THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2000 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 95, No. 2, 2000 ISSN 0002-9270/00/$20.00 PII S0002-9270(99)00818-7
A Cost-Effectiveness Analysis of Prescribing Strategies in the Management of Gastroesophageal Reflux Disease Lauren B. Gerson, M.D., M.Sc., Anthony S. Robbins, M.D., Alan Garber, M.D., Ph.D., John Hornberger, M.D., M.Sc., and George Triadafilopoulos, M.D. Division of Gastroenterology and Departments of Medicine and Health Research and Policy, Stanford University School of Medicine, Stanford, California; VA Palo Alto Health Care System, Palo Alto, California; and Division of Medical Economics, Roche Pharmaceuticals, Palo Alto, California
OBJECTIVE: Patients who have uncomplicated gastroesophageal-reflux disease (GERD) typically present with heartburn and acid regurgitation. We sought to determine the cost-effectiveness of H2-receptor antagonists (H2RAs) and proton-pump inhibitors (PPIs) as first-line empiric therapy for patients with typical symptoms of GERD. METHODS: Decision analysis comparing costs and benefits of empirical treatment with H2RAs and PPIs for patients presenting with typical GERD was employed. The six treatment arms in the model were: 1) Lifestyle therapy, including antacids; 2) H2RA therapy, with endoscopy performed if no response to H2RAs; 3) Step up (H2RA-PPI) Arm: H2RA followed by PPI therapy in the case of symptomatic failure; 4) Step down arm: PPI therapy followed by H2RA if symptomatic response to PPI, and antacid therapy if response to H2RA therapy; 5) PPI-on-demand therapy: 8 wk of treatment for symptomatic recurrence, with no more than three courses per year; and 6) PPI-continuous therapy. Measurements were lifetime costs, quality-adjusted life years (QALYs) gained, and incremental cost effectiveness. RESULTS: Initial therapy with PPIs followed by on-demand therapy was the most cost-effective approach, with a costeffectiveness ratio of $20,934 per QALY gained for patients with moderate to severe GERD symptoms, and $37,923 for patients with mild GERD symptoms. This therapy was also associated with the greatest gain in discounted QALYs. The PPI-on-demand strategy was more effective and less costly than the H2RA followed by PPI strategy or the other treatment arms. The results were not highly sensitive to cost of therapy, QALY adjustment from GERD symptoms, or the success rate of the lifestyle arm. However, when the success rate of the PPI-on-demand arm was ⱕ59%, the H2RA-PPI arm was the preferred strategy. CONCLUSION: For patients with moderate to severe symptoms of GERD, initial treatment with PPIs followed by on-demand therapy is a cost-effective approach. (Am J Gastroenterol 2000;95:395– 407. © 2000 by Am. Coll. of Gastroenterology)
INTRODUCTION Symptoms of gastroesophageal-reflux disease (GERD), such as heartburn and acid regurgitation, are common. Surveys have found that 44% of all Americans suffer from heartburn at least once per month (1), 30% at least once per week (2), and up to 7% daily. Heartburn is often chronic (3), and may diminish quality of life (4). In addition, lifetime drug therapy is costly. Management of heartburn commences with lifestyle modification and prescription of antacids or H2-receptor antagonists (H2RAs). After an 8-wk trial of H2RA therapy for relief of heartburn, patients may be managed in several ways. If improved, they may remain on H2RAs either indefinitely or intermittently, with therapy resumed upon symptom recurrence. If not improved, patients may be treated with proton-pump inhibitors (PPIs, so-called “step up” therapy) or be referred for an endoscopic examination. If they fail empirical PPI therapy, most patients will undergo endoscopic examination. The term “step down” therapy refers to starting with a PPI for 8 wk, followed by H2RAs if the patient responds to a PPI, and then stepping down to antacids if the patient is asymptomatic on H2RAs; this is an additional approach. Another therapeutic option, used less frequently, is to perform initial endoscopy to define disease severity and to screen for Barrett’s esophagus. Although PPIs cost more than H2RAs at usual dosages, they may be more cost-effective because they are more effective in healing both erosive (5–14) and nonerosive disease (15), thus decreasing the number of office visits and diagnostic tests. Patients found to have erosive disease on endoscopy are usually treated with maintenance PPI therapy. However, recent studies suggest that patients who have nonerosive disease do not differ from those with erosive disease in terms of symptom severity, duration, or response to antisecretory therapy (16, 17). Therefore, regardless of endoscopic findings, patients with nonerosive disease may be treated with PPIs. Traditionally, to receive PPIs, patients with nonerosive disease must fail H2RA therapy and dem-
396
Gerson et al.
onstrate pathological reflux on 24-h ambulatory esophageal pH testing. Even after healing of erosive esophagitis with either PPIs or H2RAs, multiple studies have shown that nearly all patients will relapse by 6 months if they do not receive maintenance therapy (18). Daily PPI maintenance therapy reduces the 6-month and 12-month relapse rates to 15–30% (19 –21). After initial treatment with a PPI, approximately one-half of the patients who are stepped down to an H2RA will relapse (22, 23). Intermittent therapy with PPI given three times per week, or on the weekends, is not as effective and is associated with relapse rates roughly equivalent to those for daily H2RAs (24, 25). Patients who require chronic medical therapy for symptom relief may also consider surgery with a laparoscopic Nissen fundoplication. However, postoperative complications, which can include dysphagia and the gas-bloat syndrome, can occur in 10 –30% of patients postfundoplication (26, 27). Patients with nonerosive disease also may require chronic maintenance therapy. Two randomized double-blind placebo studies in such patients showed the superiority of PPI over placebo (28, 16). Intermittent or on-demand treatment (defined as an 8-wk course of treatment given as needed for symptom recurrence) for patients with GERD is a promising approach that has not been well studied. Natural history studies of GERD demonstrate that often patients take their medication intermittently (3). Recently, the first clinical trial of intermittent therapy for GERD comparing omeprazole and ranitidine was published (29). This double-blind multicenter European study randomized patients with nonerosive or mild-moderate erosive disease to 2– 4 wk of omeprazole or ranitidine therapy for symptomatic recurrence of heartburn symptoms over the course of a year. Although PPI therapy was significantly superior to H2RA in terms of symptomatic relief by the second week, the percent of patients who were able to tolerate intermittent therapy was not affected by initial treatment at randomization. Fifty percent of patients from both PPI and H2RA groups completed intermittent treatment, whereas approximately onequarter of patients from each group required maintenance therapy. Although an 8-wk course of therapy is usually required for healing of erosive esophagitis, the study provides evidence that intermittent therapy is an effective clinical strategy for the majority of GERD patients. In the current era of cost consciousness, there is a growing need to demonstrate that available treatments not only improve quality of life, but also are cost-effective. Several cost-effectiveness studies regarding the chronic management of GERD have been published (30 –38). The major drawbacks to these studies have been they study exclusively patients who have erosive esophagitis, and thus disregard the remaining 50% of patients who have nonerosive symptomatic disease; they require endoscopy to determine disease severity before treatment; they omit quality-of-life assessments; they include surgery only for patients who have failed medical therapy, not accounting for patients who
AJG – Vol. 95, No. 2, 2000
choose surgery to stop taking medications; they do not consider prokinetic therapy; 24-h ambulatory esophageal pH studies and/or manometry are not included as part of the GERD evaluation; they calculate costs over 1 yr, rather than over the patient’s lifetime; and none of the prior models include the options of step-up or step-down therapy. In addition, some of the studies used a modified Delphi technique to obtain probabilities and costs. Finally, the option of treatment on demand (treatment for symptomatic recurrence) has not been considered. To determine the most cost-effective therapy for the empirical management of patients with GERD, we constructed a decision model using a 40-yr-old patient with heartburn as our base case, and modeled total costs and QALYs over the course of that individual’s lifetime.
