Prediction of Hospital Readmission Using the CORE and CORE+ Scores in Persons With COPD

Background: Identifying persons with COPD at high risk for hospital readmission provides opportunities for efficient and appropriate care to lower readmission risk. This study examined 30-d and 60-d hospital readmission prediction of the COPD-readmission (CORE) score and a newly developed CORE+ score. The relationship between CORE and CORE+ scores and ICU admission, endotracheal intubation, and in-hospital noninvasive ventilation (NIV) use was explored. Methods: A retrospective cohort study evaluated participants with spirometry-confirmed COPD from 2 Midwestern academic hospitals. The CORE score variables included eosinophil blood count, FEV₁/FVC (<0.70) and FEV₁ (≤40% of predicted), triple inhaler therapy, previous hospitalization, and presence of neuromuscular disease. Out-of-hospital NIV use and Charlson comorbidity index were added to compose the CORE+ score. Researchers assessed associations between variables and outcomes with chi-square test or Fisher exact test, compared results of CORE and CORE+ scores with Wilcoxon signed-rank test, assessed each score's 30-d and 60-d readmission predictive power with multiple logistic regression, and evaluated predictive accuracy with AUC of receiver operating characteristic using alpha < 0.05. Results: Of 391 participants, the study found a 22% 30-d, all-cause readmission rate and a 16% 60-d, all-cause readmission rate. CORE+ score had better predictive accuracy than the CORE score for 30-d readmission (area under the curve 0.81 [95% CI 0.76-0.86]; AUC 0.73 [95% CI 0.66-0.79], P < .001) and 60-d readmission (AUC 0.77 [95% CI 0.71-0.83]; AUC 0.75 [95% CI 0.69-0.81], P < .001). Participants who used in-hospital NIV had higher median CORE+ scores (P = < .001). Conclusions: CORE and CORE+ scores demonstrated good to very good predictive accuracy for 30-d and 60-d readmission, respectively. Moreover, this study demonstrated a linear relationship between in-hospital NIV use and CORE+ score.