Results of validation of the calculator for dynamic risk assessment of heart failure decompensation in patients with low left ventricular ejection fraction

Objective: to develop and validate a calculator program for assessing the risk of acute decompensation of heart failure in outpatients with heart failure with reduced left ventricular ejection fraction (HFrEF). Material and methods. A single-center randomized clinical retrospective-prospective study included 652 patients with HFrEF followed at a Heart Failure Center, aged 62.5 (61; 68.5) years; 545 (83.6 %) were men. Of these, 106 patients were selected by simple random sampling to form the validation group. In this group, monthly telephone interviews were conducted by a paramedic from the chronic heart failure center using a structured questionnaire focused on symptoms of fluid retention. The data were entered into a computer-based calculator program, which generated recommendations (emergency hospitalization, unscheduled cardiology visit, or continued follow-up) displayed on the monitor. Validation was performed by retrospectively comparing the calculator’s recommendations with actual adverse events over one year of follow-up. Cases of emergency hospitalization and unscheduled medical visits were recorded. Results. For identifying indications for unscheduled visits, the calculator showed a sensitivity of 85.5 %, specificity of 99.5 %, and area under the ROC curve (AUC) of 0.925. For identifying indications for any medical contact, sensitivity was 73.5 %, specificity 99.7 %, and AUC 0.866. Overall performance for determining the need for any medical contact was: sensitivity 73.5 %, specificity 99.7 %, AUC 0.866. Conclusions. The developed calculator program effectively identifies cases requiring medical attention in various forms, and its recommendations for unscheduled visits or emergency hospitalization are justified in the vast majority of cases.