Interobserver and intraobserver reliability of the Rome II criteria in children

Abstract

Background: Functional gastrointestinal disorders are common in children. It has been suggested that the diagnosis of these conditions should be based on the "pediatric Rome II" criteria. The interobserver reliability for the DSM-IV, another symptom-based criteria is considered almost perfect in multiple studies. There are no studies assessing the reliability of the Rome II criteria in children.

Objectives: To evaluate the reliability of the pediatric Rome II criteria.

Methods: Interobserver reliability-Ten pediatric gastroenterologists and 10 fellows in pediatric gastroenterology were provided with 20 clinical vignettes, the Rome II criteria, and a list of 15 possible diagnoses. Each of the raters was instructed to select one or more diagnoses for each vignette. Intraobserver reliability-The specialists were provided with the same set of vignettes 4 months later.

Results: Average percentage of agreement coefficient: 45% (specialists), 47% (fellows). In order to correct for possible agreement by chance, we calculated the kappa coefficient, a measure of pairwise agreement corrected for chance. Specialists: k = 0.37 (p < 0.0001), trainees: k = 0.41, (p < 0.0001). Physicians with a special interest in functional gastrointestinal disorders (k = 0.37, p < 0.0001), and other specialists (k = 0.38, p < 0.0001). Analysis of data in pain and constipation diagnosis subgroups revealed even lower kappa (constipation: k = 0.2, p < 0.0001; pain: k = 0.3, p < 0.0001). Intraobserver agreement: k = 0.63 (p < 0.0001).

Conclusion: The interobserver reliability of the Rome II criteria among pediatric gastroenterologists and fellows is low. Further validation of the criteria is necessary.