Clinical and economic evaluation of laparoscopic surgery compared with medical management for gastro-oesophageal reflux disease: 5-year follow-up of multicentre randomised trial (the REFLUX trial)

Article history paragraph text

The research reported in this issue of the journal was funded by the HTA programme as project number 97/10/99. The contractual start date was in May 2007. The draft report began editorial review in October 2011 and was accepted for publication in July 2012. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors' report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Adrian Grant has received salary support from the NIHR as director of the NIHR Programme Grants for Applied Research programme.

Copyright statement

© Queen's Printer and Controller of HMSO 2013. This work was produced by Grant et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Background

In the Health Technology Assessment (HTA)-commissioned REFLUX trial, laparoscopic fundoplication for people with chronic symptoms of gastro-oesophageal reflux disease (GORD) was shown to significantly improve reflux-specific and general health-related quality of life (HRQoL) at least up to 12 months after surgery. However, cost-effectiveness was uncertain without more reliable information about longer-term costs and benefits. Here, we report the findings from longer-term follow-up of the REFLUX trial.

Objective

To evaluate, at 5 years after surgery, the clinical effectiveness, cost-effectiveness and safety of a policy of relatively early laparoscopic surgery compared with continued medical management among people with GORD symptoms that are reasonably controlled by medication and who are judged suitable for both surgical and medical management.

Methods

Results

At entry to the original trial, participants had been taking GORD medication for a median of 32 months and had a mean age of 46 years, and 66% were men; the randomised groups had been well balanced. Responders at 5 years were older, had been on medication for a shorter time prior to trial entry and had higher baseline quality-of-life scores than non-responders; however, the randomised groups of responders were similar in baseline characteristics. Primary analyses were based on the ‘intention-to-treat’ (ITT) principle, with secondary per-protocol analyses based on those who, at 1 year, had received their allocated treatment.

By 5 years, 63% (n = 112) of the 178 randomised surgery participants and 13% (n = 24) of the 179 randomised medical management participants had actually received fundoplication (equivalent figures in the preference groups were 85% and 3%). There had been a mixture of clinical and personal reasons for those allocated surgery not receiving it, sometimes related to long waiting times. A total or partial wrap procedure had been performed depending on surgeon preference; perioperative complications had been uncommon with no deaths associated with surgery.

At each year, there were significant differences in the REFLUX score (a third of a SD; p < 0.01 at 5 years) favouring the randomised surgical group, reflecting differences in general discomfort (particularly), wind and frequency, nausea and vomiting, and activity limitation subscores. SF-36 and EQ-5D scores also favoured the randomised surgical group, especially SF-36 norm-based general health, but differences attenuated over time and were generally not statistically significant at 5 years [EQ-5D difference (ITT) 0.047, 95% confidence interval (CI) −0.013 to 0.108; p = 0.13]. The lower the REFLUX score and hence the worse the symptoms at trial entry, the larger the benefit observed after surgery. Post hoc exploratory analyses showed that those randomly allocated to medical management who subsequently had surgery had worse symptoms (lower baseline scores) than those who continued on medical management as allocated; following surgery, the scores of these patients markedly improved and this explains, at least in part, why differences in outcome between the randomised groups became less marked over time.

The preference surgical group also had low REFLUX scores at baseline. These scores improved substantially after surgery and at 5 years they were slightly better than those in the preference medical group.

Overall, 4% (n = 16) of the total 364 in the study who had fundoplication had a subsequent reflux-related operation, of whom two had a further (i.e. third) operation. Reoperation was most often conversion to a different type of wrap or a reconstruction of the same wrap. There were only two cases of reversal of the fundoplication and neither was in the randomised comparison. In total, 3% (n = 12) of those who had fundoplication required surgical treatment for a complication directly related to the original surgery, including oesophageal dilatation (n = 4) and repair of incisional hernia (n = 3). Patterns of ‘difficulty swallowing’, flatulence and ‘wanting to vomit but being physically unable to do so’ – all problems that have previously been associated with anti-reflux surgery – were similar in the two randomised groups.

Conclusions

After 5 years' follow-up, a policy of relatively early laparoscopic fundoplication among patients for whom reasonable control of GORD symptoms requires long-term medication and for whom both surgery and medical management are suitable continues to provide better relief of GORD symptoms with associated better quality of life. Complications of surgery were rare. Despite being initially more costly, a surgical policy is likely to be more cost-effective for such patients suffering from GORD who were eligible for the REFLUX trial.

Trial registration

This study is registered as ISRCTN15517081.

Funding

Funding for this study was provided by the Health Technology Assessment programme of the National Institute for Health Research.

Gastro-oesophageal reflux disease

The lower oesophagus, at its junction with the stomach, normally acts as a sphincter to prevent the contents of the stomach flowing back up the oesophagus. When the sphincter does not work adequately, the acid stomach contents leak, or ‘reflux’, into the oesophagus. The commonest symptom that this causes is heartburn, a burning sensation in the chest or throat. GORD has been defined through an international consensus process as ‘a condition which develops when the reflux of stomach contents causes troublesome symptoms and/or complications’; in this consensus, symptoms were considered ‘troublesome’ ‘if they adversely affected a patient's well-being’. 6

Symptoms caused by gastro-oesophageal reflux are common: between 20% and 30% of a ‘Western’ adult population experience heartburn and/or reflux intermittently. 7–9

Treatment of GORD includes both medical and surgical management, the options depending on the severity of symptoms. The majority of people with reflux have only mild symptoms and require little, if any, medication. The simplest is self-administered antacids with advice to alter lifestyle factors such as dietary modification, smoking cessation and weight reduction. A minority have severe symptoms and develop overt complications, despite full medical therapy, and require surgical intervention. Among the remainder, control of symptoms requires regular or continuous acid-suppression therapy using either histamine receptor antagonists (H₂RAs) or proton pump inhibitors (PPIs); initial high-dose therapy may be followed by maintenance treatment using these drugs either intermittently or continuously at a reduced dose sufficient to suppress symptoms. It is from this intermediate group of patients with significant disease requiring maintenance medical treatment that most of the treatment costs for the health service arise.

Laparoscopic fundoplication

Interest in surgery as an alternative to long-term medical therapy for GORD has been considerable since the introduction of the minimal access laparoscopic approach in the early 1990s. 10 Randomised trials conducted comparing laparoscopic with open surgery showed similar improvement in symptoms but with clear benefits of the laparoscopic approach in terms of recovery and fewer postsurgical complications.11 As a consequence, surgery was suggested as an alternative to long-term maintenance medical treatment with anti-reflux drugs.

The operative method, whether using an open or a laparoscopic approach, involves performing a fundoplication by wrapping the fundus of the stomach around the lower oesophagus to create a high-pressure zone, thus reducing gastro-oesophageal reflux. The wrap created can be either complete (360°) or partial. Many operative variants have been described. The commonest operation is a 1-cm complete wrap fashioned over a large bougie, the so-called ‘short-floppy Nissen’. 12,13 There has been debate about the use of a partial rather than a total fundoplication. The partial approach has a number of potential advantages (such as fewer postoperative complications) but several controlled studies have shown broad equivalence between the two approaches;14 for the purpose of this study they were therefore regarded as equivalent. Although fundoplication is reported to produce resolution of reflux symptoms in upwards of 90% of patients, like all surgery it carries risks and can have side effects. There is also uncertainty about the durability of benefit and frequency and severity of side effects following surgical therapy. Long-term follow-up to 12 years after open reflux surgery suggested attenuated but continuing better control of reflux symptoms; however, other symptoms such as difficulties swallowing (dysphagia), rectal flatulence and inability to belch or vomit were more common in surgical patients. 15 An important objective of this study was to determine if the long-term pattern of symptoms following laparoscopic surgery was similar.

Medical management

Proton pump inhibitors, sometimes supplemented with prokinetics or alginates, are the most effective medical treatment for moderate to severe GORD. Once started on PPIs, the majority of patients with significant GORD remain on long-term treatment. 16 It is estimated that around 1% or more of the UK adult population are prescribed PPI maintenance therapy. 17–19 The cost to the NHS of medical management of GORD is considerable. In England alone, the cost of PPIs is estimated to be £220M per year. 20 Of this budget, most of this prescribing occurs within the primary care setting. 21,22

Although PPIs are generally considered safe, there is increasing acknowledgement of their possible adverse effects. 23,24 Gastric acid suppression predisposes to enteric infections and the sustained hypergastrinaemia resulting from PPI use causes rebound acid hypersecretion and the development of acid-related symptoms if the drug is stopped. Acute severe hypomagnesaemia has been recognised relatively recently as a rare adverse reaction to PPIs; the mechanism underlying it is not known. The clinical significance of impaired vitamin B₁₂ and iron absorption due to PPIs is uncertain; there is also controversy about the risk of fractures and pneumonia and about the occurrence and significance of gastric mucosal atrophy and intestinal metaplasia, which have been seen in Helicobacter pylori-positive patients taking PPIs. Drug–drug interactions have also been a cause for concern,25 although unequivocal evidence of their occurrence does not in itself establish clinical significance.

For the purpose of this study, medical therapy was taken to mean long-term therapy with PPIs (or H₂RAs if intolerant to PPIs).

Rationale for the study design

The original study design was based on the belief that decisions about the management of GORD should be made using unbiased, statistically precise comparisons of alternative policies. At study entry all patients fulfilled three criteria: they were on long-term acid suppression with PPIs; they had symptoms that were thought to be adequately controlled; and they were suitable in terms of fitness and comorbidity for either surgical or continuing medical treatment for their GORD. At the time that the study was planned, the consensus opinion of clinicians was that these three criteria identified GORD patients for whom surgical and continuing medical treatment could be considered equally acceptable treatment options and that, consequently, the comparison should be undertaken in patients meeting these criteria.

The most likely sources of bias were in the ways in which the groups being compared were selected; how their outcomes were assessed; and how the management was actually delivered. This is the basis for using a pragmatic randomised controlled trial (RCT) design. Random allocation protected against selection bias. Confining the trial to those with no clear treatment preference limits biased patient-centred assessment of outcome, and pragmatic comparison of alternative policies [with intention-to-treat (ITT) analysis] avoids bias introduced by individual cases of non-compliance. This approach had limitations, however, and for this reason we chose to incorporate two parallel, non-randomised preference groups.

