Running cohort diagnostics in a random sample

[daniprietoalhambra]

Team,
Sorry if I'm asking an elephant-in-the-room question, but has anybody suggested the possibility of making Cohort Diagnostics capable to run in a random sample of a whole database? I can see how we might need to run this tool in a whole database for very rare conditions etc, but for other things that are really common (eg type 2 diabetes) we could improve performance and cut cost quite dramatically by running in a random sample of the patient table