A bilateral oophorectomy at the time of elective hysterectomy is often performed to prevent ovarian cancer. The assumption that endogenous estrogen can be easily replaced with supplemental medication fosters the decision for routine oophorectomy. Published reports on the use of postmenopausal estrogen indicate that compliance is less than perfect. This fact could affect the overall outcome. Decision analysis techniques with Markov cohort modeling were used to evaluate the policy of elective bilateral oophorectomy. Results from studies judged methodologically sound were combined to determine values representing the influence of estrogen on coronary heart disease, breast cancer, and osteoporotic fracture. The decision tree also explicitly incorporated patient compliance. When compliance with estrogen therapy is assumed to be perfect, oophorectomy yields longer life expectancy than retaining the ovaries. When actual drug-taking behavior is considered, retaining the ovaries results in longer survival. This analysis highlights the importance of including the effects of patient compliance with treatment recommendations when the impact of a health policy decision such as prophylactic surgery is assessed.