4 Evidence and interpretation
The Appraisal Committee (appendix A) considered evidence from a number of sources (appendix B). Methadone and buprenorphine are licensed for use in both detoxification and maintenance therapy. The main focus of the Assessment Group's report and the manufacturer's submission was use of the technologies in maintenance therapy.
4.1.1 Thirty-one systematic reviews met the inclusion criteria of the Assessment Group. The reviews included evidence from randomised controlled trials (RCTs) and other types of study. Many of the studies were included in several of the reviews. The Assessment Group identified an additional 27 RCTs published since 2001. Most of the systematic reviews and RCTs were of moderate to good quality. Of the 27 RCTs, 16 were conducted in the USA, three in Australia, three in Iran, two in the Netherlands, two in Austria and one in Norway.
4.1.2 Most of the evidence reported is for men aged 30–49 years, in good health, who met DSM-III or -IV criteria for opioid dependence, had no serious psychiatric or medical comorbidities and had not undergone therapy for drug misuse in the months before maintenance therapy was started. Pregnant women and all people younger than 18 years were excluded from most trials.
4.1.3 Most studies were undertaken in outpatient or inpatient settings or specialised treatment centres, and very few were conducted in community settings. Various delivery options were reported, but generally delivery of methadone maintenance therapy (MMT) and buprenorphine maintenance therapy (BMT) was characterised by fixed doses of medication, supervised consumption (no take-home medication), discharge of people who missed 3 consecutive days of treatment, limited adjuvant psychosocial therapy, no rewards for treatment compliance, intensive monitoring, limited length of treatment and relatively short periods of follow up (in most cases up to 1 year).
4.1.4 Most trials used a fixed-dose design, in which all those included were given a fixed dose of methadone or buprenorphine. Methadone doses were in the range 50–150 mg/day and buprenorphine doses were in the range 1–15 mg/day. Some recent studies have used a flexible-dosing design (in which a person's dose is adjusted during treatment as necessary). The Assessment Group judged this to be a better reflection of current practice in the UK, where each person receives a flexible individualised dose of methadone or buprenorphine.
4.1.5 The two main outcomes reported in the included systematic reviews and RCTs were retention on treatment and illicit use of opioids, the latter being reported in a variety of ways (for example, proportion of people taking illicit opioids, or the mean rate of diamorphine intake assessed by self-report methods and/or urinalysis), making meta-analysis more difficult. Limited data were available for HIV-related outcomes, side effects/adverse events and mortality, and non-health outcomes (that is, crime and employment).
4.1.6 The results from the meta-analyses showed that fixed-dose MMT has superior levels of retention on treatment compared with placebo or no treatment. One meta-analysis (n = 505), which used doses of 20–50 mg/day of methadone compared with no therapy, gave a relative risk (RR) of remaining on treatment of 3.05 (95% confidence interval [CI] 1.75 to 5.35). In another systematic review (n = 348), which pooled the results from trials that used daily doses in the range 20–97 mg methadone, the RR of remaining on treatment was 3.91 (95% CI 1.17 to 13.2).
4.1.7 The results from the meta-analyses showed that fixed-dose MMT resulted in lower rates of illicit opioid use compared with placebo or no treatment. One systematic review (n = 246), which compared 60 mg methadone daily with no therapy, gave an RR of illicit opioid use (self-reported) of 0.31 (95% CI 0.23 to 0.42). Another systematic review (n = 347), comparing doses of 50 mg or more methadone with placebo, resulted in an RR of illicit opioid use of 0.82 (95% CI 0.69 to 0.98).
4.1.8 There were fewer self-reported adverse events with MMT compared with placebo or no therapy, although this difference was not statistically significant (RR 0.59, 95% CI 0.33 to 1.04). Three systematic reviews of non-randomised studies reported the effects of methadone on HIV-related outcomes. HIV risk behaviour or risk scores and seroconversion rates (development of antibodies) were in general better in the MMT groups compared with no therapy. The results showed no statistically significant differences between MMT and BMT for the self-reported outcomes of number of sex partners and frequency of unprotected sex.
