Studies With Endoscopic Databases

G&H Can you describe the history and scope of the CORI database?

GH The Clinical Outcomes Research Initiative (CORI) was initiated in 1995 under the auspices of the American Society for Gastrointestinal Endoscopy (ASGE). By 2004, approximately 400 physicians at 75 practice sites were submitting 120,000 procedural reports annually and these numbers continue to grow. Close to one million reports are currently stored as part of the National Endoscopic Database. The philosophy behind CORI is to capture information on endoscopies performed across the nation in a wide variety of settings and to develop a truly representative picture of endoscopic practice.

G&H How is CORI organized and what kind of information is collected?

GH There is a central data repository at the Oregon Health and Science University in Portland, which also serves as a sort of headquarters. Practicing endoscopy sites apply to this office to be included as participants in the database. Sites must commit to the parameters of the program, which include reporting at least 95% of procedures performed and providing funding for on-site software installation, in order to be considered. The executive board of the program reviews all applicants and accepts the vast majority, giving preference to those with relatively high volume of procedures and coverage of areas and socio-economic groups not already represented. Once a site is accepted, a representative from the Oregon office visits the site and coordinates the necessary software installation. This software enables practicing physicians to enter data from every procedure they perform. The data entered include basic patient demographics, indications for the procedure, sedation method, findings of the procedure, and details of any endoscopic treatment provided during the procedure. Funding for CORI is provided the ASGE, the National Institutes of Health, and the Agency for Healthcare, Research, and Quality, along with a number of corporate sponsors.

G&H Are there other comparable procedural databases currently in use?

GH CORI is the largest endoscopy database in the United States. In terms of nongastroenterologic practice, the Veterans Administration (VA) has a comprehensive database, which captures all patient and practice information from their medical centers. Many large health maintenance organizations have their own patient databases as well.

G&H Could you describe your own research utilizing CORI?

GH Our group utilized information from CORI to describe variations in endoscopic practice across the nation. We looked at the proportion of procedures in which small bowel biopsies were taken when there was a clinical indication to screen for celiac disease. We found a marked disparity in small bowel biopsy practice across the nation, when the endoscopy sites were categorized into academic centers, VA medical centers, and community practice groups, with VA centers demonstrating the highest rate of biopsy performance. Another study described the proportion of patients with gastric ulcers in whom biopsies were performed to exclude malignancy. Again, we found fairly marked differences among the three site types, with community centers recording highest biopsy rates. Finally, a third study looked at the proportion of patients with chronic diarrhea who underwent colonic biopsy for microscopic colitis and again demonstrated disparity in biopsy rates, with patients seen in academic centers being biopsied most frequently.

G&H Have you utilized database technology to conduct any other studies?

GH We conducted another study at the Mayo Clinic utilizing the institutional endoscopic database to assess colonoscopy performance among clinicians in the Division of Gastroenterology and provide feedback. The advantage of performing this study in the Mayo setting is that it is a very large division with 50–60 colonoscopists. This diverse cohort of endoscopists composed a heterogeneous group with different levels of experience.

Feedback was provided via e-mail to each endoscopist on their cecal intubation rate, insertion time, which is the time taken to reach the cecum, and withdrawal time, which is the time from reaching the cecum to completion of the procedure, as well as the phase of the procedure during which inspection is performed. The aim of this study was to assess the impact of feedback on endoscopy performance.

Following introduction of feedback, there was an improvement in cecal intubation rates. Average insertion time decreased postfeedback, whereas withdrawal time remained constant. This implied that more procedures were completed successfully and in a shorter period of time following feedback, and the stability of withdrawal time indicated that the inspection phase of the procedure was not compromised. For this study, e-mails were distributed individually. However, with the available database at Mayo, this process could be automated.

G&H How can the findings from these studies be applied in general endoscopic practice?

GH One of the key components in improving endoscopic quality is to enhance uniformity of practice. The first step in achieving uniformity is to identify those practices in which disparity exists. If gastroenterologists are aware of these disparities, steps can then be taken to first arrive at a consensus regarding best practice and then to collectively strive to adhere to these practices 100% of the time.

G&H What are the pros and cons of a large, comprehensive database like CORI versus an institutional database?

GH The biggest strength of CORI is that it is representative of diverse endoscopic settings across the whole country. Historically, most research has come from academic centers but academic centers only comprise a small proportion of the total number of medical centers across the country. CORI reflects endoscopic practice outside academic centers and in the future it will continue to be helpful in monitoring endoscopic practice patterns across the nation.

The challenge in creating databases is to strike the correct balance between developing a resource that is sufficiently user-friendly to promote physician compliance in entering all the required information and, on the other hand, still capturing sufficient information for future use as a research tool. In the current climate, with physicians busier than ever, data entry needs to be as time-efficient as possible in order for endoscopists to participate. Conversely, data capture needs to be comprehensive enough to serve its purpose of facilitating meaningful research. CORI does an admirable job in this regard.