Comparison of a composite symptom score assessing both symptom frequency and severity with a score that assesses frequency alone: a preliminary study to develop a practical symptom score to detect gastro-oesophageal reflux disease in a resource-poor setting

Abstract

BACKGROUND: There is a need for a practical instrument to screen for gastro-oesophageal reflux disease (GORD) in epidemiological studies. OBJECTIVES: To develop a practical score to detect GORD and compare assessment of both symptom frequency and severity with frequency alone. METHODS: One hundred patients with upper gastrointestinal symptoms and 150 volunteers with no such past history faced an interviewer-administered questionnaire assessing seven symptoms, graded for frequency and severity. Two scores were generated. Score 1, the sum of frequency of symptoms and score 2, the sum of products of frequency and severity of each. Internal consistency, test-retest reliability and criterion validity against 24-h pH monitoring were assessed. Cut-off scores were generated by receiver operating characteristic curves using scores of half the volunteers and patients selected randomly and validated on the other half. RESULTS: Cut-off scores and area under the curve for score 1 were >/=10.5 and 0.93, and score 2 were >/=12.5 and 0.93, respectively. The sensitivity and specificity of diagnosing the disease in the remaining participants using score 1 was 89.7 and 92.4% and score 2 was 93.8 and 94.0%, respectively. The instrument had good internal consistency (Cronbach alpha = 0.73) and reliability (intraclass correlation coefficient of scores 1 and 2 were 0.94 and 0.95, respectively). Score 2 showed better correlation with 24-h pH monitoring parameters (Spearman's rank correlation, P = 0.01). CONCLUSION: Our score is valid, reliable and can detect GORD with high sensitivity and specificity. A score assessing both frequency and severity of symptoms correlates better to an objective measure of GORD.