A national colorectal cancer screening programme started in England in 2013, offering one-off flexible sigmoidoscopy to all men and women aged 55 years in addition to the biennial faecal occult blood testing programme offered to all individuals aged 60-74 years. We analysed data from six pilot flexible sigmoidoscopy screening centres to examine factors affecting the adenoma detection rate (ADR). We did a retrospective analysis of flexible sigmoidoscopy screening procedures performed in individuals aged 55 years at six pilot sites in England as part of the National Health Service Bowel Scope Screening programme. ADR (number of procedures in which at least one adenoma was removed or biopsied, divided by total number of procedures) was calculated for each site and each endoscopist. Multiple regression models were used to examine the variation in ADR with withdrawal time and extent of examination, and the effect of other factors including comfort and bowel preparation on extent of examination. The analysis included 8256 procedures done between May 7, 2013, and May 6, 2014. The overall ADR was 9·1% (95% CI 8·5-9·8; 755 of 8256 procedures), varying from 7·4% (6·2-8·9) to 11·0% (9·1-13·4) by screening centre. The ADR was 11·5% (95% CI 10·6-12·5; 493 of 4299 procedures) in men and 6·6% (5·9-7·4; 262 of 3957 procedures) in women (p<0·0001). On multivariate analysis, factors associated with adenoma detection were male sex (relative risk 1·69, 95% CI 1·46-1·95; p<0·0001) and a withdrawal time from the splenic flexure of at least 3·25 min in negative procedures (1·22, 1·00-1·48; p=0·045). However, increasing the withdrawal time to 4·0 min or more did not increase the likelihood of adenoma detection (1·22, 0·99-1·51; p=0·057). Procedures not reaching the splenic flexure were associated with lower chance of adenoma detection (eg, 0·77, 0·66-0·91; p=0·0015 for procedures reaching the descending colon), but there was no additional benefit associated with reaching the transverse colon (0·83, 0·67-1·02; p=0·069). Women (0·83, 0·80-0·87; p<0·0001), individuals with adequate (0·79, 0·76-0·83; p<0·0001) or poor (0·58, 0·51-0·67; p<0·0001) bowel preparation (compared with good bowel preparation), and those with mild (0·82, 0·76-0·88; p<0·0001) or moderate or severe (0·58, 0·51-0·66; p<0·0001) discomfort (compared with no discomfort) were less likely to have a procedure reaching the splenic flexure. Key performance indicators for flexible sigmoidoscopy screening should be defined, including standards for insertion and withdrawal times, optimal depth, and bowel preparation. ADR could be improved by recommending a withdrawal time from the splenic flexure of at least 3·25 min (ideally 3·5-4·0 min). None.