Publication Date




Embargo Period


Degree Type


Degree Name

Doctor of Philosophy (PHD)


Epidemiology (Medicine)

Date of Defense


First Committee Member

Tulay Koru Sengul

Second Committee Member

Noella A. Dietz

Third Committee Member

Edward J. Trapido

Fourth Committee Member

David J. Lee


Prostate cancer is the most common solid malignancy diagnosed in American men. One in six men in the US is expected to be diagnosed with prostate cancer in his lifetime. Despite the large number of men diagnosed every year with prostate cancer, there is no consensus concerning the best and most appropriate treatment. Moreover, most men with early stage prostate cancer have a life expectancy comparable to similarly aged men without prostate cancer and will die with prostate cancer rather than from prostate cancer. Without evidence supporting one treatment over another, it is important to understand the factors that inform treatment selection and what impact they may have on survival. The Florida Cancer Data System was used to obtain data from men (n=118,533) diagnosed with prostate cancer between the years 2001 and 2009. US Census data was used to construct area-based socioeconomic measures of education and socioeconomic status. Census tract based Rural-Urban Commuting Area Codes were used to determine urban vs. rural residence. Approximately 73% of the men in this study were Non-Hispanic White, 13% Non-Hispanic Black, 11% Hispanic and 3% were categorized as Non-Hispanic Other or “Unknown” race/ethnicity. Most (81%) men were diagnosed with localized (early stage) prostate cancer. Roughly 10% of the cases were Regional/Distant (advanced stage) and 9% were not staged. Men diagnosed with localized prostate cancer received surgery (34%), radiation (32%), hormonal therapy (22%) or watchful waiting (12%) as their initial treatment. Binary logistic regression models were fitted to predict late stage prostate cancer. For men diagnosed with localized prostate cancer, polytomous logistic regression models were fitted to predict initial treatment. Lastly, survival analysis was conducted using the Kaplan-Meier method and Cox proportional hazard regression models were used to model time-to-event outcome (overall survival/all-cause mortality). The regression models were adjusted for age at diagnosis, insurance, education, marital and smoking status, SES, urban/rural residence, initial treatment and tumor grade. The results of the analyses show that sociodemographic factors were associated with stage of cancer diagnosis, initial treatment selection, and overall survival for men with prostate cancer. Non-Hispanic Black men were more likely to present with late stage disease (OR=1.16; 95% CI: 1.09, 1.23) as were men who were not married (OR=1.24; 95% CI: 1.18, 1.30) and who currently smoked (OR=1.36; 95% CI: 1.28, 1.46). Variation existed in the initial treatments for men diagnosed with localized prostate cancer. Non-Hispanic Black (OR=0.66; 95% CI: 0.61, 0.71) and Hispanic men (OR=0.85; 95% CI: 0.79, 0.92) were less likely to receive surgery or radiation therapy, respectively. Non-Hispanic Black men with the lowest socioeconomic status were less likely to receive radiation therapy over watchful waiting (OR=0.66; 95% CI: 0.50, 0.87). Compared to men with private insurance, men with no insurance were more likely to receive watchful waiting over surgery (OR=2.04; 95% CI: 1.75, 2.38), radiation therapy (OR=2.32; 95% CI: 1.96, 2.7) or hormonal therapy (OR=1.43; 95% CI: 1.22, 1.69). Men with public insurance were less likely to receive surgery (OR=0.88; 95% CI: 0.84, 0.93) and more likely to received hormonal therapy (OR=1.18; 95% CI: 1.12, 1.25) over watchful waiting. Lastly, men with less than a high school education were more likely to select watchful waiting (OR=1.11; 95% CI: 1.01, 1.2) over surgery. Cox proportional hazard regression models showed differences in overall survival based upon initial treatment and sociodemographic factors. Although surgery had better 5-year survival rates than watchful waiting (82% vs. 76%), there was no statistically significant difference in overall mortality risk between the two treatments. All of the sociodemographic predictors of interest (race/ethnicity, socioeconomic status, education level, smoking status and marital status) were significant predictors of overall mortality risk for men diagnosed with localize prostate cancer. With an eye toward reducing healthcare costs and improving efforts to eliminate health disparities, it is imperative that we understand what external factors affect screening behaviors for early detection, initial treatment selection, and survival for men diagnosed with prostate cancer.


prostate cancer; florida cancer data system; initial treatment