To the Editor.
—In his response1 to our commentary on the coronary-prone behavior pattern,2 Ray H. Rosenman, MD, seems to have missed the point.We noted that Friedman and Rosenman had originally described five categories of behavior (Al, A2, B3, B4, and C).2 When these five individual categories did not show sufficient discriminating power for coronary risk in the Western Collaborative Group Study, they were consolidated and reduced to the two coarse A and B types for which a distinction could be demonstrated. Because we believe that "human behavior is far too complex to be suitably described in the two categories of A or B,"2 we speculated about reasons why the original five types—although each important—might have yielded false-negative results in the Western Collaborative Group Study. The population may have been too constrained in its white-collar male volunteers; the behavior was identified with structured questionnaires and
Horwitz RI, Horwitz SM. Coronary-Prone Behavior Pattern. Arch Intern Med. 1987;147(2):205. doi:10.1001/archinte.1987.00370020025014