With the data of the German National Health Survey the question is discussed, if self-reports on diseases, gained in a survey, are useful for estimating prevalences within the German population. Self-reports on myocardial infarction serve as an example for examining the reliability and plausibility of such data. The prevalence rate for myocardial infarction in this survey is compared with prevalence rates of other data sources as the MONICA-register and surveys. Because these other data sources have some problems and deficits according to their validity too, the value of self-reports on diseases from the National Health Survey is seen positively for establishing a health reporting system.