Finding the Path to Nosocomial COVID-19

Abstract

Background: Among a multitude of factors contributing to the current COVID-19 pandemic is a high known rate of human-to-human transmission. This effect is magnified in hospitals by the high-risk environment when treating COVID-19 patients. Healthcare systems have attempted to address exposure risk with infection control policies with unknown benefit. Yet few reports describe hospital-acquired COVID-19 (HAC). Examining a select few cases could give insight into the incidence, mechanism, and diagnosis of HAC. Methods: Bulk EHR data was acquired for all patients who tested positive for COVID-19 at our institution from 03/01/2020 to 06/01/2020. Data were analyzed using R v4.0.0 and RStudio v1.2.5042. Positive tests with high probability of nosocomial infections were identified as patients with two initial negative COVID-19 tests who then tested COVID-19 positive no sooner than 48 hours after the initial negative test result along with a documented change in clinical status. Patient charts were then reviewed to examine possible etiologies of nosocomial infection in context of infection control policies. Demographics, hospital stay data, and outcomes were also reviewed. Results: Nine patients were identified on initial data extraction with high likelihood of HAC. Four of nine after chart review had a significant documented clinical change confirming HAC. Six of nine patients (66%) expired during their hospital stay. Five of the six patients became positive on the same unit and floor (non-COVID), with three of these patients during a four-day timespan. Chart review revealed flaws in patient and provider cohorting, PPE allocation, and COVID-19 testing resources. Discussion: The goal of this project was to identify a subset of patients with HAC and systematically map points of failure throughout the hospital course. Key points of failure in the identified patients include lapses in cohorting (no Person Under Investigation stratification, lack of resources to support proper cohorting) and lack of COVID-19 care provider cohorting (ED and ICU personnel treating all categories of patients). Decreased access to PPE also coincided with nosocomial infection in certain units. Limited COVID-19 testing and retesting could have increased exposure. Reviewing database testing and chart review, three parameters were found to have good value for selecting patients with HAC: an initial negative test, a positive test after set number of days, and an associated change in clinical status during the time period of new positive test result. Further studies should be done to determine the criteria with the best characteristics to identify HAC.