4.1.9 A meta-analysis of observational studies that compared the number of deaths (per person years of exposure) in people in and out of methadone treatment reported an RR of 0.25 (95% CI 0.19 to 0.33), indicating that people who were not taking methadone or were discharged from treatment were four times more likely to die than those on treatment. The base rates (for those out of methadone treatment) in the included studies included showed a wide variation.
4.1.10 The level of criminal activity decreased in people on MMT compared with those on placebo or no therapy. One study reported a reduction in criminal activity in the MMT group that was not statistically significant (RR 0.39, 95% CI 0.12 to 1.25) and two studies reported effect sizes of 0.54 and 0.70. (Effect sizes are calculated by subtracting the mean of the control group from the mean of the treatment group and dividing by the standard deviation. Conventionally, effect sizes of 0.2 are considered 'small', 0.5 'medium', and 0.8 'large'.)
4.1.11 One systematic review of randomised studies reported retention on treatment for various doses of buprenorphine compared with placebo or no therapy. Five RCTs (n = 1131) used doses of less than 5 mg buprenorphine, resulting in an RR of 1.50 (95% CI 1.19 to 1.88). Four RCTs (n = 887) used a dose of 6–12 mg, resulting in an RR of 1.74 (95% CI 1.06 to 2.87). Four RCTs (n = 728) used a dose of 18 mg, resulting in an RR of 1.74 (95% CI 1.02 to 2.96).
4.1.12 One small RCT (n = 40), included in an unpublished systematic review, reported a reduction in mortality in people on BMT (16 mg) compared with those on placebo and counselling treatment over a 12-month period (RR 0.05; 95% CI 0 to 0.79). No studies comparing BMT with placebo or no treatment reported data on illicit opioid use (self-reported or urinary confirmed), adverse events, HIV risk behaviour or crime.
4.1.13 Four meta-analyses of RCTs showed that fixed doses of MMT had retention on treatment superior to that of comparable fixed doses of BMT. One study (n = 540) compared 50–80 mg methadone with 6–12 mg buprenorphine, giving a hazard ratio (HR) of 1.26 (95% CI 1.01 to 1.57). Another systematic review (n = 211) compared doses of up to 35 mg methadone with up to 5 mg buprenorphine, resulting in an RR of 1.47 (95% CI 1.10 to 2.00).
4.1.14 Four systematic reviews of RCTs compared self-reported illicit opioid use between people on fixed doses of MMT and people on fixed doses of BMT. A high fixed dose of MMT (50 mg or more) was more effective than a low fixed dose of BMT (less than 8 mg) with an RR of 0.29 (95% CI 0.16 to 0.53). Results were mixed for comparisons of lower fixed doses of MMT (less than 50 mg) and higher fixed doses of BMT (8 mg or more).
4.1.15 A recently updated and unpublished Cochrane systematic review of seven RCTs directly compared flexible-dosing MMT with flexible-dosing BMT in 976 illicit-opioid-dependent people. No further RCTs comparing flexible-dose MMT and BMT were identified by the Assessment Group's searches. The daily equivalent doses in these flexible-dosing trials were 20–120 mg/day for methadone and 2–16 mg/day for buprenorphine. Treatment retention was higher for flexible MMT compared with flexible BMT dosing (pooled HR 1.40, 95% CI 1.15 to 1.69) although there was no statistically significant difference in illicit opioid use for BMT compared with MMT (standardised mean difference –0.12, 95% CI –0.26 to 0.02 ).
4.1.16 In the assessment report, the rates of occurrence in four categories of serious adverse events per 100 patient years in treatment are taken from the 'National evaluation of pharmacotherapies for opioid dependence' 2004 report, which had access to individual-patient-level data. A total of 10 serious adverse events were reported among the 420 people treated with methadone, and 20 were reported among the 492 treated with buprenorphine. A pooled RCT analysis showed no significant difference in the rate of serious adverse events with MMT compared with BMT.
