Methods: This study is a primary analysis of a mortality database that is used in tracking patients who expired in the medicine wards and intensive care units from October 2012 to May 2016 at the Ronald Reagan and Santa Monica UCLA Medical Centers. The data elements were obtained through chart abstraction as well as during the RMR meetings with the primary medical team. Descriptive statistics and logistic regression was performed. The study sample is 498 patients, 299 males (60.0%) and 199 females (40.0%), with a mean age of 69.6 (range 21-105, SD 16.51). The dependent variable is the identifiable ACP issue and covariates include demographic and clinical characteristics of the sample, such as, age and length of stay, primary and secondary reason for mortality.
Results: In a sample of 498 mortality cases, issues with ACP were identified in 112 cases (22.49%). Goals of care discussions took place at multiple outpatient and inpatient settings: outpatient (12.7%), hospital admission (50.0%), clinical deterioration (62.7%), and the day of death (18.7%). From the total sample, 131 (26.3%) were Do Not Resuscitate/Do Not Intubate (DNR/DNI) at the time of admission and 422 (84.7%) were DNR/DNI at the time of death. In addition, 352 (70.7%) were placed on comfort care with 196 (39.4%) receiving palliative care consultation.
The results of logistic regression examined characteristics of mortality cases that are more likely to have ACP issues. Strong associations were seen for female gender (OR: 1.601; 95% CI: 1.011-2.536; p=.045), longer length of stay (OR: 1.014; 95% CI: 1.003-1.025; p=.016) and primary reason for mortality infection (OR: 2.02; 95% CI: 1.144-3.567; p=.015) or acute medical (p=.04). All other variables were not significantly associated with ACP issues, including older age, transfer patient, preventability, UCLA patient with primary care physician, goals of care discussion in the outpatient setting, DNR/DNI at admission and time of death, and palliative care consult.
Conclusion: The findings of this primary data analysis demonstrate some characteristics of mortality cases that will more likely have ACP issues. Gender (being female), length of stay (longer), and primary reason for mortality (acute and infection) were all strongly associated with the cases that had ACP issues. Further study could investigate the etiology of the gender association, whether improved ACP reduces lengths of stay, and the potential benefit of more universal ACP.