Is experience a good teacher? How interns and attending physicians understand patients' choices for end-of-life care

Background. Recent studies have shown that physicians do not accurately assess patients' health status or treatment preferences. Little is known, however, about how physicians' levels of training or experience relate to their abilities to assess these preferences. To better understand this phenomenon, the authors compared the abilities of medical interns and attending physicians to predict the choices of their adult patients for end- of-life care. Methods. 230 seriously-ill adult inpatients were surveyed about their desires for cardiopulmonary resuscitation, their current quality of life, and their attitudes toward six other common adverse outcomes. The medical intern and attending physician who cared for these patients were asked to estimate the patient's responses for all of the same items. Agreement was assessed using the kappa statistic. Results. Compared with interns, attending physicians had known patients longer, had talked with patients more frequently about prognosis, and felt they knew more about their patients' preferences (all p < .0001). Despite this, the attending physicians were no more accurate than the interns in assessing patients' preferences. Both interns and attending physicians had only a fair understanding of patients' preferences for cardiopulmonary resuscitation or their quality of life (kappa statistics 0.32 to 0.47), and even less understanding of their willingness to tolerate adverse outcomes (kappa statistics -0.03 to 0.37). Conclusions. For this cohort of seriously ill patients, neither medical interns nor their attending physicians were consistently accurate in assessing patients' preferences, and attending physicians were not more accurate than medical interns. Attending physicians should not assume that they can infer patients' preferences any better than the interns caring for these hospitalized patients.