4.1.17 Comparison of data from population cross-sectional studies suggests that the level of mortality with BMT may be lower than that with MMT, although other authors have commented that these data were unlikely to capture all related deaths.
4.1.18 Higher doses of MMT (for example, 60 mg or more) were found to be more effective than doses of less than 50 mg for improving retention on treatment (for example, 60–109 mg compared with 1–39 mg resulted in an RR of remaining on treatment of 1.36 [95% CI 1.13 to 1.63]). Doses of MMT higher than 50 mg were more effective than doses of less than 50 mg in reducing self-reported illicit opioid use (for example, 50 mg or more compared with less than 50 mg resulted in an RR of 0.82 [95% CI 0.78 to 0.95]). Higher doses of MMT (60–109 mg) were also associated with a statistically significantly lower number of illicit-opioid-positive urine tests compared with much lower doses of MMT (1–39 mg). However, high-dose MMT (60–109 mg) produced a non-significantly lower number of illicit-opioid-positive urine tests than moderate-dose MMT (40–59 mg).
4.1.19 Both MMT and BMT appeared to be similarly effective whether delivered in primary care or in outpatient clinics. Although the evidence on treatment modifiers was limited, adjunct psychosocial and contingency interventions (for example, financial incentives for illicit-opioid-free urine samples) appeared to enhance the effects of both MMT and BMT.
4.1.20 The results from the meta-analyses showed that fixed-dose MMT has higher levels of retention on treatment and lower rates of self-reported illicit opioid use compared with placebo or no treatment. Higher fixed doses of MMT are more effective than lower fixed doses. There is evidence, primarily from non-randomised observational studies, that fixed-dose MMT reduces mortality, HIV risk behaviour and levels of crime compared with no therapy.
4.1.21 Meta-analyses show that fixed-dose BMT has higher levels of retention on treatment compared with placebo or no treatment, with higher fixed doses of BMT being more effective than lower fixed doses. One small RCT has shown that the level of mortality with fixed-dose BMT is statistically significantly less than that with placebo.
4.1.22 A number of RCT meta-analyses show that fixed doses of MMT are associated with higher rates of retention on treatment than similar fixed doses of BMT. High fixed doses of MMT are more effective than lower-fixed-dose BMT at preventing illicit opioid use, but results are mixed for lower-fixed-dose MMT and higher-fixed-dose BMT.
4.1.23 In the studies analysed, rates of retention on treatment with flexible-dose MMT are superior to those with flexible-dose BMT, although there is no statistically significant difference in illicit opioid use.
4.2.1 Eleven published economic evaluations met the Assessment Group's inclusion criteria for review.
4.2.2 Eight studies assessed the cost effectiveness of MMT, one assessed the cost effectiveness of BMT and two compared the cost effectiveness of BMT directly with that of MMT. The studies reported results using a range of outcome measures. The Assessment Group reported that direct comparison of the incremental cost-effectiveness ratios (ICERs) between the studies was not possible because of differences in the approaches to modelling, time horizons, comparators and perspectives, country of origin, sources of preference weights and effectiveness data used.
4.2.3 Although most of the included papers were considered to be of high quality, none used all of the appropriate parameters, effectiveness data, perspectives and comparators required to make their results generalisable to the NHS and personal social services (PSS).
4.2.4 No economic evaluations were submitted by the manufacturers of methadone oral solution.
4.2.5 The manufacturer of buprenorphine (Schering-Plough) submitted a cost-effectiveness analysis of BMT compared with MMT for opioid-dependent people over a 1-year time horizon. Cost effectiveness was assessed as the incremental cost per quality-adjusted life year (QALY) using a decision-tree-based model. Costs were calculated from an NHS and PSS perspective. Both simple one-way and probabilistic sensitivity analyses were undertaken.
4.2.6 The model was designed to estimate the cost effectiveness of BMT in three scenarios: BMT compared with no treatment for the 20% of opioid-dependent people seeking maintenance treatment who are unable to take methadone for 'clinical reasons' (as stated by the manufacturer); BMT compared with MMT for the remaining 80% of opioid-dependent people; and maintenance therapy (methadone and buprenorphine) compared with drug-free treatment for all opioid-dependent people.
