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

Abstract

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.

Graphical Abstract

Heartburn is a common symptom that comprises over one-half of referrals to gastroenterologists each year.¹ When patients fail to respond to an empiric trial of proton pump inhibitor (PPI) therapy, upper endoscopy is useful to reliably identify erosive reflux disease that requires long-term PPI therapy to induce healing and prevent downstream complications.^2,3 Over 50% of patients with reflux symptoms presenting in primary care settings have negative endoscopy,^4,5 while some European investigators show that the rate of endoscopy-negative patients might be as high as 75%.^6,7 When erosive findings are absent on upper endoscopy, ambulatory reflux monitoring is capable of identifying whether heartburn symptoms are related to nonerosive reflux disease (NERD), reflux hypersensitivity, or functional disease (ie, not related to gastroesophageal reflux disease [GERD]).⁸

Recent guidelines leverage robust systematic reviews of placebo-controlled trials to advise management tailored to the physiologic cause, also known as phenotype-guided therapy.^9–11 This approach aims to perform testing once, optimize treatment immediately, and minimize the need for repeated downstream testing. Nevertheless, most general gastroenterologists question the added value of this test-centric paradigm in light of its perceived cost compared with the sheer prevalence of heartburn.¹² As a result, patients seeking advice from gastroenterologists for heartburn rarely undergo ambulatory reflux monitoring.¹³ Further, insurers increasingly place prior authorization restrictions on upper endoscopy even when it is warranted by guidelines, perhaps because the expense of endoscopy is questionable given that heartburn is not considered an alarm symptom of cancer.¹⁴ Recognizing these valid concerns, evaluation and management strategies for heartburn vary significantly among gastroenterologists and remain poorly standardized in practice. As a result of this variation in care, many patients receive rotating courses of PPIs and antacids, report treatment dissatisfaction, and undergo repeated endoscopic and radiographic examinations in the quest to find a cure.^13,15

We performed an economic analysis from patient and insurer perspectives aiming to identify the optimal cost-effective diagnostic strategy for patients with typical symptoms of heartburn failing PPI therapy and without alarm features in the general gastroenterology community. In addition to optimizing diagnostics, we also compared discontinuing PPI therapy compared with optimizing PPI therapy as a next best step in management when erosive findings are absent.

Materials and Methods

We developed a decision analytic model that follows health outcomes and costs associated with treatment recommendations based on varying levels of routine diagnostic reflux testing (Figure 1). Our base-case patient was a commercially insured, middle-aged individual presenting to general gastroenterology for typical heartburn complaints without alarm symptoms having failed an empiric PPI trial. This study adhered to the CHEERS (Consolidated Health Economic Evaluation Reporting Standards) checklist and methodologic guidance from the Second Panel on Cost-Effectiveness in Health and Medicine.¹⁶

Figure 1.

Full model diagram including all potential diagnostic avenues. EGD, esophagogastroduodenoscopy.

We modeled 4 evaluation and management strategies: (1) PPI optimization without testing, (2) endoscopy with PPI optimization to positively identify erosive reflux disease and then optimize the PPI for all patients before considering discontinuation, (3) endoscopy with PPI discontinuation to positively identify erosive reflux disease and trial discontinuation of PPI therapy when erosive findings are absent, and (4) combined endoscopy/ambulatory reflux with PPI discontinuation to phenotype management based on presence of erosive reflux disease, NERD, or functional heartburn.

All index testing was assumed to be performed off PPI therapy across strategies. We defined PPI optimization on changing the PPI, increasing from standard-dose to high-dose PPI, or increasing the PPI dosing frequency from once daily to twice daily. This was accomplished by modeling outcomes from formal PPI trials in which adherence was rigorously monitored, with variation on the preferred method of PPI optimization accommodated in sensitivity analysis. Beyond strategies that included PPI optimization for patients without erosive reflux disease, we also assumed that patients with erosive reflux disease positively identified on endoscopy in any strategy should receive an optimized PPI. For patients that successfully discontinued PPIs when offered, we assumed that continuous PPIs would not be resumed. We also assumed that patients with predominantly functional heartburn identified on ambulatory reflux testing would receive neuromodulation therapy as an example treatment, if such patients remained symptomatic upon discontinuing PPIs. These assumptions represent core constructs that align with either guidelines or practical experience and do not limit the ability of individual physicians to tailor their practice according to individual patient needs.