MATERIALS AND METHODS General Approach We constructed decision trees of five alternative treatments for GERD and performed all calculations using Excel (39) (Microsoft Excel, Microsoft Corporation, Redmond, WA) spreadsheets. The reference treatment was lifestyle modification, against which we compared pharmacological treatments to compute incremental cost and effectiveness. The six treatments thus were (1) Lifestyle Arm: lifestyle modification, which served as the reference treatment; (2) H2RA Arm: the current standard practice for patients who have heartburn, which starts with empirical H2RAs, followed by an endoscopic examination if symptoms are not controlled, and prescribes PPI therapy only if erosive disease is confirmed; (3) H2RA-PPI (Step Up) Arm: Empirical H2RA therapy followed by an 8-wk course of PPI therapy if symptoms are not controlled on H2RA therapy, and endoscopy if there is symptomatic failure on the PPI (Fig. 1); (4) Step-Down Arm: 8-wk course of PPI followed by 8-wk course of H2RA if symptomatic relief obtained on the PPI. Patients who respond to H2RAs are stepped down to antacid therapy. Endoscopy is performed if patients fail to respond to PPI therapy (Fig. 2); (5) PPI-On Demand Arm: empirical treatment with an 8-wk course of PPI therapy administered on demand when GERD symptoms recur. Patients in this group require at most three 8-wk courses (24 wk) of medication per year. Patients failing on demand therapy (recurrence of symptoms earlier than 2 months without medication) receive continuous PPI therapy, and endoscopy is performed only if symptoms recur on daily PPI therapy; (6) PPI-Continuous Arm: empirical treatment with continuous daily maintenance PPI therapy, without performance of endoscopy unless there is symptomatic failure. Patients failing PPI therapy in any treatment arm were required to undergo endoscopy. If erosive disease was demonstrated, or if nonerosive disease was found and followed by a confirmatory 24-h ambulatory esophageal pH test, patients could progress to more complex therapy to remain
AJG – February, 2000
Cost-Effectiveness of GERD Management
397
Figure 1. Decision tree for the H2RA-PPI (step-up) arm. Circles represent chance nodes. Squares represent decision nodes. The percentage of patients responding to each intervention is listed below each arm of the tree. Patients who fail a PPI agent at any point are referred for endoscopy. Patients who fail to respond to once-daily PPI therapy progress to a higher PPI dose. Prokinetic therapy is given if a patient remains symptomatic on twice-daily PPI therapy. Motility studies and/or surgery are performed if patients elect to undergo surgery or fail medical management. Triangles represent end nodes, where patients remain for the duration of the study.
symptom-free (e.g., higher doses of PPI, addition of prokinetic agents, or surgery). The lifestyle and H2RA arms were based on current standards of practice. The step-up and step-down arms are alternative approaches that would theoretically decrease costs because they use the response to medical therapy as an indicator of GERD, and therefore decrease the need for endoscopy and other diagnostic tests. The two PPI treatment models were hypothetical, in that few GERD patients are treated empirically with PPIs, mainly because of their relatively high cost (Table 1). For each treatment, we modeled costs and effectiveness separately for the first year after diagnosis and for all subsequent years. Costs related to diagnosis, surgery, and unsuccessful pharmacological treatments were incurred only during the first year after diagnosis. By the end of that first year, patients would have
received any necessary diagnostic tests and/or treatment interventions. Model Assumptions We assumed that patients who failed a particular regimen did so within 4 wk of finishing that course of therapy. Although patients could fail a regimen as early as 8 wk after the start of the study, we used 1-yr relapse rates from the literature regardless of when a patient failed therapy. Patients diagnosed with nonulcer dyspepsia (normal upper endoscopy followed by a normal 24-h ambulatory esophageal pH test) did not receive any further medication after this diagnosis was made. Although these patients often receive H2RAs or PPIs, there is no established standard for chronic medical therapy for this subset of patients. We assumed that patients remained in a steady state after
398
Gerson et al.
AJG – Vol. 95, No. 2, 2000
Figure 2. Decision tree for the step-down therapy arm. Circles represent chance nodes. Squares represent decision nodes. The percentage of patients responding to each intervention is listed below each arm of the tree. Patients who have a response to 8 wk of PPI therapy are stepped down to H2RA therapy for 8 wk. If they remain asymptomatic, patients are then placed on antacids. In the case of symptomatic relapse, a patient can receive the therapy that was successful in the previous arm of the tree. Patients who fail PPI are referred for an endoscopic examination. The endoscopy subtree refers to all of the decisions and treatments that follow endoscopy and is shown completely in the H2RA-PPI arm tree (Fig. 1).
the first year, so that patients who did not respond to any therapy were assumed to stay in their symptomatic state on high-dose PPIs and prokinetic agents indefinitely. If patients’ symptoms were well controlled on medication, they remained on that therapy indefinitely. Patients who were well after surgery were assumed to remain symptom free without medications indefinitely. Although the duration of symptomatic relief from a laparoscopic Nissen fundoplication is not known, recent studies suggest that patients may remain symptom free up to 20 yr after an open fundoplication (40, 41). Patients with a poor surgical outcome in our model were assumed to have an impaired quality of life indefinitely and were assumed to remain on high-dose PPI therapy and prokinetic agents indefinitely. Estimation of Probabilities Estimates of responses to empirical treatment and to surgery, and of the outcomes of diagnostic tests, were based on published articles, on the esophageal motility database at Stanford University Hospital, and on expert opinion (Table 2). Our estimate of the success rate of on-demand PPI therapy (the percentage of patients who could remain symptom free without medication for 8 wk) was based on published cumulative relapse rates at 2 months for erosive and nonerosive disease. For patients with erosive disease who are taken off of PPI therapy, approximately 15% will relapse at 30 days, and 40% by 60 days (18). In nonerosive disease patients, the relapse rate is 20% at 2 months for patients on placebo, compared with ⬍5% for patients on omeprazole 20 mg taken on demand (mean number of capsules, 0.43 per day) (50, 51). In addition, a randomized trial (29) comparing intermittent omeprazole with intermittent ranitidine therapy in patients with nonerosive or mild erosive disease demonstrated that approximately 25% of patients taking omepra-
zole or ranitidine intermittently would relapse by 60 days off of therapy. Assuming a case mix of patients with 50% nonerosive and 50% erosive disease, we therefore estimated that 20% of the patients would require PPI therapy earlier than 8 wk after the initial course of medication. Our estimate of the average annual success rate for lifestyle modification (5%) was based mainly on our clinical observations. The only published prospective trial (58) comparing 15 patients treated with surgical therapy with 36 patients who received lifestyle therapy (consisting of antacid use, elevation of the head of the bed, weight loss, and bland diet) showed that 17% of the medically treated patients were asymptomatic or had mild GERD symptoms during the first year of follow-up. However, this response rate fell to 11% by 3 yr. The total number of patients in the study was small, and only patients with nonerosive disease responded to conservative treatment. Therefore, we derived our estimate of 5% from the low probability of long-term symptom relief from interventions designed to change behavior, such as those that encourage cessation of smoking (59), reduction in alcohol use, loss of weight (60), and elevation of the head of the bed (61) (which has been questioned as a strategy, because most reflux occurs during the daytime hours (62). We estimated that, in any given year, 5% of patients treated with lifestyle modification would be symptom free; this 5% is a steady-state rate, rather than a cumulative one, as many patients who have made behavior changes will relapse in any given year, and others who have previously been unsuccessful will make the changes. Estimation of Utilities for Each Health State Scores on general health surveys have demonstrated that patients with esophagitis have a decreased quality of life (4,
