Abstract
Background
Readmission rates following hospitalisation for AECOPD are high. The ability to accurately identify patients at a high risk of readmission could help clinicians effectively direct resources and interventions.
Objective
To identify predictors of readmission in patients surviving hospitalisation for AECOPD.
Method
Clinical data from consecutive patients surviving hospitalisation for AECOPD were collected. All variables associated (p<0.10) with outcome (readmission to hospital, or death at home without readmission, within 90 days of discharge) on univariate analysis were entered in to a multivariate logistic regression analysis.
Results
824 patients were recruited: mean (SD) age = 72.3 (10) years; 54.2% were female; mean (SD) FEV1 = 44.0 (17.4) % predicted; and median (IQR) length of stay = 6 (4 to 11) days.
37.3% of patients were readmitted or died within 90-days. The strongest three predictors of outcome were: stable state dyspnoea (measured using the extended MRC Dyspnoea Scale); the number of hospitalisations in the preceding year; and recent unexplained weight loss. The full regression model (table 1) showed good discrimination for 90-day readmission (AUROC = 0.751, 0.717 to 0.783).
Conclusion
Hospital readmission is common and implementation of these simple prognostic indices may help identify and manage those at a high risk of poor outcome.
- © 2012 ERS