Effect of Social Networks on Cancer-Screening Behavior of Older Mexican-American Women

BACKGROUND: Previous studies have determined that Hispanic women and, in particular, Mexican-American women have the lowest rates of cancer screening of any race and ethnic group in the United States. In the development of cancer control strategies for this population, little attention has been given to factors that encourage Mexican-American women to seek cancer preventive care. Recent studies suggest that social networks can have a positive influence on cancer-screening participation. PURPOSE: We determined the extent to which differences in social networks account for variations in breast and cervical cancer-screening practices among low-income Mexican-American women. METHODS: The data analyzed in this study were obtained from a baseline survey of knowledge, attitudes, and cancer-screening practices conducted prior to implementation of community interventions designed to improve Pap smear and mammography screening in low-income Mexican-American women 40 years old and older living along the U.S.-Mexican border in El Paso County, Texas. A random selection of 1300 households served as a sampling frame to identify Mexican-American women 40 years old and older for personal interviews. Of the 549 households identified as having at least one eligible female, 450 women completed the personal interviews that provided the data for this study. Personal interviews solicited information on age, income, marital status, place of birth, education, health insurance coverage, Pap smear- and mammogram-screening practices, and six questions relating to social network. A social network score was assigned to each woman by summation of the following six variables: number of confidants, number of close friends, number of close relatives, frequency of contact with these close friends or relatives per month, church membership, and church attendance. Women were grouped into three linear strata of social network (low, medium, high) based on tertiles of the scores. Cancer-screening history was also ascertained during the interview. RESULTS: The 2-year prevalence of Pap smear and mammography use increased with social network. For each gain in social network level (low, medium, high), adjusted for sociodemographic factors, the odds ratio for Pap smear use was 1.33 (95% confidence interval = 1.02-1.73); it was 1.40 (95% confidence interval = 1.02-1.93) for mammogram use. Of the six social network components, the number of close friends was the most important predictor of mammography (P = .002) and Pap smear (P = .025) screening. CONCLUSION: Social networks appear to be an important determinant of cancer-screening behavior among low-income, older Mexican-American women.