June’s Paper of the Month examines existing predictive models, assesses their relevance, and discusses the barriers to their clinical implementation.
ECCO topical review on predictive models on IBD disease course and treatment response
Kirchgesner J, Verstockt B, Adamina M, Allin KH, Allocca M, Bourgonje AR, Burisch J, Doherty G, Dulai PS, El-Hussuna A, Misra R, Noor N, Pittet V, Powell N, Rodríguez-Lago I, Restellini S. ECCO topical review on predictive models on IBD disease course and treatment response. J Crohns Colitis. 2025 May 4:jjaf073. doi: 10.1093/ecco-jcc/jjaf073. Epub ahead of print. PMID: 40319340.
What is known about the subject?
Predictive models are becoming increasingly important in clinical research. However, the development of any predictive model should be based on an appropriate methodology that follows the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis [TRIPOD] guidelines [1]. It is important to measure the model's performance appropriately and assess its incremental benefit compared with real-life decision making. Inflammatory Bowel Disease (IBD) is a case in point.
IBD poses a clinical challenge due to its variable progression and response to treatment [2]. Despite the development of predictive models, their clinical application remains limited due to validation and methodological inconsistencies.
What the study adds?
The ECCO topical review addresses the key methodological shortcomings that have limited the clinical utility of predictive models in IBD, such as poor internal and external validation, limited transparency, and overly narrow predictor variables. In response, the authors set out a series of best practices designed to strengthen the development of models. These include robust calibration, rigorous discrimination assessment, and the use of pre-registration protocols to improve reproducibility. Importantly, the review broadens the scope beyond statistical performance, emphasising the practical feasibility of models, such as their cost-effectiveness, ease of integration into clinical workflows, and acceptability to clinicians and patients alike. In order to bridge the gap between research and practice, the authors advocate the embedding of validated prediction tools within care pathways through clinical decision support systems and impact studies that assess their practical benefit in routine care.
Implications for colorectal practice
This current ECCO topical review delivers a clear message to colorectal surgeons managing IBD: predictive models are no longer optional, but rather essential tools with which to optimize surgical timing, patient selection, and coordination with the multidisciplinary team. The review calls for a replacement of individual, experience-based decision-making with evidence-based, standardized approaches that can be consistently applied across clinical settings. This shift is intended to ensure more equitable, reproducible care and to minimise variability in outcomes that often arise from surgeon-dependent practice patterns.
References
- Collins, G.S., Reitsma, J.B., Altman, D.G. et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD Statement. BMC Med 13, 1 (2015). https://doi.org/10.1186/s12916-014-0241-z
- Open Source Research Collaborating Group (#OpenSourceResearch) , Biological Treatment and the Potential Risk of Adverse Postoperative Outcome in Patients With Inflammatory Bowel Disease: An Open-Source Expert Panel Review of the Current Literature and Future Perspectives, Crohn's & Colitis 360, Volume 1, Issue 3, October 2019, otz021, https://doi.org/10.1093/crocol/otz021