I'm tasked with obtaining some data about chronic disease on an human population. One of these is to identify people with a chronic disease from the data set. Since chronic diseases definition varies somewhat, I'm attempting to use ICD codes to separate those with a chronic conditions from those that do not. Would be asking which group of codes should be safely used an on topic question? Would questions about general diseases analysis on human populations?
I think asking about mining databases using ICD codes could be on-topic. Having said that, I don't perceive much in the way of community knowledge surrounding that at the moment (perhaps I'll be pleasantly surprised though!), so I'm somewhat pessimistic regarding how much help you'll receive.
Asking which codes are best associated with which disorder is getting fairly off-topic, since that's moving quickly into the realm of the clinician (after all, you end up needing to know a bit about the local billing practices and disease prevalence to be able to judge whether the labeled ICD codes really do correspond to the appropriate condition).
"General human disease analysis" is largely going to be off-topic unless it specifically references a bioinformatic technique or method. Much of that will be standard wet-lab sorts of things, which will be quite off-topic here.