AJG – February, 2000
Cost-Effectiveness of GERD Management
Table 1. Costs Used in GERD Model* Diagnostic or Therapeutic Maneuver H2RA (generic ranitidine 150 mg, $1.48 per pill) empiric 8-wk course H2RA (ranitidine 150 mg b.i.d.) 1-yr continuous maintenance H2RA (ranitidine 150 mg b.i.d.) 1-yr demand maintenance† PPI (omeprazole 20 mg q.d., $3.87 per pill) empiric 8-wk course PPI (omeprazole 20 mg q.d.) 1-yr continuous maintenance PPI (omeprazole 20 mg q.d.) 1-yr demand maintenance† PPI (omeprazole 20 mg b.i.d.) 8-wk course PPI (omeprazole 20 mg b.i.d.) 1-yr continuous maintenance PPI (omeprazole 20 mg b.i.d.) plus prokinetic agent (propulsid average of 10 mg b.i.d. and 20 mg q.i.d., $0.75 for 10 mg propulsid and $1.50 for 20 mg propulsid) 8-wk course PPI (20 mg b.i.d.) plus prokinetic agent (propulsid average of 10 mg b.i.d. and 20 mg q.i.d) 1-yr continuous maintenance Upper endoscopy with biopsy (includes facility fee and pathology)‡ Prolonged acid reflux test‡ Esophageal motility test‡ Laparoscopic Nissen fundoplication (includes 2-day hospital stay)‡
Cost (1999 Dollars) $
165.76
$ 1080.40 $
497.28
$
216.70
$ 1412.55 $
650.10
$ 433.44 $ 2825.10 $
643.44
$ 4193.85
$ 1444.86 $ 144.48 $ 144.83 $11,064.87
* Costs obtained from 1999 Drug Topics Red Book (68). † Defined as three courses of medication per year with 8 wk of medication per course. ‡ Data sources for costs are the 1999 Medicare RBRVS: The Physicians’ Guide, the 1999 Medicare fee schedule for area 16, and Billing Department, Stanford University Hospital. For upper endoscopy, the professional fee is obtained using the following Current Procedural Terminology (CPT) Codes: 43239 (endoscopy with biopsy, $238.61), 99141 (conscious sedation, $65.30), and 88305 (pathology, $64.95). The 1999 facility fee at Stanford University Hospital for an upper endoscopy is $1076. The CPT codes for the other tests include 91033 (24-h ambulatory esophageal pH test, $144.48), 91011 (esophageal manometry, $144.83), and 43324 (laparoscopic Nissen, $1064.87). Each day in the hospital after the fundoplication costs $5000. GERD ⫽ gastroesophageal reflux disease; H2RA ⫽ histamine2-receptor antagonist; PPI, proton pump inhibitor.
63). Symptoms of unhealed erosive esophagitis have been shown to reduce general well-being to levels lower than those reported by patients who have angina or mild heart failure (64). Using the standard-gamble utility assessment method, where perfect health has a utility of 1.0 and death 0, prior studies in patients with angina pectoris demonstrated utility scores of 0.97 for patients with Class I disease, compared with 0.88 for patients with Class III/IV disease (65). Measurements of utilities in patients with differing severities of GERD found that after 1 wk of symptoms, all patients had lower scores on the SF-36 compared with age-matched population, with mean standard gamble scores of 0.90 compared with a mean score of 0.97 using the time trade-off method (66). Because the disutilities associated with gastroesophageal reflux disease are not well established, we elected to test a range of 0.90 – 0.99, with 0.98 set as the utility for mild GERD symptoms (meaning that pa-
399
tients would be willing to face a 2% risk of death in a gamble with a probability of 98% that perfect health is the outcome) and 0.95 set as the utility for moderate to severe GERD symptoms. These utilities reflect previously published findings in patients with mild and moderate symptoms and endoscopy-negative reflux disease using Psychological General Well-Being (PGWB) index scores (67). Because patients with a poor surgical outcome suffer from dysphagia, gas-bloat syndrome, and recurrent GERD symptoms, we estimated the utility of that state to be 0.95, the same utility as for moderate to severe GERD symptoms. Estimation of Time Spent in Various Health States For the pharmacological treatments, we estimated (based on published articles) that the symptoms of most patients responding successfully to therapy with H2RA or PPI would remit after 2 wk of therapy (6, 7). Based on average waiting times at Stanford University Hospital, we estimated that patients would wait 2 wk for either upper endoscopy, 24-h ambulatory pH monitoring, or a test for esophageal motility, and 4 wk for laparoscopic Nissen fundoplication. On-demand therapy was given as a possible maintenance therapy for patients who initially responded to H2RAs in the H2RA or H2RA-PPI arm, and for patients who failed H2RAs and responded to PPIs in the H2RA or H2RA-PPI arm. However, on-demand therapy was not prescribed in the step-down arm because this arm of therapy has traditionally focused on changing patients to the less expensive drug, rather than decreasing the usage of the effective medication. Patients in the PPI on-demand arm received this therapy as the first line of treatment. In all of these arms, the treatment given was 8 wk of medication followed by 8 wk without medication. If a patient’s symptoms recurred earlier than 8 wk, then the patient was advanced to the next level in the decision tree. Estimation of Costs We used standard doses of ranitidine (150 mg p.o. b.i.d.) and omeprazole (20 mg p.o. q.d.) to calculate the costs of medical therapy. Drug costs were based on the average wholesale price (AWP) listed in the 1999 Drug Topics Red Book (68) (Table 1). The generic cost for ranitidine was used throughout the model. We calculated the cost for cisapride therapy by averaging the costs between doses of 10 mg b.i.d. and 20 mg q.i.d., because many patients are unable to tolerate 20 mg four times a day, or require lower doses to control their symptoms. We estimated costs for diagnostic and therapeutic procedures using current procedural terminology (CPT) codes and the 1999 Medicare Fee Schedule (69) for Area 16. The facility (technical) fee for an upper endoscopy was obtained from the Endoscopy Suite at Stanford University Hospital. The cost for a 2-day inpatient hospital stay for a laparoscopic Nissen fundoplication was estimated based on average charges billed by Stanford University Hospital at approximately $5000 per day. We estimated that the cost of the extra physician visits associated
400
Gerson et al.
AJG – Vol. 95, No. 2, 2000
Table 2. Probabilities Used in GERD Model Health Outcome, Diagnostic/Therapeutic Maneuver, or Decision 8-wk empiric course of H2RA Patients symptom-free on demand* H2RA therapy Patients initially controlled on H2RA or PPI who relapse† on maintenance H2RA therapy Endoscopy 24-h ambulatory pH monitoring 8-wk course of PPI Success rate of on-demand PPI therapy Patients initially controlled on PPI who relapse‡ on maintenance and require high-dose PPI Patients controlled on high-dose PPI who relapse† and require addition of prokinetic agent Patients not controlled on high-dose PPI who are controlled after addition of prokinetic agent Patients controlled on high-dose PPI and prokinetic agent willing to take medication indefinitely Esophageal motility study Patients who choose to undergo surgery who have failed all medications who are asymptomatic on medication Patients maintained on medication who develop “cold feet” and cancel surgery Surgery successful
Outcome/Decision 1 (%) Symptom-free on H2RA (0.50) On-demand maintenance (0.50) Relapse on H2RA (0.50) Shows erosive GERD (0.50 Positive (0.80) Symptom-free on PPI (0.85) Symptom-free at 2 months (0.80) Relapse on PPI (0.20) Relapse high-dose PPI (0.15) Symptom-free (0.10) Willing to take indefinitely (0.90) Supportive (0.70) Choose surgery (0.80) (0.10) Develop “cold feet” (0.30) Good-quality outcome (0.90)
Low Value
High Value
0.34
0.70
References (5–14) (29, 42)
0.55
0.87
(20–22, 24, 43)
0.30
0.65
(44–46, 17)
0.61
1.0
(47–49)
0.64
0.92
(5–14)
0.58
0.98
(18, 29, 50, 51)
0.11
0.30
(19–21)
0.71
0.96
(6, 10, 11, 52) (22, 53) (53)
(54) (54) (54) 0.70
0.97
(19, 26, 27, 55–57, 84)
* On-demand therapy defined as an 8-wk course of medication taken no more than three times per year. † Defined as symptomatic relapse within 12 months. ‡ Presence of heartburn at least three times a week or more despite daily PPI therapy. The low and high values for the probabilities tested in the model were determined from the cited references. Abbreviations as in Table 1.
