Sir,

We thank Prof. Wiwanitkit for his comments on our article “Starreveld scoring method in diagnosing paediatric constipation”.

Regarding the first question we indeed used strict criteria to diagnose children with functional constipation (Iowa criteria) and made an effort to distinguish these children from those who have functional nonretentive faecal incontinence (FNRFI) and functional abdominal pain (FAP), using the at that time valid Rome II criteria [1]. According to the Rome II criteria, constipation in these conditions (FNRFI and FAP) is ruled out and therefore their abdominal radiographs were used as controls [2].

Regarding the second question we indeed used four observers. Although in most reports three raters are used we used four raters to cover the complete range of experience in rating abdominal radiographs (student, radiology fellow, radiologist and paediatric radiologist). We are not aware of statistical hurdles in this respect. Differences in experience do indeed result in different repeated scoring; however in our investigation, to our surprise, the student did best. The student was trained to apply the Starreveld score on two occasions; the other raters used the score on a basis “as described”.

Finally, regarding the third question, Prof. Wiwanitkit could be right; however, our groups will become too small to give a statistically sound and definite answer.