A Prospective Study to Improve the Yield of Esophagogastroduodenoscopy at a Tertiary Hospital in India

ABSTRACT

Background: Esophagogastroduodenoscopy (EGD)is an efficient tool for diagnosis, screening and therapy. However, it must be used efficiently, to maximize value for costs and reduce complications. Many centers have adopted an open-access referral policy, resulting in increased costs, waiting times and clinical workload. When EGD is used for screening of common symptoms like dyspepsia by medical and paramedical personnel, the yield is less than 50%.We planned a study to differentiate disease from non-diseased findings and improve the diagnostic yield of EGD. Methods: The clinical history and examination of 150 patients of a tertiary care hospital in north India over 2 years wasrecorded. EGD when indicated by any clinician or desired by a patient was performed by a consultant, using topical anaesthesia.The outcome of the endoscopy was categorized as positive, if there was significant finding. Less severe and equivocal findings such as gastritis, duodenitis were not considered positive for the purpose of this study. Logistic regression (forward LR score) was used; the coefficient of regressionwas used to assign a score for each symptom. Results: Pain was the most common symptom; in 110 patients (77.3%) the endoscopy showed no significant findings. Significant findings were seen in 34 patients giving a diagnostic yield of 22.7%. Clinical features like weight loss, hematemesis, melena, dysphagia, anemia, abdominal distention, ascites, and abdominal lump significantly discriminated and pointed towards a positive endoscopic finding. Pain was not a good discriminating factor; dysphagia, presence of ascites and the presence of abdominal lump independently predict significant endoscopic findings. At a total score of 4 or less endoscopy could be avoided in 75 (50%) of the patients. A score of 5 yielded a sensitivity of 82% and specificity of 71%. For picking up a positive finding. All 11 patients with malignancy had a score >5. Conclusions: Our findings suggest selection of patients for EGD can be improved based on a scoring system. It also provides a basis for prospective studies which can lead to better use of resources in future.