Background: Efforts to detect breast cancer in its early stages are necessary to reduce breast cancer-associated mortality. This study evaluated the impact of a multicomponent intervention implemented between 1987 and 1990 to increase a community's utilization of breast cancer screening by women over 50 years of age.
Methods: The study used a pretest/post-test two-community design, with one community assigned as the intervention community and the other as the comparison. The intervention consisted of a comprehensive physician involvement component and a community education effort. To assess the overall impact of the interventions, we measured women's participation in screening via random digit dial telephone surveys at three time points, each approximately 18 months apart.
Results: Over the course of the study, there were dramatic improvements in breast cancer screening participation in both communities. However, the intervention city showed more improvement in selected variables than did the comparison community in the early phases of the project between baseline and midpoint. These included increased advice by physicians to have mammograms, increased awareness that screening is necessary in the absence of symptoms, increased awareness that many women over 50 have mammograms, decreased perception of barriers to clinical breast exam, and an increase in the proportion of women having a clinical breast exam. In addition, significantly fewer women in the intervention city than in the comparison city reported never having had a mammogram at midpoint.
Conclusions: The findings demonstrate limited impact of a community intervention during a period of increasing adoption of mammography screening, in part, due to this rapidly rising secular trend. Additionally, increased activities in the comparison community were documented. Therefore, as incidence of screening increases, targeted activities aimed at population subgroups are warranted, and evaluation designs need to include multiple comparison groups or broader geographic random samples.