with each of the H2RA and PPI arms was equivalent to the costs associated with the lifestyle treatment. The lifestyle treatment would actually require a substantial outlay of resources, such as for referrals to pharmacologically based smoking-cessation programs or for medical treatments for weight loss. Because the costs for the extra physician visits in the H2RA and PPI arms cancel out the costs of the lifestyle treatment, neither value is included in the model. Modeling For each treatment, we computed the expected number of QALYs accumulated during the first year after diagnosis and in each subsequent year. We obtained this quantity by computing (for each distinct pathway of diagnostic and therapeutic procedures) the product of the utility for each health state (qi), the fraction of a year spent in that state (ti), the probability of being in that health state (pi), and then summing these products (qi ⴱ ti ⴱ pi) over all health states for each treatment. After computing the expected QALYs for each treatment arm for year 1 and for each subsequent year,
we computed the incremental cost-effectiveness (C/E) ratio using the following formula. We assumed that all patients were 40 yr of age at diagnosis and were under medical care for their GERD until they died:
冘
c 2i c 1i i⫺ 共1 ⫹ r兲 共1 ⫹ r兲 i i⫽1
冘
p 2iq 2i p 1iq 1i i⫺ 共1 ⫹ r兲 共1 ⫹ r兲 i i⫽1
n
Incremental CE ratio ⫽
i⫽1 n
i⫽1
冘 n
冘 n
,
(1) where p2i is the probability that the individual is alive at year i, q2i is the expected utility, c2 is the expected cost during the ith interval for the intervention, and r is the real (inflationadjusted) discount rate of 3% applied to both costs and health outcomes. The quantities with a “1” subscript are the analogous quantities for the lifestyle treatment. The intervals in this model are single years of the patient’s life,
AJG – February, 2000
Cost-Effectiveness of GERD Management
401
Table 3. Incremental Cost-Effectiveness Analysis*
Strategy
Discounted Incremental Cost (From Age 40 Yr Until Death)
QALY ⫽ 0.95, Base case, severe GERD Lifestyle $ 0 PPI-on-demand $26,167 H2RA-PPI $27,846 H2RA $29,965 Step down $37,641 PPI, continuous $41,112 QALY ⫽ 0.98, Base case, mild GERD Lifestyle $ 0 PPI-on-demand $26,167 H2RA-PPI $27,846 H2RA $29,965 Step down $37,641 PPI, continuous $41,112
Discounted QALYs (From Age 40 Yr Until Death)
Discounted Incremental Cost Effectiveness (From Age 40 Yr Until Death)
23.66 24.91 24.37 24.42 24.43 24.65
Reference treatment $20,934 Dominated by PPI-on-demand Dominated by PPI-on-demand Dominated by PPI-on-demand Dominated by PPI-on-demand
24.38 25.07 24.55 24.66 24.66 24.76
Reference treatment $37,923 Dominated by PPI-on-demand Dominated by PPI-on-demand Dominated by PPI-on-demand Dominated by PPI-on-demand
* Incremental cost-effectiveness ratios were obtained by comparing treatments in a stepwise fashion by order of cost and effectiveness. The ratio for PPI-on-demand compares that strategy to the lifestyle arm, the H2RA-PPI strategy is compared to the PPI-on-demand arm, and so on. Because the PPI-on-demand arm costs less and is equal or greater in effectiveness compared to the other arms, it is the dominant strategy. Abbreviations as in Table 1.
beginning when the patient is 40 yr old and extending until death. We estimated the probabilities of death during each interval using the most recent available (1992) life tables for the U.S. population (both genders, all ethnic groups) (70).
RESULTS The results of the cost-effectiveness analysis are shown in Table 3. The PPI-on-demand strategy is the most costeffective approach, with a discounted incremental cost-effectiveness ratio of $20,934 per QALY gained for patients with moderate to severe GERD symptoms (QALY level of ⱕ0.95, Fig. 3) and $37,923 per QALY gained for patients with mild GERD symptoms (QALY level of ⱖ0.98). This strategy dominated the H2RA arm, step-up and step-down arms, and the PPI-continuous arm. The alternative arms produced fewer or equal QALYs at a higher cost. The PPI-on-demand strategy remained dominant regardless of the QALY level. The reasons for the effectiveness of the PPI-on-demand strategy are partially demonstrated in Table 4. Applying a
Figure 3. Cost-effectiveness ratio for PPI-on-demand therapy as a function of QALY level. The PPI-on-demand strategy remains dominant regardless of the QALY decrement from GERD symptoms. However, at a level of ⱕ0.99, the cost-effectiveness ratio falls within the acceptable range of ⱕ$50,000/QALY.
hypothetical cohort of 1000 patients to each treatment arm, the number of patients in treatment-related endpoints was calculated for the three least expensive treatment strategies. The H2RA category refers to patients who are successfully treated with H2RAs for GERD. The PPI category refers to all patients who receive PPI therapy, regardless of whether they ultimately receive a Nissen fundoplication or combination therapy with a prokinetic agent. The largest component of the costs is procedural related. The PPI-on-demand and step-up therapy arms have the lowest costs because significantly fewer endoscopies are performed compared with the H2RA arm. Though the costs for the first year of therapy are slightly less in the H2RA-PPI arm, the subsequent years of therapy cost less in the PPI-on-demand arm and are associated with a greater number of QALYs, thus yielding the more favorable cost-effectiveness ratio for the PPI-on-demand arm. Sensitivity analyses were performed on the cost of medication, the efficacy of H2RA and PPI therapies, the efficacy of PPI-on-demand therapy, the length of on-demand therapy, the annual success rate of the lifestyle arm, the effectiveness of Nissen fundoplication, and the costs of office visits. Costs of Medication We compared results varying the price difference of the two medications from a scenario where the prices of the two medications ranged from $0.33 each (i.e., equal cost) to a tenfold higher cost for PPI ($0.33 for the H2RA vs $3.63 for the PPI). In the first scenario, the PPI-on-demand was dominant with an incremental cost-effectiveness ratio (comparing PPI-on-demand to lifestyle therapy) of $2197/QALY gained. Once the price difference was fourfold or greater (when the H2RA cost $0.33 and the PPI $1.20 per pill) the H2RA-PPI strategy cost the least but was associated with
402
Gerson et al.
