Use of the ADHERE Risk Model as a Predictor of Risk of in-Hospital Worsening Heart Failure in a Cohort

Background: Patients admitted with acute decompensated heart failure (HF) are subject to developing worsening episodes that require more complex interventions. The Acute Decompensated Heart Failure National Registry (ADHERE) risk model was developed in the United States to predict the risk of in-hospital worsening HF. Objective: To use the ADHERE risk model in the assessment of risk of in-hospital worsening HF and to determine its sensitivity and specificity in hospitalized patients. Methods: This cohort study was conducted at a Brazilian public university hospital, and data from 2013 to 2020 were retrospectively collected. P values < 0.05 were considered statistically significant. Results: A total of 890 patients with a mean age of 74 +/- 8 years were included. The model showed that, in the group of 490 patients at risk, 254 (51.8%) developed in-hospital worsening HF. In the group of 400 patients not at risk, only 109 (27.2%) experienced worsening HF. The results demonstrated a statistically significant curve (area under the curve = 0.665; standard error = 0.018; P < 0.01; confidence interval = 0.609 to 0.701), indicating good accuracy. The model had a sensitivity of 69.9% and a specificity of 55.2%, with a positive predictive value of 52% and a negative predictive value of 72.7%. Conclusions: In this cohort, we showed that the ADHERE risk model was able to discriminate patients who in fact developed worsening HF during the admission period, from those who did not.