In a one-year randomized controlled trial, we assessed the effectiveness of a computerized mammography reminder system as a component of a program to increase the use of screening mammography in three health care organizations serving inner-city women in Detroit, Michigan (two sites of a health department, one HMO site, and two sites of a private hospital). Four thousand four hundred and one women older than 40 who had visited a study site in the preceding year were randomly assigned to one of two treatment groups. Limited intervention (LI) included physician and staff breast cancer control education, facilitated mammography appointment scheduling procedures, and elimination of out-of-pocket patient cost for mammography (at three of five sites). Full intervention (FI) included all components of limited intervention plus an additional series of "cues-to-action." These included a mammography reminder form inserted in the medical record of women who were due to have mammography, intended to increase physician referral for mammography appointments, and patient reminders intended to increase completion of mammography among referred women. During the one-year intervention period 2,725 randomized women visited a study site. The 6-month mammography appointment rates among FI women vary from 38% to 65% and the FI rate exceeds the LI rate at each site with differences from 13% (95% CI, 6 to 20) to 29% (21 to 38). The annual completed mammography rate among FI women extends from 43% to 64% and exceeds the LI rate at each site by 12% (5 to 19) to 25% (16 to 34). After age-adjustment, the mammography intervention effect sizes among the five sites were not significantly different. The average increase in FI compared to LI was 18%. The computerized reminder system is effective in increasing the use of mammography in each of the study institutions and the major effect is on physician referral for mammography.