AJG – Vol. 95, No. 2, 2000
Table 4. Analysis of 1000 Patients and Treatment-Related Endpoints Treatment Arm
Patients on H2RA*
Patients on PPI†
Endoscopy
Surgery
Year 1 Cost‡
Year 2 Cost‡
H2RA H2RA-PPI PPI-on-demand
200 200 None
702 780 All
780 216 89
5 5 2
$1877 $1244 $1263
$1149 $1083 $1010
* Patients successfully treated with H2RA therapy. † Patients treated with a PPI either alone, with a prokinetic, before surgery, or after a poor surgical outcome. ‡ Discounted marginal costs for a year of therapy. For the PPI-on-demand arm, the Year 1 costs of $1263 included $917 for drug costs, $265 for diagnostic tests, and $81 for surgery. For the H2RA arm Year 1 costs, $888 was for drugs, $778 for tests, and $211 for surgery. The costs for Year 2 and subsequent years until death remain constant. Abbreviations as in Table 1.
fewer QALYs than the PPI-on-demand arm. The cost-effectiveness ratio of H2RA-PPI compared with lifestyle was $11,949. However, because the PPI-on-demand strategy was more effective than the H2RA-PPI despite a slightly higher cost, the PPI-on-demand is preferred over the H2RAPPI arm by extended dominance. The incremental costeffectiveness ratio for the PPI-on-demand compared with the H2RA-PPI in this scenario was $724. So at a price difference of fourfold or greater between the two drugs, the H2RA-PPI strategy is no longer dominated by the PPI-ondemand strategy, but the PPI-on-demand arm is still associated with the most favorable cost-effectiveness ratio (Fig. 4). Success Rate of the PPI-on-Demand Therapy As explained in the Methods section, we calculated that 80% of the patients would be able to tolerate absence of PPI for 8 wk, assuming that there would be an equal mix of patients with erosive and nonerosive disease. We performed sensitivity analysis on this figure, varying the success rate of on-demand PPI therapy from 30% to 70%. The PPI-on-
demand arm was the most cost-effective strategy when the success rate was 60 –70%, with a cost-effectiveness ratio ranging from $47,741 for the former to $29,655 for the latter (comparing PPI-on-demand to lifestyle therapy to obtain the cost-effectiveness ratios). At a demand success rate of 59%, the PPI-on-demand arm was the least expensive, but the H2RA-PPI arm was associated with more QALYs and therefore was the preferred strategy. At a PPI-on-demand success rate of 59%, the cost-effectiveness ratio for PPI-on-demand was $50,478 (comparing PPI-on-demand to lifestyle), but the ratio was $14,600 when the H2RA-PPI arm was compared to PPI-on-demand. Therefore, at a success rate of ⱕ59%, the H2RA-PPI arm would be the preferred regimen. This is the only section of the sensitivity analysis where the PPI-on-demand strategy is no longer favored (Fig. 5). Success Rate of Nissen Fundoplication We varied the success rate of a Nissen fundoplication from 70% to 90%. There were no changes in the results of our findings even if 30% of the patients treated with a Nissen fundoplication had postoperative complications. Length of On-Demand Therapy The length of the on-demand therapy was changed from the base case of 24 wk to 32 wk and 46 wk. At 46 wk, the
Figure 4. The effect of PPI/H2RA cost ratio on the cost-effectiveness ratio. The PPI-on-demand strategy is preferred when the cost ratio is between 1 and 3 (i.e., the PPI costs three times as much as the H2RA). When the ratio is ⱖ4, the PPI-on-demand is still preferred by extended dominance over the H2RA-PPI arm. } PPI-demand; 䊐 H2RA-PPI.
Figure 5. Cost-effectiveness ratio variation according to PPI-ondemand success rate. At a success rate of ⱕ59%, the PPI-ondemand arm is no longer the preferred strategy. Though the PPIon-demand costs less than the H2RA-PPI at a success rate of 59%, the latter is more effective and therefore has a more favorable cost-effectiveness ratio. } PPI-demand; 䊐 H2RA-PPI.
AJG – February, 2000
H2RA-PPI arm was less expensive but also less effective than the PPI-on-demand arm. The PPI-on-demand was the preferred strategy, with a cost-effectiveness ratio of $2930 per QALY (comparing H2RA-PPI to PPI-on-demand therapy). Varying Efficacy of the H2RA Arm We set the efficacy rate of the H2RA at 70% (the high value, Table 2) and then varied the efficacy of on-demand PPI therapy from 50% to 80%. In the first case, where the PPI-on-demand was 50% effective and the H2RA was 70% effective, the PPI-on-demand therapy was less expensive than the H2RA-PPI arm, but the H2RA-PPI arm was more effective, and therefore the preferred strategy. The incremental cost-effectiveness ratio was $105,744 for the PPIon-demand arm (comparing PPI-on-demand to lifestyle therapy) and $1026 for the H2RA-PPI arm (comparing PPIon-demand to H2RA-PPI therapy). At a success rate of ⱖ60%, the PPI-on-demand therapy was the most cost-effective strategy. Costs of Office Visits The total costs of office visits for all treatment arms were uncertain. We assumed that the number of office visits would be similar among all treatment arms, and we performed sensitivity analyses to assess the effects of the patients in the drug arms seeing their physicians more often than patients in the lifestyle arm. Even if office visits cost $1000 more per year, the PPI-demand strategy is still the most cost effective. We believe that the cost lies close to the lower end of the range, because typically patients who are taking PPIs see their doctors less often than do patients treated with lifestyle modification (who have more visits for therapy modification and symptom control). Lifestyle Arm Success Rate Another uncertain variable in the model was the annual success rate of the lifestyle arm. Although few physicians treat GERD without drugs, there is a subset of GERD patients who elect a trial of conservative measures before they take medications; for this study, we assumed a 5% annual success rate for lifestyle modification. Using success rates of 10% and 15% did not change the results significantly. At a success rate of 50%, the PPI-on-demand arm remained the preferred strategy.
DISCUSSION Our analysis of empirical treatment strategies for patients with heartburn reveals that an 8-wk course of PPIs for initial symptom relief, followed by repeat courses of therapy initiated when symptoms recur, is more cost-effective than step-up or step-down strategies, or continuous therapy with either H2RAs or PPIs. The cost-effectiveness ratio for ondemand PPI treatment of $20,934 per QALY falls well within the range generally reported as acceptable (71), and compares favorably to other accepted medical interventions,
Cost-Effectiveness of GERD Management
403
such as HMG-CoA reductase-inhibitor therapy for male heart-attack survivors who are aged 55– 64 yr and have cholesterol levels ⬎250 mg/dl ($2158 per QALY) (72), screening for hypertension in men ⬎40 yr old ($27,519 per QALY) and women ($42,222 per QALY) (73), and breastcancer screening with mammography in women aged 55– 65 yr ($41,008 per QALY) (74). For symptoms of mild GERD the cost-effectiveness ratio was $37,923. Once the decrement in QALYs from GERD symptoms reached ⱕ0.99, the C/E ratio fell into the range of ⱕ$50,000 or less (Fig. 3). We believe that our model adds several new contributions to the previously published cost-effectiveness studies for the treatment of GERD. First, whereas the prior studies evaluated short time horizons (ⱕ1 yr), our analysis used a lifetime perspective. Because GERD is a chronic disease, we believe that an analysis of lifetime costs is appropriate and more accurate. One limitation to our model using this perspective, however, is that we did not include repeat Nissen fundoplication(s) for patients undergoing surgical therapy, because the probability of undergoing a Nissen in our study was so low (approximately 10⫺7). It is not known at this time how long patients can expect symptomatic relief from a Nissen fundoplication. In one follow-up study, 17.5% of the patients required medication for recurrent GERD symptoms 5 yr after the Nissen (75). A recent retrospective study suggested that 2.5% of patients undergoing Nissen fundoplication will require repeat surgical repair or revised fundoplication (76). Second, only a few of the prior studies used QALYs as a utility measurement, and obtained their estimates through a modified Delphi process. Because GERD is a chronic disease that adversely affects quality of life, we feel that the use of utilities is justified in this setting. By using a wide range of utility measurements, we are able to assess the costeffectiveness of treatment strategies by degree of disutility from chronic GERD symptoms. Third, prior models either did not allow patients to switch from H2RAs to PPIs without endoscopy, often did not include the option of Nissen fundoplication, or sent patients without response to PPI for a Nissen without consideration of promotility therapy. By including the options of high-dose PPI therapy, followed by promotility agents and then a Nissen fundoplication, we believe that our model defines further therapeutic options available to the physician and patient more completely than prior models. Fourth, none of the models included patients with nonerosive disease, a patient group that deserves attention because the severity of their symptoms can be equal to or greater than that of patients with erosive disease. Finally, this is the first model to evaluate on-demand therapy for GERD, a strategy that is effective for a majority of patients. Surprisingly, although the PPI-on-demand strategy is the most cost effective, it is the least frequently used. In many settings, administration of H2RAs is the first line of therapy, and documentation of erosive disease with diagnostic tests is required before PPI therapy is initiated. One of the main
404
Gerson et al.