Outcomes and Costs

Model inputs are described in Table 1. Efficacy, tolerability, and safety were modeled on synthesized estimates provided in systematic reviews and network meta-analyses when available. Single-center studies were used that represented the only available published data in some cases, given that our model needed to consider outcomes within specific subgroups with or without erosive esophagitis, with or without ambulatory reflux monitoring, and across different ambulatory reflux monitoring results. Efficacy was defined on the ability of therapy to achieve complete heartburn response over at least 4-weeks’ follow-up, as a common and representative endpoint across GERD and related clinical trials. There were no major safety or tolerability concerns with PPIs in underlying trials warranting their inclusion. Costs were inflated to 2022 U.S. dollars and identified from national datasets including the Centers for Medicare and Medical Services National Average Drug Acquisition Cost database of wholesale commercial drug prices,¹⁷ national observational data on nonprescription medical/procedural healthcare costs for patients with persistent heartburn symptoms compared with an age- and sex-matched nonheartburn control individuals, and observational studies reporting work productivity losses with persistent heartburn.

Table 1.

Model Inputs

Description	Base-Case Value	Minimum	Maximum	Distribution	Reference(s)
Outcomes
Likelihood of complete heartburn relief with proton pump inhibitors to erosive reflux	72.0%	60.0%	80.0%	Beta distribution on primary relative risk estimates	Weijenborg et al (2012)19
Likelihood of complete heartburn relief with proton pump inhibitors to treat nonerosive reflux	46.0%	30.0%	60.0%	Beta distribution on primary relative risk estimates	Barberio et al (2023)20
Likelihood of complete heartburn relief with proton pump inhibitors to treat nonerosive reflux with elevated acid exposure time	54.9%	30.0%	60.0%	Beta: n = 312	Food and Drug Administration (2020)21
Likelihood of erosive esophagitis on endoscopy for heartburn failing empiric proton pump inhibitor	24.5%	0.0%	30.0%	—	Ronkainen et al (2005)7
Likelihood of an acid exposure time over 4.0% on a 48-h study among patients with heartburn failing an empiric proton pump inhibitor trial that have nonerosive reflux, tested off proton pump inhibitors	62.9%	50.0%	70.0%	Beta: n = 132	Yadlapati et al (2022)8
Likelihood of an acid exposure time over 4.0% on a 96-h study among patients with heartburn failing an empiric proton pump inhibitor trial that have nonerosive reflux, tested off proton pump inhibitors	62.9%	50.0%	70.0%	Beta: n = 132	Yadlapati et al (2022)8
Likelihood of discontinuing proton pump inhibitor with nonerosive reflux and acid exposure time over 4.0% on a 96-h study	24.1%	10.0%	30.0%	Beta: n = 83	Yadlapati et al (2022)8
Likelihood of discontinuing proton pump inhibitor with nonerosive reflux and acid exposure time at or below 4.0% on a 96-h study	40.8%	30.0%	50.0%	Beta: n = 49	Yadlapati et al (2022)8
Likelihood of discontinuing proton pump inhibitor with nonerosive reflux and acid exposure time over 4.0% on a 48-h study	21.7%	10.0%	30.0%	Beta: n = 83	Yadlapati et al (2022)8
Likelihood of discontinuing proton pump inhibitor with nonerosive reflux and acid exposure time at or below 4.0% on a 48-h study	44.9%	30.0%	50.0%	Beta: n = 49	Yadlapati et al (2022)8
Quality-adjusted life years
Health utility associated with esophagitis off of proton pump inhibitors	0.87			Mean: 0.87 SD: 0.18	Parker et al (2011)38
Health utility associated with esophagitis on proton pump inhibitors	0.93			Mean: 0.87 SD: 0.18	Parker et al (2011)38
Health utility associated with nonerosive reflux symptoms off of proton pump inhibitors	0.90			Mean: 0.90 SD: 0.12	Gerson et al (2005)18
Health utility associated with nonerosive reflux symptoms on proton pump inhibitors	0.94			Mean: 0.94 SD: 0.09	Gerson et al (2005)18
Costs
Cost of representative neuromodulatora	$0.07	$0	$0.07	—	NADAC Database17
Cost of representative proton pump inhibitora	$0.41	$0	$0.41	—	NADAC Database17
Cost of endoscopy (CPT 43239 and APC 5301 with moderate sedation CPT 99152)	$962.43	$0	$5000	—	Physician Fee Schedule39
Cost of ambulatory reflux monitoring (CPT 91035)	$496.85	$0	$5000	—	Physician Fee Schedule39
Annual healthcare cost to manage refractory heartburn	$3783.58	$0	$5000	—	Brook et al (2007)40
Work productivity impairment due to heartburn	6%	0%	10%	—	Dean et al (2003)37
Mean annual wage	$46,800	$0	$100,000	—	U.S. Bureau of Labor Statistics41
Number of gastroenterology office visits per year for persistent heartburn	1.7	0.0	12.0	—	Wahlqvist et al (2008)15
Half-day cost of childcare to attend clinic (accounting for 25% of households having children)	$14.50	$0	$14.50	—	U.S. Census Bureau42 Cost of Care Survey43
Transportation to/from medical visits	$20	$0	$20	—	Muennig (2008)44

