Published in:
Open Access 01-12-2013 | Research article
Endoscopist specialty is associated with colonoscopy quality
Published in: BMC Gastroenterology | Issue 1/2013
Login to get accessAbstract
Background
Some studies have shown that endoscopist specialty is associated with colorectal cancers missed by colonoscopy. We sought to examine the relationship between endoscopist specialty and polypectomy rate, a colonoscopy quality indicator. Polypectomy rate is defined as the proportion of colonoscopies that result in the removal of one or more polyps.
Methods
A cross-sectional study was conducted of endoscopists and their patients from 7 Montreal and 2 Calgary endoscopy clinics. Eligible patients were aged 50–75 and covered by provincial health insurance. A patient questionnaire assessed family history of colorectal cancer, history of large bowel conditions and symptoms, and previous colonoscopy. The outcome, polypectomy status, was obtained from provincial health administrative databases. For each city, Bayesian hierarchical logistic regression was used to estimate the odds ratio for polypectomy comparing surgeons to gastroenterologists. Model covariates included patient age, sex, family history of colorectal cancer, colonoscopy indication, and previous colonoscopy.
Results
In total, 2,113 and 538 colonoscopies were included from Montreal and Calgary, respectively. Colonoscopies were performed by 38 gastroenterologists and 6 surgeons in Montreal, and by 31 gastroenterologists and 5 surgeons in Calgary. The adjusted odds ratios comparing surgeons to gastroenterologists were 0.48 (95% CI: 0.32–0.71) in Montreal and 0.73 (95% CI: 0.43–1.21) in Calgary.
Conclusions
An association between endoscopist specialty and polypectomy was observed in both cities after adjusting for patient-level covariates. Results from Montreal suggest that surgeons are half as likely as gastroenterologists to remove polyps, while those from Calgary were associated with a wide, non-significant Bayesian credible interval. However, residual confounding from patient-level variables is possible, and further investigation is required.