[R] Calculating NOEL using R and logistic regression - Toxicology

[Jarno Tuimala]

Dear Danielle,

At least in industrial toxicology (my original background) the recent
tendency has been to use benchmark dose (BSD) approach instead of NOEL
or NOAEL approach due to various problems with the definition and
estimation of NO(A)EL. In R this can be achieved using the packages
drc and bmd as already mentioned by Drew Tyre.

Here's a short code example that gives you the BMD at 1% level:

library(drc)
library(bmd)
data(daphnids)
d24<-daphnids[daphnids$time=="24h",]
fit <- drm(no/total~dose, weights = total, data = d24, fct = LL.2(),
type = "binomial")
ED(fit, 1)
bmd(FIT, 0.01)

See the following documents for more information, if you're interested
in using BSD instead of NO(A)EL:

http://www.cdpr.ca.gov/docs/risk/bmdquant.pdf
http://www.efsa.europa.eu/en/efsajournal/doc/1150.pdf

Jarno