Health Policy

Challenges of using nationally representative, population-based surveys to assess rural cancer disparities

Zahnd, W.E.; Askelson, N.; Vanderpool, R.C.; Stradtman, L.; Edward, J.; Farris, P.E.; Petermann, V.; Eberth, J.M.
Jan-12-2019

Abstract

Population-based surveys provide important information about cancer-related health behaviors across the cancer care continuum, from prevention to survivorship, to inform cancer control efforts. These surveys can illuminate cancer disparities among specific populations, including rural communities. However, due to small rural sample sizes, varying sampling methods, and/or other study design or analytical concerns, there are challenges in using population-based surveys for rural cancer control research and practice. Our objective is three-fold. First, we examined the characterization of “rural” in four, population-based surveys commonly referenced in the literature: 1) Health Information National Trends Survey (HINTS); 2) National Health Interview Survey (NHIS); 3) Behavioral Risk Factor Surveillance System (BRFSS); and 4) Medical Expenditures Panel Survey (MEPS). Second, we identified and described the challenges of using these surveys in rural cancer studies. Third, we proposed solutions to address these challenges. We found that these surveys varied in use of rural-urban classifications, sampling methodology, and available cancer-related variables. Further, we found that accessibility of these data to non-federal researchers has changed over time. Survey data have become restricted based on small numbers (i.e., BRFSS) and have made rural-urban measures only available for analysis at Research Data Centers (i.e., NHIS and MEPS). Additionally, studies that used these surveys reported varying proportions of rural participants with noted limitations in sufficient representation of rural minorities and/or cancer survivors. In order to mitigate these challenges, we propose two solutions: 1) make rural-urban measures more accessible to non-federal researchers and 2) implement sampling approaches to oversample rural populations.