Optimizing the Management Algorithm for Heartburn in General Gastroenterology: Cost-Effectiveness and Cost-Minimization Analysis.

Background And Aims: Heartburn is the most common symptom seen in gastroenterology practice. We aimed to optimize cost-effective evaluation and management of heartburn.

Methods: We developed a decision analytic model from insurer and patient perspectives comparing 4 strategies for patients failing empiric proton pump inhibitors (PPIs): (1) PPI optimization without testing, (2) endoscopy with PPI optimization for all patients, (3) endoscopy with PPI discontinuation when erosive findings are absent, and (4) endoscopy/ambulatory reflux monitoring with PPI discontinuation as appropriate for phenotypic management. Health outcomes were respectively defined on systematic reviews of clinical trials. Cost outcomes were defined on Centers for Medicare and Medicaid Services databases and commercial multipliers for direct healthcare costs, and national observational studies evaluating healthcare utilization. The time horizon was 1 year. All testing was performed off PPI.

Results: PPI optimization without testing cost $3784/y to insurers and $3128 to patients due to lower work productivity and suboptimal symptom relief. Endoscopy with PPI optimization lowered insurer costs by $1020/y and added 11 healthy days/y by identifying erosive reflux disease. Endoscopy with PPI discontinuation added 11 additional healthy days/y by identifying patients without erosive reflux disease that did not need PPI. By optimizing phenotype-guided treatment, endoscopy/ambulatory reflux monitoring with a trial of PPI discontinuation was the most effective of all strategies (gaining 22 healthy days/y) and saved $2183 to insurers and $2396 to patients.

Conclusions: Among patients with heartburn, endoscopy with ambulatory reflux monitoring (off PPI) optimizes cost-effective management by matching treatment to phenotype. When erosive findings are absent, trialing PPI discontinuation is more cost-effective than optimizing PPI.