reasons that the PPI-on-demand strategy is most cost effective is that this strategy eliminates the need to perform diagnostic tests that drive up costs, but do not necessarily change the course of therapy. For example, for a cohort of 1000 patients (Table 4), approximately 780 patients would require endoscopy in the H2RA arm, compared with 216 in the H2RA-PPI arm, and 89 in the PPI arm. We found that PPI-on-demand therapy raised QALYs by 1.25, compared with lifestyle modification, translating to a gain of 456 quality-adjusted days for patients using the PPI treatment strategy, a gain greatly exceeding the QALYs gained by screening for hypertension in men ⬎40 yr of age (1–20 quality-adjusted days gained) (60), mammography in women 50 –59 yr of age (11.7 days of life saved per person screened) (77), or hormone-replacement therapy in women aged 55– 65 yr (14 quality-adjusted days gained) (78). In addition, the number of quality-adjusted days gained represents an average effect in a representative population (40yr-old patients who present with heartburn); the benefits might be greater for patients who are older or who have more severe disease. In our model, the PPI-on-demand strategy is most cost effective because it eliminates the diagnostic testing usually required before PPI therapy is initiated. This model has as the most important clinical outcome symptom relief, rather than healing demonstrated by endoscopy. The endoscopic healing rates after 8 wk of therapy in patients with erosive esophagitis taking H2RAs range from 32% to 60%; the percentages for relief of symptoms are similar (5–14). Healing and symptom relief reported in patient taking PPIs ranges from 64% to 92%—superior to those seen with H2RAs for both nonerosive and erosive disease (44). The use of PPIs has been shown to relieve symptoms rapidly, often during the first week of treatment. Our model implies that an endoscopic examination is not necessary for a patient who has uncomplicated GERD, because medical therapy can be guided by symptomatic response, and because patients with either nonerosive or erosive disease can be treated cost effectively with the same medical regimen. Although an endoscopic examination might not be needed to direct medical therapy, prior studies recommend that all patients who have chronic GERD symptoms have endoscopy to screen for the presence of Barrett’s esophagus, a premalignant metaplastic change found in 8 –20% patients with GERD (45). Screening for Barrett’s esophagus complicates the decision about endoscopy in GERD patients because screening for the condition and adenocarcinoma is controversial from an economic perspective (79). In a prospective study, the odds ratio for developing Barrett’s was 3.0 in patients with symptoms for 1–5 yr when compared with patients with symptoms for ⬍1 yr, and increased to 6.4 when the duration of symptoms was ⬎10 yr (80). Therefore, most physicians will perform endoscopy to screen for Barrett’s if GERD symptoms have been present for at least 5 yr. We did not include the cost of such screening in our decision model because we chose a 40-yr-old patient as our base
AJG – Vol. 95, No. 2, 2000
case, and did not consider duration of symptoms as an entry criterion. If we had included the cost of such a screening endoscopy for all patients, however, the cost would have been applied to all of the treatment arms, and there would have been no net effect on our analysis. Further studies are needed to determine when screening should be performed and whether it is appropriate to screen all patients with GERD for Barrett’s esophagus. Once the diagnosis of Barrett’s has been made, endoscopic surveillance every 5 yr for dysplasia in this subgroup of patients has been shown to have an incremental cost-effectiveness ratio of $27,400 per QALY gained (81). One endoscopic exam to screen for Barrett’s esophagus is sufficient, as it is highly unusual for patients with uncomplicated GERD on initial endoscopic examination to demonstrate stricture, ulceration, or Barrett’s esophagus on later examinations, even when they report a change in symptoms (82). In addition to the risk of Barrett’s esophagus, GERD has been associated with an increased risk of esophageal adenocarcinoma (83). Whether on-demand therapy would provide as much of a protective effect as continuous therapy for the development of adenocarcinoma is not known at this time. We performed sensitivity analyses on the cost difference between the two therapies, the success rate of the on-demand therapy strategy, the length of on-demand therapy, costs of office visits, and success rate of the lifestyle arm. The PPI-on-demand strategy remained dominant in all scenarios, except when the success rate of the on-demand arm fell to ⱕ59%. In this situation, the PPI-on-demand arm cost less but the H2RA-PPI arm was slightly more effective (by 0.05 QALYs) and therefore had a preferable cost-effectiveness ratio. This was the only case in our model when the H2RA-PPI arm was the preferred strategy. The option of over-the-counter (OTC) H2RAs was not addressed explicitly in our model. After several weeks of OTC therapy, many patients consult their physicians, either because their symptoms persist or because it may be possible to obtain reimbursement for a prescribed drug but not for an OTC drug. There appears to be emerging evidence now that some patients will be able to tolerate intermittent H2RA therapy (29). We considered intermittent H2RA therapy in the H2RA and H2RA-PPI arms after patients had successfully responded to H2RA therapy. Intermittent therapy was not modeled in the step-down arm, where the goal was to transition patients to the less expensive medication. However, at the point that patients consult a physician, they most likely will have failed intermittent H2RA therapy and, at that point, PPI-on-demand treatment is the most cost-effective strategy. In summary, our decision analysis indicates that, for patients who see their physicians for medical therapy for heartburn, on-demand therapy with PPIs is the most costeffective medical regimen that is also associated with the highest quality of life. Physicians and health-care plans should consider using PPIs instead of H2RAs as first-line treatment for GERD patients.
AJG – February, 2000
Reprint requests and correspondence: Lauren B. Gerson, M.D., M.Sc., 111-GI, VA Palo Alto Health Care System, 3801 Miranda Avenue, Palo Alto, CA 94304. Received Mar. 26, 1999; accepted Oct. 8, 1999.
Cost-Effectiveness of GERD Management
19. 20.
