Predictors of short-term COPD readmission

Abstract

COPD readmissions have a great impact on patients' quality of life and mortality. Our goal was to identify factors related to 60-day readmission. We conducted a prospective observational cohort study with a nested case-control study, with 60 days of follow-up after the index admission. Patients readmitted were matched, by age, baseline forced expiratory volume in 1 s and month at admission, with patients admitted in the same period but not readmitted at 2 months. Data were collected on sociodemographic and clinical characteristics and health-related quality of life data at the index admission and events from discharge to readmission within 60 days. Conditional logistic (60-day readmission) and Cox (days to readmission) regression models were constructed. Both multivariable analyses identified the following as predictors: any admission in the preceding 2 months (OR: 2.366; HR: 1.918), hematocrit at ED arrival ≤ 35% (OR: 2.949; HR: 1.570), pre-existing cardiovascular disease (valvular disease or myocardial infarction) (OR: 1.878; HR: 1.490); NIMV at discharge (OR: 0.547; HR: 0.70); no appointment with a specialist after discharge (OR: 5.785; HR: 3.373) and patient-reported need for help at home (OR: 2.978; HR: 2.061). The AUC for the logistic model was 0.845 and the c-index for the Cox model was 0.707. EuroQol EQ-5D score before the admission was correlated with a lower risk of readmission (OR: 0.383; HR: 0.670). As conclusions, we have identified factors related to 60-day readmission and summarized the findings in easy-to-use scoring scales that could be incorporated into the daily clinical routine and may help establish preventive measures to reduce future readmissions.Registration: Clinical Trial Registration NCT03227211.

Keywords: Chronic obstructive; Clinical prediction rule; Cohort studies; Hospital readmission; Nested case–control study; Pulmonary disease.