Including those with a clear preference for one policy or the other allows broader extrapolation and generalisability. Study of this group may give insights into the reasons for preference and hence give pointers to patient choices after the study. 26 Furthermore, preference may influence outcome and, if so, this may also help when making treatment decisions. 26,27 A third reason for the parallel, non-randomised preference groups28 was that the addition of data from the preference groups may reduce imprecision around the estimates from the randomised comparison and this may be particularly useful for rare events, such as complications that can be confidently ascribed to one or other treatment. (The limitation is that the preference groups are not derived by random allocation, and hence the comparisons are exposed to the biases of non-randomised studies.)

Reliable comparisons within and between randomised and preference groups require valid measurement of treatment outcome. Although there were a number of quality-of-life (QoL) tools available, none was sufficiently specific to assess the spectrum of gastrointestinal symptoms associated with the treatment of GORD, particularly those due to surgery. For this reason we developed and validated a new outcome measure (the REFLUX questionnaire). We have continued to use this as the primary outcome measure in the longer-term follow-up reported here. Details of the REFLUX questionnaire and its derivation have been described elsewhere. 1,4

Gastro-oesophageal reflux disease and its management represent a very significant call on NHS resources. Although clinical effectiveness, acceptability and safety will be important determinants of future policy, the issues of cost and resource use may be over-riding. This is the reason for the economic evaluation component of this study. Policy should be guided by both assessment of the relative cost-effectiveness of alternative policies and assessment of the impact that possible policy changes would have for the NHS and for patients with GORD.

The cost of laparoscopic fundoplication appears to be equivalent to the cost of 2–3 years of maintenance treatment with PPIs, although it is acknowledged that the costs of PPIs are falling. 29 The costs of surgery are related largely to two factors: the incidence of complications/length of hospital stay and the number of patients requiring long-term medical interventions after surgery.

We addressed cost-effectiveness in our report of the first phase of the REFLUX trial. 1 We reported both a within-trial cost-effectiveness analysis based on the results up to 12 months after surgery and an extended cost-effectiveness model that explored a number of scenarios beyond 12 months. The within-trial analysis related the extra mean costs associated with the surgical policy to the estimated increase in mean quality-adjusted life-years (QALYs) associated with surgery up to that time. The incremental cost-effectiveness ratio (ICER) was around £19,000 when the ITT analysis was used. Taking into account the uncertainties around the estimates of both costs and utilities, it was calculated that the chance that the surgical policy would be cost-effective at a threshold of £20,000 per QALY was 46%. This indicated considerable uncertainty at thresholds that are currently commonly applied to costs per QALY. The limitations of the within-trial analysis were discussed in detail in the earlier report, in particular that it ignored costs and benefits that accrued after 1 year.

The economic model was designed to address the limitations of the within-trial analysis. It explored a range of scenarios of varying lifetime benefits and costs, and analyses gave a wide range of incremental costs per QALY of £1000–44,000, again indicative of wide uncertainty. The factors contributing most to this uncertainty were the projected health-related quality-of-life (HRQoL) parameters and the long-term uptake of medication following surgery.

Thus, although data available up to a time equivalent to 1 year after surgery provided promising evidence that surgical management might well be cost-effective, there was too much uncertainty, especially about longer-term costs and benefits, to provide clear guidance for decision-makers. This was the justification for the longer-term follow-up to 5 years reported here.

Original study design

The study had two complementary components:

1. a multicentre, pragmatic30 RCT (with parallel non-randomised preference groups) comparing a laparoscopic surgery-based policy with a continued medical management policy to assess their relative clinical effectiveness

2. an economic evaluation of laparoscopic surgery for GORD to compare the cost-effectiveness of the two management policies, identify the most efficient provision of future care and describe the resource impact that various policies for fundoplication would have on the NHS.

Eligible patients who consented to participate in the RCT were randomly allocated to either laparoscopic surgery or continued medical management. Those patients who had a strong preference for one or other of the two treatment options could be recruited to the preference study. Clinical history was recorded at study entry. Participants completed health status questionnaires at the time of recruitment to the study and then at specified times equivalent to 3 and 12 months and then 2, 3, 4 and 5 years after surgery.

Approval for this study was obtained from the Scottish Multicentre Research Ethics Committee and the appropriate Local Research Ethics Committees.

Clinical centres

Clinical centres were based on local partnerships between surgeons with experience of laparoscopic fundoplication and gastroenterologists, with whom they shared the secondary care of patients with GORD. Centres were eligible if they included:

• a surgeon who had performed at least 50 laparoscopic fundoplication operations

• one or more gastroenterologists who agreed to collaborate with the surgeon(s) in the trial.

Study population

Eligible patients were those for whom care had been provided by a participating clinician who was uncertain which management policy (surgical or medical) was better. In addition, patients had to have documented evidence of GORD (based on endoscopy and/or manometry/24-hour pH monitoring) as well as symptoms for >12 months requiring maintenance PPI therapy for reasonable symptom control. Patients who were intolerant to PPIs and therefore required H₂RA therapy to control their symptoms were also eligible. Patients who were morbidly obese [body mass index (BMI) > 40 kg/m²] or who had Barrett's oesophagus of > 3 cm or evidence of dysplasia, a paraoesophageal hernia or an oesophageal stricture were all excluded.

Eligible patients who did not want to take part in the randomised trial because of a strong preference for one type of management or the other were invited to take part in the preference arm of the study. For logistical reasons and to maintain a balance between the sizes of the randomised and the preference groups, we aimed to cap the numbers of participants recruited to the preference arms to 20 per arm in each centre.

All participants gave their informed consent.

Health technology policies being compared

Study registration (and treatment allocation when randomised)

The treatment allocation for participants in the randomised component of the trial was computer generated; it was stratified by centre, with balance in respect of other key prognostic variables – age (18–49 years or 50+ years), sex (male or female) and BMI (≤ 28 or > 29 kg/m²) – by a process of minimisation. Randomisation was organised centrally at the Health Services Research Unit, University of Aberdeen, and was independent of all clinical collaborators.

Clinical management

Participants who were allocated to surgical management were invited to a consultation with the collaborating surgeon. During this consultation, the surgeon confirmed that there was no contraindication to surgery and discussed the operation in more detail, before arranging an operation date. The surgeon recorded intraoperative details on specially designed study forms. All other in-hospital data collection was the responsibility of the local study nurse. In all respects, other than the trial interventions, clinical management was left to the discretion of the clinician responsible for care. This continued to be the case in the extended follow-up phase, which is the focus of this report, with GPs monitoring subsequent care needs throughout the follow-up period.

Data collection

Follow-up by postal questionnaire was first performed at 3 months after surgery, or at an equivalent time among those who did not have surgery, and then annually. The questionnaire used for the follow-up at 2–5 years was similar to the questionnaire that had been used in the earlier phase of the trial up to 12 months after surgery. Non-responders received up to two reminder telephone calls or letters to encourage return of their postal questionnaires. On occasion, and at the participants' convenience, a shortened version of the questionnaire was completed over the telephone.

From around half-way through the 5-year follow-up, participants were sent a £5 gift voucher with their final postal reminder to compensate for their time in completing the questionnaire. This decision was taken based on the findings of a systematic review of the effects of incentives on postal questionnaire return32 and specific randomised trials that evaluated the use of vouchers. 33–35

All data were sent to the trial office in Aberdeen for processing. A random 10% sample of all data was double-entered to check accuracy and no significant errors were identified. Extensive range and consistency checks further enhanced the quality of the data.

The principal study outcome measure

The primary outcomes for measuring the differences in effects between medical and surgical management were:

• a ‘disease-specific’ measure incorporating assessment of reflux and other gastrointestinal symptoms and the side effects and complications of both therapies (the REFLUX questionnaire was developed specifically for this study4)

• NHS costs including treatments, investigations, consultations and other contacts with the health service.

The secondary outcome measures were:

• HRQoL – measured using the European Quality of Life-5 Dimensions (EQ-5D)36 and Short Form questionnaire-36 items (SF-36)37

• patient costs, including loss of earnings, reduction in activities and the costs of prescriptions and travel to health care

• other serious morbidity, such as operative complications

• (further) anti-reflux surgery

• mortality.

An example of the annual questionnaire used for collecting this information is provided in Appendix 1.

Sample size

The original aim was to recruit 600 participants to the randomised trial to give 80% power to identify a difference between the two groups of 0.25 of a standard deviation (SD) in respect of the disease-specific instrument and other continuous variables such as EQ-5D and SF-36, using a significance level of 5%. Based on the same arguments, it was planned that 300 people would be recruited to each arm of the preference study. The cost savings of a surgical policy largely depend on the number of patients managed surgically who no longer require PPI treatment, and a trial with 300 surgically managed patients would have estimated this proportion to within about 5% with 95% statistical confidence.

However, prompted by a lower rate of recruitment than expected, this target was revised in January 2003 in consultation with the Data Monitoring Committee (DMC) and representatives of the HTA programme. It was agreed that a larger benefit (0.3 of a SD) was clinically plausible based on improvements seen after surgery in the accruing literature among more severely affected people (who were not eligible for the trial). This was calculated to require 196 in each group to give 80% power (2p = 0.05).

Statistical considerations

This report describes analyses of annual questionnaire data up to 5 years after surgery (or an equivalent time if managed medically). As a general rule, in the tables and analyses presented in this report, the participants in the randomised groups are separate from those in the preference groups. A sizeable group of patients allocated to surgery did not receive surgery. Therefore, to investigate the potential influence of this non-compliance with allocation, summary statistics in the results tables are given for four main analysis populations (comprising eight groups of participants):

1. Randomised ITT population (groups that were randomised to either surgery or medical management).

2. Per-protocol (PP) population (groups that were either randomised to surgery and received surgery in the first year or randomised to medical management and did not receive surgery in the first year).

3. Preference ITT population (groups that preferred either surgery or medical management at recruitment).

4. Preference PP population (groups that either preferred surgery at recruitment and received surgery in the first year or preferred medical management and did not receive surgery in the first year).

