Risk prediction models for inadequate bowel preparation in colonoscopy: a systematic review and meta-analysis.
Yuanyuan Zhang, Jiawei Qian, Shiyu Peng et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review and meta-analysis assessed the performance, methodological quality, and applicability of prediction models for inadequate bowel preparation before colonoscopy. It included 19 models; most used logistic regression, and reported inadequate-preparation incidence ranged from 6.3% to 33.0%. Across 17 validated models, pooled discrimination was an AUC of 0.73 (95% CI: 0.70-0.77). All included studies were judged at high risk of bias, mainly because of inadequate reporting in the analysis domain. Diabetes and constipation were common predictors, while behavioral factors related to preparation were rarely included.
Key findings
- The pooled AUC across 17 validated models was 0.73 (95% CI: 0.70-0.77), indicating moderate discrimination.
- All included studies were judged at high risk of bias, primarily because of inadequate reporting in the analysis domain.
- Diabetes and constipation were the most frequently identified predictors; behavioral preparation factors were rarely incorporated.
- Reported inadequate bowel preparation incidence ranged from 6.3% to 33.0%.
Why it matters
Risk models could help identify patients more likely to have inadequate bowel preparation, which may affect colonoscopy outcomes. However, the pooled discrimination was moderate and the evidence base had substantial methodological concerns. The limited inclusion of behavioral factors also highlights a potential gap in how current models represent preparation-related risks.
What to keep in mind
All included studies were judged at high risk of bias, particularly because of inadequate reporting in the analysis domain.