80 Association Of COVID-19 “Safer-At-Home” Orders With ED-To-ED Interfacility Transfers

Abstract

Study Objectives: With the reduction in ED volumes during the COVID-19 pandemic, we sought to examine the association of “Safer at Home” lockdown orders with ED-to-ED interfacility transfers. Methods: We conducted a retrospective observational analysis using hospital electronic administrative data of all interfacility ED-to-ED transfers to a single, quaternary care adult ED from January 1, 2018 to September 30, 2020 in Middle, Tennessee. “Safer at Home” orders were issued March 23, 2020 in Davidson County, TN in response to the COVID-19 pandemic mandating citizens stay inside their homes unless engaged in “essential activities.” We defined the post-lockdown period as March 23, 2020 to September 30, 2020. We sought to identify transfers that may not require in-person evaluation and may be amenable to other modalities (eg, telehealth). Called “potentially avoidable transfers,” (PATs), they are defined as ED-to-ED interfacility transfers discharged from the ED or admitted to the hospital for <24 hours without a procedure. To operationalize this definition, we constructed a multivariable logistic regression model to examine whether this lockdown order was associated with higher odds of a transfer being a PAT. We adjusted for seasonality, time since start of the study, patient demographics including age (per 10 years), sex, race, arrival emergency severity index triage acuity, mode of arrival (helicopter vs. ground), timing of arrival (ie, business hours), clinical condition using AHRQ’s clinician condition software (CCS), and rurality. Results: During the study period there were 20,978 ED-to-ED transfers meeting eligibility criteria and, of those, 4,806 (23%) met PAT criteria. In the 7 months post-lockdown PATs were generally down trending when compared month-to-month across years. The first month post-lockdown saw a decrease in transfers and PATs (17% and 28%, respectively) but this was not sustained. In the multivariable model, there was a significant seasonal effect. After adjusting for seasonality, the lockdown was not associated with PATs (adjusted odds ratio [aOR] 0.99, 95% CI 0.2, 5.2). The following were associated with increased odds of being a PAT: lack of insurance and arrival during non-business hours. The following were less likely to be PATs: female sex, increasing age, specific diagnostic categories (digestive, infectious, and circulatory conditions), and arrival by nonambulance. Conclusions: In this single center study, there was no effect of the COVID-19 lockdown orders on PATs. However, after adjusting for covariates and seasonality, we did identify a considerable seasonal effect and an overall downward trend in PATs over time. These findings do not address the appropriateness of the transfer but whether in-person evaluation may be amenable to telehealth or other potential means. Generalizability of this single center study should be examined in other settings along with reasons for the potential downward trend.