The primary outcome measure (REFLUX QoL score) and secondary outcome measures (SF-36, EQ-5D, REFLUX symptom scores, anti-reflux surgery and use of reflux-related drugs) were analysed using general linear models. The analyses adjusted for the minimisation covariates (age, BMI and sex) and where appropriate (defined by significant at the 5% significance level) also adjusted for baseline measures and baseline measures by treatment interaction. A secondary, pre-stated subgroup analysis explored the differential effects of surgeon's preferred operative procedure on the primary outcome measure. All analyses were reported with 95% confidence intervals (CIs).

The primary analysis of the randomised groups was by ITT. The ITT approach sustains the integrity of the randomisation and gives the least biased estimate of effectiveness of the two forms of management. Given that a sizeable minority of the randomised surgical participants did not receive surgery, we were also interested in estimating the efficacy of the initial treatment received as a secondary comparison (i.e. commonly known as a PP analysis). In an open trial design a PP analysis can have substantial selection bias. To minimise the effects of selection bias we used the method of ‘adjusted treatment received’ as described by Nagelkerke et al. 38 and others. 39,40 The method used a two-stage least-squares approach whereby treatment randomised was regressed onto treatment received and the residuals from that model were used as an independent variable in a second model, together with the treatment received, to estimate the effects on the various primary and secondary outcome measures.

For the preference study, only the primary outcome was analysed statistically. The analysis compared the preference surgical group with the preference medical group and adjusted for the minimisation factors. As described above, for logistical reasons and to maintain balance between the randomised and preference groups, we capped the number of preference participants at 20 per group per centre. The study design was not therefore a true comprehensive cohort. We did consider modelling differences between the randomised and preference groups; however, it is not universally accepted that formal modelling is appropriate in this context. In this case we knew from the randomised arms that there was a strong interaction between treatment effects and baseline REFLUX QoL, and in addition we knew that there was a large difference in QoL between preference arms at baseline (and patient demographics such as age and sex). We therefore decided that formal modelling of the arms would add little to the comparison given the large confounding between preference groups.

Sensitivity analyses

The sensitivity of the primary outcome analysis result was investigated using two approaches – the effect of excluding a large centre and the effects of missing data. In the first approach the largest recruiting centre, Aberdeen, was excluded and the analysis as described above was rerun. Second, previous work demonstrated that the primary outcome was likely missing at random (MAR) or missing completely at random (MCAR) and that a repeated measures analysis (using all available data) was an appropriate statistical method for analysing data up to 12 months. 41 We therefore used a repeated measures analysis on the primary outcome across all of the follow-up data (12 months to 5 years) to investigate the effect of incorporating a profile of measures for each participant. No further imputation for missing values was necessary.

Data monitoring

During recruitment, an independent DMC met on three occasions and each time saw no reason to recommend any fundamental changes to the protocol. The committee did not meet after recruitment was completed.

Recruitment to the trial

Participants were recruited in 21 clinical centres, all within the UK (their locations are listed on the left-hand side of Table 1). Recruitment to the trial was open from March 2001 until the end of June 2004, although not all centres enrolled over the total period because of the staggered introduction of centres and early closure for logistical reasons in a few places. 1

A total of 357 participants were recruited to the randomised component: 178 allocated to surgery and 179 allocated medical management. 453 participants agreed to join the preference component: 261 choosing surgery and 192 choosing medical management. Table 1 shows recruitment by centre. Around one-fifth of the randomised participants were enrolled in Aberdeen; no centre contributed > 10% of participants in the preference component.

Analysis populations

Throughout the analyses presented later in this chapter, the participants in the randomised component are kept separate from those in the preference component (other than for rare surgical events). The numbers of participants in each of the four main analysis populations are shown in Table 2. All 357 who joined the randomised component are in the randomised ITT population; only the 280 within this group who actually received their allocated management over the first year are in the randomised PP population. All 453 participants who joined the preference component are in the preference ITT population; the 407 of these who, by the end of the first year, were managed as originally chosen were in the preference PP population.

Trial conduct

The derivation of the main study groups and their progress through the stages of follow-up in the trial are shown in Figure 1. This is in the form of a CONSORT (Consolidated Standards of Reporting Trials) flow diagram. In total, 1078 patients were considered for trial entry and 200 of these were found not to meet one or more of the eligibility criteria. Of the 68 patients eligible for the study but not recruited, 51 declined to participate, six were subsequently deemed inappropriate for the study by the surgeon responsible for care and the remaining 11 were missed.

FIGURE 1.

The CONSORT diagram.

Details of the clinical management actually received are described later in this chapter.

The mean (SD) time intervals in months between the receipt by the trial office of each subsequent annual postal questionnaire are shown in Table 3; all were near 12 months, as would be expected. There was, however, a difference between the randomised groups in the time interval between the 1-year and the 2-year questionnaires (mean 12.2 months surgical group vs 13.9 months medical group). In part, this was due to more late returns in the medical management group – the median intervals were closer: 12.00 and 13.00 months respectively. As described previously,1 early follow-up was adjusted to be at a time equivalent to 3 and 12 months after surgery. The adjustments in the medical group to match this could be only approximate and this is the explanation for the difference that remained between the randomised groups. An advantage of long-term follow-up to 5 years is that any difference in the timing of follow-up becomes proportionately smaller over time.

More details of the response rates to the annual questionnaires are provided in Table 4. The overall rates of return of annual follow-up questionnaires (years 1–5) were 89.5%, 77.7%, 76.7%, 69.8% and 68.9% of the study participants. Seven participants are known to have died up to the end of the 5-year follow-up; equivalent response rates among those not known to have died are 89.8%, 77.9%, 77.0%, 70.2% and 69.5%. There were no substantive differences in response rates between the groups.

Three participants died before the 1-year follow-up was reached: two in the preference surgery group and one in the randomised medical group. None of these participants actually had surgery. Four died subsequently; there is no evidence linking these deaths to trial participation.

Description of the groups at trial entry

Baseline characteristics of groups compared at 5 years

There were differences in baseline characteristics between those who had completed a questionnaire at 5 years and those who had not (Table 8). For example, responders had a higher mean age (47.9 years vs 43.6 years), had been on prescribed medication for a shorter period at recruitment to the REFLUX trial (50.5 months vs 60.2 months) and had higher QoL scores at baseline (measured on the disease-specific REFLUX instrument, EQ-5D and SF-36).

However, the baseline characteristics of those in the randomised surgical and randomised medical groups who completed a questionnaire at 5 years were very similar, with the only notable difference being in BMI (Table 9). The mean baseline BMI among responders in the randomised surgical group was higher (29.0 kg/m²) than that for responders in the randomised medical management group (27.7 kg/m²). As described in Chapter 2, these results confirmed that a repeated measures analysis assuming no differential loss to follow-up could be considered.

Surgical management

Table 10 summarises the use of surgery in the four study groups over the full 5-year follow-up period. At the end of the first year, 111 participants (62.4%) randomised to surgery had actually undergone fundoplication. Over the next 4 years, one more member of this group had fundoplication, bringing the total to 112 (62.9%). In the randomised medical group, 10 participants (5.6%) had fundoplication in the first year, with a further 14 participants having fundoplication in subsequent years, bringing the total at 5 years to 24 (13.4%). In the preference surgical group, 218 participants (83.5%) had fundoplication in the first year, with four more in the period up to 5 years, taking the percentage to 85.1%. Surgical management applied to only three participants (1.6%) in the preference medical group in the first year, with a further three being operated on in the subsequent 4 years (total 3.1%).

Information about the reasons why participants allocated surgery did not receive it in the first year is available for 47. For 25 of these 47, this was a clinical decision, most commonly the surgeon deciding that surgery was not appropriate; most of the other 22 changed their minds about surgery for a variety of work- or home-related reasons. A further 20 withdrew for unknown reasons. There is no doubt, however, that a number of these participants suffered long delays before being formally offered surgery, and this was an important factor in their eventual decision to choose not to have surgery after all. The trial was conducted at a time when there was great pressure on surgical services in the NHS, with long delays for elective surgery for non-life-threatening benign conditions being common. Indeed, the average time between trial entry and surgery in the trial was 8–9 months. 1

Details of the surgery received by the 111 participants (62.4%) randomised to surgery and the 218 preference participants (83.5%) who actually received surgery in the first year, the perioperative complications that they experienced and their hospital stay have been reported previously but are summarised in Appendix 2 for completeness. There were no perioperative deaths.

Table 11 shows the numbers of those who had fundoplication who subsequently had a second reflux-related operation during the 5 years of follow-up. Overall, this applied to 16 participants (4.4%) among the 364 who had a first operation: five (4.5%) in the randomised surgery group; one (4.2%) in the randomised medical group; eight (3.6%) in the preference surgery group; and two (33.3%) in the preference medical group. In total, five of the 16 operations were reconstructions of the same wrap, three were repairs of hiatus hernia only, six were conversions to a different type of wrap and two were reversals of the fundoplication. Two of these 16 participants had a third reflux-related operation; both were in the preference surgery group – one a reconstruction of the same wrap and one a repair of hiatus hernia only.

Late postoperative complications

Table 12 describes late postoperative complications among those participants who had surgery, in each of the study groups and overall. Of the total 364 who had fundoplication, 12 (3.3%) had a late complication: four (1.1%) were oesophageal dilatations/stricture dilatations; three (0.8%) were repairs of incisional hernias; and five (1.4%) were a heterogeneous group of other complications as detailed in the table.

Medication

Figure 2 summarises reported use of any PPI medication in the previous 2 weeks across the follow-up time points of the trial. Full details are provided in the tables in Appendix 3. From the time of the first annual follow-up onwards, rates in both medical groups were consistently around 80%. The rates in the randomised surgical ITT group at the first, second and third annual follow-ups were approximately 36–38%, rising to 43% in the fifth year. The extent to which these rates reflected medication taking among those allocated to surgery and who had fundoplication (rather than those who did not have surgery) can be gauged from the randomised surgery PP group: 7.3% (3 months), 12.5% (1 year), 15.1% (2 years), 19.6% (3 years), 23.9% (4 years) and 25.6% (5 years).

FIGURE 2.

Use of PPI medication at baseline and at follow-up points up to 5 years.