Reproduced from the original publication for the underlying model and updated to include 2022 year costs and additional inputs for the comprehensive algorithm.²³ Clinical response was derived from primary relative risk estimates in systematic reviews and network meta-analyses anchored on pooled placebo responder rates. Representative proton pump inhibitors and neuromodulators have similar costs, and our findings therefore do not anchor on specific proton pump inhibitors or neuromodulators; for modeling purposes, lansoprazole 30 mg and nortriptyline 25 mg were used. APC, Ambulatory Payment Classification; CPT, Current Procedural Terminology; NADAC, National Average Drug Acquisition Cost.

^aFor cost purposes, the representative neuromodulator was nortriptyline 25 mg and representative proton pump inhibitor was lansoprazole 30 mg as lower-cost examples within respective drug classes.

Effectiveness

We modeled treatment effectiveness on the increase in quality-adjusted life years (QALYs) found with resolution of heartburn symptoms mapped on valid and reliable health utilities.¹⁸ For PPI optimization, health states were modeled directly on the likelihood of clinical response.^19–21 We recognized that PPI discontinuation does not reflect responder status. For example, patients with intermittent GERD-related heartburn may not need daily PPI. Or, patients with functional heartburn might discontinue PPI and remain symptomatic. Therefore, we first modeled whether patients would discontinue PPI⁸ and then assigned patients to health states based on known physiology and the resulting on- or off-PPI status.¹⁸ As with most benign illnesses, effectiveness gains are small because an improbable complete loss of QALYs would represent death. In contrast, a seemingly small 0.03-QALY gain represents an additional 11 complete healthy days without heartburn per year.

Analysis

The base-case analysis evaluated outcomes and costs accumulated over a 1-year time horizon to inform cost-effectiveness from the patient perspective and cost savings from an insurer perspective. This time horizon aligns with usual timeframes for annual deductible, premium, and copay determinations by health insurers. In cost-minimization analyses to ascertain cost savings, health gains are not considered. In cost-effectiveness analysis, health gains are assigned a dollar value that can outweigh additional upfront costs such as diagnostic testing for heartburn. In contemporary economic studies, an incremental cost-effectiveness ratios below the willingness-to-pay threshold of $100,000 per QALY gained is considered cost-effective. Probabilistic sensitivity analysis was performed on 100,000 simulated patients to evaluate robustness of model estimates. Sensitivity analyses were conducted across ranges for model inputs to identify the most important patient and economic factors that influence cost-effective decision-making. A 3% annual discount rate was used. Analysis was performed with TreeAge Pro 2022 R2 (TreeAge Software Inc, Williamstown, MA).

