We performed a chart review study to determine whether physicians provided less vigorous treatment for elderly patients with breast cancer than for younger patients. Seven hospitals, varying with respect to size, medical school affiliation, and patient socioeconomic status participated in this study. Patterns of breast cancer care were compared with a set of branching logic standards in the form of a criteria map. A chart-based comorbidity index was developed to control for the influence of coexisting diseases on cancer management. Logistic regression analysis of 374 breast cancer cases revealed that both patient age and comorbidity status significantly and independently affected treatment. Among stage I and II patients with no or very mild comorbid disease, 96% of those aged 50 to 69 years received appropriate surgery vs 83% of patients aged 70 years and older. We conclude that physicians may manage patients with this highly treatable disease according to chronologic age without regard for physiological condition and that this age bias may result in a less favorable prognosis than could be achieved using currently recommended therapy.