BACKGROUND: The effectiveness of mammography screening could be improved if factors that influence nonattendance were better understood. METHODS: We examined attitudes, beliefs, and knowledge in relation to nonattendance in a population-based mammography screening program, using a case-control design. Data were collected from November 1997 to March 1998 through telephone interviews with 434 nonattenders and 515 attenders identified in a population-based mammography register in central Sweden. The questions asked drew primarily upon the components constituting the Health Belief Model. RESULTS: Multivariate analysis showed that nonattendance was most common among women within the highest quartile of perceived emotional barriers, compared to women within the lowest quartile (OR = 4.81; 95% CI 2.96-7.82). Women who worried most about breast cancer were more likely to attend than those who worried least (OR = 0.09; 95% CI 0.02-0.31). Women with the highest scores of perceived benefits were more likely to attend than women with the lowest ones (OR = 0.35; 95% CI 0.08-0.75). Other factors associated with nonattendance were less knowledge about mammography and breast cancer, lack of advice from a health professional to participate, and very poor trust in health care. CONCLUSIONS: Our findings suggest that increased participation in outreach mammography screening programs can be achieved through enhancement of breast cancer awareness and possibly by reducing some of the modifiable barriers. mammography; mass screening; breast cancer; attitudes; Sweden.