The Impact of Bowel Cleansing on Follow-up Recommendations in Average-Risk Patients With a Normal Colonoscopy

Stacy B Menees, MD, MS; Eric Elliott, MPH; Shail Govani, MD; Constantinos Anastassiades, MD; Stephanie Judd, MD; Annette Urganus, MPH; Suzanna Boyce, MPA; Philip Schoenfeld, MD, MSEd, MSc (Epi)


Am J Gastroenterol. 2014;109(2):148-154. 

In This Article


Demographic Data

Between 1 January 2009 and 31 December 2009, 4,527 out-patient colonoscopies were performed for an indication of "average-risk" and "screening" at the Ann Arbor VAHCS and University of Michigan. After applying exclusion criteria, 1,387 normal colonoscopies remained for analysis. These 1,387 colonoscopies were performed by 56 different gastroenterologists, excluding GI fellows. A total of 18 GI fellows participated in 18.7% of these endoscopic procedures. The majority (89.7%) of procedures were classified as technically "not difficult." The mean age of the subjects was 56.7±7.1 years; mean body mass index was 28.2±5.7 kg/m2, and the subjects were predominantly Caucasian (77.6%) and male (50.9%). Demographic differences between groups based on preparation quality (Table 1a) and recommendation type (Table 1b) are provided.

Follow-up Recommendations for Repeat Colonoscopy

Recommendations that were inconsistent with guidelines were given in 23.9% (332/1,387) of average-risk patients with a normal screening colonoscopy (Table 2). Preparations rated as excellent/good had the highest frequency (84.7%) of recommendations for repeat colonoscopy in 10 years. Of the preparations, 25% rated as fair were recommended for repeat colonoscopy in 10 years. Preparations rated as poor had a high frequency (65.9%) of being recommended to have repeat colonoscopy in ≤1 year.

Factors Associated With Follow-up Recommendations Inconsistent With Guidelines

Crude estimates for predictors of recommendations inconsistent with guidelines are given in Table 3, whereas Table 4 demonstrates the effect of predictors after adjustment for all data collected. Bowel preparation quality with a rating of fair or poor was associated with an 18-fold and 2.3-fold increase in the odds of receiving a recommendation inconsistent with guidelines, respectively, along with age ≥70 (odds ratio=2.2; 95% confidence interval 1.2–4.1).