P168 Development of predictive model for early detection of Inflammatory Bowel Disease in Korea
Lee, C.(1)*;Koh, S.J.(1);Choi, K.(1);Lee, H.J.(1);Im, J.P.(1);Kim, J.S.(1);
(1)Seoul National University College of Medicine, Department of Internal Medicine and Liver Research Institute, Seoul- Korea, Korea- Republic Of;
Background
Diagnostic delay in inflammatory bowel disease (IBD) was frequent and may be associated with poor outcomes. The aim of this study was to develop a predictive model for early diagnosis of IBD.
Methods
From March 2020 to January 2022, a 15-item Red Flags questionnaire was administered to all patients who visited outpatient clinic with gastrointestinal symptoms suspected of IBD. All patients underwent additional diagnostic tests to confirm IBD. We estimated a model that could predict the early diagnosis of IBD using questionnaire and laboratory test results.
Results
A total of 186 patients were enrolled in this study and 114 (61.2%) patients were finally confirmed with IBD. The logistic regression identified 4-item independently associated with a diagnosis of IBD. 1) Perianal fistula/abscess or perianal lesions (apart from hemorrhoids) (OR, 2.6; p=0.044). 2) Chronic abdominal pain/discomfort (>3 months) (OR, 0.3; p=0.013). 3) Rectal bleeding (OR, 7.4; p<0.001). 4) First-degree relative with confirmed IBD (OR, 12.0; p=0.033). A receiver operating characteristic curve was used to evaluate predictability of the model for diagnosis of IBD and each area under the curve was as follows. 1) Questionnaire only : AUC 0.831 (95% confidence interval; 0.766-0.896), 2) Questionnaire and serum laboratory tests including blood cell counts, erythrocyte sedimentation rate and C-reactive protein : AUC 0.885 (95% C.I.; 0.834-0.936), 3) Questionnaire, serum laboratory tests and fecal calprotectin : AUC 0.895 (95% C.I.; 0.845-0.944). This model confirmed that the predictive value was particularly excellent for early diagnosis of Crohn's disease. 1) Questionnaire only : AUC 0.891 (95% C.I.; 0.818-0.964), 2) Questionnaire and serum laboratory tests : AUC 0.960 (95% C.I.; 0.914-1.000), 3) Questionnaire, serum laboratory tests and fecal calprotectin : AUC 0.975 (95% C.I.; 0.944-1.000).
Conclusion
The predictive model for early diagnosis of IBD has a high predictive value. The model allows for early detection of IBD from non-IBD patients.