Results

Costs and effectiveness outcomes are detailed in Table 2 and Figure 2.

Table 2.

Cost-Effectiveness of Algorithmic Treatment Strategies for Chronic Constipation Among Patients Referred to General Gastroenterology That Fail an Initial Empiric Trial of Osmotic Laxative or Soluble Fiber Supplementation

Strategy	Total Cost ($/y)	Total Effectiveness (QALY Gained/y)	Incremental cost ($)	Incremental Effectiveness (QALY)	ICER ($ per QALY Gained)
Insurer perspective
PPI optimization without testing	$3784	0.89	Reference	Reference	Reference
EGD with PPI optimization	$2775	0.92	$1020 saved	+0.03	Dominates PPI optimization without testing
EGD with PPI discontinuation trial for nonerosive heartburn	$3588	0.95	$196 saved	+0.06	Dominates PPI optimization without testing ICER of $27,073 per QALY gained compared with EGD with PPI optimization
EGD and concurrent 96-h ambulatory reflux monitoring off PPI for nonerosive heartburn with phenotype guided management: PPI optimization for NERD and PPI discontinuation for functional heartburn	$1601	0.93	$2183 saved	+0.04	Dominates endoscopy and PPI optimization without testing
EGD and concurrent 48-h ambulatory reflux monitoring off PPI for nonerosive heartburn with phenotype guided management after a PPI discontinuation trial	$1601	0.95	$2183 saved	+0.06	Dominates endoscopy and PPI optimization without testing
EGD and concurrent 96-h ambulatory reflux monitoring off PPI for nonerosive heartburn with phenotype guided management after a PPI discontinuation trial	$1601	0.95	$2183 saved	+0.06	Dominates all strategies
Patient perspective
PPI optimization without testing	$3128	0.89	Reference	Reference	Reference
EGD with PPI optimization	$1507	0.92	$1621 saved	+0.03	Dominates PPI optimization without testing
EGD with PPI discontinuation trial for nonerosive heartburn	$1887	0.95	$1241 saved	+0.06	Dominates PPI optimization without testing ICER of $12,651 per QALY gained compared with EGD with PPI optimization
EGD and concurrent 96-h ambulatory reflux monitoring off PPI for nonerosive heartburn with phenotype guided management: PPI optimization for NERD and PPI discontinuation for functional heartburn	$708	0.93	$2420 saved	+0.04	Dominates endoscopy and PPI optimization without testing
EGD and concurrent 48-h ambulatory reflux monitoring off PPI for nonerosive heartburn with phenotype guided management	$762	0.95	$2366 saved	+0.06	Dominates endoscopy and PPI optimization without testing
EGD and concurrent 96-h ambulatory reflux monitoring off PPI for nonerosive heartburn with phenotype guided management	$732	0.95	$2396 saved	+0.06	More cost-effective than optimizing PPI for NERD without trying PPI discontinuation first, and dominates all other strategies

Relative cost-effectiveness of prescription drug or pelvic floor physical therapy strategies depends on the primary goals of treatment and likelihood of treatment success. EGD, esophagogastroduodenoscopy; ICER, incremental cost-effectiveness ratio; NERD, nonerosive reflux disease; PPI, proton pump inhibitor; QALY, quality-adjusted life year.

Figure 2.

Cost-effectiveness of diagnostic testing strategies for typical GERD symptoms in general gastroenterology. Endoscopy with ambulatory reflux monitoring is the most cost-effective routine strategy to manage patients with typical heartburn failing empiric PPI without alarm symptoms. Analyses are presented from (A) insurer perspectives and (B) patient perspectives.

Costs and Outcomes of PPI Optimization Without Testing

PPI optimization without testing costs insurers $3784/y and costs patients $3128/y due to healthcare expenses and lower work productivity associated with limited symptom relief. This results in a QALY loss of 0.11/y, equivalent to a loss of 40 healthy days over the course of 1 year.

