A total of 21 articles were included.
Five trials (1,585 patients), including 3 RCTs (617 patients), assessed the effect of endoscopy on patient outcomes.
Six trials (2,487 patients), including 3 RCTs (617 patients), assessed the effect of endoscopy on medical resource use.
Five trials (1,933 patients), including 4 prospective follow-up studies with retrospective analysis of H. pylori screening strategy, were used to determine if a negative endoscopy in H Pylori-positive dyspepsia resulted in reduced medical resource use.
Four trials (2,347 patients) assessed the cost-effectiveness of endoscopy in dyspepsia.
The kappa value for agreement was 0.70 at the abstract stage and 0.88 for application of the inclusion criteria.
Effect of endoscopy on patient outcomes: the definition of dyspepsia was not uniform. None of the included studies used standardised, reliable and valid instruments to assess patients symptoms or health-related quality of life. The 3 RCTs scored either 2 or 3 on the Jadad scale and reported no difference in symptoms in the endoscopy group compared with empiric H2 antagonist therapy, or upper gastrointestinal radiography after 6 months or 1 year. One RCT reported increased satisfaction and fewer sick leave days with prompt endoscopy when compared with ranitidine.
Effect of endoscopy on patient outcomes in H. pylori-positive dyspepsia: 4 studies evaluated alternative management strategies, taking H. pylori status into account. None of the studies used a randomised design to assess the effect of the different strategies on dyspepsia symptoms, anxiety, quality of life, or satisfaction with care.
Effect of endoscopy on medical resource utilisation: the 3 RCTs reported conflicting results. One study reported no advantage of endoscopy over empiric H2 antagonist; one trial reporting fewer office visits and lower total costs for endoscopy compared with radiography; and one study reported fewer physician visits and lower total drug costs for prompt endoscopy compared with ranitidine.
Effect of a negative endoscopy on medical resource utilisation in H. pylori-positive dyspepsia: 5 studies screened patients for H. pylori status, which was followed by endoscopy for those older than 45 years of age who tested H. pylori-positive or had recently used non-steroidal anti-inflammatory drugs. Eleven studies assessed the impact of endoscopy on resource use and reported that empiric therapy reduced the use of endoscopy. Only four of these studies defined dyspepsia. The results as to whether the reduction in endoscopy was offset by increased utilisation of other resources were conflicting.
Effect of endoscopy on clinical decision-making: 2 studies found decision-making to be improved after endoscopy, while two reported only minimal changes as a result of endoscopy.
Effect of endoscopy on cost-effectiveness: 4 of the 5 studies reported empiric H. pylori therapy and the test-and-treat strategy to be of greater cost-effectiveness or cost-utility than initial endoscopy , while one study reported no difference in costs between strategies. Some models were sensitive to the costs of endoscopy and medical therapy. The probabilities of therapy failure varied across the studies.