A validated web-based tool to display individualised Crohn's disease predicted outcomes based on clinical, serologic and genetic variables

Siegel CA1,2, Horton H3, Siegel LS4,5, Thompson KD1,2, Mackenzie T2, Stewart SK1, Rice PW5, Stempak JM6, Dezfoli S3, Haritunians T3, Levy A3, Baek M3, Milgrom R6, Dulai PS1,2, Targan SR3, Silverberg MS6, Dubinsky MC7, McGovern DP3. Aliment Pharmacol Ther. 2015 Nov 15. doi: 10.1111/apt.13460. [Epub ahead of print]
Author information

1Dartmouth-Hitchcock Inflammatory Bowel Disease Center, Dartmouth-Hitchcock Medical Center, Lebanon, NH, USA. 2Department of Medicine, Geisel School of Medicine at Dartmouth, Hanover, NH, USA. 3F. Widjaja Foundation Inflammatory Bowel and Immunobiology Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA, USA. 4Siegel Environmental Dynamics, Hanover, NH, USA. 5Climate Interactive, Washington, DC, USA. 6Zane Cohen Centre for Digestive Diseases, Division of Gastroenterology, Mount Sinai Hospital, University of Toronto, Toronto, ON, Canada. 7Icahn School of Medicine at Mt. Sinai, New York, NY, USA.


BACKGROUND: Early treatment for Crohn's disease (CD) with immunomodulators and/or anti-TNF agents improves outcomes in comparison to a slower 'step up' algorithm. However, there remains a limited ability to identify those who would benefit most from early intensive therapy.

AIM: To develop a validated, individualised, web-based tool for patients and clinicians to visualise individualised risks for developing Crohn's disease complications.

METHODS: A well-characterised cohort of adult patients with CD was analysed. Available data included: demographics; clinical characteristics; serologic immune responses; NOD2 status; time from diagnosis to complication; and medication exposure. Cox proportional analyses were performed to model the probability of developing a CD complication over time. The Cox model was validated externally in two independent CD cohorts. Using system dynamics analysis (SDA), these results were transformed into a simple graphical web-based display to show patients their individualised probability of developing a complication over a 3-year period.

RESULTS: Two hundered and forty three CD patients were included in the final model of which 142 experienced a complication. Significant variables in the multivariate Cox model included small bowel disease (HR 2.12, CI 1.05-4.29), left colonic disease (HR 0.73, CI 0.49-1.09), perianal disease (HR 4.12, CI 1.01-16.88), ASCA (HR 1.35, CI 1.16-1.58), Cbir (HR 1.29, CI 1.07-1.55), ANCA (HR 0.77, CI 0.62-0.95), and the NOD2 frameshift mutation/SNP13 (HR 2.13, CI 1.33-3.40). The Harrell's C (concordance index for predictive accuracy of the model) = 0.73. When applied to the two external validation cohorts (adult n = 109, pediatric n = 392), the concordance index was 0.73 and 0.75, respectively, for adult and pediatric patients.

CONCLUSIONS: A validated, web-based tool has been developed to display an individualised predicted outcome for adult patients with Crohn's disease based on clinical, serologic and genetic variables. This tool can be used to help providers and patients make personalised decisions about treatment options.

© Copyright 2013-2017 GI Health Foundation. All rights reserved.
This site is maintained as an educational resource for US healthcare providers only. Use of this website is governed by the GIHF terms of use and privacy statement.