REFERENCES 1. Heartburn across America: A Gallup Organization national survey. Princeton NJ: Gallup Organization, 1988. 2. Nebel OT, Forbes MF, Castell DO. Symptomatic GE reflux: Incidence and precipitating factors. Dig Dis Sci 1976;21: 953– 6. 3. Isolauri J, Luostarinen M, Isolauri E, et al. Natural course of gastroesophageal reflux disease. 17–22 year follow-up of 60 patients. Am J Gastroenterol 1997;92:37– 41. 4. Revicki DA, Wood M, Maton PN, et al. The impact of gastroesophageal reflux disease on health-related quality of life. Am J Med 1998;104:252– 8. 5. Robinson M, Sahba B, Avner D, et al. A comparison of lansoprazole and ranitidine in the treatment of erosive oesophagitis. Aliment Pharmacol Ther 1995;9:25–31. 6. Vantrappen G, Rutgeerts L, Schurmans P, et al. Omeprazole (40 mg) is superior to ranitidine in short-term treatment of ulcerative reflux esophagitis. Dig Dis Sci 1988;33(5):523–9. 7. Zeitoun P, Desjars De Keranroue N, Isal JP. Omeprazole versus ranitidine in erosive oesophagitis. Lancet 1987;621–2. 8. Porro GB, Pace F, Peracchia A, et al. Short-term treatment of refractory reflux esophagitis with different doses of omeprazole or ranitidine. J Clin Gastroenterol 1992;15(3):192– 8. 9. Sandmark S, Carlsson R, Fausa O, et al. Omeprazole or ranitidine in the treatment of reflux esophagitis: Results of a double blind randomised multicentre study. Scand J Gastroenterol 1988;23:625–32. 10. Dehn TCB, Shepherd HA, Colin-Jones D, et al. Double blind comparison of omeprazole (40 mg od) versus cimetidine (400 mg od) in the treatment of symptomatic erosive reflux esophagitis, assesed endoscopically, histologically, and by 24 h pH monitoring. Gut 1990;31:509 –13. 11. Havelund T, Lauresen LS, Skoubo-Kristensen E, et al. Omeprazole and ranitidine in treatment of reflux oesophagitis: Double blind comparative trial. Br Med J 1988;296:89 –92. 12. Bardhan KD, Hawkey CJ, Long RG, et al. Lansoprazole versus ranitidine for the treatment of reflux oesophagitis. Aliment Pharmacol Ther 1995;9:145–51. 13. Bate CM, Keeling PWN, O’Morain C, et al. Comparison of omeprazole and cimetidine in reflux oesophagitis: Symptomatic, endoscopic, and histological evaluations. Gut 1990;31: 968 –72. 14. Feldman M, Harford WV, Fisher RS, et al. Treatment of reflux esophagitis resistant to H2-receptor antagonists with lansoprazole. A new H⫹/K⫹-ATPase inhibitor: A controlled, doubleblind study. Am J Gastroenterol 1993;88:1212–7. 15. Havelund T, Laursen LS, Lauritsen K. Efficacy of omeprazole in lower grades of gastro-oesophageal reflux disease. Scand J Gastroenterol 1994;29(suppl 201):69 –73. 16. Laursen LS, Bondesen S, Hansen J, et al. Omeprazole 10 mg or 20 mg daily for the prevention of relapse in gastro-esophageal reflux disease? A double-blind comparative study. Gastroenterol 1992;102:A109. 17. Johansson KE, Ask P, Boeryd B, et al. Oesophagitis, signs of reflux, and gastric acid secretion in patients with symptoms of gastro-oesophageal reflux disease. Scand J Gastroenterol 1986;21:837– 47. 18. Hetzel DJ, Dent J, Reed WD, et al. Healing and relapse of
21. 22. 23. 24.
25.
26. 27. 28.
29.
30. 31. 32. 33.
34. 35. 36.
37.
38.
405
severe peptic esophagitis after treatment with omeprazole. Gastroenterology 1988;95:903–12. Lundell L. Long-term treatment of gastro-oesophageal reflux disease with omeprazole. Scand J Gastroenterol 1994;29(suppl 201):74 – 8. Hallerback B, Unge P, Carling L, et al. Omeprazole or ranitidine in long-term treatment of reflux esophagitis. Gastroenterology 1994;107:1305–11. Gough AL. Lansoprazole is superior to ranitidine 300 mg bid in prevention of relapse of reflux oesophagitis. Gut 1994;35: S58 (abstract). Vigneri S, Termini R, Leandro G, et al. A comparison of five maintenance therapies for reflux esophagitis. N Engl J Med 1995;333:1106 –10. Gough AL, Long RG, Cooper BT, et al. Lansoprazole versus ranitidine in the maintenance treatment of reflux oesophagits. Aliment Pharmacol Ther 1996;10:529 –39. Dent J, Yeomans ND, Mackinnon M, et al. Omeprazole v ranitidine for prevention of relapse in reflux esophagits. A controlled double blind trial of their efficacy and safety. Gut 1994;35:590 – 8. Sontag SJ, Robinson M, Roufail W, et al. Daily omeprazole surpasses intermittent dosing in preventing relapse of eosophagitis: A US multi-center double-blind study. Aliment Pharmacol Ther 1997;11:373– 80. Luostarinen M. Nissen fundoplication for gastro-esophageal reflux disease: Long-term results. Annales Chirurgiae et Gynaecologiae 1995;84:115–20. Rantanen TK, Halme TV, Luostarinen ME, et al. The long term results of open antireflux surgery in a community-based health care center. Am J Gastroenterol 1999;94:1777– 81. Venables TL, Newland RD, Patel AC, et al. Maintenance treatment for gastro-oesophageal reflux disease. A placebocontrolled evaluation of 10 mg omeprazole once daily in general practice. Scand J Gastroenterol 1997;32:627–32. Bardhan KD, Muller-Lissner S, Bigard MA, et al. Symptomatic gastro-oesophageal reflux disease: Double blind controlled study of intermittent treatment with omeprazole or randitine. Br Med J 1999;318:502–7. Bate CM. Cost-effectiveness of omeprazole in the treatment of reflux oesophagitis. Br J Med Econ 1991;1:52– 61. Bate CM, Richardson PDL. A one year model for the costeffectiveness of treating reflux oesophagitis. Br J Med Econ 1992;2:5–11. Hillman AL, Bloom B, Fendrick AM, et al. Cost and quality effects of alternative treatments for persistent gastroesophageal reflux disease. Arch Int Med 1992;152:1467–72. Jonsson B, Stalhammer NO. The cost-effectiveness of omeprazole in intermittent and maintenance treatment of reflux oesophagitis—the case of Sweden. Br J Med Econ 1993;6: 111–26. Lindberg G. Omeprazole vs ranitidine in reflux oesophagitis. Pharmacoeconomics 1994;5(suppl 3):27–34. Bloom B. Cost and quality of treating erosive esophagitis: A re-evaluation. Pharmacoeconomics 1995;8:138 – 46. Harris RA, Kupperman M, Richter JE. Proton pump inhibitors or histamine-2 receptor antagonists for the prevention of recurrence of erosive reflux esophagitis: A cost-effectiveness analysis. Am J Gastroenterol 1997;92:2179 – 87. Zagari M, Villla K, Freston JW. Proton pump inhibitors versus H2-receptor antagonists for the treatment of erosive gastroesophageal reflux disease: A cost-comparative study. Am J Man Care 1995;1:247–55. Sadowski D, Champion M, Goeree R, et al. Health economics of gastroesophageal reflux disease. Can J Gastroenterol 1997; 11(suppl B):108B–12B.
406
Gerson et al.
