The CONFIDENT machine learning prognostic model for all-cause mortality outperformed the PREDICT-HFpEF score in the validation cohort (C-index 0.72 vs 0.67; P=0.036).
Observational (n=1,208)
Yes
Does the CONFIDENT machine learning model improve prediction of all-cause mortality and HF hospitalization compared to existing risk scores in patients with HFpEF?
A machine learning-based prognostic model using multimodal real-world data outperformed existing risk scores in predicting mortality and HF hospitalization in patients with HFpEF.
Effect estimate: C-index 0.72 (95% CI 0.65-0.78)
Absolute Event Rate: 0.72% vs 0.67%
p-value: p=0.036
Abstract Aims Heart failure with preserved ejection fraction (HFpEF) is a heterogeneous condition with high morbidity and mortality. Accurate risk stratification is important for advancing drug development and improving clinical care. Methods and Results CONFIDENT is an observational, multi-cohort study across three centers in Europe and the US. Patients with HFpEF, according to the HFA-PEFF criteria with ≥ 2 years of follow-up, were included from 2013 to 2022. Data include electronic health records, lab tests, echocardiography, and electrocardiography. We developed machine learning-based prognostic models to predict all-cause mortality and heart failure (HF) hospitalization. Model performance was compared to validated risk score and validated in an external cohort. A total of 1208 patients were included in the study. The mean age was 72±12 and the mean BMI 32±9 kg/m2. The 2-year risk of HF hospitalization and all-cause mortality ranged from 13 to 44% and 9 to 19%, respectively. The all-cause mortality prognostic model achieved fair discrimination with a C-index of 0.68 95% CI 0.62-0.74, and 0.71 95% CI 0.64-0.78 in the training cohorts, and a good discrimination of 0.72 95% CI 0.65-0.78 in the validation cohort, but performed better than the PREDICT-HFpEF score (C-index: 0.66 95% CI 0.54-0.72, p-value = 0.006; 0.65, 95% CI 0.55-0.72, p-value 0.001 and 0.67 95% CI 0.59-0.73, p-value = 0.036, respectively). Similar results were observed when compared to the Meta-Analysis Global Group In Chronic Heart Failure Risk Score (MAGGIC). The HF hospitalization model also outperformed both comparators, including MAGGIC + natriuretic peptide. Conclusion CONFIDENT prognostic models for all-cause mortality and HF hospitalization using routinely collected variables can reliably predict outcomes and potentially facilitate personalized care and trial recruitment strategies in HFpEF.
Fudim et al. (Thu,) conducted a observational in Heart failure with preserved ejection fraction (HFpEF) (n=1,208). CONFIDENT machine learning-based prognostic models vs. PREDICT-HFpEF score and MAGGIC score was evaluated on All-cause mortality (C-index 0.72, 95% CI 0.65-0.78, p=0.036). The CONFIDENT machine learning prognostic model for all-cause mortality outperformed the PREDICT-HFpEF score in the validation cohort (C-index 0.72 vs 0.67; P=0.036).