Date of Award


Document Type

Campus Access Dissertation


Health Services and Policy Management

First Advisor

Sudha Xirasagar


Background: Colorectal cancer mortality ranks third among cancer-related cause of death in United States. Medical care is a tremendous burden for cancer patients. It is well established that colorectal cancer screening is cost effective and life saving. Recent guidelines by the American Cancer Society (ACS) recommend that adults aged over 50 years receive colorectal cancer screening. Colonoscopy is the most preferred method for colorectal cancer screening. However, little is known about the impacts of distance, rural residence and area socioeconomic status and the likelihood of colonoscopy utilization among patients of African American PCPs, particularly when the PCPs are trained in colonoscopy.

Data collection: The study used data collected for a project funded by the National Cancer Institute, through the South Carolina Cancer Disparities Community Network (one of 25 Cancer Network Programs). We conducted patient chart reviews at seven colonoscopy-trained and five untrained primary care physician offices (selected by convenience sampling) in South Carolina. Consecutive patient charts of patients aged ≥50 years were retrieved using the billing databases. We conducted chart reviews of 2,272 patients aged ≥50 years who were established patients of 12 primary care physicians. Based on inclusion and exclusion criteria, 677 patients of trained PCPs were eligible to study the distance's effect and 1196 patients of colonoscopy-trained and untrained PCPs for the rurality and area SES.

Data Analysis: ARC GIS software was used to calculate distance between patients' zip code centroids and the endoscopy center where their PCP performed colonoscopy. Distance is analyzed as a continuous variable. The Rural Urban Commuting Areas (RUCA 2.0) classification system was used to classify patients' residence zip code into urban and rural. The U.S. Census 2000 was used to retrieve for each zip code the area-based SES data, including percent non-African American population, percent households above federal poverty level, percent households with any adult aged ≥ 25 years who have completed college, median housing value, percent households with dividend, rental or interest income, percent employed persons aged ≥ 16 years in non-blue collar occupations. A composite SES index was calculated as documented by Rosenberg et al. The composite scores were classified into four quartiles (highest=highest area SES). We controlled for patient demographics, insurance source, duration of patient-PCP relationship, and presence of serious co-morbidities.

Results: Distance is not a barrier for the colonoscopy use among patients of trained PCPs (OR=1.006; 95%CI: 0.997-1.016). Rural patients were more likely to use the procedure than the urban patients (OR=1.966; 95%CI: 1.438 - 2.688). Area SES was adversely associated with colonoscopy utilization. As compared to the highest quartile, the OR was 0.649 for the first quartile (95%CI: 0.438 - 0.961), 0.632 for the second quartile (95%CI: 0.417 - 0.958), and 0.527 for the third quartile (95%CI: 0.352 - 0.788).

Conclusion: A possible reason for the lack of association between distance and colonoscopy completion is motivational effects of PCPs trained in colonoscopy. A similar reason may be operating among rural patients' higher odds of colonoscopy relative to urban patients.