The physician -patient interaction and colorectal cancer screening in Ontario

The objective of this study was to determine factors that influence patient completion and physician recommendation of a colorectal cancer (CRC) screening test.;Thirty-three percent of the general population subjects reported that they had ever had a CRC screening test. Of the nonscreeners, 55% reported that they intended to have a test and 45% reported that they did not intend of have a test. The strongest factor associated with the most recent completion of a fecal occult blood test (FOBT) within two years or a lower endoscopy within ten years compared to being a nonscreener that intended to have a test, was a CRC screening discussion with their physician [Adj OR 10.5 (95% CI 7.1--15.5)]. Sixty-two percent of the physician subjects reported that they routinely recommended FOBT for CRC screening. Only 15% reported that they routinely recommended colonoscopy and 6% for flexible sigmoidoscopy as primary screening modalities. A CRC screening discussion between patient and physician was more likely to have occurred if the physician's self-reported practice was the routine recommendation of CRC screening [OR 2.0 (95% CI 1.3--3.3)], if the patient had at least one physician visit within the previous year [OR 11.5 (95% CI 2.6--50.1)], if there was a positive family history of CRC [OR 4.4 (95% 2.5--7.7)], or if the length of clinic visits was longer.;CRC screening continues to be underutilized despite increased awareness. Both patient and physician factors are important for CRC screening adherence.;A telephone survey was conducted with 1002 general population subjects between the ages of 50 and 75 years by a random digit dialing approach of Ontario households. Sixty-nine percent of the respondents provided contact information for their current primary care physicians. Multiple wave mailed questionnaires were sent to the identified physicians and 472 responded. The data were analyzed using descriptive and univariate analyses. Adjusted odds ratios were generated using multiple logistic regression.