Individual patient records from the National Hospital Discharge Survey for 1988 and 1998 comprising approximately 10 million cases were the basis for a binary logistic regression model to predict coronary artery bypass graft. Patterns in 1988 and in 1998 indicated a dramatic and pernicious gender discrepancy in medical decisions involving bypass surgery. Overall, men were twice as likely to have bypass as were women with comparable medical profiles. This result was particularly evident for women in their 40s and 50s, when angina was a primary diagnosis, and in the presence of any comorbid condition (e.g., diabetes or hypertension). Failure to pursue proactive case management of women with heart disease costs women's lives.