OBJECTIVES: To evaluate the psychometric properties of the Diary for Irritable Bowel Syndrome Symptoms-Constipation (DIBSS-C), one of three subtype-specific measures developed by the Patient-Reported Outcome Consortium’s Irritable Bowel Syndrome (IBS) Working Group, for qualification by the United States (US) Food and Drug Administration to support primary and secondary endpoints in IBS clinical trials.
METHODS: Observational data were collected from 108 adults with IBS-C at 10 clinical sites across the US. Using a handheld electronic device, participants completed the DIBSS-C for 17 consecutive days; data related to bowel movements (BMs) were collected on an event-driven basis and abdominal symptoms were rated each evening. Global status items and the Gastrointestinal Symptom Rating Scale-IBS (GSRS-IBS) were completed on Days 10 and 17 and the IBS-Symptom Severity Scale (IBS-SSS) on Day 17. Data collected on Days 4 through 17 (Week 1 and Week 2) were analyzed to assess item-level performance, internal consistency, test-retest reliability, construct validity, and discriminating ability.
RESULTS: The abdominal symptom items were highly correlated with each other and the subscale comprised of these items produced high estimates of internal consistency (Cronbach’s alpha Week 1: 0.98 and Week 2: 0.96) and test-retest reliability (intraclass correlation coefficient: 0.93). As anticipated, test-retest reliability evidence was not as strong for the frequency-based BM-related outcomes. Key construct validity hypotheses were supported for both abdominal and BMrelated items (i.e., correlations with GSRS-IBS and IBS-SSS items addressing similar concepts were higher than others). While all comparisons were statistically significant for the abdominal symptoms, evidence for the discriminating ability of BM-related outcomes was strongest for known-groups comparisons based on constipation severity.
CONCLUSIONS: Overall, the results provide strong support for the reliability and validity of the DIBSS-C. Data from a recent clinical trial are expected to confirm these findings, support the measure’s responsiveness, and identify thresholds to support the interpretation of changes in scores.
