Search in the Abstract Database

Search Abstracts 2011

* = Presenting author

P121. Inter-observer agreement for Crohn's disease sub-phenotypes using the Montreal Classification: How good are we? A multi-centre Australasian study

K. Krishnaprasad1,2, J.M. Andrews2, R. Gearry3, T. Florin4, I. Lawrance5, P.A. Bampton6, R. Leong7, G. Mahy8, J. Doecke9, G. Radford-Smith1,10

1Queensland Institute of Medical Research, Brisbane, Australia; 2Royal Adelaide Hospital, Adelaide, Australia; 3University of Otago, Christchurch, New Zealand; 4Mater Adult Hospital, Brisbane, Australia; 5Fremantle Hospital, Fremantle, Australia; 6Flinders Medical Centre, Adelaide, Australia; 7Concord Hospital, Sydney, Australia; 8Townsville Hospital, Townsville, Australia; 9CSIRO Preventative Health Flagship, Brisbane, Australia; 10Royal Brisbane & Women's Hospital, Brisbane, Australia

Aim: Crohn's Disease (CD) exhibits significant clinical heterogeneity. Classification systems attempt to describe this; however, their utility and reliability is dependent on Inter-Observer Agreement (IOA). We therefore formally sought to evaluate IOA using the Montreal Classification (MC).

Methods: Data from 35 CD patients, diagnosed from January 2004 onwards, were randomly contributed from 6 Australian IBD centres (5–7 cases per site). De-identified records were presented to a panel of 13 expert practitioners from 8 member centres of the Australia and New Zealand Inflammatory Bowel Disease Consortium (ANZIBDC). The practitioners (clinicians and nurses) classified the cases using the MC and forwarded their data for central blinded analysis. Data on smoking, surgery, medications and a prediction for future disease outcome was also tested. IOA was evaluated using Kappa statistics, with pre-specified outcomes of κ > 0.8 excellent; 0.61–0.8 good; 0.41–0.6 moderate and <0.4 poor.

Results: Of the 35 cases, 22 (63%) were male, average age was 33 yrs (19–69) and 7 had resections, with average disease duration of 3.5 yrs (1–6). With regard to IOA at diagnosis; 10 complete rating sets are available on 27–35 subjects (missing data on some). Agreement was excellent for age (A1-A3), κ = 0.84; good for disease location (L1-L3), κ = 0.73; only moderate for disease behaviour (B1-B3), κ = 0.54; and good for the presence of perianal disease (p), κ = 0.6. At last follow-up, 9 complete rating sets are available for 26–27 cases. IOA was good for disease location, κ = 0.68; and only moderate for disease behaviour, κ = 0.46; but excellent for the presence of perianal disease, κ = 0.88. IOA for the use of Immunosuppression ever during follow-up and for the presence of stricture were both good (κ = 0.79 and 0.64 respectively).

Conclusion: IOA is generally robust amongst this panel of experts; however some areas are less consistent than others. The reasons for this deserve further attention to ensure a uniform approach to classification or perhaps better initial recording of clinical data. Omissions and inaccuracies reduce the value of clinical data when comparing cohorts across different centres, and potentially impact on the ability to translate recent genetic discoveries into novel tools that aim to improve clinical practice.