A Composite symptom score using frequency and severity correlates better to an objective measure of Gastro-Oesophageal Reflux Disease (GERD) than one scoring frequency of symptoms alone

Abstract

INTRODUCTION: The prevalence of GERD is increasing. Community prevalence in Sri Lanka is unknown. There is lack of a practical screening instrument to use in an epidemiological setting. OBJECTIVE : To develop a practical clinical score to screen for GERD in the community and assess whether a score using both symptom frequency and severity correlated better with an objective measure of GERD than one using only symptom frequency. METHODOLOGY: A cross-sectional validity study was performed in 100 GERD patients and 150 healthy controls comparable in age and gender. Ethical clearance was granted. GERD was diagnosed by upper gastro-intestinal endoscopy, including patients with all grades of oesophagitis. All subjects faced a GERD-specific interviewer-administered questionnaire with seven upper gastro-intestinal symptoms (heartburn, acid regurgitation, chest/abdominal pain, abdominal distension, dysphagia, cough, belching). Each symptom was graded using a 5-item Likert scale for frequency (never, monthly, 2-4 times per week, weekly, daily) and a 4-item scale for severity (no effect, mild, moderate, severe) and two scores generated. Score 1 being the sum of frequency of symptoms while score 2 was the sum of products of frequency and severity of each symptom. All GERD patients underwent 24h ambulatory pH monitoring. Face and content validity were assessed by expert consultation and literature review, internal consistency by Cronbach alpha statistics, reliability by intra class correlation coefficient estimation and concurrent validity by comparison of scores with 24 hour pH monitoring values as the gold standard. Cut-off values were determined by constructing receiver-operating characteristic curves. RESULTS: For both scores, mean scores of cases were significantly higher than controls (p<0.001) Cut-off score for score 1 was ≥ 10.50 (sensitivity 92.0%; specificity 78.7%; area under the curve 0.937 respectively). Cut-off score for score 2 was ≥ 12.50 (sensitivity 90.0%; specificity 78.0%; area under the curve 0.929 respectively). Intra class correlation coefficient for score 1 and 2 were 0.94 and 0.82 respectively. There was good correlation between both symptom scores and 24-h pH metry parameters (Spearman rank correlation, p=0.01), but score 2 showed a significantly better correlation (correlation of Total reflux time pH<4 with score 1 and score 2 was 0.491 and 0.651; p=0.001, and of Demeester score with score 1 and score 1 was 0.590 and 0.747; p<0.001). CONCLUSION: Our GERD questionnaire is valid, reliable and showed better correlation with an objective test when both severity and frequency of symptoms were scored rather than frequency of symptoms alone.