To the Editor.
—Dr Novack and colleagues1 made assertions about physicians' decision making and how it can be improved that seemed to go beyond the available evidence. First, their statements about what influences physicians' decision making seemed to be based on studies of association that may have been confounded. For example, to support their statement "physicians'effectiveness... depends on a variety of personal factors," they cited the studies by Nightingale and Grant that showed associations between physicians' responses to a written instrument designed to assess their response to risks and their patterns of utilization, eg, the length of time they conduct cardiopulmonary resuscitation.2 However, the work of Nightingale and Grant could not determine whether risk attitudes directly affected decision making or were simply a marker for something else that did. Likewise, to support the statement "sex thus affects... clinical decisions," Novack and colleagues cited a study in which the authors concluded, "it