BACKGROUND: Crohn's disease is a heterogeneous disease, and several classification systems have been developed to classify the patients in more homogeneous groups. Our aim was to assess the intra- and interobserver variation when classifying patients according to the widely used Vifenna classification.

METHODS: Ten randomly selected Crohn's disease cases were presented to 11 Danish gastroenterologists with a special interest in inflammatory bowel diseases. Clinical details, together with endoscopic, radiologic, and pathologic reports, were presented to the participants as a PowerPoint slide show, sent by e-mail with a data collection form. The experts were asked to classify the cases according to the Vienna classification and to evaluate intraobserver variation; the participants classified the patients 3 times. The strength of agreement was calculated using kappa statistics.

RESULTS: Classification of the patients according to age gave a kappa value of 1.00. The intraobserver kappa value was good, with an average kappa value of 0.75 (range, 0.42-0.86) for location and 0.77 (range, 0.53-1.00) for behavior. The mean overall interobserver kappa value was 0.64 (range, 0.12-1.00), which improved slightly between the first and third rounds. When classifying according to location and behavior, most patients were classified in 2 or 3 different ways, and in no patients was there full agreement among the observers for both location and behavior.

CONCLUSIONS: In this study, we found an overall good interobserver agreement when using the Vienna classification, although when looking at individual cases, there was some disagreement.