RESUMEN
OBJECTIVE: To evaluate the utility of the substituted judgement standard in terminally ill patients by determining agreement between patients, family proxies and physicians. BACKGROUND: Several studies have addressed the utility of substituted judgement, showing conflicting data whether surrogates are accurate predicting patient's wishes. METHODS: Patients with acquired immunodeficiency syndrome, congestive heart failure, chronic liver disease admitted to the San Juan Veterans Affairs Medical Center in Puerto Rico from November 1997 to February 1998 were evaluated. A questionnaire presented three hypothetical situations on withholding and withdrawal of life-support and CPR. The percent agreement was used as a measure of concordance between choices made by physician, surrogate and patients. RESULTS: Twenty patients met inclusion criteria (5 chronic liver disease, 9 heart failure, 6 AIDS). Relatives had a higher percent of agreement as compared to physicians in all vignettes. Even though, none did better than chance in predicting patient's wishes (k < 0.4). There was a tendency for relatives not to provide a wanted life-support measure, and for physicians to provide an unwanted life-support measure. CONCLUSION: The poor agreement between patients and surrogates suggests that substituted judgement is not an accurate tool to make end-of-life decisions. These findings, although similar to previous published studies, are unique because the direction of discrepant responses is opposite to the findings of studies published elsewhere. These results could reflect religious, cultural and socioeconomic differences.