BACKGROUND: Although abnormal screening mammograms deleteriously affect the psychological well-being of women during the time immediately surrounding the tests, their long-term effects are poorly understood. PURPOSE: To characterize the long-term effects of false-positive screening mammograms on the behavior and well-being of women 40 years of age or older. DATA SOURCES: English-language studies from the MEDLINE, Web of Science, EMBASE, CINAHL, PsycINFO, and ERIC databases through August 2006. STUDY SELECTION: Studies were identified that examined the effects of false-positive results of routine screening mammography on women's behavior, well-being, or beliefs. DATA EXTRACTION: Two investigators independently coded study characteristics, quality, and effect sizes. DATA SYNTHESIS: 23 eligible studies (n = 313,967) were identified. A random-effects meta-analysis showed that U.S. women who received false-positive results on screening mammography were more likely to return for routine screening than those who received normal results (risk ratio, 1.07 95% CI, 1.02 to 1.12). The effect was not statistically significant among European women (risk ratio, 0.97 CI, 0.93 to 1.01), and Canadian women were less likely to return for routine screening because of false-positive results (risk ratio, 0.63 CI, 0.50 to 0.80). Women who received false-positive results conducted more frequent breast self-examinations and had higher, but not apparently pathologically elevated, levels of distress and anxiety and thought more about breast cancer than did those with normal results. LIMITATIONS: Correlational study designs, a small number of studies, a lack of clinical validation for many measures, and possible heterogeneity. CONCLUSIONS: Some women with false-positive results on mammography may have differences in whether they return for mammography, occurrence of breast self-examinations, and levels of anxiety compared with women with normal results. Future research should examine how false-positive results on mammography affect other outcomes, such as trust and health care use.
Brewer et al. (Tue,) studied this question.