Table 13 allows further exploration of the reasons for the rise in medication use in the randomised surgery group. It distinguishes those reporting taking medication at the end of the first year of follow-up from those who indicated that they were not taking medication at that time. It shows that around 10–20% of those taking medication at the end of the first year did not report medication use at subsequent annual follow-up. Among those not taking medication at the first annual follow-up in the surgical groups, around 10% rising to around 20% reported medication use at subsequent annual follow-up. This contrasts with the rates in the medical groups, with around 50–60% of those not taking medication at the end of the first year reporting anti-reflux drug use in subsequent annual follow-up. The pattern of type of PPI used changed over the course of the study. Although lansoprazole had been the most commonly used PPI at trial entry, omeprazole use increased over time to become the predominant PPI.

Outcome

Health status

Full details of the health status and QoL measures at each time point of follow-up are in the tables in Appendix 4. Details of the statistical testing of the health status and QoL scores can be found in the next section of this chapter.

REFLUX score

Figure 3 summarises changes in the disease-specific REFLUX score over the follow-up period. From this it can be seen that the scores at all time points are highest (indicating fewest symptoms) in the randomised surgical and preference surgical groups. However, the differences between the surgical and medical groups narrow over time. This is due principally to the scores in the randomised medical group improving over the first 3 years and, to a lesser extent, those in the preference medical group improving over the latter end of the follow-up period. The scores for the five components of the measure are summarised graphically in Figures 4–8. These show that the overall difference between the groups is principally due to the ‘general discomfort’ component and, to a lesser extent, the ‘nausea and vomiting’ and ‘activity limitations’ components.

FIGURE 3.

Mean REFLUX QoL score at baseline and at follow-up points up to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 4.

Mean REFLUX QoL general discomfort symptom score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 5.

Mean REFLUX QoL wind and frequency symptom score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 6.

Mean REFLUX QoL nausea and vomiting symptom score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 7.

Mean REFLUX QoL activity limitation symptom score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 8.

Mean REFLUX QoL constipation and swallowing symptom score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

Short Form questionnaire-36 items

The pattern of SF-36 scores, both for the composite physical and mental scores and for the individual dimensions (Figures 9–16), was similar to that seen for the REFLUX score, although more compact. Differences narrowed over the 5 years of follow-up, with the ‘general health’ dimension showing the clearest differences between the surgery and the medical management groups.

FIGURE 9.

Mean SF-36 norm-based physical functioning score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 10.

Mean SF-36 norm-based role physical score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 11.

Mean SF-36 norm-based bodily pain score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 12.

Mean SF-36 norm-based general health score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 13.

Mean SF-36 norm-based vitality score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 14.

Mean SF-36 norm-based social functioning score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 15.

Mean SF-36 norm-based role emotional score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

FIGURE 16.

SF-36 norm-based mental score at baseline and follow-up points to 5 years (score range 0–100; the higher the score, the better the patient felt).

European Quality of Life-5 Dimensions

Figure 17 graphically displays the EQ-5D scores over the course of the follow-up period. The pattern is similar to that seen for the REFLUX score although differences are less marked and only clearly seen over the first 3 years.

FIGURE 17.

EQ-5D at baseline and follow-up points to 5 years.

Statistical analyses

Discussion

Follow-up to 5 years after laparoscopic surgery described here provides clear evidence of sustained improvement in GORD symptoms, as judged by the REFLUX QoL scores. Differences between the groups as randomised did tend to diminish over the course of the study; nevertheless, the analyses at 5 years (the primary end point) showed highly statistically significant results with effect sizes of the order of 0.6 of a SD.

This report concentrates on the data collected annually at a time equivalent to between 2 and 5 years post surgery. Data were collected through self-complete postal questionnaires, backed up by postal and telephone reminders and occasional completion of the questionnaire over the telephone. The response rate did drop over time, from 90% at 1 year to around 70% at 5 years. The principal reason for not obtaining a follow-up questionnaire was a loss of contact, such as following a home move; the second most common reason was a decision by a participant to decline further follow-up. The category of ‘non-responder’ accounted for only around 8% of those without a follow-up questionnaire. Response analysis showed that responders at 5-year follow-up had a higher mean age, had been prescribed anti-reflux medication for a shorter period of time at recruitment and had higher QoL at baseline. However, the characteristics of responders and non-responders at 5 years were similar across the two randomised groups.

Randomised trials, such as the REFLUX trial, that compare surgery with medical management are challenging to mount because of the stark contrast between the treatments compared. As described in the previous report of this study, recruitment was not easy and it is to the credit of the many staff in the 21 centres involved in the trial that this was accomplished successfully. A second challenge was that, after randomisation, a sizable proportion of participants did not receive the treatment to which they had been allocated – again, reflecting the contrasts in the treatments. We explored the impact of this in a number of ways.

Figure 18 shows the results of a supplementary analysis of the group randomly allocated surgery stratified by whether or not they actually had surgery. It shows that those who had surgery started from a lower REFLUX QoL baseline score (had worse symptoms) than those who did not undergo surgery, and then had a sharp rise in score following the operation such that their scores were consistently higher than those who did not actually have fundoplication. To put this another way, the improvement seen among those who had surgery was greater than that in the randomised group overall.

FIGURE 18.

Mean REFLUX QoL scores for (a) all randomised to surgery, (b) those randomised to surgery who had fundoplication and (c) those randomised to surgery who did not have surgery.

Figure 19 shows a similar supplementary analysis of the group allocated medical management stratified by whether or not they in fact had surgery in the first year. This shows that those who had fundoplication (the lowest line) had more severe symptoms of GORD (low REFLUX QoL scores) at the time of trial entry, worse even than the preference surgical group. In contrast, those solely managed medically had relatively high baseline scores. Scores among those randomised to medical management who had surgery improved markedly over the course of the follow-up, such that by years 4 and 5 the scores in the two strata were similar. This indicates that much of the narrowing of the scores in the ITT groups over the 5 years can be explained by surgery in the randomised medical group.

FIGURE 19.

Mean REFLUX QoL scores for (a) all randomised to medical management, (b) those randomised to medical management who did not have surgery and (c) those randomised to medical management who had fundoplication.

We assessed more formally the extent to which surgery in the randomised medical management group might have affected the results by undertaking adjusted treatment received analyses. We decided to base these on treatment status at the first year follow-up point. We chose this partly to be consistent with our previous report of the results up to 1 year and partly because we considered that those who had surgery after that time point were likely to be highly selected. To put this another way, we were concerned that a PP analysis up to 5 years would be particularly prone to bias. The adjusted treatment received analyses, as expected, indicated larger effects of surgery – with differences in score around 25–50% higher. As illustrated by the preference groups in this study, the proportion of those recommended surgery and willing to have it who subsequently go on to have fundoplication is likely to be higher in everyday practice. Hence, we would argue that the results of the adjusted treatment received analyses are likely to provide a better estimate of the benefits of a policy of laparoscopic fundoplication as would apply in the health service.

The principal concern about laparoscopic fundoplication is possible risks associated with the surgery. We described intra- and postoperative surgical outcomes in our previous report. 1 Among the 329 patients in the randomised surgical and preference surgical groups who had fundoplication in the first year, there were no major surgical complications. Two patients (0.6%; 95% CI 0.1% to 2.2%) required conversion to an open procedure; eight (2.4%; 95% CI 1.2% to 4.7%) had a visceral injury; and one (0.3%; 95% CI < 0.1% to 1.7%) had a blood transfusion. Three were admitted to a high-dependency unit, but none to an intensive care unit. The 5-year follow-up provides information about longer-term risks. We are aware of seven deaths among trial participants; however, none has an apparent link to the trial. Twelve (3.3%) of the total of 364 participants who had a fundoplication had a late complication: four were oesophageal dilatations/stricture dilatations, three had repairs of incisional hernias and five were a heterogeneous group of other complications (see Table 12). Sixteen (4.4%) of those who had fundoplication required further surgery (see Table 11): five reconstruction of the same wrap, six conversion to another type of wrap, three repair of hiatus hernia only and two reversal of fundoplication. These, albeit uncommon, complications need to be taken into account when surgery is being considered.

Proton pump inhibitor use in the randomised medical group was consistently around 80%, although these participants were not always the same people at each follow-up. In our questionnaire, we chose to ask about anti-reflux drug use over the preceding 2 weeks as we thought that a recollection over a longer period would be unreliable. Nevertheless, taking of PPIs seems to be dynamic (patients stopping and restarting) and rates of use at any time over a longer period would likely have been higher. We did observe more visits to GPs in the medical groups for reflux-related reasons during the 5 years of follow-up but are not able to say whether this was due to routine reassessments or because symptom control was less stable or inadequately controlled in the medical group.

The pattern of PPIs used did change over the course of the study. At baseline, the commonest PPI was lansoprazole, but omeprazole superseded this over the course of the trial. Much of this change occurred in the first year and hence could be a consequence of the review of medical management that was part of the trial management for those randomised to medical management.

The larger number of overnight hospital admissions in the randomised medical management group was largely, but not totally, explained by the minority who went on to have surgery; as discussed in Chapter 5 describing the economic evaluation, this was the principal driver of extra resource use by the medical group during the longer-term follow-up.

Despite the methodological challenges alluded to above, the study, through the data presented here, has successfully addressed the first of the objectives of this longer-term follow-up: to assess whether or not short-term clinical benefits, principally in terms of symptom control, are sustained – they are, albeit attenuated. In the next chapter we consider the REFLUX trial in the context of the three other randomised trials that have been conducted worldwide comparing laparoscopic fundoplication with medical management, and assess whether or not the results of the REFLUX trial are consistent with those of the other trials.

Introduction

The REFLUX trial is one of four randomised trials that have compared laparoscopic surgery with medical management of GORD. Although the REFLUX trial has similarities to the other trials, its design is the most pragmatic42 and this is reflected in significant differences in comparison with the other trials. The characteristics of the four trials are summarised in some detail in Appendix 5; key similarities and differences in characteristics between the REFLUX trial and the other trials will be highlighted here. This overview draws heavily on the relevant Cochrane review,43 two of whose authors are authors of this report, but incorporates reports published since the Cochrane review, identified primarily through an updated search using a similar strategy to the one described in the Cochrane review.

