According to the good old truism, the last and crucial proof of the pudding is in the eating thereof; and so, the last and crucial test of a therapeutic agent is its consumption by a patient. There is, however, one essential difference: When the pudding is eaten, with a sense of satisfaction, we know that it was a good, or at least an eatable pudding.
If the patient improves after taking a remedy, we do not yet know that he improved on account of the remedy. The post hoc type of reasoning or logic is not respectable; but it is all too apt to creep in unawares, unless one takes great precautions indeed.
Clinical evidence needs especially to be on its guard against this pitfall, for the conditions of disease never remain constant; nor is it possible to foresee with certainty the direction which they are going to take. It
SOLLMANN T. THE CRUCIAL TEST OF THERAPEUTIC EVIDENCE. JAMA. 1917;LXIX(3):198-199. doi:10.1001/jama.1917.02590300038011