In defense of Pierre Louis who pioneered the epidemiological approach to good medicine.

Pierre Louis, in his 1836 publication, "Researches on the effects of bloodletting in some inflammatory diseases," concluded that there were "narrow limits to the utility" of bloodletting for the treatment of pneumonia. Louis's conclusions were based on an observed excess mortality of patients bled rapidly post disease onset compared with patients bled later on. Doctors had used bloodletting for at least 2,000 years for a wide variety of purposes. Louis's approach was a first-step toward what we would call today an evidence-based medicine. A book denouncing "Bad Medicine" throughout history suggests that Louis inappropriately analyzed his data, overestimated the alleged benefits of venesection on the evolution of pneumonia and therefore contributed to promote this dangerous treatment. I show in this article that these criticisms are inaccurate. In particular, Louis was aware of the potential confounding effect of age and interpreted his data accordingly. Accusing Louis of "bad medicine" misses the historical importance of his struggle to improve clinical medicine using population thinking and group comparison.