A colleague of mine is about to do a validation study of some diagnosis given. In this study either the diagnosis is correct or it's not, i.e. it's a binomial trial. We assume that about 90% of the diagnoses are correct given. My question is how do we calculate the minimum sample size?