There is growing evidence that stress contributes to cardiovascular disease. Chronic stress contributes to the atherosclerotic process through increased allostatic load, which is mediated by the neuroendocrine and immune systems (sympathetic nervous system and hypothalamus-pituitary adrenal axis) and related chron-

Digital echocardiography makes computer‐processed estimates of ejection fraction feasible for clinical use but increases physician reading time. The practicabilities were examined by three novice fellows and four experienced attendings. Ejection fraction was visually estimated from playback of videotape or cine‐loop displays. Ejection fraction was also estimated from single tracings of endocardium, digitized and applied to biplane Simpson's rule, and expressed in whole units. Differences between fellows' and attendings' visual estimates were close to 0 ± 6.4 median standard deviation. The 95% confidence intervals for reproducing visual and computer‐processed ejection fractions ranged from 15% to 46% of mean ejection fraction; for comparing the two methods, from 7% to 36%. Intraobserver reading errors varied widely and with one observer, systematically, and were independent of experience, but dependent on the quality of signals. Computer‐processed readings of ejection fraction should be reserved for images of reasonable quality and for confirming visually estimated ejection fractions between the lower limits of normal (45%‐50%) to moderately severely depressed (25%‐30%), when accuracy is clinically relevant or when a serial change is at the confidence limits of the reader and needs verification. (ECHOCARDIOGRAPHY, Volume 8, January 1991)