63 Incidence and Determinants of COVID-19 Emergency Department Revisits

ABSTRACT

Study Objectives: Emergency department (ED) revisits are associated with significant resource utilization. Accordingly, revisits serve as an important quality measure for emergency care. In recent times, EDs have been challenged by critical resource constraints in the setting of the COVID-19 pandemic. When appropriate, medically stable COVID-19 patients are discharged home rather than admitted for further care. However, the natural history of COVID-19 is not well understood and patients may quickly progress to requiring medical attention. To our knowledge, ED revisits have not been previously characterized in the setting of COVID-19. We aim to quantify the incidence of, as well as determine risk factors for, ED revisits for COVID-19 patients. Methods: We conducted retrospective study of 323 reverse-transcription polymerase chain reaction-confirmed COVID-19 patients who presented to a single academic tertiary-care institution from March 15 to April 15 of 2020. Demographic and clinical information was abstracted from the electronic medical record. Predictor variables (age, history of hypertension, diabetes, asthma, chronic obstructive pulmonary disease, current tobacco or marijuana use) were selected based on current knowledge of risk factors for severe COVID-19 illness. All return visits to the ED within 28 days of index ED presentation were classified as revisits. Multivariable logistic regression models were used to identify independent demographic and clinical risk factors for ED revisits. We also performed exploratory univariable analyses of a subset of 179 patients who had measured serum biomarkers (absolute neutrophil count (ANC), alanine aminotransferase (ALT), ferritin, C-reactive protein, D-dimer, lactate dehydrogenase (LDH)) in order to identify potential biochemical risk factors for ED revisits. Results: Of the 323 patients studied, 98 were discharged from the ED during their index visit and 225 were admitted to the hospital. Among those discharged, 25/98 (25.5%) returned within 28 days of index ED presentation. Median time to revisit was 3 days (interquartile range (IQR) 2 to 7). Among those admitted during their index visit (median hospital length of stay 6 days), 26/225 (11.6%) returned within 28 days of index ED presentation. Median time to revisit for this group was 14.5 days (IQR 5 to 22). Cumulative incidence of ED revisits was 15.8% (95% CI 12.2 to 20.2). Patients with and without ED revisits were similar across demographic and clinical variables examined, with the exceptions of tobacco or marijuana use and history of COPD. Both tobacco or marijuana use (odds ratio (OR) 2.9, 95% CI 1.1 to 7.6) and history of COPD (OR 3.1, 95% CI 1.1 to 8.8) were found to be independent risk factors for ED revisits. In our exploratory analysis of patients with biomarker data, ANC (OR 0.808, 95% CI 0.689 to 0.948), ALT (OR 0.973, 95% CI 0.953 to 0.993), and LDH (OR 0.996, 95% CI 0.992 to 0.999) were found to be associated with ED revisits. Conclusion: The incidence of ED revisits in our COVID-19 cohort was 15.8% (95% CI 12.2 to 20.2). Risk factors for revisits included current tobacco or marijuana use and history of COPD. Preliminary study suggests the utility of serum biomarker data in helping to stratify revisit risk. In future analysis we will determine the reasons for ED revisits as well as develop a model for identifying those at risk for ED revisits.