Objectives: To determine whether ethnic and other social factors affect how frequently do-not-resuscitate (DNR) orders are written, the timing of DNR orders, or patient involvement in the DNR decision.
Design: Retrospective cohort.
Methods: Patients who died in one urban teaching hospital on the medicine, cardiology, or family practice service during 1988 were eligible; 288 were included in the analyses. Chi-square tests and logistic regression were used to examine frequency of DNR orders and patient involvement; analysis of variance and linear regression were used to examine timing of the DNR orders.
Results: Non-whites were more likely than whites to have DNR orders (OR 1.76; 95% CI, 1.09-2.84) but timing of the DNR order did not vary significantly by race/ethnicity. Patients who spoke English fluently were more likely to be involved in the DNR decision than those who did not (OR 1.28; 95% CI, 1.01-1.61). Patients with documented human immunodeficiency virus were more likely than uninfected patients to have DNR orders (OR 3.51; 95% CI, 1.36-9.02), to be involved in the decision (OR 10.11; 95% CI, 4.87-21.00); and to have DNR orders written earlier (P = 0.02). Alcoholic patients were more likely than non-alcoholics to have DNR orders (OR 1.17; 95% CI, 1.04-1.33).
Conclusions: Ethnic and other social factors do appear to play a role in DNR decisions. It needs to be determined if these differences are due to patient preferences or clinician characteristics.