> "gallon li" <gallon.li at gmail.com> wrote in
> news:54f7e7c30801272341v3082e988qa6363991df956488 at mail.gmail.com: 
> 
>> Does anybody know if there is such a function to estimate the
>> distribution for interval censored data?
>>
>> survfit doesn't work for this type of data as I tried various
>> references. 
> 
> My reading of R-help posts by Therneau and Lumley is that interval 
> censoring is only available for parametric estimates in the survival 
> package.
> 
> It's not KM, but perhaps this will be of interest:
> 
> "Turnbull's Nonparametric Estimator for Interval-Censored Data"
> Suely Ruiz Giolo
> 
> <http://www.est.ufpr.br/rt/suely04a.pdf>
> 
> Giolo cites KLEIN, J. P, MOESCHBERGER, M. Survival Analysis. New York: 
> Springer Verlag, 1997, for the algorithm.
> 
> This MIT-CSAIL report says they have an available R package that handles 
> interval censoring and time-dependent covariates:
> 
> http://dspace.mit.edu/bitstream/1721.1/33957/1/MIT-CSAIL-TR-2006-055.pdf
> 

-- 
Robert Gentleman, PhD
Program in Computational Biology
Division of Public Health Sciences
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N, M2-B876
PO Box 19024
Seattle, Washington 98109-1024
206-667-7700
rgentlem at fhcrc.org