Introduction Caecal intubation rate (CIR) is an important quality assurance measure in colonoscopy. There are many factors known to affect this including; sex of patient, BMI, age, timing of list and volume of colonoscopies performed by endoscopist. We analysed the factors that predict colonoscopy completion in a high throughput UK Centre.

Methods All patients that had colonoscopy over the course of a year were identified from the Sheffield Teaching Hospitals database (Infoflex). Potential determinants of CIR analysed include: patient factors (age, sex and indication for procedure); procedural factors (IV sedation used, endoscopist: gastroenterologist vs surgeon vs nurse). Individual factors were assessed using χ² test and subsequently multivariate analysis to identify independent risk factors for non-completion.

Results 5227 colonoscopies were performed between September 2008 and August 2009 at Sheffield teaching hospitals. 50 of these were excluded from analysis due to incomplete information leaving a total of 5177 colonoscopies to analyse. The mean age of patients was 61.39 with an almost equal sex split (47.4% male). Seven nurse endoscopists performed 2380 procedures (mean 340/person), 28 gastroenterologists performed 1754 procedures (mean 63/person), 19 surgeons performed 858 procedures (mean 45/person). The overall CIR was 89.9%. Factors associated with non-completion on univariate analysis include; Female sex (352/2692 incomplete in females vs 168/2415 in males, OR 1.9 p=<0.001) age >64 (295/2573 incomplete in ≤64 vs 236/2737 in those >64, OR 1.3 p =<0.001). Procedural indications with a significantly different CIR included: abdominal pain (76/579 with vs 448/4598 without, OR 1.3 p=0.01) anaemia (125/1034 with vs 399/4143 without, OR 1.3 p=0.01) weight loss (39/198 with vs 485/4940 without, OR 1.7 p=0.009), non-surveillance (43/703 in surveillance patients vs 481/4474 in non-surveillance, OR 1.8 p=<0.001). Use of iv sedation reduced CIR (243/2082 with vs 281/3095 without, OR 1.3 p=0.002). Endoscopist factors associated with non-completion include non-nurse endoscopist (218/2380 incomplete for nurses vs 306/2797 in non-nurses, OR 1.2 p=0.04) Multivariate analysis identified several independent factors for non-completion (Abstract 002).

Conclusion Our study has identified new independent factors that predict CIR. These include; patient factors (Indication of weight loss) and endoscopist factors (nurse endoscopist). We feel nurse endoscopist is likely to be a surrogate marker for volume of colonoscopies performed. Perhaps this indicates fewer endoscopists should be performing more procedures.