Related Experiment Video
Updated: Aug 16, 2026
Structured Approach to Colonoscopy Technique Optimization: A Single-Center Experience with Novice Endoscopists
Published on: July 11, 2025
Risk prediction models for inadequate bowel preparation before colonoscopy: A systematic review and meta-analysis
Hui-Ling Yan1, Jia-Yu Fu2, Mao-Ting Huang1
1School of Nursing, Hunan Engineering Research Center for Early Diagnosis and Treatment of Liver Cancer, Hunan Province Key Laboratory of Tumor Cellular & Molecular Pathology, Cancer Research Institute, Hengyang Medical School, University of South China, Hengyang, Hunan Province, China.
Risk prediction models for inadequate bowel preparation (IBP) show moderate performance but suffer from significant bias and limited external validation. Standardized reporting and independent validation are crucial for improving colonoscopy quality.
Area of Science:
- Gastroenterology
- Clinical Epidemiology
- Health Informatics
Background:
- Inadequate bowel preparation (IBP) compromises colonoscopy safety, efficiency, and diagnostic accuracy.
- Existing multivariable prediction models for IBP risk have shown variable performance.
Purpose of the Study:
- To systematically evaluate the predictive performance and methodological quality of existing risk prediction models for IBP before colonoscopy.
Main Methods:
- Systematic search of multiple databases (PubMed, Embase, etc.) up to December 2025.
- Data extraction using CHARMS; risk of bias and applicability assessed with PROBAST.
- Pooled analysis of AUCs with random-effects models; heterogeneity assessed using I².
Main Results:
- Included 31 studies with 46 prediction models; common predictors: constipation, diabetes, age, BMI, prior colorectal surgery.
- Pooled AUCs: 0.76 (internal validation) and 0.72 (external validation), with substantial heterogeneity (>90%).
- Limited external validation (10 studies, 15 cohorts); high risk of bias in most studies (analysis domain); low applicability concerns.
Conclusions:
- IBP prediction models demonstrate moderate to good average discrimination but are limited by heterogeneity, bias, and insufficient external validation.
- Standardized outcome definitions and reporting are needed.
- Large, multicenter, independent external validations are essential to enhance clinical utility.
