Relative Cost-effectiveness of Management Strategies for Uninvestigated Dyspepsia

We read the article by Wechsler et al.1Wechsler E.V. et al.Clin Gastroenterol Hepatol. 2023; https://doi.org/10.1016/j.cgh.2023.01.003Abstract Full Text Full Text PDF Google Scholar The authors constructed a health economic model dealing with the initial management of uninvestigated dyspepsia. This issue has occupied clinical trialists and evidence synthesis experts for the last 30 years. Overwhelming evidence points to noninvasive strategies, such as testing for Helicobacter pylori and treating individuals testing positive with eradication therapy (so-called “test and treat”) or empirical acid suppression, being the most cost-effective, compared with prompt endoscopy.2Ford A.C. et al.Gastroenterology. 2005; 128: 1838-1844Abstract Full Text Full Text PDF PubMed Scopus (108) Google Scholar, 3Ford A.C. et al.Aliment Pharmacol Ther. 2008; 28: 534-544Crossref PubMed Scopus (45) Google Scholar, 4Eusebi L.H. et al.BMJ. 2019; 367: l6483Crossref PubMed Scopus (23) Google Scholar In addition, the cause of symptoms in most individuals is functional dyspepsia, with fewer than 1% having upper gastrointestinal cancer.5Nasseri-Moghaddam S. et al.Clin Gastroenterol Hepatol. 2022; https://doi.org/10.1016/j.cgh.2022.05.041Abstract Full Text Full Text PDF Scopus (2) Google Scholar Therefore, national guidelines no longer recommend endoscopy in all patients with dyspepsia,6Black C.J. et al.Gut. 2022; 71: 1697-1723Crossref PubMed Scopus (7) Google Scholar,7Moayyedi P. et al.Am J Gastroenterol. 2017; 112: 988-1013Crossref PubMed Scopus (277) Google Scholar other than in countries where the incidence of gastric cancer is highest,8Talley N.J. et al.J Gastroenterol Hepatol. 1998; 13: 335-353Crossref PubMed Scopus (52) Google Scholar because it is more expensive and no more effective, and the pick-up rate for upper gastrointestinal malignancy is low. We were, therefore, surprised to read that, according to the model, prompt endoscopy maximizes cost-effectiveness and cost-satisfaction in uninvestigated dyspepsia. The authors base much of the model’s assumptions on our recent network meta-analysis.4Eusebi L.H. et al.BMJ. 2019; 367: l6483Crossref PubMed Scopus (23) Google Scholar However, they seem to have conducted their own analyses using crude data from the included trials themselves, rather than using the pooled analyses we published in the paper. They highlight that satisfaction was higher among individuals assigned to prompt endoscopy, although only 6 of the 15 trials in the network meta-analysis assessed this.4Eusebi L.H. et al.BMJ. 2019; 367: l6483Crossref PubMed Scopus (23) Google Scholar In addition, according to their model, annual management costs seemed to be almost equivalent in the “test and treat” and prompt endoscopy arms. With “test and treat” ranked above prompt endoscopy for the relative risk of remaining symptomatic at 12 months in the network meta-analysis,4Eusebi L.H. et al.BMJ. 2019; 367: l6483Crossref PubMed Scopus (23) Google Scholar a cost per endoscopy in their model of $962, and 95% of patients in the network meta-analysis assigned to prompt endoscopy undergoing endoscopy, compared with only 24% of those allocated to “test and treat,” it is unclear how this can be the case. In fact, this conclusion contradicts previous cost-effectiveness analyses of “test and treat” versus both prompt endoscopy and empirical acid suppression, based on individual patient-level data,2Ford A.C. et al.Gastroenterology. 2005; 128: 1838-1844Abstract Full Text Full Text PDF PubMed Scopus (108) Google Scholar,3Ford A.C. et al.Aliment Pharmacol Ther. 2008; 28: 534-544Crossref PubMed Scopus (45) Google Scholar where “test and treat” cost significantly less ($389 less per patient) than prompt endoscopy,2Ford A.C. et al.Gastroenterology. 2005; 128: 1838-1844Abstract Full Text Full Text PDF PubMed Scopus (108) Google Scholar but was of a similar cost per patient to empirical acid suppression.3Ford A.C. et al.Aliment Pharmacol Ther. 2008; 28: 534-544Crossref PubMed Scopus (45) Google Scholar In 1 of these analyses, prompt endoscopy was only cost-effective versus “test and treat” if the willingness to pay per patient symptom-free at 12 months was $180,000. The model also does not have face validity, because it is hard to understand how “test and treat” can be as effective as endoscopy and cost less, yet endoscopy be more cost-effective. Cost-effectiveness models can be useful to estimate the benefits of one management strategy versus another. However, when there are already unequivocal findings from the literature, based on individual patient data meta-analysis and network meta-analysis of all available trials, the added value of such an approach is questionable. This is especially true when most randomized trials comparing endoscopy with “test and treat” have come to the conclusion that the latter is more cost-effective.