Randomized controlled trial of a structured intervention to facilitate end-of-life decision making in patients with advanced cancer.

Abstract

This study tested the efficacy of an intervention on end-of-life decision making for patients with advanced cancer. One hundred twenty patients with metastatic cancer who were no longer being treated with curative intent (and 87 caregivers) were randomly assigned to the intervention (n = 55) or treatment as usual (n = 65). Primary outcome measures were the proportion of patients with do-not-resuscitate (DNR) orders, timing of DNR orders, and place of death. Secondary outcome measures were completed at study enrollment, 3 weeks later, and 3 months later, including patients' knowledge, mood, and caregiver burden. High, but equivalent, rates of DNR orders were observed in both groups. In per-protocol analyses, DNR orders were placed earlier for patients who received the intervention (median, 27 v 12.5 days; 95% CI, 1.1 to 5.9; P = .03) and they were more likely to avoid a hospital death (19% v 50% (95% CI, 11% to 50%; P = .004). Differences between the groups over time were evident for estimates of cardiopulmonary rehabilitation (CPR) success rates (P = .01) but not knowledge of CPR (P = .2). There was no evidence that the intervention resulted in more anxious or depressive symptoms. Caregivers experienced less burden in terms of disruption to schedule if the patient received the intervention (P = .05). An intervention, consisting of an informational pamphlet and discussion, was associated with earlier placement of DNR orders relative to death and less likelihood of death in hospital. There was no negative impact of the intervention on secondary outcomes, although the sample may have been too small to detect differences.