Endoscopy Performed Off PPI for Patients Failing a PPI Trial

We modeled 2 strategies that offered endoscopy to patients failing empiric PPI therapy. In both strategies, endoscopy was performed off PPI to identify and treat patients with erosive reflux disease. As a result, strategies that involved endoscopy were less costly and more effective than PPI optimization without testing (Table 2).

Choosing PPI Optimization or PPI Discontinuation for Patients Without Erosive Findings on the Index Endoscopy

In the absence of erosive findings, many patients still have nonerosive reflux as a cause of heartburn failing empiric PPI. By optimizing the PPI as a routine best next step for patients without erosive findings on endoscopy, insurer costs would decrease $1020/y and patient costs would decrease $1621/y compared with PPI optimization without testing. And, health gains would increase by 0.03 QALYs (ie, 11 healthy days added over 1 year).

Many patients have functional heartburn that is unrelated to GERD. If a trial of PPI discontinuation were offered in the absence of erosive findings, insurer costs would decrease $196/y and patient costs would decrease $1241/y. These cost decreases associated with offering PPI discontinuation would be smaller compared with routinely offering PPI optimization, because the subset of patients with nonerosive reflux as a cause of heartburn would experience care delays. In contrast, health gains would be higher by offering PPI discontinuation rather than PPI optimization for patients without erosive findings on endoscopy (0.06 QALYs [ie, 22 healthy days added over 1 year]), because the PPI discontinuation trial identifies patients that do not need continuous heartburn or antireflux treatment.

For patients without erosive findings on endoscopy, PPI discontinuation was more cost-effective than PPI optimization to both insurers (incremental cost-effectiveness ratio = $27,073 per QALY gained) and patients ($12,651 per QALY gained). This was because PPI discontinuation was able to identify and change management for patients that had functional heartburn (ie, heartburn for reasons other than erosive GERD). In contrast, PPI optimization did not delay effective GERD treatment for the subset of patients with nonerosive reflux. As a result, PPI optimization resulted in smaller health gains and was less cost-effective but was also more cost saving overall compared with PPI discontinuation. Absent ambulatory reflux monitoring, neither strategy fully stratified nonerosive reflux from functional heartburn but instead relied on an empirical approach for the combined population without erosive findings on endoscopy.

Adding Ambulatory Reflux Monitoring to Endoscopy Performed Off PPI for Patients Failing a PPI Trial

For patients failing empiric PPI therapy, guidelines advise combining ambulatory reflux monitoring with concurrent endoscopy for patients without endoscopic evidence of erosive reflux disease. Ambulatory reflux monitoring stratifies patients with GERD-related heartburn from functional heartburn. Following this practice reduces costs to insurers by $2183, reduces costs to patients by $2396, and adds 0.06 QALYs or 22 healthy days over 1 year, compared with PPI optimization without testing. By comprehensively identifying and stratifying positive care plans based on the presence of erosive reflux disease, NERD, or functional heartburn as well as response to a trial of PPI discontinuation, combined endoscopy with concurrent ambulatory reflux monitoring was the least expensive and most effective strategy for insurers and patients.

Trialing PPI discontinuation was also more cost-effective than trialing PPI optimization for patients with true NERD (defined by elevated acid exposure time without erosive findings on endoscopy). This is because PPI discontinuation identified patients with NERD that did not need continuous PPI, whereas PPI optimization did not.

Sensitivity Analyses to Identify Determinants of the Optimal Diagnostic Testing Approach

From a patient perspective, there was no scenario in which patients would prefer endoscopy alone compared with combining endoscopy with ambulatory reflux monitoring (Supplemental Figures 1–6). Further, we did not identify a scenario in which patients or insurers would prefer PPI optimization without testing on the basis of cost-effectiveness (Figure 3).

Figure 3.

Probabilistic sensitivity analyses demonstrate that endoscopy with ambulatory reflux monitoring is preferred across contemporary willingness-to-pay ranges from (A) insurer and (B) patient perspectives.

