Abstract 16663: Psychometric Summary of the Heart Failure Symptom Survey

Abstract

Introduction: The Heart Failure Symptom Survey (HFSS) was adapted from the Cardiac Symptom Survey to measure common symptoms experienced by heart failure (HF) patients. The HFSS assesses four components (frequency, severity, interference with physical activity and enjoyment of life) for 14 symptoms including shortness of breath at rest, with activity or when lying down or during the night; extremities swelling; bloating; fatigue; chest pressure; irregular heartbeat; worsening cough; dizziness; difficulty sleeping; forgetfulness; and feeling down. To date the formal psychometric analysis and validation of the HFSS has not been published. Hypothesis: Determine reliability and validity of the HFSS using baseline data from multiple HF studies. Methods: Data were merged from several HF studies from four research universities. The HFSS items were summed to compute scores for symptom frequency, severity, interference with physical activity, interference with enjoyment of life, and a total score of all 4 components. Internal consistency was evaluated using Cronbach’s alpha for reliability as well as computing the number of factors from exploratory factor analysis Scree test. Initial external validity was explored by comparing New York Heart Association (NYHA) class to the four component and overall total scores. Results: For the 623 patients in the merged dataset, ages ranged from 22 to 99 with a mean of 65.0 +/- 13.8; 61.0% were male; 69.0% were White, 28.9% were African American; 88.4% had reduced ejection fraction &lt=40%; 41.0% were NYHA class II and 33.3% were class III. Internal reliability estimates for each component were: frequency (Cronbach’s alpha=0.892), severity (alpha=0.913), interference with physical activity (alpha=0.915), and interference with enjoyment of life (alpha=0.921) and overall total (alpha=0.976). For each component and the total scale, a single factor was indicated by the Scree test. The four component and total scores were all correlated moderately with NYHA class (Spearman’s rho, 0.337 to 0.352). Conclusions: The HFSS showed high reliability across multiple HF samples and positive association with NYHA class reflecting validity in relation to heart failure clinical severity useful for tailoring symptom management.