Correlation with unknown numbers

#1
I am looking at a data set that can correlate prior doses of drug A with the amount of time (in days) before developing infections after stopping drug A. We hypothesize that people with lower dose will have longer latency before developing infections. But we have a lot of people (especially with very small doses) who never developed infections. How should I code those people so that they can be included in the study? Do I just put an arbitrarily large number?
 

Karabiner

TS Contributor
#2
You could check Cox regression. I.e. a survival analysis, which can handle "censored" cases.

With kind regards

Karabiner