[R] survival analysis: interval censored data

[Ruth Arias]

hallo terry:

I attached araceae data set, 

when I use this:

surara<-survfit(Surv(time,time2,event)~categoria)

Call: survfit(formula = Surv(time, time2, event) ~ categoria)

            records n.max n.start events median 0.95LCL 0.95UCL
categoria=C      94    63       0     21     NA      NA      NA
categoria=E     111    77       0     26     NA      NA      NA
> summary(surara)
Call: survfit(formula = Surv(time, time2, event) ~ categoria)

                categoria=C 
 time n.risk n.event entered censored survival std.err lower 95% CI upper 95% CI
 2006     63       5       0       23    0.921  0.0341        0.856        0.990
 2007     35       2      30        1    0.868  0.0483        0.778        0.968
 2008     62       5       1        3    0.798  0.0536        0.700        0.910
 2009     55       4       0        5    0.740  0.0570        0.636        0.861
 2010     46       5       0       41    0.660  0.0611        0.550        0.791

                categoria=E 
 time n.risk n.event entered censored survival std.err lower 95% CI upper 95% CI
 2005     71       7       3        0    0.901  0.0354        0.835        0.973
 2006     67       2       0       22    0.875  0.0391        0.801        0.955
 2007     43       1      36        1    0.854  0.0432        0.774        0.943
 2008     77       5       0        8    0.799  0.0469        0.712        0.896
 2009     64       4       1        3    0.749  0.0502        0.657        0.854
 2010     58       7       0       51    0.658  0.0545        0.560        0.774

but whe I included type=interval, 

> suraraint<-survfit(Surv(time,time2,event,type='interval')~categoria) # falta arreglar lo del intervalo!!!
> summary(suraraint)
Call: survfit(formula = Surv(time, time2, event, type = "interval") ~ 
    categoria)

                categoria=C 
 time n.risk n.event survival std.err lower 95% CI upper 95% CI
 2004  95.00   13.14    0.862  0.0354        0.795        0.934
 2007  31.86    7.19    0.667  0.0695        0.544        0.818
 2008   1.67    1.67    0.000     NaN           NA           NA

                categoria=E 
 time n.risk n.event survival std.err lower 95% CI upper 95% CI
 2004  112.0   18.47    0.835  0.0351        0.769        0.907
 2005   40.5    1.06    0.813  0.0401        0.738        0.896
 2007   37.5    7.46    0.651  0.0620        0.540        0.785

it does not survival calculed for very year

I have a one-year interval between each census



________________________________
De: Terry Therneau <therneau en mayo.edu>
Para: Ruth Arias <rueu_ac en yahoo.es>
CC: r-help en r-project.org
Enviado: miércoles 28 de septiembre de 2011 16:00
Asunto: Re:  survival analysis: interval censored data

You have still not given me enough information to reproduce your
problem.  "Why doesn't it include all years?"  I have no way of knowing,
since we have no data.

--- begin included message --
halo david

when I use type= 'interval'

Call: survfit(formula = Surv(ingreso, fecha, estado, type = "interval")
~ 
    categoria)

and when I use just 

Call: survfit(formula = Surv(ingreso, fecha, estado) ~ categoria)

I don t know why when I use type = "interval" it does not survival
calculed for very year


regards
------------ pr?xima parte ------------
An embedded and charset-unspecified text was scrubbed...
Name: araceae.txt
URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20110928/e361216e/attachment.txt>