A growing body of evidence demonstrates the existence of important racial and ethnic differences in health care. Although many studies report on disparities in health care for acute and chronic conditions, 1-5 relatively few address racial disparities in the ICU, and those report confl icting fi ndings. Some have reported that black patients receive fewer medical interventions, 6,7 have shorter lengths of stay, and use fewer resources. 7 In contrast, others have found increased resource utilization, including larger numbers of ICU admissions and higher numbers of medical interventions for minority than for white patients. 8,9 Background: There is confl icting evidence about the infl uence of race/ethnicity on the use of intensive care at the end of life, and little is known about the infl uence of socioeconomic status. Methods: We examined patients who died in the ICU in 15 hospitals. Race/ethnicity was assessed as white and nonwhite. Socioeconomic status included patient education, health insurance, and income by zip code. To explore differences in end-of-life care, we examined the use of (1) advance directives, (2) life-sustaining therapies, (3) symptom management, (4) communication, and (5) support services. Results: Medical charts were abstracted for 3,138/3,400 patients of whom 2,479 (79%) were white and 659 (21%) were nonwhite (or Hispanic). In logistic regressions adjusted for patient demographics, socioeconomic factors, and site, nonwhite patients were less likely to have living wills (OR, 0.41; 95% CI, 0.32-0.54) and more likely to die with full support (OR, 1.59; 95% CI, 1.30-1.94). In documentation of family conferences, nonwhite patients were more likely to have documentation that prognosis was discussed (OR, 1.47; 95% CI, 1.21-1.77) and that physicians recommended withdrawal of life support (OR, 1.57; 95% CI, 1.11-2.21). Nonwhite patients also were more likely to have discord documented among family members or with clinicians (OR, 1.49; 95% CI, 1.04-2.15). Socioeconomic status did not modify these associations and was not a consistent predictor of end-of-life care. Conclusions: We found numerous racial/ethnic differences in end-of-life care in the ICU that were not infl uenced by socioeconomic status. These differences could be due to treatment preferences, disparities, or both. Improving ICU end-of-life care for all patients and families will require a better understanding of these issues.