Implementation of Uniform Processes of Care Across an Academic Health System During the Initial COVID-19 Pandemic Surge and Their Association with Outcomes

Introduction Disparities in care delivery and outcomes are common in healthcare in the United States. The SARS-CoV-2 pandemic in the spring of 2020 in the United States and around the world resulted in a surge in the need for acute and critical care services for patient with acute respiratory disease. Many individual hospitals and health systems were unprepared for this surge of patients with a novel and acute respiratory disease which may have exacerbated pre-existing disparities. To prepare for this challenge the Yale New Haven Health System developed a response to the SARS-CoV-2 pandemic in 2020 which was multifactorial including: 1) Implementation of a uniform COVID management protocol across the care continuum, 2) Precise criteria for hospital and Intensive Care Unit (ICU) and Stepdown Unit (SDU) admission, 3) Augmented ICU and SDU bed availability, 4) Implemented load balancing across the entire health system. To understand the impact of these interventions we reviewed and compared mortality across the Yale-New Haven Health System both between hospitals and to national data. We also analyzed administration of medications to understand local adherence to the COVID-19 management protocol implemented during the initial wave of the pandemic.

Methods This investigation is an observational, retrospective study of 3,112 patients infected with SARS-CoV-2 during the first wave of the pandemic in southern Connecticut and Rhode Island. All COVID-19 admissions to the Yale New Have Health System from March through June of 2020 were included. Patients all received care at a hospital within the Yale New Haven Health System which has 2693 beds across 7 campuses in southern Connecticut and Rhode Island. The primary outcome was in-hospital mortality for patients with COVID-19. Demographics were extracted as well as specific data associated with process of care including timing of administration of Tocilizumab, aspirin, and corticosteroids. Transfers between hospitals within the health system were identified. Mortality rates were compared between the central tertiary care hospital and the smaller community and community teaching hospitals using logistic regression to adjust for patient factors.

Results Analysis of process of care metrics including time to Tocilizumab, aspirin, and corticosteroids shows adherence of recommended processes of care across Yale New Haven Health System. The overall mortality rate of 15.9% was lower than published national comparators. Hospital mortality rates compared between the central tertiary care center and smaller hospitals within the system were similar when adjusted for multiple patient factors including race and ethnicity.

Conclusions In this investigation of COVID-19 outcomes in an academic health system with geographic and social diversity, we find that the observed low mortality rate was consistent across the health system. We propose that this is in part related to consistency of care and structural factors such as load balancing. We believe that these findings highlight the potential value of implementing uniform processes designed to reduce noise and bias in clinical judgment.