The role of area-level influences on prostate cancer grade and stage at diagnosis

• Published in: Preventive medicine 2004-09, Vol.39 (3), p.441-448
• Main Authors: Klassen, Ann C., Curriero, Frank C., Hong, Jean Hee, Williams, Chyvette, Kulldorff, Martin, Meissner, Helen I., Alberg, Anthony, Ensminger, Margaret
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; African Americans - statistics & numerical data; Age Factors; Aged; Aged, 80 and over; Census; Confidence Intervals; Disparities; Hierarchical modeling; Humans; Male; Maryland - epidemiology; Middle Aged; Multicenter Studies as Topic; Neoplasm Staging - methods; Odds Ratio; Predictive Value of Tests; Prognosis; Prostate cancer; Prostatic Neoplasms - diagnosis; Prostatic Neoplasms - epidemiology; Prostatic Neoplasms - pathology; Registries; Retrospective Studies; Risk Factors; Sensitivity and Specificity; Social factors; Socioeconomic Factors; Survival Analysis
• ISSN: 0091-7435; 1096-0260
• DOI: 10.1016/j.ypmed.2004.04.031

Background. This research explores area-level social influences on prostate cancer, to test whether area-level influences explain disparities in U.S. prostate cancer burden. Methods. The authors geocoded 23,993 1992–1997 Maryland prostate cancer cases, and linked cases to 1990 census data. The authors examined the effect of 17 area-level social variables, measured at block group, tract, and county, modeling individual and multilevel predictors of later stage and higher tumor grade. Results. Younger age, black race, higher grade or ungraded tumors, and earlier year of diagnosis were associated with later stage. Block group percentage of white-collar workers (O.R. = 0.93, 95% C.I. = 0.89, 0.98), and county resources (O.R. = 0.94, 95% C.I. = 0.89, 0.98), were protective of later stage. Older age, black race, and earlier year of diagnosis were associated with higher grade. Block group income was protective for white men (O.R. = 0.92, 95% C.I. = 0.87, 0.96), but for all men, county resources increased risk of higher grade (O.R. = 1.23, 95% C.I. = 1.16, 1.31). Conclusions. Social resources did not significantly reduce racial differences. Results suggest tumor biology is related to relative resources, with better outcomes associated with greater small-area wealth in low-resource counties, but stage at diagnosis is associated with absolute resources, with better outcomes associated with higher small-area social class in high-resource counties.