Beyond patient-level barriers: Evaluation of rural residency and area-level deprivation and clinical trial (CT) participation.

Abstract

e19068 Background: Residents of rural and socioeconomically deprived areas have worse cancer outcomes and are less likely to participate in CTs. It is unknown whether these area-level attributes predict CT participation after accounting for individual sociodemographic variables. Methods: We combined data from two SWOG national survey studies. S0316, a multi-center prospective survey, was combined with data from a large web-based survey. Both studies examined CT treatment decision making from diagnosis in patients with common cancers. Zip codes of residences were classified as rural or urban using Rural-Urban Continuum Codes (urban, 1-3, vs. rural, 4-9) and a 3-level ordinal variable (urban, 1-3, vs. rural, 4-7, vs. very rural, 8-9). We identified socioeconomically deprived areas (SDAs) as those with an Area Deprivation Index (ADI) in the upper quartile. We also examined areas by ADI quartile. Multivariable logistic regression was used to evaluate the association of rural residency and area-level socioeconomic deprivation with CT participation after adjusting for important individual-level factors (age, sex, race/ethnicity, income, and education), stratified by study and cancer type. Results: Among 7080 patients, 1299 (18.5%) were from rural areas, 653 (9.4%) were from SDAs; and 715 (10.1%) participated in a CT. Patients had breast (56.1%), prostate (21.8%), lung (13.6%), and colorectal (8.4%) cancer. In univariate analysis, rural patients were 23% less likely to participate in a CT (OR=0.77, 95% CI: 0.62-0.95, p=.016); in multivariate analysis, results were similar (OR=0.80, 95% CI: 0.64-0.99, p=.037). Very rural residents were 34% less likely to participate than urban patients (OR=0.66, 95% CI, 0.55-0.80, ordinal p=.031). There was no statistically significant association between residence in SDAs and CT participation (p>.30). Rural patients were more likely to have >1 comorbid conditions (45.4% vs. 39.5%, p=.001), to be concerned about how to pay for their care (40.6% vs. 32.4%, p<.0001), and to travel farther for care (median 50.0 vs. 12.0 miles, p<.0001). Conclusions: Our findings are the first to show that rural residents are less likely to participate in CTs even after accounting for confounding individual factors. These results agree with prior observations that rural patients must travel longer distances for care, including for CT care. Reducing the travel burden for rural cancer patients could improve their CT participation and the generalizability of CT results to all patients.