Loudin Michael, Johnson Kimberly, Lum Joshua, Amy Laird, Jack Wiedrick, Kian Keyashian, A Model for Identifying Actionable Findings on Computed Tomography in Crohn’s Disease Patients in the Emergency Department, Journal of Digestive Disorders And Diagnosis, Volume 1, Issue 3, 2017, Pages 1-10, ISSN 2574-4526, https://doi.org/10.14302/issn.2574-4526.jddd-17-1688.
(https://oap-researcharticles.org/jddd/article/552)
Abstract: Patients with inflammatory bowel disease (IBD) frequently visit the emergency department (ED). The use of cputed tomography (CT) scans in this population has drastically increased in recent years and may confer an increased risk of malignancy. Records were obtained for IBD patients aged 18 or older who visited our institutional ED with a gastrointestinal chief complaint and who had a CT scan ordered by an ED physician. A predictive model for identifying a clinically actionable finding (CAF) on CT scan was created using logistic regression carried out on a predetermined set of variables. Data were available on 156 Crohn’s disease (CD) patients contributing 350 visits and 63 ulcerative colitis (UC) patients contributing 114 total visits. CAF was identified at 108/350 (30.9%) of visits in CD patients and 33/114 (29.0%) of visits in UC patients. History of CAF (OR 11.6, CI 4.54-29.6) and a platelet count above 400,000/mL (OR 3.42, CI 1.56-7.50) were the strongest predictors of CAF. History of psychiatric illness (OR 0.67, CI 0.35-1.29) and diarrhea (OR .043, CI 0.23-0.83) were associated with a lower likelihood of CAF. A prediction model was created that was able to detect 94.4% of CAF cases while correctly predicting CAF non-cases 35% of the time. This model holds promise as a tool to reduce imaging in this population.
Keywords: Inflammatory bowel diseases; Crohn disease; ulcerative colitis; X-ray computed tomography; Emergency Department