4.2.7 The model included data on people retained on treatment at specified time points up to 6 months, and then followed those retained on treatment at 6 months for a further 6 months. It was assumed that people not retained on treatment returned to their pre-treatment habits however long they had been taking maintenance therapy. The data for retention on treatment and dosing for the initial 13 weeks were based on one RCT, which compared flexible dose regimes of BMT and MMT. Data on retention between 13 and 26 weeks and between 6 months and 1 year were based on two open-label stages from the same RCT. Health-related utility values were based on results from a published study and included an adjustment factor from another published study. Data on resource use and costs were derived from several studies. The use of healthcare resources was assumed to be the same for people treated with methadone or buprenorphine.
4.2.8 When BMT was compared with no treatment for the 20% of people who could not have MMT, BMT was shown to be more expensive and slightly more effective than no treatment (ICER £30,000 per additional QALY gained).
4.2.9 For people who could be treated with either therapy, BMT was dominated by MMT, as BMT was slightly more expensive than MMT and yielded marginally fewer QALYs. However, the difference in QALYs was very small (0.00055) and given the parameter uncertainty in the model, the difference in benefit is highly uncertain.
4.2.10 The analysis of maintenance treatment (with either drug) compared with no treatment resulted in an ICER of £12,600 per additional QALY gained. However, the Assessment Group expressed concerns about this result because of the method of analysis, which excluded buprenorphine.
4.2.11 The manufacturer noted that the better retention on treatment for methadone compared with buprenorphine in the pivotal trial did not translate into incremental improvements in the QALYs for methadone. Deterministic sensitivity analyses showed that the model was sensitive to the proportion of patients retained on buprenorphine and methadone at induction, 6 weeks, 13 weeks and 6 months. It was also sensitive to changing the health-related utility values at 12 months for buprenorphine or methadone.
4.2.12 The Assessment Group developed a decision tree with Monte Carlo simulation to assess the cost effectiveness of BMT and MMT compared with drug-free therapy, and of BMT compared with MMT. The model estimated costs and outcomes from an NHS and PSS perspective for a 12-month period for the three strategies. Maintenance therapy was assumed to be a flexible-dosing regimen, and the mean daily dose was assumed to be constant from week 13 onwards. The average cost of dispensing drugs was based on assumptions of supervised self-administration 6 days a week for the first 3 months, then unsupervised self-administration 6 days a week from 3 to 6 months, and unsupervised self-administration three times a week from 6 to 12 months. In addition to drug costs, estimates of resource use included counselling sessions, monitoring of treatment, GP visits, emergency department visits, inpatient hospital stays, outpatient mental health appointments and inpatient mental health admissions.
4.2.13 Data on retention on treatment at 2, 6, 13 and 25 weeks and 12 months were included in the model. The data for retention on treatment in the model were taken from a systematic review that identified seven trials that compared methadone and buprenorphine in flexible dosing (pooled HR of 1.40, 95% CI 1.69 to 1.15). The Assessment Group model also took into account urinalysis data, as some people still misuse drugs when in a maintenance programme. Data on the percentage of retained patients who were drug free were taken from the combined analysis of opioid-negative urine samples from two studies. It was assumed that patients not retained on treatment returned to their pre-treatment habits however long they were taking maintenance therapy, and that 89% of those not retained on treatment would be using opioids (based on data from a UK cohort study). Data from the 'National treatment outcome research study' (NTORS) were used to inform estimates of the proportion of drug-dependent people who were injecting.
4.2.14 Health outcomes were expressed as QALYs. In the absence of published data on quality of life associated with drug misuse, the Assessment Group obtained health-related utility data from a panel of members of the public. The Assessment Group assumed that people not retained on treatment returned to their pre-treatment habits irrespective of their period of MMT or BMT, and used the same estimated QALY for those not retained on treatment for MMT and BMT.
