M1032 Factors Associated with Practice of Colorectal Cancer Screening Among Primary Care Physicians in Hong Kong: A Cross-Sectional Study

vs use of colonoscopy for symptoms (diagnostic) or higher risk patients (surveillance), may contribute to the mortality difference if such differences resulted in diagnosis at a more advanced stage. Our aim was to explore if Caucasian patients were more likely to undergo primary screening colonoscopies vs. diagnostic or surveillance colonoscopies than were nonCaucasian patients. Methods: We randomly sampled 650 veterans from the national VA administrative databases (Austin Information Technology Center (AITC)), aged 50 and older, who had undergone outpatient colonoscopy between October 2005 and September 2006 in VA facilities using ICD-9 procedure codes. We then abstracted colonoscopy indication and race from their electronic medical records through VistAWeb. Colonoscopy indication was classified as either screening, diagnostic or surveillance. A chi-squared test was used to examine differences in colonoscopy indication by the dichotomous race variable (Caucasian vs. Non-Caucasian). Results: In our sample 84% of veterans were Caucasian and 16% were non-Caucasian (90% of non-Caucasians were African American). We deleleted 81 individuals from the analysis due to missing data for race. Our sample size for this analysis was 569. The indications for Caucasians and non-Caucasians were not significantly different (p=0.72) and were 32% and 34% screening, 22% and 33% diagnostic, and 45% and 41% surveillance. Furthermore, the distribution of the indication patterns was similar in the group missing race data. Conclusions: During the study period about a third of the colonoscopies in this national VA sample were for primary CRC screening. There is no evidence that Caucasian patients weremore likely to have had colonoscopy for primary screening, versus for symptoms or surveillance, than were minority patients. These results support the equity of the VA system, but do not explain the cancer mortality difference.