Sir.—With regard to Dr Brown's article entitled "Berkson Fallacy Revisited" (Am J Dis Child 130:56,1976), we would like to repeat the comments we made in The Lancet concerning similar criticism by others of our paper on reserpine and breast cancer.1.2
Our paper on reserpine and breast cancer1 has drawn a number of critical comments. The general lack of validity of the criticisms, together with the importance of the matter, have prompted us to comment.
The main criticisms appear to be the following: (1) the exclusion of patients who were hospitalised for diseases which are, or might be, associated with reserpine use produced an inappropriate lowering of the rate of reserpine use among control subjects; (2) hospital controls are not representative of the general population and are, therefore, inappropriate....
Criticism (1) reflects lack of understanding of one of the principles of comparability in case-control studies. The cases of