4.2.15 For the reference case, the analysis of MMT compared with no treatment resulted in an ICER of £13,700 per additional QALY gained. BMT was dominated by MMT. The analysis of BMT compared with no treatment resulted in an ICER of £26,400 per additional QALY gained.
4.2.16 An additional non-reference case analysis was also conducted which included costs to the criminal justice system and to victims of crime. Costs to victims of crime included the costs of increased security measures and the direct costs of material or physical damage. Results for the non-reference case were that all strategies were dominated by MMT, and that BMT was dominant over no treatment.
4.2.17 A number of sensitivity analyses were conducted on the reference and non-reference cases. With regard to administration of buprenorphine, a sensitivity analysis was conducted assuming that from week 1 to 13 buprenorphine was delivered under supervision on alternate days, and that from week 14 to 52 it was delivered unsupervised on alternate days. BMT was still dominated by MMT. However, the ICER for BMT compared with no treatment was reduced to £24,000 per QALY gained.
4.2.18 Two sensitivity analyses were also carried out on the utility values. The first of these considered the published utility values that had also been used in the manufacturer's analysis. However, instead of using a health-related utility value for a specific point of time, the overall QALY value for both strategies (while on treatment) was used. For the 'no treatment' and 'drop-out from treatment' health states, the Assessment Group assumed a utility value of 0.505. A further analysis was done using the utility values from a large published study that compared MMT with methadone plus diamorphine. Using the utility values from the manufacturer's submission, the analysis resulted in BMT no longer being dominated by MMT, but the ICER was £108,300 per QALY gained, because of the very small positive difference in QALYs. Using the utility values from the large published study, the ICER for MMT versus no treatment was £16,400 per QALY gained, and BMT was still dominated by MMT. Comparing BMT with no treatment, the values used in the manufacturer's submission resulted in an ICER of £27,500 per QALY gained. Using the utility values from the large published study, the ICER for BMT compared with no treatment was £31,600 per QALY gained.
4.2.19 The final sensitivity analysis examined the impact of excluding the costs to the victims of crime, to produce an evaluation from a societal perspective with costs to the criminal justice system only. In this analysis, MMT was no longer dominant over no treatment, and instead had an ICER of £25,000 per QALY gained. BMT was still dominated by MMT. Comparing BMT with no treatment, BMT was no longer dominant and had an ICER of £37,800 per QALY gained.
4.3.1 The Committee reviewed the data available on the clinical and cost effectiveness of methadone and buprenorphine, having considered evidence on the nature of the condition and the value placed on the benefits of methadone and buprenorphine by people with opioid dependence, those who represent them, and clinical experts. It was also mindful of the need to take account of the effective use of NHS resources.
4.3.2 The Committee considered the evidence on the clinical effectiveness of MMT and BMT for maintenance therapy in the management of opioid dependence. The Committee acknowledged that the clinical trials showed that people on methadone or buprenorphine were retained longer in treatment compared with those on placebo. The Committee also acknowledged that the observational and trial data showed that people on methadone or buprenorphine were less likely to die than those on placebo or no therapy. For people on methadone, there was also a reduction in the use of illicit opioids compared with those on placebo. For the comparison of methadone with buprenorphine, the Committee noted that the trials showed that people on methadone were retained longer in treatment compared with those on buprenorphine. For illicit opioid use while in treatment, there were no statistically significant differences between the two drugs. The Committee noted that there was uncertainty around the risk of mortality in the published research, and heard from the experts about the potential increased risk of death for people using methadone compared with buprenorphine, and the potential increased risk of death for other people when diversion (where the medication is forwarded on to others for non-prescription uses) of methadone occurs. The Committee considered the importance of supervision of both methadone and buprenorphine and noted that the Assessment Group's model assumed supervised administration of the drugs for 6 days a week for the first 3 months, which is in line with the Department of Health guidelines.
