Abstract B50: Concordance between self-reported and medical record breast cancer treatment data by income and education.

While the use of self-reported treatment data is common in cancer disparities research, it is unknown whether their accuracy varies by study participant's income or education. Using data collected from a population-based case-control study of contralateral breast cancer, we compared the concordance between participant responses to a telephone questionnaire and medical record data obtained from all facilities where a participant received breast cancer treatment. This analysis included women age 40 – 79 diagnosed with contralateral breast cancer (n=256) or only a single breast cancer (n = 533) between 1990 and 2005. Concordance was evaluated by calculating the percent agreement, Kappa statistic and its 95% confidence interval (CI) for the agreement between self-report and medical record data for the type of breast cancer surgery received, receipt and start date (month and year) of chemotherapy and radiation, and type of hormonal therapy received of all 789 participants stratified by highest level of education received (≤ high school, some college, ≥ college graduate) and annual household income (<$35,000, $35,000-$69,000, ≥$70,000) at the time of diagnosis. Concordance between self-reported and medical record data for receipt of mastectomy, chemotherapy, and radiation therapy was high and did not vary according to education or income strata (all κ > .90). However, some variation with respect to treatment start date was observed. For radiation start date, concordance was lower among women reporting an income of <$35,000 compared to those with an income ≥$70,000 (κ = 0.47, 95% CI 0.45–0.48 vs. κ = 0.60, 95% CI 0.58–0.62, respectively), and lower among women with ≤ high school education compared to college graduates (κ = 0.51, 95% CI 0.50'0.53 vs. κ = 0.60, 95% CI 0.58'0.61 respectively). These same patterns were seen with chemotherapy start dates. These findings suggest a high degree of accuracy regardless of income or education for self-reported receipt of various forms of breast cancer treatment. There was overall less concordance with respect to start dates of radiation and chemotherapy with evidence that concordance was the poorest among women with less income and less education. Studies of socioeconomic disparities in breast cancer treatment that require accurate dates of treatment initiation may require ascertainment of medical record data to avoid potential bias.

Citation Format: Jean A. McDougall, Kathleen E. Malone, Mei-Tzu C. Tang, Christopher I. Li. Concordance between self-reported and medical record breast cancer treatment data by income and education. [abstract]. In: Proceedings of the Fifth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2012 Oct 27-30; San Diego, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2012;21(10 Suppl):Abstract nr B50.