39. Microsoft Excel. Redmond, Washington: Microsoft Corporation, 1997. 40. Shirazi SS, Schulze K, Soper RT. Long-term follow-up for treatment of complicated chronic reflux esophagitis. Arch Surg 1987;122:548 –52. 41. Luostarinen M, Isolauri J, Laitiner J, et al. Fate of Nissen fundoplication after 20 years. A clinical, endoscopical, and functional analysis. Gut 1993;34:1015–20. 42. Kuster E, Ros E, Toledo-Pimentel V, et al. Predictive factors of the long term outcome in gastro-oesophageal reflux disease: Six year follow up of 107 patients. Gut 1994;35:8 –14. 43. Lundell L, Backman L, Ekstrom P, et al. Prevention of relapse of reflux esophagitis after endoscopic healing: The efficacy and safety of omeprazole compared with ranitidine. Scand J Gastroenterol 1991;26:248 –56. 44. Bate CM, Green JRB, Axon ATR, et al. Omeprazole is more effective than cimetidine for the relief of all grades of gastrooesophageal reflux disease-associated heartburn, irrespective of the presence or absence of endoscopic oesophagitis. Aliment Pharmacol Ther 1997;11:755– 63. 45. Spechler SJ, Goyal RK. Barrett’s esophagus. Pathophysiology, diagnosis, and management. New York: Elsevier, 1985. 46. Knill-Jones RP, Card WI, Crean GP, et al. The symptoms of gastroesophageal reflux and of oesophagitis. Scand J Gastroenterol 1984;19(suppl 106):72– 6. 47. Mattioli S, Pilotti V, Spangaro M, et al. Reliability of 24-hour home esophageal pH monitoring in diagnosis of gastroesophageal reflux. Dig Dis Sci 1989;34:71– 8. 48. Masclee AAM, De Best ACAM, De Graaf R, et al. Ambulatory 24-hour pH-metry in the diagnosis of gastroesophageal reflux disease. Scand J Gastroenterol 1990;25:225–30. 49. Schenk BE, Kuipers EJ, Klinkenberg-Knol EC, et al. Omeprazole as a diagnostic tool in gastroesophageal reflux disease. Am J Gastroenterol 1997;92:1997–2000. 50. Lind T, Havelund T, Carlsson R, et al. The effect of omeprazole (OME) 20 mg and 10 mg daily on heartburn in patients with endoscopy negative reflux disease (ERND) treated on an on-demand basis. Gastroenterology 1996;110:A178. 51. Lundell L. New information relevant to long-term management of endoscopy-negative reflux disease. Aliment Pharmacol Ther 1997;11(suppl 2):93– 8. 52. Klinkenberg-Knol EC, Festen HPM, Jansen JBMJ, et al. Long-term treatment with omeprazole for refractory reflux esophagitis: Efficacy and safety. Ann Int Med 1994;121: 161– 6. 53. Expert Opinion (Includes authors of paper and staff surgeons at Stanford University Hospital). 54. Motility database, Stanford University Esophageal Motility Laboratory. All motility studies performed from 1993–1997 for the indication of heartburn were analyzed. Motility studies supportive of surgery included patients with normal motility, non-specific motility disorder, nutcracker esophagus, and diffuse esophageal spasm. Studies not supportive of surgery were patients with achalasia, scleroderma or other connective tissue disorders, and studies showing low amplitude of esophageal peristalsis in the esophageal body. 55. DeMeester TR, Bonavina L, Albertucci M. Nissen fundoplication for gastroesophageal reflux disease. Ann Surg 1986; 204:9 –20. 56. Shirazi SS, Schulze K, Soper RT. Long-term follow-up for treatment of complicated chronic reflux esophagitis. Arch Surg 1987;122:548 –52. 57. Luostrarinen M. Nissen fundoplication for reflux esophagitis. Long-term clinical and endoscopic results in 109 of 127 consecutive patients. Ann Surg 1993;217:329 –37. 58. Behar J, Sheahan DG, Biancani P, et al. Medical and surgical
AJG – Vol. 95, No. 2, 2000
59.
60. 61. 62. 63.
64. 65. 66. 67.
68. 69. 70. 71. 72.
73. 74. 75. 76. 77. 78. 79. 80. 81.
management of reflux esophagitis. A 38-month report on a prospective clinical trial. N Engl J Med 1975;293:263– 8. Berenson MM, Sontag S, Robinson MG, et al. Effect of smoking in a controlled study of raniditine treatment in gastroesophageal reflux disease. J Clin Gastroenterol 1987;9: 499 –503. Naslund E, Granstrom L, Melcher A, et al. Gastro-oesophageal reflux before and after vertical banded gastroplasty in the treatment of obesity. Europ J Surg 1996;162:303– 6. Pollmann H, Zillessen E, Pohl J, et al. Effect of elevated head position in bed in therapy of gastroesophageal reflux. Zeitschrift fur Gastroenterologie 1996;34(suppl 2):93–9. De Caestecker JS, Blackwell JN, Pryde A, et al. Daytime gastro-oesophageal reflux is important in oesophagitis. Gut 1987;28:519 –26. Mcdougall NI, Johnston BT, Kee F, et al. Natural history of reflux oesophagitis: A 10 year follow up of its effect on patients symptomatology and quality of life. Gut 1996;38: 481– 6. Dimenas E. Methodological aspects of evaluation of quality of life in upper gastrointestinal diseases. Scand J Gastroenterol 1993;28(suppl 99):18 –21. Nease RF, Kneeland T, O’Connor GT, et al. Variation in patient utilities for outcomes of the management of chronic stable angina. JAMA 1995;273:1185–90. Ethiopia A, Gregor JC, Preiksaitis HG, et al. An evaluation of utility measurement in gastroesophageal reflux disease (GERD). Gastroenterology 1998;114:G0477. Dimenas E, Carlsson G, Glise H, et al. Relevance of norm values as part of the documentation of quality of life instruments for use in upper gastrointestinal disease. Scan J Gastroenterol 1996;31(suppl 221):8 –13. Drug Topics Red Book. Montvale, NJ: Medical Economics, 1997. American Medical Association. Medicare RBRVS: The Physician’s Guide. Chicago, IL: AMA, 1997. National Center for Health Statistics. Vital Statistics of the United States 1992, Vol. II, Section 6, Life Tables. Washington DC: U.S. Public Health Service, 1996. Garber AM, Phelps CE. Economic foundations of cost-effectiveness analysis. J Health Econ 1997;16:1–31. Goldman L, Weinstein MC, Goldman PA, et al. Cost-effectiveness of HMG-CoA reductase inhibition for primary and secondary prevention of coronary heart disease. JAMA 1991; 265:1145–51. Littenberg B, Garber AM, Sox HC Jr. Screening for hypertension. Ann Int Med 1990;112:192–202. Eddy DM. Screening for breast cancer. Ann Int Med 1989; 111:389 –99. Klingler PJ, Hinder RA, Floch NR, et al. Five-year follow-up of patients after laparoscopic Nissen fundoplication. Gastroenterology 1998;114:G0734. El-Serag HB, Sonnenberg A. Outcome of erosive reflux esophagitis after Nissen fundoplication. Am J Gastroenterol 1999;94:1771– 6. Kattlove H, Liberati A, Keeler E, et al. Benefits and costs of screening and treatment for early breast cancer. JAMA 1995; 273:142– 8. Tosteson AN, Weinstein MC. Cost-effectiveness of hormone replacement therapy after the menopause. Baillieres Clinical Obstetrics and Gynaecology 1991;Dec, 5(4):943–59. Wright TA, Gray MR, Morris AI, et al. Cost effectiveness of detecting Barrett’s cancer. Gut 1996;39:574 –9. Lieberman DA, Oehlke M, Hlefand M, and the GORGE Consortium. Risk factors for Barrett’s esophagus in community-based practice. Am J Gastroenterol 1997;92:1293–7. Provenzale D, Kemp A, Arora S, et al. A guide for surveillance
AJG – February, 2000
of patients with Barrett’s esophagus. Am J Gastroenterol 1994; 89:670 – 80. 82. Ben Rejeb M, Bouche O, Zeitoun P. Study of 47 consecutive patients with peptic esophageal stricture compared with 3880 cases of reflux esophagitis. Dig Dis Sci 1992; 37:733– 6.
Cost-Effectiveness of GERD Management
407
83. Lagergren J, Bergstrom R, Lindgren A, et al. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999;340:825–31. 84. Rattner DW, Brooks DC. Patient satisfaction following laparoscopic and open antireflux surgery. Arch Surg 1995;130: 289 –94.