Predicting the Prevalence of Cardiometabolic Risk Factors When Clinical Data Are Limited