Threshold Value of Endoscopy and Ambulatory Reflux Testing to Insurers

From the insurer perspective, allowing coverage for endoscopy among patients with heartburn failing an empiric PPI trial appears cost saving, as long as endoscopy cost does not exceed $1971 and endoscopy is performed off PPI (Supplemental Figures 7–9). Covering ambulatory reflux monitoring at the time of the index endoscopy would increase cost savings to insurers below combined reimbursement of $3642 (Figure 4). These valuations require local adjustment for geographic variability in reimbursement and healthcare utilization. This analysis anchored on endoscopy with PPI optimization, rather than endoscopy with PPI discontinuation, because PPI optimization was the more cost-saving strategy in primary analysis.

Figure 4.

Sensitivity analysis demonstrates that insurer preferences depend on the cost of endoscopy and ambulatory reflux monitoring. Endoscopy with a trial of PPI optimization is used for this analysis due to cost savings compared with trialing PPI discontinuation among patients without erosive findings.

Discussion

Typical GERD symptoms lead many patients to seek gastroenterology care.¹ For patients with persistent symptoms despite an empiric PPI trial, ambulatory reflux monitoring appears cost-effective to incorporate alongside endoscopy to match diagnosis and treatment to the appropriate physiology. For patients without erosive findings that do not undergo ambulatory reflux monitoring, a trial of PPI discontinuation appears more cost-effective than a trial of PPI optimization as a next step in treatment, yet PPI optimization appears more cost saving. Together, these findings suggest that an algorithmic approach to comprehensively stratify erosive and NERD from functional heartburn combined with a trial of PPI discontinuation for patients without erosive findings provide value to patients and insurers.

As with other common ailments in gastroenterology such as irritable bowel syndrome, constipation, and dyspepsia,^22–25 insurance coverage is a major barrier to care delivery for patients with heartburn. As a result, prior authorizations are a top policy focus of both the American Gastroenterological Association and American College of Gastroenterology.^26,27 Beyond coverage for endoscopy, ambulatory reflux monitoring may not be covered until the findings on the index endoscopy are known.²⁸ Paradoxically, a second endoscopy to then place an ambulatory reflux monitoring device would all but ensure that overall costs exceed the value-based threshold price for diagnostic GERD testing established in our study.

Beyond insurance coverage, gastroenterologists strive to offer testing that is clinically necessary and cost-effective. Endoscopy is most useful to identify erosive reflux disease.¹¹ Gastroenterologists are then left to choose between a cost-saving trial of PPI optimization to benefit the subset of patients with NERD and a cost-effective trial of PPI discontinuation to identify patients that do not need continuous PPI therapy. The question is then what to do when the patient fails the chosen management approach. Ambulatory reflux testing maximizes cost-effectiveness and cost savings by establishing the physiology to answer this question.¹⁰ Patients that hold the PPI for endoscopy can continue to hold the PPI for 2–3 weeks as a treatment trial, with subsequent care guided by combined results of diagnostics and PPI discontinuation.^8,29,30 For example, patients with functional heartburn on ambulatory reflux monitoring can be safely managed with neuromodulation even if they fail a trial of PPI discontinuation. Patients with NERD and high symptom-associated probability failing PPI discontinuation can be managed with optimized PPI. Finally, patients that successfully discontinue PPI therapy but have NERD on ambulatory reflux monitoring can consider on-demand or as-needed therapy, while other patients that have functional heartburn and succeed in discontinuing PPI therapy need no further reflux evaluation.

Our study suggests that nuance of ambulatory reflux monitoring matters less than whether ambulatory reflux monitoring is performed at all. For example, the cost-effectiveness of 96-hour monitoring was technically superior to 48-hour monitoring,³¹ and yet the cost savings differed by <$44 with no significant difference in health outcomes among these options. This compares with increased expenditure of $1602 to insurers, $2410 to patients, and 22 healthy days lost by not performing ambulatory reflux monitoring at all. In practice, these findings are important to mitigate the need for repeat endoscopy for the sole purpose of adding ambulatory reflux monitoring.³²