The three comparable trials

The Anvari et al. trial44–46 is a publicly funded single-centre trial conducted in Canada, led by upper gastrointestinal surgeons. It is the smallest of the four trials (104 randomised). The two intervention policies were standardised and the surgery was undertaken by only four surgeons (Table 23). Reflecting this, nearly all participants – unlike in the REFLUX trial – were managed in the way allocated. Like the REFLUX trial, its primary outcome was a GORD-related QoL instrument (the GERSS or Gastro-Esophageal Reflux Symptom Score), and HRQoL was measured with the same instruments as in REFLUX (SF-36 and EQ-5D). The first report described the trial up to 12 months after surgery,44 and recent papers have reported 3-year results45 and an economic evaluation. 46 At 3 years, participants in the medical group were offered surgery and a large proportion (42%) accepted; hence, although further follow-up is reported to be ongoing, it will be of limited usefulness in comparing laparoscopic surgery with medical management.

The LOng-Term Usage of esomeprazole versus Surgery for treatment of chronic GERD (LOTUS) trial47–50 is the largest of the four trials (554 randomised). The study was funded by a pharmaceutical company, AstraZeneca, and the reports all include authors based in the company. The trial involved 39 centres in 11 European countries and was led by an upper gastrointestinal surgeon. The trial is described as ‘not designed as a superiority or equivalence trial but, rather, was an exploratory study to estimate the efficacy of laparoscopic anti-reflux surgery and PPI treatment in PPI responders’. Unlike in the REFLUX trial, all participants had shown response to PPI treatment in a run-in phase, and both clinical management policies were strictly standardised (see Table 23).

The method by which the total fundoplication approach was standardised has been described in detail. 50 In the medically managed group, the only PPI used was esomeprazole, initially at the standard dose of 20 mg. Both the surgical and medically treated patients were followed up by the investigators at 6-monthly intervals and symptoms were assessed using the Gastrointestinal Symptoms Rating Scale (GSRS) questionnaire. In the medically treated group, esomeprazole could be increased to 40 mg once a day and then to 20 mg twice a day if symptom control was insufficient. Another key difference from the REFLUX trial was that the primary outcome measure was ‘treatment failure’. A single definition of treatment failure could not be used for both trial groups; rather, this was specifically defined for each group (including in the medical group need for escalation of medication and in the surgical group, need for regular medication). The concern is that the thresholds for these may not reflect similar levels of GORD. A GORD-specific QoL instrument (Quality of Life in Reflux and Dyspepsia or QOLRAD54) was among the secondary outcomes but was given relatively little emphasis in the reporting of the trial. No HRQoL instruments were used and there was no economic evaluation. Although the main analysis was said to be carried out on an ITT basis, it seems that the 40 people allocated surgery who did not receive it were excluded from analyses. Results were first reported after 3 years' follow-up47 and recently 5-year data have been published. 48

The Mahon et al. trial51–53 was a two-centre UK trial led by and involving two upper gastrointestinal surgeons. It is not clear how the main trial was funded but supplementary funds were provided by Jansen Pharmaceuticals ‘for physiological studies’ and by Ethicon Endo-Surgery for the economic analysis. 52 In total, 217 people were randomised; the sequence was ‘computerised’ but the randomisation process and extent of concealment were not described. The two surgeons used a similar Nissen fundoplication method (see Table 23) and there was the option of four different PPI regimens depending on what PPI a participant had been taking prior to the trial. A range of outcome measures were reported and these included a gastrointestinal symptom score (GSRS) and a HRQoL measure [Psychological General Well-Being Index (PGWI)]. 55 All those allocated to medical management were offered surgery after 1 year (and apparently this was made clear to potential participants before trial entry) and the majority [54/94 (57%)] then had surgery. The 1-year follow-up was thus essentially the end of this randomised trial, even though a further follow-up has been reported. 53

Gastro-oesophageal reflux disease-related quality-of-life and symptom scores

Data available for each of the trials that describe GORD QoL or symptom scores at 1, 3 and 5 years' follow-up are summarised in Tables 24–26. Although it is not possible to combine data because different instruments (or subscales of instruments) were used in the trials, the results are consistent.

At 1 year there are eligible data from all four trials (see Table 24). In each case there are highly statistically significant differences all favouring the surgically managed groups. As mentioned above, the randomised element of the Mahon et al. trial51–53 ended at 1 year but data at 3 years are available for the other three trials (see Table 25). Again, all favour the surgical group and this was statistically significant in both the LOTUS47–50 and the REFLUX1–3 trials.

Only the LOTUS and (now) the REFLUX trial have reported 5-year follow-up. GORD-related QoL scores significantly favour the surgical groups in both trials (see Table 26).

Health-related quality of life

No general HRQoL measure has been reported for the LOTUS trial. 47–50 Data for the other three trials are shown in Tables 27–29. The SF-36 was used in the Anvari et al. trial44–46 as it was in the REFLUX trial. 1–3 Unfortunately, it is reported only as the two summary component scores, physical (PCS) and mental (MCS), plus the ‘general health’ domain score. For comparability, in Tables 27 and 28 the same score formats are shown for the REFLUX trial but it should be borne in mind that the eight domain scores shown in Chapter 3 for the REFLUX trial are more informative.

At 1 year, in both trials, the PCS and MCS favour the surgical group, although only the difference in the PCS in the REFLUX trial1–3 is statistically significant. Both trials showed marked differences in the ‘general health’ domain score. There was also a statistically significant difference favouring surgery in the Mahon et al. trial51–53 (based on the PGWI).

Although EQ-5D data were collected in the Anvari et al. trial,44–46 they were not reported in a way that allows interpretation. At baseline, scores were markedly lower in the surgery group [mean 0.68 (SD 0.28) vs 0.76 (SD 0.21)] and the reason for this imbalance is not clear. At 1 year the equivalent results were 0.79 (SD 0.23) compared with 0.81 (SD 0.19), that is, still lower in the surgery group. As shown in Table 27, in the REFLUX trial,1–3 the mean 1-year EQ-5D score was higher in the surgery group (p = 0.07).

At 3 years, the report of the Anvari et al. trial mentions collection of the SF-36 ‘every 3 months’ but the only data reported are for the ‘general health’ domain score. This, as in the REFLUX trial, significantly favours the surgical group (see Table 28). There is no mention of collection of EQ-5D data in the 3-year follow-up of the Anvari et al. trial. At 5 years, the only data describing generic HRQoL are from the REFLUX trial (as the LOTUS trial has not included a measure) (see Table 29).

Individual symptoms of gastro-oesophageal reflux disease or its management

Data describing individual symptoms are available for all trials, although only dysphagia was reported in the Mahon et al. trial. 51–53

Surgical complications

Like all procedures involving surgery under general anaesthesia, laparoscopic fundoplication carries risks. Table 30 summarises intra and early postoperative complications reported in the four trials.

Discussion

Of the four trials, the REFLUX trial is the most pragmatic in design. It involved a large proportion of UK centres where laparoscopic anti-reflux surgery is undertaken and the surgery was undertaken by NHS upper gastrointestinal surgeons within these centres, all of whom had experience of carrying out the procedure. The exact method of fundoplication was left to the discretion of the surgeon, so he or she was comfortable with the approach. After surgery and, in the medically treated patients, after optimisation of their PPI medication, care of the participants was the responsibility of GPs. The principal measure of outcome was a patient-reported disease-specific QoL measure. Unlike the other trials, the REFLUX trial was coordinated from an accredited trials unit, local recruitment was led by gastroenterologist/gastrointestinal surgeon partnerships rather than by gastrointestinal surgeons alone, and the trial was publicly funded through the HTA programme rather than by industry.

In respect of potential benefits of surgery, the four trials appear to be consistent. All show significantly better relief of GORD symptoms for as long as the length of their current follow-up. (Surprisingly, the LOTUS trial report48 does not draw attention to this but, judged on data describing the QOLRAD reported in an e-table, there are significant differences between the groups in all dimensions of this instrument, favouring surgery.) Data available describing the principal symptoms of GORD (heartburn and regurgitation) show large differences, again favouring surgery. Only limited data are available from generic QoL measures, and much of this is from the REFLUX trial; although differences are less marked than for the GORD-related QoL instruments, they are consistent with benefit from surgery.

The four trials are broadly consistent in respect of intraoperative and early postoperative complications: a small number of operations are converted to an open procedure, a small number of laparoscopic procedures have associated visceral injuries, a small number of people have problems postoperatively and a small number require dilatation of the wrap. The REFLUX trial suggests that 4.5% have reoperations and the other trials are broadly consistent with this. None of the trials had a reported perioperative death. Data from the Finnish Registry56 suggest a mortality of 0.1%, but this is based on a single case among 1162 people who had laparoscopic fundoplication; furthermore, the registry included all cases of fundoplication and hence went beyond the sorts of patients recruited to the REFLUX trial.

The other trials, particularly the LOTUS trial, show higher rates of dysphagia and flatulence following laparoscopic fundoplication than in the medically managed group. As mentioned above, a small number of participants in the REFLUX trial did have a dilatation procedure, presumably because of difficulty swallowing, but this was not reflected in responses to the REFLUX questionnaire, suggesting that there were only a few isolated cases of dysphagia following surgery in this trial. Similarly, there were no significant differences in flatulence in the REFLUX trial.

Hence, taking all four trials together, it is now possible to give a clear picture of most of the potential benefits and risks of laparoscopic fundoplication, at least up to 5 years. There are, however, differing resource implications of surgery and medical management. In the next chapter we explore whether or not the benefits of surgery in patients with established GORD requiring long-term PPI therapy for reasonable control and suitable for either clinical policy (average age around 45 years) are sufficient to outweigh any differences in costs.

Systematic review of existing cost-effectiveness evidence

The aim of this systematic review is to identify any existing cost-effectiveness studies that compare laparoscopic fundoplication with medical management for GORD. A previous HTA report included a review of the evidence available from 1995 to December 2005 and identified three relevant studies (described below). 1 The updated search focuses on the period from December 2005 to April 2011. The methods used to identify studies and the results of the systematic search are discussed in the sections below.

Within-trial economic evaluation

Follow-up data from the REFLUX trial up to 5 years after surgery are now available. These economic data represent the longest follow-up of randomised patients currently available. These data can help to inform the question regarding the sustainability of initial improvement in HRQoL following surgery. This section describes the updating of the cost-effectiveness analysis using these data to reduce the level of uncertainty about the cost-effectiveness of surgery and thus its role in the NHS.

