Abstract
BACKGROUND: Previously developed screening colonoscopy algorithms based on diagnostic and endoscopy procedural variables have not been sufficiently accurate for use in epidemiological and health services research. OBJECTIVE: To increase understanding of the administrative health database variables that could help to discern screening and nonscreening colonoscopy. METHODS: A qualitative study using physician focus groups was conducted in Montreal (Quebec), Calgary (Alberta) and Toronto (Ontario). Specialty-specific focus group sessions were held among family physicians and gastroenterologists - the physicians responsible for referring patients to and performing screening colonoscopy, respectively. Interview guides were developed to better understand physician clinical and billing practices. Discussions were audiotaped, transcribed verbatim and analyzed using the constant comparative approach. RESULTS: Forty family physicians and seven gastroenterologists participated in five focus group sessions. Patient variables included demographics (age) and medical history (colorectal cancer risk factors⁄symptoms, medication for colorectal cancer risk factors⁄symptoms, gastrointestinal disorders, severe disease). Clinical practice variables included timing of the colonoscopy (evenings, weekends, holidays, during hospitalization; same-day endoscopist consultation and colonoscopy), use of services (hospitalization, annual examination, transfer from other facility) and procedure use patterns (large bowel or other medical⁄surgical procedure before and subsequent to colonoscopy). However, wide variability in clinical and billing practices will likely preclude the development of a reasonably accurate screening colonoscopy algorithm. Physicians suggested adding a screening colonoscopy code to the administrative health data. CONCLUSIONS: Failure to acknowledge the limitations of the provincial administrative health databases to identify screening colonoscopy may lead to incorrect conclusions and the establishment of inappropriate health care policies.