Family involvement in end-of-life hospital care

OBJECTIVES: To examine whether the end-of-life treatment provided to hospitalized patients differed for those who had a family member present at death and those who did not. DESIGN: A retrospective cohort analysis. SETTING: An urban community hospital. PARTICIPANTS: All 370 inpatients who died during a 1-year period. MEASUREMENTS: Medical records were examined for whether life-support treatments were provided or withdrawn, occurrence and timing of do-not-resuscitate (DNR) orders, and use of comfort measures such as narcotics and sedation. RESULTS: DNR orders were written for 85% of patients. For patients who had a DNR order written, the average time from the DNR order to death was 4.8 days. Only 26% of patients had one or more treatments withdrawn. Sixty-seven percent of patients received narcotics before death, and 22% received sedatives. Patients aged 75 and older and African Americans were less likely to have a family member present at death. After adjusting for age and ethnicity, patients who had a family member present at death were more likely to have DNR orders written, to have treatments withdrawn, and to receive narcotics before death. Patients with a family member present at death also had a shorter time to death after DNR orders were written. CONCLUSION: The presence of a family member at death appears to be an indirect measure of family involvement during patients' hospitalization. Family involvement before death may reduce the use of technology and increase the use of comfort care as patients die.