Trends and racial disparities in breast cancer screening

Objective: (i) To describe the national trend of mammogram screening, clinical breast examination and breast cancer screening from 2000 to 2010 by race. (ii) To estimate racial disparities for mammogram screening, clinical breast examination and breast cancer screening from 2000 to 2010. Methods: Medical Expenditure Panel Survey (MEPS) data from 2000 to 2010 was used for the analysis. All females with age ≥40 years were included in the study. Outcome variables were mammogram screening (MS), clinical breast examination (CBE) and breast cancer screening (BCS). MS and CBE were defined as those females who received MS and CBE respectively in previous one or two years. BCS was defined as those who received MS and CBE (both) in previous one or two years. Main independent variable was race categorized as non-Hispanic whites (NHW), non-Hispanic blacks (NHB), Hispanics and others. Descriptive analysis was conducted to describe the national trends in MS, CBE and BCS for each year from 2000 to 2010 by race. Bivariate and multivariate logistic regression was conducted to identify racial disparities in MS, CBE and BCS; and non-linear Blinder Oaxaca decomposition was conducted to decompose disparities in explained and unexplained component. Results: Final cohort consisted of 79,068 females (weighted sample size= 764,361,258). MS, CBE and BCS rates were almost stable throughout the decade. In multivariate logistic regression, NHB (OR: 1.27, 95% CI: 1.16-1.40, p<.0001) and Hispanics (OR: 1.27, 95% CI: 1.16-1.40, p<.0001) reported higher odds of receiving MS. For BCS use, NHB (OR: 1.31, 95% CI: 1.21-1.42, p<.0001) and Hispanics (OR: 1.22, 95% CI: 1.12-1.32, p<.0001) reported higher odds of receiving screening for BCS. For CBE, NHB (OR: 1.43, 95% CI: 1.29-1.58, p<.0001) reported higher odds and Hispanics (OR: 0.81, 95% CI: 0.74-0.9, p<.0001) reported lower odds of receiving CBE compared to NHW. Non-linear decomposition was conducted to decompose disparity between NHW and Hispanics. For MS, total disparities explained by observed covariates were found to be 175.19%, for CBE, it was found to be 281.07% and for BCS, it was 160.48%. In our study, we found insurance to be the biggest driver of disparities, followed by usual source of care in all three types of screening. Conclusion: Trends for MS, CBE and BCS were stable from 2000 to 2010 raising concerns about impact of guidelines revised during this period. Bivariate and multivariate logistic regression indicated that a disparity does exist in all three types of screening. In our study, we found insurance and usual source of care were contributing maximum towards disparities. In order to enhance screening among Hispanics, national program/interventions should target these factors.