Performance of crisis standards of care guidelines in a cohort of critically ill COVID-19 patients in the United States

Abstract

Many US states published crisis standards of care (CSC) guidelines for allocating scarce critical care resources during the COVID-19 pandemic. However, the performance of these guidelines in maximizing population benefit has not been well tested. In 2,272 adults with COVID-19 requiring ICU admission drawn from the STOP-COVID multicenter cohort, we tested three approaches to CSC algorithms: SOFA scores grouped into ranges, SOFA score ranges plus comorbidities, and a hypothetical approach using raw SOFA scores not grouped into ranges. We found that area under receiver operating characteristic (AUROC) curves for all three algorithms demonstrate only modest discrimination for 28-day mortality. Adding comorbidity scoring modestly improves algorithm performance over SOFA scores alone. The algorithm incorporating comorbidities has modestly worse predictive performance for Black compared to White patients. CSC algorithms should be empirically examined to refine approaches to the allocation of scarce resources during pandemics and to avoid potential exacerbation of racial inequities.

Health Professional vs Layperson Values and Preferences on Scarce Resource Allocation

“…38 Exploration of the potential for racial bias in these algorithms has been undertaken in limited degree, demonstrating a possible difference in model performance across racial groups. 39 Ongoing concerns about how minoritized communities may be disproportionately affected are raised throughout recent literature and our findings continue to demonstrate lack of consensus in both SRA policy and among critical care resource use in general. [40][41][42][43] While determination of how racial bias should be mitigated was not directly addressed in our study, we did find that both HCPs and laypersons felt less strongly about deprioritizing people with a shorter expected lifespan if a patient were also a member…”

Simulation of New York City’s Ventilator Allocation Guideline During the Spring 2020 COVID-19 Surge

“…In this simulation study, a crisis period requiring ventilator triage per NYVAG was defined as once ventilator demand exceeded 95% prepandemic ventilators (250 patients) . Patients’ SOFA scores were queried from the health record while ventilated, with missing SOFA subscores presumed normal (ie, scored as 0) . Race and ethnicity were included in study data because of growing concern that existing crisis standards of care protocols might worsen existing social disparities; data were based on self-report by the patient or identified unknown if unable or unwilling to self-identify.…”

“…[12][13][14][15] Despite these potential shortcomings, NYVAG is operationally similar to many other current guidelines. 7,9,[16][17][18] While other studies may have simulated NYVAG [19][20][21][22][23][24][25][26][27][28] or other CSC, [29][30][31][32] most perform a limited analysis of how resources might be allocated to patients not already receiving that resource (front-end triage) and do not consider potential reallocation after a time trial (back-end triage).…”

Treating COVID-19: Evolving approaches to evidence in a pandemic

“…Despite the significant scientific advances during this time, entire domains of inpatient care remained under-investigated, such as methods to prevent in-hospital transmission (e.g., the efficacy of negative airflow rooms) or algorithms to ration scarce resources during crisis standards of care and their potential to exacerbate racial, ethnic, or socioeconomic inequities. 14 …”