|
normal = pmax(1 - stats::pnorm(d, mean = p$mu, sd = p$sigma), 0), |
I don't quite know how this package works so I apologise if this is of topic but I was looking at the code and I think that here may want double interval censored discretisation as both the primary and secondary events are censored (they are days or weeks I think). Using continuous directly will I think induce a bias in the SD and using some other approximate discretisation will induce some other kind of problem!
We have a bunch of tooling for this and related research that might be of interest but no pressure: https://primarycensored.epinowcast.org/
censcast/R/los_spec.R
Line 30 in 4c4945a
I don't quite know how this package works so I apologise if this is of topic but I was looking at the code and I think that here may want double interval censored discretisation as both the primary and secondary events are censored (they are days or weeks I think). Using continuous directly will I think induce a bias in the SD and using some other approximate discretisation will induce some other kind of problem!
We have a bunch of tooling for this and related research that might be of interest but no pressure: https://primarycensored.epinowcast.org/