Revalidation of description of constipation in terms of recall bias and visual scale analog questionnaire

The present study was designed to identify a cut-off value to define subjective and relatively objective criteria of constipation using the visual scale analog questionnaire (VSAQ) in healthy subjects. In addition, the importance of recall bias when evaluating constipation was investigated by repeating the questionnaire and ensuring the subjects maintained diaries.

Seven hundred and sixty healthy hospital personnel were questioned by means of a standard questionnaire. Subjects were initially asked if they were constipated (self-reported) and their daily defecation frequencies. Severity of the parameters of constipation, the consistency of defecation in the form of hard stools, straining and incomplete evacuation were also investigated using a VSAQ (0-10). Subjects were asked to complete a standard form about their daily bowel habits in the subsequent 7 days (diary). At the end of this series, the questionnaire forms completed at the beginning were readministered. Using the criteria of functional constipation, the prevalence of self-reported, symptom-based (>/=2 criteria) and diary-based (>/=2 criteria in the diary) were defined.

Of the subjects, 48.5% (369/760) completed diaries regarding their bowel habits and completed the questionnaire for the second time (198 female, 171 male; mean age 31.6 +/- 7.1 years). According to only interrogation, 29.8% of subjects reported that they were constipated; however, this number increased to 39.6% when symptom-based constipation (>/=2 criteria) was considered. Significant agreement was observed between the results of self-reported constipation in form I and II, and symptom-based and diary-based constipation (concordance = 77.7-98.6%, k = 0.47-0.97). Furthermore, 98.1% of the subjects who reported that they were not constipated scored 3 on the VSAQ; conversely, 91.8% who accepted being constipated scored >3 for the same question. A total of 76.1% subjects who had symptom-based constipation scored 3 on the VSAQ, 97.3% of those who had <2 criteria scored 3. When asked 'Are you constipated?' 1.2% of subjects with none of the criteria for diary-based constipation, and 10.7% of subjects who had one criteria scored >3 on the VSAQ. Also, 91.8% of those with three criteria and 100% of those with four criteria had a score >3 on the VSAQ for the same question.

The prevalence of constipation in the questionnaire form based on self-reported, symptom-based and diary-based criteria were highly compatible with the result obtained on readministration. Recall bias was negligibly low. Also, the present results suggest that the diagnosis of constipation is more accurate when >2 criteria are present. In addition, the VSAQ seems to be sensitive enough to differentiate subjects with constipation from those without, when a score of 3 has been chosen as the cut-off value for discrimination. However, this sensitivity was less in the group who stated they were constipated.