Screening
Using facility zip code level mammography utilization to overcome biases that plague survey data on screening utilization Garth Rauscher* Garth Rauscher Rauscher Rauscher Rauscher University of Illinois School of Public Health, Division of Epidemiology and Biostatistics
Introduction: National and local survey data on mammography screening suffer from self-report inaccuracies (mostly in the form of over-reporting) and low survey participation, where participants are more likely than the general population to have been previously screened. As a result, survey data overestimate screening and may mask disparities in screening. We obtained population-level prevalence estimates of screening mammography in Chicago, while avoiding the low response rates and over-reporting issues that plague survey data. Methods: For calendar year 2019, we requested zip code-level counts of screening mammograms for sites within and around the city of Chicago. Counts were aggregated to the zip code and divided by the number of women aged 40-79 to obtain approximate utilization rates. Next, we linked zip code compositional characteristics (e.g., segregation, disadvantage) and estimated prevalence differences (PD) associations between zip code characteristics and screening utilization. We also compared estimates to those from the BRFSS and HCS. Results: Utilization estimates were lower than those obtained from survey data. Mean zip code utilization rate was 29% when assigning all screening mammograms to women aged 40-79, versus 48.5% in the past 12 months using BRFSS data. Higher utilization was associated with greater zip code-level income, education, and private health insurance. Predominantly minority zip codes had utilization rates that were approximately 10 percentage points lower than those of non-minority zip codes (95% CI: -0.18, -0.02); this disparity was statistically eliminated when zip code socioeconomic status was added to the models (PD=-0.01, 95% CI: -0.16, 0.15). Updated results will be presented. Conclusion: Self-reported data on mammography utilization provides an overly optimistic picture of screening utilization and masks existing disparities in utilization.