Background and Aims Crohn's disease (CD) is an inflammatory bowel disease (IBD) caused by a combination of genetic, clinical, and environmental factors. Identification of CD patients at high risk of requiring surgery may assist clinicians to decide on a top-down or step-up treatment approach. Methods We conducted a retrospective case-control analysis of a population-based cohort of 503 CD patients. A regressionbased data reduction approach was used to systematically analyse 63 genomic, clinical and environmental factors for association with IBD-related surgery as the primary outcome variable. Results A multi-factor model was identified that yielded the highest predictive accuracy for need for surgery. The factors included in the model were the NOD2 genotype (OR=1.607, P=2.3×10 −5 ), having ever had perianal disease (OR=2.847, P=4×10 −6 ), being post-diagnosis smokers (OR=6.312, P=7.4×10 −3 ), being an ex-smoker at diagnosis (OR=2.405, P=1.1×10 −3 ) and age (OR=1.012, P=4.4×10 −3 ). Diagnostic testing for this multi-factor model produced an area under the curve of 0.681 (P=1×10 −4 ) and an odds ratio of 3.169, (95 % CI P=1×10 −4 ) which was higher than any factor considered independently.Conclusions The results of this study require validation in other populations but represent a step forward in the development of more accurate prognostic tests for clinicians to prescribe the most optimal treatment approach for complicated CD patients.