Study Objectives
High volumes of critically ill patients amidst the COVID19 pandemic prompted the development of crisis standards of care (CSC) to guide resource allocation should demand exceed supply. Racial equity in CSC has been discussed widely. This study explores the utility and racial equity implications of CSC when prioritizing critically ill patients for scarce resources across a major metropolitan area.
Methods
This multi-site retrospective cohort study included patients admitted to an intensive care unit (ICU) within 20 miles of Boston at the height of the first COVID19 surge (April 18-21, 2020). A priority score (1-8) was calculated for each newly admitted patient based on state CSC using the Sequential Organ Failure Assessment (SOFA) or modified SOFA, as well as predicted 1- or 5-year mortality. Attending emergency physicians reviewed charts to determine likely near-term mortality and reported their confidence (0-100) in this decision. Descriptive statistics were used to characterize the study cohort. Fixed effects linear regression was used to model the effect of race on priority score.
Results
Eight hospitals each contributed between 15 and 54 patients for a total cohort size of 257, of whom 130 (50.6%) were white, 52 (20.2%) Black, and 43 (16.7%) Hispanic. Nearly half (47.1%) had a priority score of 1 and 19.8% scored 2, such that 66.9% were in the top priority category for resource allocation. 9.4% had comorbidities indicating likely 5-year mortality, while 14.8% had conditions thought to predict death within 1 year. Reviewers were uncertain about these determinations, with an average confidence of only 48.2-68.0% depending upon the comorbidity. In the fixed effects model, Hispanics had an average priority score 0.81 points lower than whites (95% CI -1.20,-0.45); no difference was found between Black patients (0.3, 95%CI -0.20,0.80) and white patients.
Conclusion
In this diverse, region-wide cohort of critically ill patients, few meaningful racial differences were identified in the prioritization of patients under existing crisis standards of care. Hispanic patients may score slightly better than whites, though this may have little real-world significance. Importantly, physicians who would make CSC-based resource allocation decisions had poor confidence in predicting near-term mortality. This raises concern both for clinician moral injury and the fairness of considering comorbid conditions in CSC.
