The value and the fallacy of statistics in the observation of disease