A Novel User-Friendly Model to Predict Corticosteroid Utilization in Newly Diagnosed Patients with Ulcerative Colitis

Background: Corticosteroid (CS) use is an important marker of poor prognosis in ulcerative colitis (UC). Our aim was to develop and validate a model to predict the risk of CS utilization over the course of disease in newly diagnosed patients with UC. Methods: Newly diagnosed patients with UC from a nationwide VA cohort were followed over time to evaluate factors predictive of CS use. Multivariate logistic regression was performed. Model development was performed in a random 2/3 of the total cohort and then validated in the remaining 1/3. The primary outcome was the use of CS for the management of UC. Candidate predictors included routinely available data at the time of UC diagnosis, including demographics, laboratory results, and index colonoscopy findings. Results: Six hundred ninety-nine eligible patients with UC were followed for a median duration of 8 years. Two hundred eighty-eight patients (41.2%) required CS use for the management of UC. Key predictors for CS utilization selected for the model were as follows: age, non-African American ethnicity, presence of hypoalbuminemia, and iron-deficiency anemia at the time of UC diagnosis, endoscopic extent, or severity of disease at index colonoscopy. Model discrimination was good (area under the receiver operator curve 0.71 [95% confidence interval, 0.66-0.76] for the model including baseline UC extent and 0.71 [95% confidence interval, 0.67-0.76]) for the model including baseline UC severity. Model calibration was consistently good in all models (Hosmer-Lemeshow goodness of fit P > 0.05). The models performed similarly in the internal validation cohort. Conclusions: We developed and internally validated a novel prognostic model to predict CS use among patients with newly diagnosed UC.