Assessing the effect of interventions on quality of life in cardiac patients lacks sensitivity because there is no specifically validated scale for measuring depression in cardiac patients. A questionnaire of 35 items (selected for face validity) was given to 246 cardiac outpatients (age 59.3 +/- 14.1 years, 159 male, 87 female). The Beck Depression Scale was then administered, followed by blinded clinical rating of depression. The item scores were subjected to common factor analysis. Internal consistency was assessed using alpha reliability coefficients and clinical validity using Spearman correlation coefficients. The final scale consisted of 26 items (alpha reliability coefficient 0.90) in 2 robust dimensions and 7 subscales. The scale correlated well with clinical rating and with the Beck Depression Scale, but without the marked skewness of the latter. The behavior of the new Cardiac Depression Scale suggests that it will be an excellent measure for studies of outcome in cardiac patients.