Results

Patient population

Complete-case analysis consisted of the patients who returned all questionnaires and completed all EQ-5D profiles. Overall, there are 172 patients in the complete-case analysis (88 randomised to medical management and 84 randomised to surgery). Table 32 shows the numbers of questionnaires returned (includes those with some missing data) and the numbers of completed questionnaires returned for each year. As expected, the number of questionnaires returned in each year of follow-up decreases with time. The return of questionnaires does not follow a monotonic pattern, that is, patients who did not return the questionnaire for one particular year may have returned a questionnaire in subsequent years. Therefore, the number of patients in the complete-case analysis is lower than the number of completed questionnaires in year 5. The large number of patients not included in the complete-case analysis because of missing data strengthens the rationale for using the multiple imputation data sets in the base case.

TABLE 32 - Numbers of questionnaires returned and completed questionnaires returned and corresponding proportions per trial arm, according to ITT analysis

Completed questionnaires means that all of the questions on health-care resource use and EQ-5D were filled in.

Year	Questionnaires returned, n (%)		Completed questionnaires,a n (%)
	Surgery	Medical management	Surgery	Medical management
1	154 (87)	164 (92)	134 (75)	147 (82)
2	128 (72)	142 (79)	121 (68)	134 (75)
3	132 (74)	134 (75)	112 (63)	119 (66)
4	126 (71)	129 (72)	114 (64)	118 (66)
5	127 (71)	119 (66)	115 (65)	113 (63)
Number of patients in complete-case analysis			88 (49)	84 (47)

Health-care resource use

Table 33 summarises yearly health-care resource use in the two trial arms according to ITT analysis. During the first year of the trial, 111 patients randomised to surgery and 10 patients randomised to medical management underwent laparoscopic fundoplication. The 111 patients who were randomised to and received surgery constituted the surgery group in the PP analysis. The 169 patients who were randomised to medical management and did not undergo surgery during the first year of follow-up constituted the medical management group in the PP analysis. In the subsequent years of follow-up there were 15 patients who underwent surgery (one patient who had been randomised to surgery and 14 patients who had been randomised to medical management). These patients are included in the overnight hospital admissions category. Patients randomised to medical management reported more hospital and GP visits than the surgery patients over the 5 years of follow-up.

TABLE 33 - Health-care resource use per year per trial arm, according to ITT analysis

N/A, not applicable.

Health-care resource	Year	Reflux-related reasons, n		All reasons, n
		Surgery (n = 178)	Medical management (n = 179)	Surgery (n = 178)	Medical management (n = 179)
Laparoscopic fundoplication (first year)	1	111	10	N/A	N/A
Hospital overnight admissions (excluding surgery in the first year)	1	4	2	8	8
	2	1	2	8	10
	3	2	9	6	10
	4	2	9	2	10
	5	0	8	1	11
	Total	9	30	25	49
Hospital day admissions	1	22	24	40	53
	2	5	4	23	24
	3	4	6	4	10
	4	12	9	13	11
	5	4	11	7	14
	Total	47	54	87	112
Visits to and from GP	1	110	103	394	376
	2	34	115	269	373
	3	38	99	381	386
	4	55	126	422	469
	5	36	119	404	370
	Total	273	562	1870	1974
Number of patients on reflux-related medication	1	58	148	N/A	N/A
	2	48	124	N/A	N/A
	3	51	113	N/A	N/A
	4	51	106	N/A	N/A
	5	56	98	N/A	N/A

Table 34 shows the costs associated with health-care use according to ITT analysis for all available cases (see Appendix 9 for corresponding table for PP). All available cases uses data from all questionnaires returned at each time point. Per annum costs and costs per category refer to all available data, that is, to all participants who returned the questionnaire for that particular year or for that particular category. Therefore, the sum of the costs per category is different from the sum of the costs per annum. Similarly, total costs for complete-case analysis do not correspond to the sum of the costs per category or to the sum of the costs per annum because complete case is a subset of all available data because of the non-monotone missing data pattern. Total costs for complete-case analysis refer to the patients who returned all questionnaires and completed all EQ-5D profiles (84 surgery patients and 88 medical management patients).

TABLE 34 - Costs associated with resource use for all available cases, discounted from year 2 at 3.5%, according to ITT analysis

CIs estimated using OLS regression.

Only surgery occurring during the first year of the trial was included here. Laparoscopic fundoplication occurring in the subsequent years of the trial has been included in reflux-related hospital admissions. For the 10 medical management patients who had surgery in the first year of the trial, the average (SD) cost was £2679 (£126). For the 111 surgery patients who had surgery in the first year of the trial, the average (SD) cost was £2798 (£501).

Returned questionnaires in each year			Mean (SD) resource-use cost (£)		Incremental mean cost (cost surgery – cost medical management) (95% CIa) (£)
Surgery	Medical management	Year	Surgery	Medical management
154	164	1	2500.75 (1697.99)	559.62 (1006.81)	1941.13 (1621.43 to 2260.83)
128	142	2	94.15 (317.63)	150.96 (356.57)	−56.81 (−138.08 to 24.46)
132	134	3	94.35 (340.33)	276.41 (894.16)	−182.05 (−345.87 to −18.24)
126	129	4	111.41 (394.00)	303.50 (1337.26)	−192.09 (−436.56 to 52.28)
127	119	5	58.38 (178.58)	234.03 (629.33)	−175.65 (−290.26 to −61.03)
Cost category
Surgery in year 1b			1734.05 (1407.58)	164.31 (644.63)	1569.74 (1342.05 to 1797.42)
Reflux-related hospital night admissions			343.82 (1176.05)	302.34 (818.41)	41.47 (−247.48 to 330.42)
Reflux-related hospital day admissions			221.67 (633.61)	250.35 (631.37)	−28.68 (−209.24 to 151.87)
Reflux-related GP visits			127.18 (178.96)	200.13 (462.53)	−72.95 (−173.26 to 27.35)
Medication			121.34 (265.05)	365.70 (517.05)	−244.35 (−361.82 to −126.89)

Patients randomised to medical management accumulate lower costs than patients randomised to surgery. Table 34 indicates that surgery patients accrued a large proportion of the total costs in the first year, and accumulated lower costs during the remaining 4-year follow-up than the medical management group. In contrast, the costs accrued by medical management patients are evenly distributed across the duration of the trial. These results suggest that the cost trend in medical management patients is steeper than in surgery patients; hence, that cumulative costs in medical management patients tend to increase at a greater rate than in surgery patients. Costs associated with surgery were the major cost driver for the surgery group. Costs associated with reflux-related medication were significantly greater for the medical management group than for the surgery group. Costs associated with admissions to hospital and GP visits were not statistically significantly different between the two groups. Surgery during years 2–5 is accounted for in the overnight hospital admissions. There were a few crossovers from medical management to surgery from year 2; hence, the difference in costs associated with overnight hospital admissions between the two treatment groups is small. These results suggest that patients undergoing surgery in subsequent years are not a major cost driver in determining the cost-effectiveness of surgery.

Quality-adjusted life-years

Table 35 summarises the EQ-5D scores reported at each follow-up point for all available cases (see Appendix 9 for the corresponding table for PP). All available cases uses data from all questionnaires returned at each time point. The surgery group appears to have better HRQoL than the medical management group, despite starting from a lower baseline EQ-5D on average (0.7201 in the medical management group and 0.7107 in the surgery group). The difference in HRQoL between the two treatment groups decreased with time. This may be due to patients randomised to medical management undergoing surgery throughout the follow-up period and/or to diminishing treatment effect over time.

TABLE 35 - Health-related quality of life (EQ-5D) for all available cases according to ITT analysis

n refers to the number of patients originally randomised to each trial arm.

CIs estimated using OLS regression.

Unadjusted for baseline EQ-5D.

Completed questionnaires returned at each time point		Follow-up	Mean (SD) EQ-5D		Difference in EQ-5D (surgery – medical management) (95% CI)b,c
Surgery (n = 178a)	Medical management (n = 179a)		Surgery	Medical management
171	173	Baseline	0.7107 (0.2581)	0.7201 (0.2545)	−0.0094 (−0.0638 to 0.0445)
149	153	3 months	0.7881 (0.2328)	0.6894 (0.3012)	0.0987 (0.0376 to 0.1597)
152	164	Year 1	0.7537 (0.2468)	0.7097 (0.2715)	0.0440 (−0.0136 to 0.1016)
122	138	Year 2	0.7619 (0.2718)	0.7172 (0.3127)	0.0447 (−0.0273 to 0.1167)
129	132	Year 3	0.8034 (0.2312)	0.7474 (0.2621)	0.0560 (−0.0043 to 0.1163)
125	127	Year 4	0.7713 (0.2438)	0.7544 (0.2719)	0.0169 (−0.0472 to 0.0810)
124	117	Year 5	0.7743 (0.2590)	0.7612 (0.2815)	0.0131 (−0.0555 to 0.0817)

Comparison of costs and quality-adjusted life-years between multiple imputation and complete case

Table 36 shows the comparison of the total costs per year between the complete-case data set and the multiple imputation results. Complete case includes only those participants who returned all questionnaires and fully completed the EQ-5D questionnaires. The similarity of both the means and the CIs provides some reassurance of the validity of the multiple imputation model. The distribution of costs and EQ-5D scores in the imputed data sets matches reasonably well the distribution of the original data (see Appendix 8 for details). Furthermore, the Monte Carlo errors are < 15% of the coefficient and CI estimates, suggesting that 100 imputations are sufficient to ensure reproducibility and statistical efficiency. 69

TABLE 36 - Comparison between the complete-case and multiple imputation data sets for costs and HRQoL, according to ITT allocation

N/A, not applicable (due to costs being calculated using the area under the curve method).

95% CI estimated with OLS regression and the ‘mim’ command.

Monte Carlo error – SD across repeated runs of the same imputation procedure with the same data. 69

Total QALYs adjusted for baseline EQ-5D.

