Predictors of post-colonoscopy emergency department use

Background and Aims: Unplanned hospital visits within 7 days of colonoscopy were recently proposed as a quality measure. It is unknown whether patient, procedure, or endoscopist characteristics predict post-colonoscopy emergency department (ED) visits. Our aim was to determine the incidence and relatedness of ED visits within 7 days of colonoscopy and to identify predictors of post-colonoscopy ED use. Methods: In this retrospective, single-center, cohort study, we evaluated outpatient colonoscopies performed at a tertiary academic medical center or affiliated facility between January 2008 and September 2013. We determined the incidence of ED visits within 7 days of colonoscopy and the relatedness of the ED visit to the procedure. We assessed for independent factors associated with ED use within 7 days using logistic regression analysis. Results: We reviewed 50,319 colonoscopies performed on 44,082 individuals (47% male, median age 59 years) by 40 endoscopists. There were 382 (0.76%) ED visits after colonoscopy, of which 68% were related to the procedure. On multivariate analysis, recent ED visit (odds ratio [OR], 16.60; 95% confidence interval [CI], 12.83-21.48; P <.001), EMR (OR, 4.69; 95% CI, 2.82-7.79; P <.001), number of medication classes (OR, 1.18; 95% CI, 1.11-1.26; P <.001), endoscopist adenoma detection rate (ADR) (OR, 1.14; 95% CI, 1.01-1.29; P =.029), and white race (OR, 0.77; 95% CI, 0.62-0.97; P =.028) were identified as independent variables associated with ED visits after colonoscopy. Conclusions: Increased patient complexity, higher endoscopist ADR, and EMR were associated with increased ED use after colonoscopy. Patients at high risk for an unplanned hospital visit within 7 days should be targeted for quality improvement efforts to reduce adverse events and cost.