Our findings assume that gastroenterologists should be capable of leveraging diagnostic tools to render a diagnosis and care plan based on the underlying physiology at the outset.^9–11 From a physiologic standpoint, gastroenterologists should be capable of managing erosive reflux disease, NERD, and functional heartburn that frequently overlap in practice. Equally, our findings also emphasize the importance of reassurance, communication skills, the patient-physician relationship, and the consultant-primary care relationship.³³ These skills are important to mitigate unnecessary testing for refractory symptoms when the diagnosis has already been made. When a patient fails to respond to appropriate therapy, our primary analysis assumed the status quo: that primary care providers and/or gastroenterologists might question the diagnosis despite objective findings and then repeat testing unnecessarily, rather than reinforcing a positive diagnosis and considering overlap functional heartburn.¹⁵ For example, many patients with erosive reflux disease experience esophagitis healing with PPI therapy but have persistent heartburn, and in such cases these patients often undergo serial endoscopy. For gastroenterologists that are comfortable with recognizing and communicating the overlap between functional heartburn and true GERD-related symptomatology with patients and referring providers, the estimated cost-effectiveness of index testing combined with confident management skills may be even greater than reported here.

Limitations of our study are inherent to its intended applicability. Algorithms provide an anchor for shared decision making that must then be tailored to individual patient needs. We did not evaluate the cost-effectiveness of performing reflux monitoring on-PPI therapy prior to off-PPI therapy for patients with a high probability of GERD.³⁴ Instead, we focused on the broader reality that ambulatory reflux monitoring may not be performed at all. We also acknowledge limited capacity for routine ambulatory reflux monitoring. This limitation is balanced against the reality that our study explores where gastroenterology practice could be, rather than where it stands today. Some might consider routine endoscopy for heartburn as a barrier to endoscopy access.³⁵ To the contrary, our findings suggest that appropriate index testing with confident management should limit the need for an inappropriate second or third endoscopy. While there was no sufficient dataset to model 24-hour catheter-based monitoring, outcomes may be favorable especially among patients with daily heartburn.^29,36 Single-center and older studies were used for model inputs when those data accurately reflected the model input and no other data were available. For example, the most appropriate study on the effects of heartburn on work productivity was published almost 20 years ago.³⁷ As is typical, we therefore report sensitivity analysis around these model inputs demonstrating robustness of predictions. Finally, we focused our study on patients with heartburn symptoms and did not aim to study patients that complain of regurgitation in the absence of heartburn.

In summary, the most cost-effective diagnostic approach to managing patients with typical GERD symptoms of heartburn that lack alarm symptoms and fail an empiric PPI is to offer endoscopy with ambulatory reflux monitoring followed by a trial of PPI discontinuation for patients without erosive findings. These findings strongly support recent clinical practice guidelines from the American Gastroenterological Association and American College of Gastroenterology, posing an opportunity for insurers to defer to gastroenterologists to guide care for one of the most common reasons for referral in gastroenterology practice.

What You Need to Know.

Background

• Over one-half of patients report dissatisfaction and/or nonresponse to treatment for typical gastroesophageal reflux symptoms (heartburn and/or regurgitation) in gastroenterology.

• For patients with typical gastroesophageal reflux symptoms failing an empiric proton pump inhibitor, guidelines advocate ambulatory reflux monitoring alongside index endoscopy; however, this practice is poorly adopted.

• Controversy exists on whether to optimize or discontinue proton pump inhibitor therapy as a best next step when erosive findings are absent.

Findings

• By identifying erosive reflux disease, performing an index endoscopy is less expensive and more effective than avoiding diagnostic evaluation entirely.

• By identifying nonerosive reflux disease and functional heartburn, adding ambulatory reflux monitoring to the initial evaluation optimizes cost-effectiveness.

• For patients without erosive reflux disease, a trial of proton pump inhibitor discontinuation is more cost-effective than optimizing the proton pump inhibitor.

Implications for clinical care

• This economic analysis demonstrates robust support for the adoption and insurance coverage to support recent practice guidelines from the American Gastroenterological Association and American College of Gastroenterology to promote cost-effective care.

Abbreviations used in this paper:

GERD

gastroesophageal reflux disease

NERD

nonerosive reflux disease

PPI

proton pump inhibitor

QALY

quality-adjusted life year