Palliative Care Consultations in Trauma Patients and Role of Do-Not-Resuscitate Orders: Propensity-Matched Study

Abstract

Objective: To delineate characteristics of trauma patients associated with a palliative care consultation (PCC) and to analyze the role of do-not-resuscitate (DNR) orders and related outcomes. Methods: Retrospective study included 864 patients from 2 level one trauma centers admitted between 2012 and 2019.  Level 1 trauma centers are designated for admission of the most severe injured patients. Palliative care consultation group of 432 patients who received PCC and were compared to matched control (MC) group of 432 patients without PCC. Propensity matching covariates included Injury Severity Score, mechanism of injury, gender, and hospital length of stay (HLOS). Analysis included patient demographics, injury parameters, intensive care unit (ICU) admissions, ICU length of stay (ICULOS), duration of mechanical ventilation, timing of PCC and DNR, and mortality. Palliative care consultation patients were further analyzed based on DNR status: prehospital DNR, in-hospital DNR, and no DNR (NODNR). Results: Palliative care consultation compared to MC patients were older, predominantly Caucasian, with more frequent traumatic brain injury (TBI), ICU admissions, and mechanical ventilation. The average time to PCC was 5.3 days. Do-not-resuscitate orders were significantly more common in PCC compared to MC group (71.5% vs 11.1%, P < .001). Overall mortality was 90.7% in PCC and 6.0% in MC ( P < .001). In patients with DNR, mortality was 94.2% in PCC and 18.8% in MC. In-hospital DNR-PCC compared to NODNR-PCC patients had shorter ICULOS (5.0 vs 7.3 days, P = .04), HLOS (6.2 vs 13.2 days, P = .006), and time to discharge (1.0 vs 6.3 days, P = .04). Conclusions: Advanced age, DNR order, and TBI were associated with a PCC in trauma patients and resulted in significantly higher mortality in PCC than in MC patients. Combination of DNR and PCC was associated with shorter ICULOS, HLOS, and time from PCC to discharge.