Heart failure prognosis: comorbidities matter

Area of Science:

• Cardiology
• Health Services Research
• Medical Informatics

Background:

• Existing heart failure (HF) risk prediction models often use a limited selection of broad comorbidity categories.
• This limits their ability to accurately capture the full spectrum of patient health status.

Purpose of the Study:

• To determine if a comprehensive assessment of individual comorbidities, rather than grouped categories, improves HF prognosis prediction.
• To evaluate the performance of a novel multimorbidity (MM) index against established risk scores.

Main Methods:

• A retrospective, observational study analyzed data from 602,050 US Veterans Affairs heart failure patients (2006-2011).
• The study used a cross-validated approach, with six-month mortality as the primary outcome.
• All International Classification of Disease (ICD) codes for comorbidities were used as independent variables, without prior categorization.

Main Results:

• The novel multimorbidity (MM) index demonstrated superior predictive accuracy, with an area under the receiver-operating curve (AROC) of 0.784.
• The MM index significantly outperformed established indices, including the Quan variant of the Charlson Index (AROC=0.656) and the Elixhauser Index (AROC=0.639).
• Predictive accuracy for chronological age (AROC=0.649) and ejection fraction (AROC=0.533) was also surpassed by the MM index.

Conclusions:

• Incorporating a comprehensive set of individual comorbidities significantly enhances the accuracy of heart failure prognostic models.
• The developed MM index shows promise for improving risk stratification in HF patients.
• Future research should focus on integrating the MM index into clinical practice for HF prognosis.