4.3.3 The clinical experts raised concerns about the generalisability of the RCTs, none of which were conducted in the UK. The Committee heard from the experts that there were a number of differences between the trials and current NHS practice (such as the dose used, and the levels of supervised delivery and psychosocial intervention). The Committee also noted that access to psychosocial care is limited and variable around the UK. The Committee also heard from the patient experts that the cost of illicit street drugs in the countries where the trials were conducted differed from the cost in the UK, and that this could affect the degree of retention in maintenance programmes. The Committee heard from the experts that despite the differences between the trials and current NHS practice, the outcomes of the trials could be generalised to opioid-dependent people in England and Wales. The Committee additionally acknowledged that in England and Wales flexible dosing is most commonly used and that programmes of supportive care are generally available.
4.3.4 The Committee considered the cost-effectiveness evidence for the comparisons of flexible doses of methadone and buprenorphine versus no treatment, and acknowledged the inclusion of costs for supervised delivery on a daily basis for each of the drugs for a minimum of 3 months, and the ongoing costs of supportive care, including psychosocial care delivered alongside these drugs. The Committee concluded that, on the basis of the evidence, both methadone and buprenorphine in flexible dosing regimens are clinically effective and cost effective, compared with no treatment, for maintenance therapy in the management of opioid dependence.
4.3.5 The Committee heard from the experts that it was not always clear which drug (methadone or buprenorphine) should be prescribed in individual cases. In some circumstances there can be clinical reasons for prescribing either methadone or buprenorphine, taking into account the person's history of opioid dependence. For people who are less opioid dependent and are planning on becoming abstinent, buprenorphine may provide greater flexibility and enable earlier detoxification. The Committee also heard that some people may have a preference for one drug over the other, which will affect their compliance with and retention in treatment. The Committee considered carefully the issue of mortality from overdose, particularly when methadone treatment is started. The Committee was also aware of the risks of diversion of these drugs to non-drug-users, especially children, in particular the high mortality risk associated with methadone in opioid-naïve people. However, the Committee considered that the current guidance, while taking account of the adverse effects of therapy in people prescribed the drugs, could not deal individually with all the issues associated with diversion. The Committee was persuaded of the importance of having both drug treatment options available, and that the decision on which was the most appropriate treatment for an individual should be made on a case by case basis. The Committee concluded that the responsible clinician, in consultation with the person, should estimate the risks and benefits of prescribing methadone or buprenorphine, taking account of the person's lifestyle and family situation (for example, whether they are considered chaotic and might put children and other opioid-naïve individuals living with them at risk).
4.3.6 The Committee was aware of the importance of supervised therapy in avoiding the risks associated with adverse effects, in particular those associated with diversion of treatment. The Committee noted that the current Department of Health guideline on supervision explicitly states that, after an initial 3-month supervision period, the level of supervision should only be relaxed when the patient's compliance is assured.
4.3.7 The Committee considered the cost-effectiveness evidence for the comparison of methadone and buprenorphine. Although methadone dominates buprenorphine for all the scenarios because it is cheaper and yields marginally more QALYs (0.067), the Committee acknowledged that in certain circumstances a person is not able to take methadone and therefore the appropriate comparator for the alternative treatment in these cases would be no treatment. The ICER in the reference case for buprenorphine versus no treatment is £26,400 per additional QALY gained.
4.3.8 Taking all these factors into account, the Committee concluded that the decision about which drug to use should be made on a case by case basis and should consider a number of clinical and patient factors, including the person's history of opioid dependence, their commitment to a particular long-term management strategy and an estimate of the risks and benefits made by the responsible clinician in consultation with the person. However, the Committee was mindful that methadone is cheaper than buprenorphine and therefore concluded that, if both drugs are equally suitable for a person, methadone should be prescribed as first choice.
4.3.9 The Committee also noted the importance of supportive care used alongside these drugs, and concluded that the delivery of both methadone and buprenorphine should be part of a programme of supportive care to ensure maximum benefit. The Committee also considered that this package of care should ideally include psychosocial care, but that methadone and buprenorphine should be provided even when psychosocial care is not available.