Year	Incremental mean cost (cost surgery - cost medical management) (95% CIa) (£)		Difference in mean EQ-5D (surgery - medical management) (95% CIa)
	Complete case	Multiple imputation	Follow-up	Complete case	Multiple imputation
N/A	N/A	N/A	Baseline	−0.0388 (−0.1083 to 0.0308)	−0.0091 (−0.0615 to 0.0433)
N/A	N/A	N/A	3 months	0.0848 (0.0122 to 0.1573)	0.0825 (0.0232 to 0.0142)
1	2197.14 (1779.67 to 2614.61)	1958.46 (1617.31 to 2299.62)	Year 1	0.0519 (−0.0198 to 0.1237)	0.0407 (−0.0150 to 0.0963)
2	−139.14 (−237.02 to −41.26)	−44.58 (−129.68 to 40.52)	Year 2	0.0467 (−0.0356 to 0.1289)	0.0445 (−0.0218 to 0.1108)
3	−193.44 (−361.94 to −24.93)	−127.42 (−306.36 to 51.52)	Year 3	0.0508 (−0.0195 to 0.1211)	0.0454 (−0.0150 to 0.1057)
4	−37.66 (−173.02 to 97.70)	−144.61 (−374.78 to 85.56)	Year 4	0.0324 (−0.0395 to 0.1044)	0.0260 (−0.0356 to 0.0875)
5	−165.12 (−304.74 to −25.50)	−123.90 (−236.56 to −11.24)	Year 5	−0.0095 (−0.0871 to 0.0680)	0.0294 (−0.0358 to 0.0945)
Total cost	1661.78 (1130.00 to 2193.55)	1517.95 (1006.49 to 2029.41)	Total QALYs (unadjusted)	0.1976 (−0.0857 to 0.4810)	0.1948 (−0.0356 to 0.4251)
Monte Carlo errorb		10.16	Monte Carlo errorb		0.0034
			Total QALYs (adjusted)c	0.3039 (0.0928 to 0.5150)	0.2160 (0.0205 to 0.4115)
			Monte Carlo errorb		0.0035

For both the complete-case and multiple imputation data sets, the participants randomised to laparoscopic fundoplication accrued greater costs but also reported greater HRQoL than participants randomised to continued medical management. The 95% CI for mean incremental QALYs crosses zero for the unadjusted for baseline estimates, whereas it remains above zero for the adjusted values. This result reflects the baseline imbalance in mean utility between treatment groups. Therefore, these results strongly indicated that surgery is associated with a greater QALY improvement than medical management. The sum of the differences in EQ-5D for the ITT groups does not correspond to the incremental mean QALYs because of the effect of discounting.

Cost-effectiveness

The results of the incremental analysis suggest that laparoscopic fundoplication is a cost-effective strategy for GORD patients eligible for the REFLUX trial (Table 37). The results for the complete-case analysis concur with those for the multiple imputation data set; across adjusted and unadjusted ICER for baseline EQ-5D, ICERs range between £5468 and £8410, well below conventional cost-effectiveness thresholds of £20,000 and £30,000 per additional QALY. For both data sets (complete case and multiple imputation), the probability of surgery being the more cost-effective intervention is > 0.82 for incremental analysis unadjusted for baseline EQ-5D and > 0.93 once incremental QALYs are adjusted for baseline EQ-5D. In the ITT analysis the ICER is higher for the multiple imputed data than for the complete case if QALYs are adjusted for baseline EQ-5D, but lower if QALYs are unadjusted. This might reflect the effect of having baseline EQ-5D in the prediction model, which would preclude the need for adjustment.

TABLE 37 - Incremental analysis for the ITT analysis at 5 years of follow-up for the complete-case and multiple imputation data sets

Probability of intervention being cost-effective calculated with semiparametric bootstrapping.

Data set	Adjustment for baseline EQ-5D?	Incremental mean costs (£) (95% CI)	Incremental mean QALYs (95% CI)	ICER (£)	Probability cost-effective at £20,000 per QALYa	Probability cost-effective at £30,000 per QALYa
Complete case	No – unadjusted QALYs	1661.78 (1130.00 to 2193.55)	0.1976 (−0.0857 to 0.4810)	8409.82	0.828	0.866
	Yes – adjusted QALYs	1661.78 (1130.00 to 2193.55)	0.3039 (0.0928 to 0.5150)	5468.36	0.989	0.996
Multiple Imputation	No – unadjusted QALYs	1517.95 (1006.49 to 2029.41)	0.1948 (−0.0356 to 0.4251)	7792.35	0.861	0.906
	Yes – adjusted QALYs	1517.95 (1006.49 to 2029.41)	0.2160 (0.0205 to 0.4115)	7027.55	0.932	0.962

Figure 20 shows how the NMB associated with laparoscopic fundoplication increases with the duration of the trial. This reflects the increase in costs associated with the medical group, which offsets the initial investment made in laparoscopic fundoplication in the surgery group.

FIGURE 20.

Net monetary benefit (incremental QALYs × £20,000 per QALY – incremental costs) over the duration of the REFLUX trial for the multiple imputation and complete-case data sets (QALYs adjusted by baseline EQ-5D).

Structural sensitivity analysis: per-protocol status for the complete case

Structural sensitivity analysis consisted of PP status at 1 year for the complete case. In the PP analysis patients are classified according to the treatment actually received at 1 year of follow-up. The PP group consists of 111 patients who were randomised to surgery and who actually had surgery during the first year of the trial and 169 patients who were randomised to medical management and who did not undergo surgery during this time period. However, complete-case data exist only for 84 medical management patients and 66 laparoscopic fundoplication patients. Appendix 9 presents detailed results for costs and HRQoL according to PP analysis. As expected, patients who actually had surgery have higher costs than patients who did not undergo surgery, regardless of their randomisation. Table 38 summarises the incremental results of the PP analysis. Similar to the ITT analysis, the surgical policy is likely to be cost-effective at conventional (NICE) thresholds for cost-effectiveness. The incremental costs are higher and the incremental QALYs lower for the PP analysis (for surgery compared with medical management) than for the ITT analysis if no adjustment is made for baseline imbalances in EQ-5D. Therefore, the ICER is also greater (surgery is less cost-effective than suggested by the ITT analysis). Once total QALYs are adjusted for baseline EQ-5D, however, the incremental mean QALYs increase substantially and the ICER is reduced. Nevertheless, the adjusted ICER in the ITT analysis is lower than that in the PP analysis by around £2000.

TABLE 38 - Incremental analysis for the PP analysis at 5 years of follow-up for the complete-case data set

Probability of intervention being cost-effective was calculated with semiparametric bootstrapping.

Adjustment for baseline EQ-5D	Incremental mean costs (95% CI) (£)	Incremental mean QALYs (95% CI)	ICER (£)	Probability cost-effective at £20,000 per QALYa	Probability cost-effective at £30,000 per QALYa
Unadjusted QALYs	2323.77 (1799.90 to 2847.65)	0.1782 (−0.1316 to 0.4879)	13,043.90	0.672	0.747
Adjusted QALYs	2323.77 (1799.90 to 2847.65)	0.3200 (0.0837 to 0.5562)	7262.85	0.957	0.983

Scenario sensitivity analysis: all general practitioner and all hospital costs for complete case

The results of the scenario analyses strengthen the case for the surgical policy (Table 39). For scenario 1, replacing reflux-related GP costs by all GP costs, the ICER increased slightly in relation to the base case. Nevertheless, the ICER remains well below conventional thresholds and the probability of surgery being cost-effective is > 0.83, for both adjusted and unadjusted analyses. In scenario 2, replacing reflux-related hospital costs by all hospital costs, medical management was ‘dominated’ by the surgical policy because of this intervention being associated with greater benefits in terms of QALYs and lower costs. For this scenario the probability of surgery being cost-effective was > 0.93.

TABLE 39 - Incremental analysis for the scenario sensitivity analysis at 5 years of follow-up for the complete-case data set

Probability of intervention being cost-effective calculated with semiparametric bootstrapping.

Sensitivity analysis	Adjustment for baseline EQ-5D?	Incremental mean costs (95% CI) (£)	Incremental mean QALYs (95% CI)	ICER (£)	Probability cost-effective at £20,000 per QALYa	Probability cost-effective at £30,000 per QALYa
Scenario 1: all GP costs	No – unadjusted QALYs	1685.60 (1103.97 to 2267.23)	0.2125 (−0.0748 to 0.4998)	7932.23	0.826	0.863
	Yes – adjusted QALYs	1685.60 (1103.97 to 2267.23)	0.3191 (0.1061 to 0.5321)	5282.36	0.987	0.994
Scenario 2: all hospital costs	No – unadjusted QALYs	−262.72 (−860.08 to 334.65)	0.2125 (−0.0748 to 0.4998)	Medical management dominated	0.930	0.928
	Yes – adjusted QALYs	−£262.72 (−860.08 to 334.65)	0.3191 (0.1061 to 0.5321)	Medical management dominated	0.999	0.999

Sensitivity analysis for the multiple imputation model: departure from missing at random assumption

The multiple imputation procedure assumes that the individuals who completed and returned all questionnaires are similar to the individuals who did not, conditional on their observed characteristics (MAR assumption). 69,79 However, this may not be the case: patients who did not return a questionnaire may have experienced worse HRQoL and accrued higher health service costs, or vice versa. Sensitivity analysis on the multiple imputation model tested how sensitive the cost-effectiveness results are to the MAR assumption. Figure 21 represents the change in NMB adjusted for baseline EQ-5D as costs and QALYs are varied in patients with missing data. The origin, marked as ‘base case’, refers to the incremental results from the multiple imputed data set (ICER = £7028 per additional QALY). The right quadrant plots NMB after increasing the total costs in steps of 10% for patients for whom there was missing data, for both treatment groups and for surgery-allocated patients. The left quadrant plots NMB after decreasing total QALYs in similar fashion. Positive values for NMB indicate that surgery is cost-effective; negative values indicate that surgery is not cost-effective for a threshold of £20,000 per additional QALY.

FIGURE 21.

Net monetary benefit (incremental QALYs adjusted for baseline EQ-5D × £20,000 per QALY – incremental costs) over variation in total costs and total QALYs in the multiple imputed data set.

The cost-effectiveness of surgery is relatively insensitive to any increase in costs; the NMB changes little if costs are increased for patients with missing data in both treatment groups and if costs are increased just for surgery-allocated patients with missing data. A similar result is observed for the reduction in total QALYs for all patients with missing data. In contrast, the cost-effectiveness of surgery is highly sensitive if it is assumed that surgery-allocated patients with missing data experience lower HRQoL than patients with complete data. A 10% decrease in QALYs for patients randomised to surgery with missing data results in NMB decreasing to negative values. This scenario shows that missing data can have an impact on the results under certain conditions. It is impossible to empirically confirm or refute the scenario from the data in the trial, but it could be considered an extreme case. It seems improbable in practice that surgical patients with poor quality of life are less likely to respond to follow-up questionnaires than similar participants undergoing medical management.

