Resting Borg score as a predictor of safe discharge of chronic obstructive pulmonary disease from the emergency department observation unit

Abstract

Background: Chronic obstructive pulmonary disease exacerbations (eCOPD) can be life-threatening and costly. Emergency department (ED) observation units (ED-Obs) offer short-term care to safely reduce preventable hospitalizations. Accurately identifying eCOPD patients who can be discharged safely will improve outcomes. Objectives: The objective were to: I) evaluate utility of conventional clinical variables as predictors of safe discharge and II) assess utility of serial resting Borg score and novel Dyspnea Assessment Score (DAS) for identifying eCOPD patients who can be safely discharged from ED-Obs. Methods: This study was carried out in a 680-bed tertiary, academic hospital with >700 annual eCOPD ED encounters and a 16-bed ED-Obs. A two-phase study of eCOPD patients admitted to ED-Obs was performed. Objective I was a retrospective study including all eCOPD admits from April 2016 to May 2017. Predictor variables (demographics, COPD severity, comorbid conditions, exacerbation severity, clinical care in ED) and outcome variables (ED-Obs disposition, ED revisits) were obtained through electronic medical records. Safe discharge was defined as home disposition from ED-Obs without 7-day revisit. A stepwise regression was performed for predictors of safe discharge. Objective II was a prospective observation study for change in every 4-hour serial resting Borg score and DAS as identifiers of safe discharge. Comparative and receiver operating characteristic (ROC) analyses were performed. A p-value of <0.05 was considered significant. Results: In Objective I, 171 patients with age, FEV1%, and body mass index of 59.8 (±9.5) years, 35 (±24)%, and 28.8 (±8) m2/kg were included. After ED-Obs treatment 78 (45.6%) were hospitalized and 93 (54.4%) were discharged home, of whom 11 (6.4%) had 7-day ED revisit. Safe discharge occurred in 82 (48%). None of the predictor variables correlated with safe discharge. In Objective II, of 38 patients included, 20 (52.6%) had safe discharge. Among others, 16 (42%) were hospitalized and two (5.2%) had 7-day ED revisit. The admission Borg scores and DASs were similar in both groups. The predisposition Borg score was significantly lower in patients with safe discharge (2.75 vs. 5.28, p < 0.001) and had the highest area under curve on ROC (0.77) for safe discharge. DAS was not significantly different between groups. Conclusions: Routine clinical variables do not identify eCOPD patients who can be safely discharged from ED-Obs. Change in resting Borg score during the course of ED-Obs treatment safely identifies patients for discharge. Prospective, external validation is needed to incorporate serial Borg scores in ED-Obs disposition decision for improved safety.