Racial segregation as an important determinant of lung cancer outcomes

The purpose of this study is to measure the extent to which racial segregation is associated with the odds of receiving treatment and five year survival rates for patients diagnosed with Non-Small Cell Lung Cancer (NSCLC). This study uses a cohort of cancer incidences between January 2000 to December 2011. Racial segregation is the exposure of interest in this study while we control for area level poverty. The patient level data were merged with Census 2000 data.  We operationalized racial segregation with dissimilarity index (D).  The D index ranging from zero to 1 indicates the departure from even racial distribution across census tracts when it approaches 1.  Data were also merged with Rural Urban Commuting Area (RUCA) codes provided by the US Department of Agriculture resulting in a variable measuring levels of rurality as all models are stratified by the levels of rurality ( rural/small town, large town, , suburban, , urban).  We employed multi-level logistic models for outcome variables; unstaged and late stage diagnoses, and receiving treatment. We also conducted the analysis of survival time until death. In all models tumor stage, tumor grade, age, race, and sex were controlled.  Our analysis provides a more comprehensive understanding of the social context of race as racial segregation is examined while community poverty and levels of rurality are accounted for.  This study contributes to the current literature of cancer epidemiology as it suggests cancer disparity among black and white may not be related to race per se, rather it could be best explained by the spatial and social construction of race