Background and study aims: Adenoma detection rates (ADRs) at screening flexible sigmoidoscopy are known to vary between endoscopists. Variability in the technique used and in the quality of bowel preparation may explain this. The aim of this study was to establish whether there is a relationship between the grading of bowel preparation and the ADR.
Materials and methods: The relationship between the ADR and assessment of bowel preparation was examined using the full United Kingdom Flexible Sigmoidoscopy Screening Trial data set (n = 38 601). The consistency of the bowel preparation classification was then investigated by six experienced endoscopists (video scorers), who examined 260 flexible sigmoidoscopy cases - 20 from each of the 13 trial endoscopists.
Results: Overall, the ADR was significantly higher in flexible sigmoidoscopy examinations with excellent or good bowel preparation ( P = 0.02). However, endoscopists with a higher ADR coded a smaller proportion of their examinations as having excellent/good preparation ( P = 0.002). Video scorers agreed with the trial endoscopists' definition of bowel preparation in 48.9 % of the readings, but they scored the quality of preparation as poorer than the trial endoscopists in 36.4 % and 40.6 %, respectively, in the intermediate-performance group (10 % < ADR < 14 %) and lower-performance group (ADR </= 10 %) in comparison with only 12.9 % in the high-performance group (ADR > or =14 %). There was a significant linear trend between the proportion scored as having poor bowel preparation and the ADR ( P < 0.001), varying from 2.7 % in the higher-performance ADR group to 13.4 % in the lower-performance group.
Conclusions: Endoscopists with a higher ADR are more likely to be critical of the quality of bowel preparation. Training in judgement processes such as non-acceptance of suboptimal bowel preparation is required in order to ensure universally high standards in screening procedures.