Do-Not-Resuscitate Orders in the Last Days of Life F. Amos Bailey, M.D., 1,2 Rebecca S. Allen, Ph.D., 3 Beverly R. Williams, Ph.D., 1,2 Patricia S. Goode, M.D., 1,2 Shanette Granstaff, M.P.H., 1,4 David T. Redden, Ph.D., 1,5 and Kathryn L. Burgio, Ph.D. 1,2 Abstract Background: The purpose of this analysis was to describe the presence and timing of do-not-resuscitate (DNR) orders for imminently dying patients in VA Medical Centers, and to examine factors associated with these processes. Methods: Data on DNR orders in the last 7 days of life were abstracted from the medical records of 1,069 veterans who had died in one of six VA hospitals in 2005. Results: Of the 1069 records, 681 (63.7%) had an active DNR order at time of death. Among these, records indicated that the order was written within the last 24 hours for 219 (32.2%), 1–2 days prior to death for 54 (7.9%), 3–7 days prior to death for 256 (37.6%), and > 7 days prior to death for 152 (22.3%). Veterans with a family member present at time of death and those who received pastoral care visits were more likely to have DNR orders. African American veterans and veterans who died unexpectedly were less likely to have DNR orders. Compared with those dying on a general medicine unit, veterans dying in the emergency department or an intensive care unit (ICU) and veterans dying during a procedure or in transit were less likely to have DNR orders. Mental health diagnoses were not associated with presence of a DNR order. Conclusion: Results suggest that the DNR process might be improved by interventions that target ICU settings, facilitate transitions to less intensive locations of care, ensure the involvement and availability of pastoral care staff, and create environments that support the presence of family members. Introduction E nd-of-life decision making is an ongoing process that, in accordance with the Patient Self-Determination Act, 1 may begin in primary care settings with advance care planning discussions between competent adults and their physicians, 2-4 and end in critical care settings with decisions made by physicians with or without patient/surrogate in- put. 5–8 Medical ethicists recommend engaging patients in discussions about end-of-life treatment preferences early in the illness trajectory, particularly to communicate about car- diopulmonary resuscitation (CPR) versus do-not-resuscitate (DNR) orders. 9 Many providers delay this decision-making process, be- lieving the time of death is well in the future and time for these discussions remains. However, medical providers are not very accurate with prognostication, with estimates of time left to live often discrepant by a factor of up to five. 10–12 Discus- sion of death during active medical care can be intrinsically uncomfortable for the provider, patient, and family. As a re- sult, these discussions usually occur very late in the illness, if at all. 13–15 Delays in establishing resuscitation code status has led to futile resuscitation attempts of seriously ill patients. It is common to see DNR orders written in the last hours or days before death, even though resuscitation would predictably not have been beneficial for weeks and even months earlier. When a patient and/or family have agreed to DNR status, physicians may be more comfortable prescribing medications for comfort and nurses less conflicted about administering these medication to patients who are imminently dying. 16,17 Thus, the DNR order can facilitate comfort care interventions and allow a natural death to occur. Further, increasing the rate of DNR orders over time has been associated with decreased prevalence of terminal hospitalizations in the last week of life. 18 Research among community-dwelling patients has found that having a DNR order was related to a preference to die at home, close proximity to death, daily incontinence, sleep 1 Birmingham/Atlanta Geriatric Research, Education, and Clinical Center (GRECC), Department of Veterans Affairs, Birmingham, Alabama and Atlanta, Georgia. 2 Department of Medicine, University of Alabama at Birmingham, Birmingham, Alabama. 3 Department of Psychology/CMHA , University of Alabama, Tuscaloosa, Alabama. 4 Department of Biostatistics, University of Alabama, Tuscaloosa, Alabama. 5 Department of Biostatistics, University of Alabama at Birmingham, Birmingham, Alabama. Accepted February 22, 2012. JOURNAL OF PALLIATIVE MEDICINE Volume 15, Number 7, 2012 ª Mary Ann Liebert, Inc. DOI: 10.1089/jpm.2011.0321 751