Starreveld scoring method in diagnosing childhood constipation

BACKGROUND: Four scoring methods exist to assess severity of fecal loading on plain abdominal radiographs in constipated patients (Barr-, Starreveld-, Blethyn- and Leech). So far, the Starreveld score was used only in adult patients. OBJECTIVE: To determine accuracy and intra- and inter-observer agreement of the Starreveld scoring method in the diagnosis of functional constipation among pediatric patients. In addition, we compared the Starreveld with the Barr scoring method. MATERIALS AND METHODS: Thirty-four constipated and 34 non-constipated children were included. Abdominal radiographs, obtained before treatment, were rated (Starreveld- and Barr) by 4 observers. A second observation after 4 weeks was done by 3 observers. Cut-off level for the Starreveld score, accuracy as measured by the area under the receiver operator characteristics curve, and inter- and intra-observer agreement were calculated. RESULTS: Cut-off value for the Starreveld score was 10. AUC for Starreveld score was 0.54 and for Barr score 0.38, indicating poor discriminating power. Inter-observer agreement was 0.49-0.52 4 (Starreveld) and 0.44 (Barr), which is considered moderate. Intra-observer agreement was 0.52-0.71 (Starreveld) and 0.62- 0.76 (Barr). CONCLUSION: The Starreveld scoring method to assess fecal loading on a plain abdominal radiograph is of limited value in the diagnosis of childhood constipation.