[R] poisson and peritonitis

Peter Dalgaard BSA p.dalgaard at biostat.ku.dk
Tue Jan 23 00:14:42 CET 2001

Troels Ring <tring at mail1.stofanet.dk> writes:

> Dear friends. I have a very simple question. In an old paper (Vonesh,Stat 
> Med 1990;9:263-271), the rate of peritonitis in 81 peritoneal dialysis 
> patients is estimated. The estimated intercept and an effect for race (0,1) 
> in a generalized linear model with poisson distribution is given. The 
> patients are observed a different length of time but time is not in the 
> model, so I cannot see how is accounted for that. Anyhow, the observed 
> number of zero incidences (35) is compared to the fitted zero incidences 
> (said to be 33.54).
> I would guess the way to do that was to let
> mm <- exp(0.00149+race*0.7888) (his intercept and effect for race)
> sum(dpois(0,mm)) but this doesn't yield but 25.64225
> I know the times, but how could I get to the results of Vonesh without 
> letting it have an effect ?
> In hope of forgiveness if I have totally misunderstood what is going on

It can only be guesswork, but one idea could be that time actually
does enter but in the form exp(0.00149+race*0.7888)*time alias
exp(0.00149+race*0.7888+log(time)) which is a glim with an offset.
Then since the offset is known it might not be given in the results?

   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

More information about the R-help mailing list