Background: This study aims to identify comorbidities through various sources and assess their short-term impact on relative survival in a cohort of heart failure (HF) patients.
Methods: Newly hospitalized HF patients were identified from hospital discharge abstracts (HDA) of Lombardy Region, Italy, from 2008 to 2010. Charlson comorbidities were assessed using the HDA and supplemented with drug prescriptions and disease-specific exemptions. A Cox model was fit for the one-year relative survival from HF.
Results: The cohort consisted of 51,061 HF patients (53% women; median age 80years). After integrating information from all sources, the prevalence rates of diabetes, chronic pulmonary disease and renal disease were 27.6%, 26.2% and 14.2%, respectively. The prevalence of comorbidity increased to 78%. Survival in the HF cohort was worse with increasing number of comorbidities and was inferior to that in the reference population. Notably, the overall performance of the relative survival models was similar regardless of the strategy used to ascertain comorbidity.
Conclusions: Comorbidities cluster in hospitalized HF patients, and increasing comorbidity burden is associated with worse survival. Integration of a comprehensive search of electronic records to supplement HDA improves the prevalence estimates of comorbidities, although it does not improve discrimination of the risk prediction models.
Keywords: Charlson index; Claims data; Comorbidity; Drug prescriptions; Hospital discharges; Relative survival.
Copyright © 2017 Elsevier B.V. All rights reserved.