The Psychometric Properties of the Brief Symptom Inventory Depression and Anxiety Subscales in Patients With Heart Failure and With or Without Renal Dysfunction

More than 5 million Americans have heart failure (HF); approximately one third have concomitant renal dysfunction. Anxiety and depressive symptoms are the most common psychological responses of these individuals and may influences outcomes; thus a reliable valid instrument to measure these is needed. This article reports a psychometric evaluation of the Brief Symptom Inventory (BSI) depression and anxiety subscales in patients with HF and with or without renal dysfunction, as these scales are commonly used in this population for research studies. This rigorous psychometric analysis used existing data from 590 patients with HF with an average ejection fraction of 35% ± 15% and average age of 63 ± 13 years. Patients were categorized as normal renal function (n = 495) or renal dysfunction (n = 95), and groups were compared and analyzed separately. Cronbach's alpha for the BSI subscales was .82 for those with normal renal function and .88 for those with renal dysfunction. Factor analysis determined that the subscales evaluated one dimension, psychological distress, in both groups. Construct validity was examined using hypothesis testing, and construct validity was supported in patients with HF and with normal renal function by significant associations of the BSI subscales with another measure of depression and a measure of perceived control. Construct validity in patients with HF and renal dysfunction was not strongly supported. Only the BSI depression subscale predicted poorer outcomes in patients with HF and with normal renal function; neither subscale was associated with event-free survival at 12 months in those with renal dysfunction. The BSI anxiety and depression subscales provide reliable and valid data in patients HF and normal renal function. Although reliability is excellent, construct validity was weak in those patients with HF and with concomitant renal dysfunction, which may reduce the validity of those data.