Validation of within-trial (5-year) analysis and exploration of the need for a long-term model

Results

Costs

Figure 22 shows the difference in costs (excluding initial surgery) in years 1–5. Mean costs are greater in the medical management arm of the trial after the first year and the CIs only just cross zero. These results are very similar to those of Table 34.

FIGURE 22.

Difference in costs (£) excluding initial surgery (mean, 95% CI).

European Quality of Life-5 Dimensions

Figure 23 shows the difference in EQ-5D at 3 months and in years 1–5. Mean HRQoL tends to be greater in the surgical group during the trial, although the CIs cross zero in some periods. These results are very similar to those of Table 35.

FIGURE 23.

Difference in EQ-5D, adjusted for baseline (mean, 95% CI).

Discussion

The results of the cost-effectiveness analysis strongly suggest that a policy of offering laparoscopic fundoplication to people with GORD who require long-term PPI treatment for symptom control is more cost-effective than continuing to manage them with PPIs (with selective use of surgery if symptoms are poorly controlled), assuming that the cost-effectiveness thresholds used by NICE (£20,000–30,000 per QALY) are appropriate for the NHS. Surgery represents a greater initial investment and lower medium-term costs, whereas costs associated with medical management remain relatively constant or slightly increase over time. The difference in HRQoL achieved with surgery is sustained over 5 years, although the results indicate that mean EQ-5D scores for surgery and medical management tended to converge (as discussed in Chapter 3, in part this reflects later surgery in patients randomised to medical management). The ICER favours surgery when incremental QALYs are both adjusted and unadjusted for baseline EQ-5D. Nevertheless, adjusted incremental QALYs are likely to be a more reliable estimate of treatment effect as they account for differences in baseline utility. Patients randomised to medical management reported higher baseline utility than patients randomised to surgery. Failure to adjust for these baseline differences could result in a biased ICER, as discussed elsewhere.62 The results from the multiple imputed data set are likely to be more accurate than the results from the complete-case analysis because of the large number of patients with incomplete data (> 50%). Therefore, multiple imputation was chosen for the base case. Nevertheless, the results are similar across the data sets and laparoscopic surgery is the more cost-effective intervention for both.

There is little uncertainty regarding the cost-effectiveness results once adjustment for EQ-5D at baseline is performed. The probability of surgery being cost-effective ranged between 0.932 and 0.999 for the base case and across the scenarios tested. Furthermore, it is clear from the results of the scenario analysis that the base-case results are robust to alternative costing assumptions. The PP analysis is used to test whether or not the ITT analysis is potentially misleading because of the dilution of treatment effect (some patients randomised to surgery did not have surgery and some patients randomised to medical management actually had surgery). The PP analysis has the advantage of mimicking clinical practice and could be thought to be more relevant to decision-makers. However, the PP analysis is not without its limitations. First, and as with any PP analysis, it is sensitive to selection bias because of breaking randomisation. Second, the PP analysis may still underestimate the effect of surgery because patients having surgery between 2 and 5 years are counted as medical management. Third, the PP analysis is actually a subset of the ITT groups, which further reduces the data set. For these reasons, the ITT results are likely to be more reliable. It is important to characterise any uncertainty in the analysis as failure to do so can result in inaccurate estimates of cost-effectiveness, particularly when costs and benefits are highly skewed. 82 In addition, any analyses of uncertainty can help to illustrate where caution should be exercised when interpreting the results of a cost-effectiveness analysis. The results of the sensitivity analyses suggest that the uncertainty is likely to be driven by HRQoL. If QALYs for randomised surgery patients with missing data are reduced, surgery may no longer be cost-effective.

For the within-trial analysis no assumptions are needed about the longer-term effectiveness and costs associated with surgery and medical management. However, the within-trial analysis has some disadvantages. First, it does not account for any differences in costs and QALYs that may be expected over the longer term (> 5 years), which could be due to differences in recurrence/relapse, medication use, NHS service utilisation or HRQoL. Second, it uses data only from the REFLUX trial and does not consider other sources of evidence. Third, only a limited range of sensitivity analyses was possible. Finally, the large proportion of missing observations required an assumption regarding the mechanism of missing data, which may have some impact on the cost-effectiveness estimates. The exploration of the need for a longer-term model aimed to address the first limitation of the within-trial analysis. A mixed model was used to examine the trend in the difference in costs and the difference in QALYs between treatment groups over time.

No evidence was found to suggest that the cost-effectiveness of surgery diminishes over a longer follow-up time. Both multiple imputation and mixed models are commonly used methods to handle missing data. Multiple imputation was used in the previous section because, by imputing missing data, it naturally allows the estimation of the total cost and total QALYs for each patient in the trial. Furthermore, it can handle correlation between several outcomes (in this case costs and QALYs) as well as correlation between outcomes over time. Mixed models do not explicitly impute missing data but adjust the estimates of the differences between treatment groups at each discrete follow-up time to take account of the missing data. The approach therefore offers an alternative method to multiple imputation to examine trends in the difference in costs and the difference in QALYs between treatment groups over time. Because the analyses using multiple imputation and mixed models agree, we can have more confidence that the results are valid and that surgery is the most cost-effective intervention.

A number of other studies have quantified the cost-effectiveness of laparoscopic fundoplication and medical management. Not all of these, however, use a common metric (such as QALYs) to measure benefits. Of those studies quantifying the benefits associated with the two treatments using QALYs, the ICER for surgery ranged from £180 to £49,000. There are a number of key differences between the methodologies used in the studies, which limit the extent to which comparisons can be made between results. Importantly, not all of the studies are based on within-trial analysis; in fact, only two are: those by Grant et al. 1 and Goeree et al. 46 The remainder use modelling techniques to either extrapolate short-term trial results over the longer term or pool available evidence to generate estimates of costs and outcomes. Comparing the results from Grant et al. 1 with those from Goeree et al. 46 we can see that there are quite significant differences in the estimates of cost per QALY, from £19,000 in Grant et al. 1 to £49,000 in Goeree et al. 46 This difference is primarily driven by the difference in QALYs. In Goeree et al. 46 the EQ-5D score is actually lower in the surgical group than in the medical management group (this is unadjusted for baseline imbalances) whereas the HUI3 score is higher for the surgical group than the medical management group. The reason for the difference between the EQ-5D and the HUI3 scores is not discussed in the paper. The cost differences in the two studies were similar. Comparing the results from Goeree et al. 46 with those from the updated trial analysis we see even starker contrast between the ICERs produced (£7028 vs £49,000). Again, this is driven by the differences in EQ-5D scores observed throughout the trial period. The EQ-5D scores in the REFLUX 5-year analysis are consistently higher in the surgery group than in the medical management group, although there is a tendency for convergence towards the end of the follow-up period. Further research is required to look at why the trials produce such different results using the EQ-5D.

Implications for health care

Extending the use of laparoscopic fundoplication to people whose GORD symptoms require long-term medication for reasonable control and who would be suitable for surgery would provide health gain that extends over a number of years. The longer-term data reported here indicate that this is highly likely to be a cost-effective use of resources. The more troublesome the symptoms, the greater the potential benefit from surgery.

Recommendations for research

The practical implications for health services of any extension of the use of laparoscopic fundoplication depend on how many patients might seek such surgery as a consequence. Most patients taking anti-reflux medication are managed in general practice. Currently, it is uncertain how many people require long-term medication for reasonable control of their GORD symptoms, how many of these would be suitable for surgery and how many would seek it; hence, it is not clear what the most efficient provision of future care might be. We therefore recommend further research to address these issues and explore the practical and resource implications of alternative policies for laparoscopic fundoplication, which include extending its use within the NHS to the sorts of patients enrolled in the REFLUX trial.

Contribution of authors

Adrian Grant (Professor, Health Services Research Trialist) was the principal grant applicant and contributed to the development of the trial protocol and the preparation of the report and had overall responsibility for the conduct of the study.

Charles Boachie (Statistician, Health Statistics) conducted the statistical analysis.

Seonaidh Cotton (Trial Co-ordinator, Health Services Research Trialist) was responsible for the day-to-day management of the trial, monitored data collection and assisted in the preparation of the report.

Rita Faria (Research Fellow, Health Economics) was involved in the cost-effectiveness section, namely conducting the economic evaluation and writing the final report.

Laura Bojke (Senior Research Fellow, Health Economics) was responsible for the cost-effectiveness section – supervising the economic evaluation and the systematic review of existing economic evidence and writing the final report.

David Epstein (Honorary Research Fellow, Health Economics) was involved in the cost-effectiveness section – conducting and supervising the economic evaluation and writing the final report.

Craig Ramsay (Professor, Health Services Research Statistician/Trialist) contributed to the grant application and the trial design and conducted the statistical analysis.

Belen Corbacho (Research Fellow, Health Economics) was responsible for the systematic review of existing health economic evidence – study selection, data extraction, validity assessment and writing the final report.

Mark Sculpher (Professor of Health Economics, Health Economics) was responsible for the economic evaluation section of the grant application and overseeing the economic evaluation.

Zygmunt Krukowski (Surgeon, Gastroenterology) advised on clinical aspects of the trial and commented on the draft report.

Robert C Heading (Honorary Professor) advised on clinical aspects of the trial design and the conduct of the trial and commented on the draft report.

Marion Campbell (Director, Health Services Research Statistician/Trialist) contributed to the development of the trial design and all aspects of the conduct of the trial and commented on the draft report.

Publications

1. Grant AM, Cotton SC, Boachie C, Ramsay CR, Krukowski ZH, Heading RC, et al. Minimal access surgery compared with medical management for gastro-oesophageal reflux disease: five year follow-up of a randomised controlled trial (REFLUX) BMJ 2013;346:f1908. DOI: http://dx.doi.org/10.1136/bmj.f1908 (published online on 18 April 2013).

2. Faria R, Bojke L, Epstein D, Corbacho B, Sculpher M, on behalf of the REFLUX trial group. Cost-effectiveness of laparoscopic fundoplication versus continued medical management for the treatment of gastro-oesophageal reflux disease based on long-term follow-up of the REFLUX trial. Br J Surg 2013; in press. DOI: http://dx.doi.org/10.1002/bjs.9190.

Disclaimer

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health.