Edinburgh Napier University

  Aims: Heart failure (HF) with preserved ejection fraction (HFpEF) poses significant diagnostic challenges due to its complex aetiology and overlapping symptoms with other HF types. The heterogeneity of HFpEF, compounded by frequent comorbidities, complicates diagnosis. This study aimed to enhance HFpEF prediction through a two‐phase approach: a simplified risk score and a decision tree model. Methods and results: In Phase 1, an 8‐point risk score based on accessible clinical parameters was developed. In Phase 2, we conducted comprehensive predictive modelling using decision tree analysis. Data from 560 HF patients were analysed. It achieved an accuracy of 63.13% (sensitivity: 62.87%, specificity: 54.24%). In Phase 2, a decision tree model using broader clinical variables improved accuracy to 73.04% (sensitivity: 53.89%, specificity: 81.17%). Conclusions: This dual framework provides tools for both quick screening and detailed risk stratification in various clinical settings.