Reproducibility, validity, and responsiveness of a disease-specific symptom questionnaire for gastroesophageal reflux disease

Abstract
The purpose of this study was to establish the reproducibility, validity, and responsiveness of a symptom questionnaire to assess patients with gastroesophageal reflux disease (GERD). A total of 300 patients with GERD completed questionnaires before and 6 months after laparoscopic Nissen fundoplication. Forty-six GERD patients who continued on omeprazole served as controls. Lower esophageal sphincter pressure, 24-h pH, and quality of life (SF36) were measured at baseline and follow-up. Reproducibility was calculated as an intraclass correlation coefficient (ICC) from a repeated-measures analysis of variance on symptom scores (SS) on two consecutive days. Validity was established by correlating SS with 24-h pH and SF36 scores. Responsiveness was calculated as the the ratio of the mean paired difference in score in the surgical group to the within-subject variability in control subjects. Reproducibility was very high, as revealed by an ICC of 0.92. Strong correlations between SS and SF36 scores at baseline and after surgery demonstrated high cross-sectional validity. Correlation between change in SS and change in pH, SF36 pain, general health, and physical health scores demonstrated longitudinal validity. The mean (95% confidence interval) paired differences in SS were 25.6 (23.7, 27.5) in the study and 2.0 (-3.2, 7.3) in the control groups, and the responsive index was 1.0. The estimated minimally important clinical difference was 7. We conclude that the symptom score is a reproducible, valid, and responsive instrument for assessing symptoms caused by GERD.