Mammography screening continues to be under-utilized, especially among women from lower socioeconomic groups. In order to determine whether having direct access to health care services has an effect on mammography use among low income women, we conducted a randomized trial of two alternative letter reminders among 1,717 women who were enrolled at two locations of a multi-site inner city health department in Detroit. All participants were 39(1/2) years of age and older and were due for a screening mammogram at randomization. A physician-directed reminder form was placed in each of the participant's medical records at the beginning of the study. In addition participants were randomized to receive either a letter directing them to visit their primary care physician, a letter directing them to contact the clinic directly to schedule a mammogram, or no letter. Study participants were predominantly African-American, two-thirds of whom were over age 50, and who had minimal health insurance coverage. During the intervention year, mammograms were completed by 179 out of 967 study women at site one (18.5%), and 90 out of 750 study women at site two (12%). A multivariate model controlling for the simultaneous effect of age, insurance type, visit history and past mammography use, showed no significant independent effect of either type of letter reminder on mammography completion during the study year. In conclusion, letters targeted at women due for screening mammograms did not have a beneficial effect on mammography utilization above and beyond that of a physician medical record reminder.