Search in the Abstract Database

Abstracts Search 2014

* = Presenting author

P452. Increasing interobserver agreement on IBD endoscopic scoring systems: results from the IGIBDEndo educational program

M. Daperno1,2, M. Comberlato3,4, F. Bossa4,5, A.G. Bonanomi4,6, G. Lombardi4,7, L. Biancone4,8, R. Cosintino4,9, R. D'Incà4,10, A. Cassinotti4,11, R. Mangiarotti4,9, A. Papa4,12, R. Pica4,13, F. Rizzello4,14, A. Armuzzi4,12, A. Orlando4,15, on behalf of IG-IBD4, 1Fondazione IBD Onlus, IBD Foundation, Torino, Italy, 2AO Ordine Mauriziano, Gastroenterology Division, Turin, Italy, 3Ospedale di Bolzano, Gastroenterology, Bolzano, Italy, 4IGIBD, Italian Group for IBD, Florence, Italy, 5Casa del Sollievo Divina Provvidenza, Gastroenterology, San Giovanni Rotondo, Italy, 6A.O. Careggi, Gastroenterology Division, Florence, Italy, 7AORN A. Cardarelli, Napoli, UOC of Gastroenterology, Napoli, Italy, 8Tor Vergata University, Internal Medicine, Rome, Italy, 9A.O.San Camillo Forlanini, Gastroenterology Unit, Rome, Italy, 10University of Padova, Gastroenterology Department, Padova, Italy, 11Luigi Sacco University Hospital, Gastroenterology Department, Milano, Italy, 12Catholic University, Internal Medicine and Gastroenterology Unit, Complesso Integrato Columbus, Rome, Italy, 13Pertini Hospital, Gastroenterology Unit, Roma, Italy, 14S. Orsola Malpighi Hospital, Internal Medicine, Bologna, Italy, 15AO Villa Sofia-Cervello, Internal Medicine, Palermo, Italy

Background

Endoscopic scoring systems are a reliable tool to obtain objective measure of inflammation in inflammatory bowel disease (IBD) and they are increasingly used in clinical trials, also as surrogate markers. Moreover their use is advocated by actual ECCO (European Crohn's and Colitis Organization) guidelines, e.g. Rutgeerts' score in post-surgical recurrence monitoring. However agreement is not always extremely high, especially among non-trial-dedicated Gastroenterologist dealing in clinical practice with IBD patients.

Aims of this study were to promote an educational program dedicated to IBD endoscopic scoring systems, and to measure pre- and post-program agreement among participants.

Methods

171 Gastroenterologists were involved in 7 meetings dedicated to education focused on Mayo endoscopic subscore for ulcerative colitis (UC), Rutgeerts' score for post-surgical Crohn's disease (CD) and CDEIS (CD index of severity) and SES-CD (simple endoscopic score for CD) for luminal CD. For UC and post-surgical CD 5 clips for each scenario was reviewed blindly by all partecipants via a dedicated iPad system before and after extensive discussion of scoring modalities, for luminal CD other 5 endoscopic clips were reviewed once for CDEIS and SES-CD. Statistical analysis included descriptive and kappa statistics, as well as coefficient of variation.

Results

Agreement differed significantly before and after discussion for 3 UC (p < 0.001) and 2 post-surgical CD (p < 0.01) clips, respectively, leading ro relevant kappa statistics variations. For SES-CD (44%) and CDEIS (48%), similar coefficient of variation were calculated, and ICC (intraclass correlation coefficient) was 0.725 and 0.740, respectively. Results are reported in detail in the Table.

Table: Measure of agreement at first and second observation of 5 endoscopic clips with relevant endoscopic scores
Agreement with correct valuep valueKappa
At first round (5 clips)At second round (5 clips) At first round (5 clips)At second round (5 clips)
UC (Mayo)75%87%<0.00010.4450.713
Post-surgical CD (Rutgeerts')86%96%<0.00010.6560.853
Luminal CD (SES-CD)0.725ICC for agreement
Luminal CD (CDEIS)0.740ICC for agreement

Conclusion

The IGIBDEndo project was focused on an unmet need: a better knowledge and diffusion of relevant endoscopic scores for IBD. The methodology resulted efficient to promote an increase in interobserver agreement, however agreement is still suboptimal, in line with data from Literature at least for Mayo endoscopic subscore. Further educational efforts are to be considered also in this field, the methodology may be exported to different clinical conditions.