Prostate Cancer Treatment Patterns in Relation to Risk Category: a Georgia Statewide Analysis Open Access

Felzer, Jamie Rachel (2012)

Permanent URL:


Introduction: Prostate cancer is the most common invasive malignancy of men in the United States. In 2011, 240,890 men were estimated to be diagnosed and 33,720 were estimated to die of this disease. Management and control of prostate cancer is complex. Although no single treatment can be considered the standard of care, it is evident that certain subgroups of men may receive better treatment than others. Rural residents, black men, those with low SES or less access to medical care have a worse prognosis. In this population-based study, we seek to determine patterns of prostate cancer treatment in the state of Georgia.

Methods : Data for men with newly diagnosed prostate cancer within a six year (2004-2009) interval were obtained from the recently expanded, statewide Georgia SEER Registry. Criteria for prostate cancer risk classification were based on the 2010 NCCN categories that incorporate disease stage, grade and the pre-diagnosis level of the prostate specific antigen (PSA). The main treatments of interest were surgery, radiation, or "any treatment" (surgery and/or radiation).

Results : Among 37,667 prostate cancer cases diagnosed in Georgia during the study interval, 37,274 (99%) had sufficient data to be classified into a risk category. 23,935 out of 37,274 men, or 64%, received either radiation or surgery. Due to the strong interactions between risk category and the covariates, no overall effect could be calculated. All models were stratified on age, while some still required further stratification. For example, radiation had additional interactions with marital status and county type. Overall, men with low risk disease were 6- to 11-fold less likely than high risk men to receive any localized prostate-directed treatment, although the magnitude of this association differed significantly by age.

Discussion : Our results clearly demonstrated that risk category plays a critical role in determining which patients receive certain treatments, but the magnitude and direction of the relationship between disease characteristics and therapy receipt varies by age, marital status, race and nature of residential area. The mechanisms by which patient (and perhaps provider) characteristics modify the association with treatment type are not clear, and need to be further explored.

Table of Contents


Variable Classification 6
Statistical Analysis 7
Unadjusted Analyses and Descriptive Statistics 9
Overview of Multivariable Analyses 10
Multivariable and Stratified Analyses for Any Prostate-Directed Treatment (Surgery and/or Radiation) 10
Multivariable and Stratified Analyses for Radiation Treatment 11
Multivariable and Stratified Analyses for Surgery 12
Strengths 16
Limitations 16
Conclusions and Future Directions 18
Table 1. Distribution of variables according to NCCN Risk Categories 23
Table 2. Stratified, Adjusted Logistic Regression of Any Treatment 25
Table 3a. Stratified, Adjusted Logistic Regression of Overall Radiation Therapy for Men Under 60 Years Old 26
Table 3b. Stratified, Adjusted Logistic Regression of Overall Radiation Therapy for Men Greater Than or Equal to 60 Years Old 27
Table 4. Adjusted, Stratified Logistic Model for Men of All Ages Receiving Surgery 28

About this Master's Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
  • English
Research Field
Committee Chair / Thesis Advisor
Partnering Agencies
Last modified

Primary PDF

Supplemental Files