POLST-Discordant Intensive Care Near the End of Life: A Retrospective Cohort Study

Importance: Patients with chronic life-limiting illness frequently use the Physician Orders for Life Sustaining Treatment (POLST) to document preferences for limiting intensive care. However, the incidence and predictors of POLST-discordant care near the end of life are not known. Objective: To describe the incidence, predictors, and intensity of POLST-discordant intensive care among patients hospitalized near the end of life. Design: Retrospective cohort study over 2010-2015. Setting: Two teaching hospitals in a large academic healthcare system. Participants: All POLST users with chronic life-limiting illness hospitalized during the last six months of life who died during 2010-2015. Exposure(s): Age, days from POLST completion to admission, history of cancer, history of dementia, admitting diagnosis of traumatic injury, and whether the patient signed his/her own POLST. Main Outcome(s) and Measure(s): POLST-discordant intensive care was defined as admission to the intensive care unit or cardiopulmonary resuscitation (CPR) among patients whose preexisting POLST orders that would ordinarily preclude such care. Among recipients of POLST-discordant intensive care, we also measured life-sustaining treatments (invasive mechanical ventilation, vasoactive infusions, new renal replacement therapies, CPR). All outcomes were measured during the last hospitalization of life. Results: We identified 1,201 eligible decedents (mean [SD] age, 70.0 [14.6] years; 41% female), of whom 268 had POLST orders for “comfort measures only” (CMO) and 476 had orders for “limited additional interventions” (LAI). Among patients with treatment-limiting POLST orders, the incidence of POLST-discordant intensive care was 36% (95%CI 33-40%) [CMO: 28%, 95%CI 23-34%) ; LAI: 41%, 95%CI 37-46%], and the incidence of POLST-discordant life-sustaining treatment was 16% (95%CI 13-18%). Compared to patients without cancer, patients with cancer were less likely to receive POLST-discordant care (CMO: RR 0.51, 95%CI 0.32-0.80; LAI: RR 0.71, 95%CI 0.55-0.91). In the comfort-only group, patients with dementia were less likely to receive POLST-discordant care than those without dementia (RR 0.44, 95%CI 0.26-0.76), and patients admitted for traumatic injury were more likely to receive POLST-discordant care than those admitted for other diagnoses (RR 1.74, 95%CI 1.13-2.69). Patient age, timing of POLST, and POLST signer (patient vs. surrogate) were not independently associated with POLST-discordant care. Conclusions and Relevance: In an academic healthcare system, one in three patients with treatment-limiting POLST orders received POLST-discordant intensive care near the end of life. Patients with cancer or dementia were less likely to receive POLST-discordant care than those without these illnesses. Patients hospitalized for traumatic injury were more likely to receive POLST-discordant care. Our findings may guide future improvements in the implementation of the POLST.