A predictive model identifies patients most likely to have inadequate bowel preparation for colonoscopy

Clin Gastroenterol Hepatol. 2012 May;10(5):501-6. doi: 10.1016/j.cgh.2011.12.037. Epub 2012 Jan 10.

Abstract

Background & aims: An inadequate level of bowel preparation can affect the efficacy and safety of colonoscopy. Although some factors have been associated with outcome, there is no strategy to identify patients at high risk for inadequate preparation. We searched for factors associated with an inadequate level of preparation and tested the validity of a predictive clinical rule based on these factors.

Methods: We performed a prospective study of 2811 consecutive patients who underwent colonoscopy examinations at 18 medical centers; clinical and demographic data were collected before the colonoscopy. Bowel preparation was classified as adequate or inadequate; 925 patients (33%) were found to have inadequate preparation. Multivariate analysis was used to identify factors associated with inadequate preparation, which were expressed as odds ratio (OR) and used to build a predictive model.

Results: Factors associated with inadequate bowel preparation included being overweight (OR, 1.5), male sex (OR, 1.2), a high body mass index (OR, 1.1), older age (OR, 1.01), previous colorectal surgery (OR, 1.6), cirrhosis (OR, 5), Parkinson disease (OR, 3.2), diabetes (OR, 1.8), and positive results in a fecal occult test (OR, 0.6). These factors predicted which patients would have inadequate cleansing with 60% sensitivity, 59% specificity, 41% positive predictive value, and 76% negative predictive value; they had an under the receiver operating characteristic curve value of 0.63. Assuming 100% efficacy of a hypothetical regimen to address patients predicted to be at risk of inadequate preparation, the rate would decrease from 33% to 13%.

Conclusions: We identified factors associated with inadequate bowel preparation for colonoscopy and used these to build an accurate predictive model.

Publication types

  • Evaluation Study

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Colonoscopy / methods*
  • Female
  • Humans
  • Male
  • Middle Aged
  • Predictive Value of Tests
  • Preoperative Care / methods*
  • Preoperative Care / standards*
  • Prospective Studies
  • Reproducibility of Results
  • Sensitivity and Specificity