Evaluation of bowel distention with a neutral contrast agent: some statistical concerns.

Editor: I read with great interest the article by Dr Megibow and colleagues (1) in the January 2006 issue of Radiology. I have some concerns about the Materials and Methods and Results sections of this article. In the Materials and Methods section, under the subtitle “Statistical Analysis,” the authors state that the two groups included in their study were compared with respect to the mean values for each segment of the gastrointestinal tract (stomach, duodenum, jejunum, and ileum) by using the Mann-Whitney test, and they add that the combined mean scores were compared at a 95% confidence level by using the Wilcoxon signed rank test. As it is known, whereas a Mann-Whitney U test is used when ordinal data are obtained in the independent groups situation, a Wilcoxon signed rank test is used when ordinal data are obtained in the paired samples situation (2). In this case, the use of a Mann-Whitney U test seems appropriate since the two groups were composed of different patients. Nevertheless, it is not clear which test was really employed by the authors. In the Results section, under the subtitle “Qualitative Analysis,” the authors report that the mean value for distention in the ileum was significantly higher in group 1 than group 2. However, they give a P value of .140 that by definition does not indicate significance at the 5% -error level. This point needs to be clarified. Finally, since the qualitative image analysis included two independent readers and their combined mean scores were used for statistical comparisons, it might have been meaningful to evaluate the interobserver agreement with statistics.