Impact of the COVID-19 pandemic on the care of hospitalised patients with do-not-resuscitate orders: a retrospective cohort study.

OBJECTIVES: Do-not-resuscitate (DNR) orders limit the use of cardiopulmonary resuscitation but are often interpreted more broadly. During COVID-19 surges, resource constraints may have increased variability in application, potentially affecting patients with DNR orders. We aimed to determine whether admission during COVID-19 surge months in 2020 versus non-surge months was associated with differences in use of life-sustaining treatments among hospitalised patients with DNR orders. METHODS: Using data from five states (Arizona, California, Iowa, Mississippi and New York), we identified non-surgical patients ≥40 years with early-DNR orders (within 24 hours of admission) admitted during COVID-19 surge versus non-surge months in 2020. In this retrospective cohort study, we determined patient-level odds of receiving mechanical ventilation (MV), renal replacement therapy (RRT) or central venous catheterisation (CVC) during COVID-19 surge versus non-surge periods and differences in hospital-level rates of each treatment. RESULTS: We identified 191 317 (13.3%) patients with early-DNR orders during COVID-19 surges and 149 022 (12.6%) in non-surge periods. Patient-level utilisation of MV, RRT and CVC was higher during surge periods (MV: 7.7% vs 5.9%, RRT: 4.1% vs 3.8%, CVC: 2.4% vs 2.0%) with similar results after accounting for case-mix differences. The range of risk-adjusted hospital rates for CVC and RRT was similar across time periods but higher for MV during surges. CONCLUSIONS: Resource strains did not lead to more restrictive care for patients with early-DNR orders during COVID-19 surges. Rather, higher rates of MV, RRT and CVC during surge periods likely reflect greater illness severity. The increased prevalence of early-DNR orders during surges warrants further investigation.