RTI International is submitting this grant application, Spatial Impact Factors and Mammography Screening project in response to the PA Studies of the Economics of Cancer Screening, Prevention, and Care, to address the National Cancer Institute's (NCI's) concerns about .barriers to preventive and screening services for cancer, particularly mammography and breast cancer. To address these issues and contribute useful insights for expanding the uptake of mammography, we will develop an innovative, robust, spatial-analytic methodology and create a transportable database of contextual variables for use in examining and explaining observed variation in health-seeking behaviors among women across different markets and geographic regions. We will use the linked SEER-Medicare database in a pilot study to examine the specific factors contributing to the wide variation in mammography screening uptake among (primarily) elderly women. Findings will provide a better understanding of cultural and environmental factors that influence access and utilization. NCI and Comprehensive Cancer Control programs can use these findings to develop more effective intervention strategies and better simulation models of cancer costs and burdens. The large, retrospective, population-based cohort of elderly women is drawn from 10 geographic regions in the SEER-Medicare database that span the urban-rural continuum and provide good coverage of various races and ethnicities, and women with and without cancers. Our dependent variable is a binary variable indicating whether a woman received at least one screening mammogram between 1995 and 1999. Explanatory variables include health market characteristics, socio-cultural characteristics, and other features of women's neighborhoods; physical transportation and environmental conditions; and characteristics of the women themselves (e.g., age, race and ethnicity, type of insurance coverage, dual eligibility, disabled status). Our research design begins with a conceptual model that allows for spatial interaction among biologic or environmental factors along the pathways to health care utilization. After using some innovative methods to create some explanatory factors and to determine a geographically relevant model specification, we will estimate a multilevel model and interpret results to determine (a) what factors influence screening behavior and where (at which geographic locations), (b) how robust findings are to level of aggregation of data (Census tract or zip code and county), and (c) whether our complex model is more useful than a simpler benchmark model using simpler covariates commonly found in the literature. This work will demonstrate the feasibility of conducting a broader study encompassing more than twice as many counties of the United States (with additions covered by the expanded SEER database).
| Mobley, Lee R; Kuo, Tzy-Mey May; Clayton, Laurel J et al. (2009) Mammography facilities are accessible, so why is utilization so low? Cancer Causes Control 20:1017-28 |
| Mobley, Lee R; Kuo, Tzy-Mey May; Andrews, Linda (2008) How sensitive are multilevel regression findings to defined area of context?: a case study of mammography use in California. Med Care Res Rev 65:315-37 |
