Point-of-care practice audits allow documentation of procedural outcomes to support quality improvement in endoscopic practice. To evaluate a colonoscopists' practice audit tool that provides point-of-care data collection and peer-comparator feedback. A prospective, observational colonoscopy practice audit was conducted in academic and community endoscopy units for unselected patients undergoing colonoscopy. Anonymized colonoscopist, patient and practice data were collected using touchscreen smartphones with automated data upload for data analysis and review by participants. The main outcome measures were the following colonoscopy quality indicators: colonoscope insertion and withdrawal times, bowel preparation quality, sedation, immediate complications and polypectomy, and biopsy rates. Over a span of 16 months, 62 endoscopists reported on 1279 colonoscopy procedures. The mean cecal intubation rate was 94.9% (10th centile 84.2%). The mean withdrawal time was 8.8 min and, for nonpolypectomy colonoscopies, 41.9% of colonoscopists reported a mean withdrawal time of less than 6 min. Polypectomy was performed in 37% of colonoscopies. Independent predictors of polypectomy included the following: endoscopy unit type, patient age, interval since previous colonoscopy, bowel preparation quality, stable inflammatory bowel disease, previous colon polyps and withdrawal time. Withdrawal times of less than 6 min were associated with lower polyp removal rates (mean difference -11.3% [95% CI -2.8% to -19.9%]; P=0.01). Cecal intubation rates exceeded 90% and polypectomy rates exceeded 30%, but withdrawal times were frequently shorter than recommended. There are marked practice variations consistent with previous observations. Real-time, point-of-care practice audits with prompt, confidential access to outcome data provide a basis for targeted educational programs to improve quality in colonoscopy practice.