Colonoscopy is considered by many professional and lay groups to be the preferred option for colorectal cancer screening and by virtue of polypectomy, it allows removal of premalignant lesions and possibly curative treatment of early stage cancers. Despite the positive attributes of screening colonoscopy, because of development of interval cancers following colonoscopy, concerns about the accuracy of colonoscopy as currently practiced have been raised. It also appears that there is relative overuse of the procedure for follow up after polyp removal. There is a relative paucity of population-based data regarding colonoscopy utilization and outcomes. The proposed study will include two components, one using a cancer-free population and one using patients with incident colorectal cancer. In the first phase, a cohort of patients undergoing colonoscopy with polypectomy will be identified through a 5% random sample of Medicare claims data. Previous claims will be searched to stratify according to a prior history of colon polyps or other risk factors, or colonic surgery. Following the index polypectomy, claims will be examined to determine the timing and outcome of repeat colonoscopy and other colonic procedures, with a particular focus on colonoscopies performed before the 3-year minimum interval. Practice patterns will be compared between patient, physician and facility characteristics. In the second phase, patients with a first diagnosis of colorectal cancer will be identified through SEER registry files. Among patients with a polyp containing early stage cancer, the outcome of treatment with endoscopic resection will be compared to surgical resection, and delayed outcomes, including the need for subsequent surgery will be measured. Factors that predict the use of endoscopic resection will be identified and propensity scores for treatment will be developed and used to compare outcomes. Among all patients with incident cancer, prior claims data will be searched to determine the use of colonoscopy. A cohort of patients with potentially missed or interval cancer will be identified and compared to other patients according to patient, physician, facility, and small area characteristics. The studies proposed in this application will provide important data about the use and potential effectiveness of colonoscopy in routine practice. Given the increased emphasis on colonoscopy as a preferred screening test in the general population, these data will help inform future policy decisions.
Colorectal cancer, the second most fatal cancer in the U.S., is largely preventable through the removal of polyps. However, there is only limited knowledge about the treatment and outcome of colorectal polyps in routine clinical practice. Using data from a large number of Medicare patients, we will study practice patterns for polyp removal via colonoscopy, evaluate its effectiveness in the treatment of early stage colorectal cancer, and estimate the potential failed detection rate at colonoscopy.
