| Summary: | Background: Risk stratification allows guiding treatment for patients with heart diseases. We aimed to provide clinical rules for identifying patients at high risk for mortality owing to heart failure. Methods: The research was a secondary analysis of data from a community cohort study that had been conducted from October 2007 to December 2011 in the United States. In total, 425 adult patients (180 women) with heart failure had been included. The mean (SD) age was 73·5 (13·1) years, and participants had been followed for 2 years. The outcome was mortality and the predictors were a variety of demographic, clinical, and laboratory variables. Results: We recognized two clusters of patients in the sample with different survival probabilities (P < 0·001) and defined clinical rules for identifying patients at high risk for mortality. Association rule mining showed that diabetes was a strong solitary risk for mortality in heart failure patients (support = 38·4%, confidence = 100%, lift = 1·9). Also, combinations of categories in age, sex, systolic blood pressure, estimated glomerular filtration rate, left ventricular ejection fraction, serum sodium, blood urea nitrogen, the cause of heart failure (ischemic vs. non-ischemic), diabetes mellitus, and increase in natriuretic peptide were able to identify patients at risk for death. Conclusion: Different approaches to risk stratification provide distinct sets of clinical pictures. We presented the results as sets of clinical pictures accessible to all healthcare professionals. The clinical pictures support recognizing patients with heart failure who are at high risk for mortality.
|
|---|