How to derive survival time on anonomised dates

survival-analysis

#1

i wonder if there is a convention and a google search is not revealing much. Date of birth has been anonomised using july for month. Obviously this leads to some negative survival times. I don’t want to simply set negative survival times to 0 and create an artificial steep drop in any KM plot. Thus if survival was between -6 and +6 i used: (random no. from uniform dist) x 12. But it’s crude and i’m doing this on the fly. Is there an approach that’s more likely to be considered ‘acceptable’ than others? The industry likes to produce ‘guidelines’ but they don’t tend to handle admin databases and thus i guess they don’t have a position on this. cheers

edit: is there any software that can anonomise/de-identify data? i think redcap does it but you need your data in redcap 1st…