Systematic evaluation of complications related to endoscopy in a training setting: a prospective 30-day outcomes study

Background: The 30-day frequency of negative outcomes after outpatient endoscopy performed by gastroenterology fellows is unknown. Methods: Questionnaires were mailed to 1000 consecutive patients 30 days after endoscopy to evaluate procedure-related negative outcomes (serious and minor adverse events) and patient satisfaction. Serious adverse events were defined as follows: oversedation requiring administration of a reversal agent, and those that resulted in a physician visit, emergency department visit, admission to the hospital, or death. Minor adverse events were defined as all problems other than serious adverse events that patients related to their endoscopic procedure. Results: The 30-day frequency of negative outcomes in the 869 patients who responded was 14.3%, of which 0.6% were serious and 13.7% were minor adverse events. The frequency of negative outcomes was 17.1% for EGD, 15.0% for colonoscopy, 24.4% for combined EGD and colonoscopy, and 7.8% for flexible sigmoidoscopy. One hundred percent of the serious adverse events were known to us, but only 16.0% of minor adverse events (p < 0.001). Multiple logistic regression identified midazolam dose (OR for each 1 mg increase in dose 4.5; 95% Cl [2.7, 7.3]; p < 0.001), treatment with warfarin (OR 3.0; 95% Cl [1.4,6.2]; p = 0.003), comorbid disease (OR 2.1; 95% Cl [1.3, 3.4]; p = 0.001), endoscopy performed in July or August (OR 2.0; 95% Cl [1.1,3.7]; p = 0.02), and age (OR for each 1 year increase in age 1.03; 95% Cl [1.01, 1.05]; p = 0.01) as independent predictors of negative outcomes. There was a significant association between negative outcomes and decreased patient satisfaction, and patients who reported negative outcomes were less likely to agree to endoscopy in the future. Conclusions: Serious adverse events were rare after endoscopy performed by gastroenterology fellows. Contacting patients 30 days after outpatient endoscopy significantly improved the detection of negative outcomes. Although the majority of negative outcomes were minor, these adverse events were associated with decreased patient satisfaction.