4 Consideration of the evidence

The Appraisal Committee reviewed the data available on the clinical and cost effectiveness of dexamethasone intravitreal implant, having considered evidence on the nature of diabetic macular oedema (DMO) and the value placed on the benefits of dexamethasone intravitreal implant by people with the condition, those who represent them, and clinical experts. It also took into account the effective use of NHS resources.

Clinical need and practice

4.1 The Committee heard from patient experts about the nature of DMO and their experience with treatment. It heard that the loss of vision has a significant impact on a person's independence; for example, it can affect their ability to drive and perform everyday activities such as dressing and making a cup of tea. The patient experts commented that the condition can disrupt employment because of attendance at regular follow‑up or monitoring appointments and, furthermore, some people may be unable to work or care for family members. The patient experts acknowledged that although an injection into the eye is unpleasant it is not painful, and they are willing to have injections to maintain their sight. They emphasized that the effect of dexamethasone intravitreal implant lasted much longer than anti‑vascular endothelial growth factor (VEGF) treatments. The Committee heard from clinical experts that although anti‑VEGF treatments have made a big difference to clinical practice the need for repeat treatment visits is problematic. The Committee noted that dexamethasone intravitreal implant is licensed for use every 6 months in line with the MEAD trials (see section 3.2) but heard that it is often given more frequently than this in practice (every 4 months). The Committee heard from the clinical and patient experts that there is a clinical need for alternative treatments for people with DMO that is unresponsive to non‑corticosteroid treatment. The Committee concluded that patients and clinicians considered dexamethasone intravitreal implant to be a valuable option that could offer longer‑term benefits than anti‑VEGF treatments to some people with DMO, and for people with DMO that is unresponsive to non‑corticosteroid treatment.

4.2 The Committee considered which people with DMO would potentially be eligible for treatment with dexamethasone intravitreal implant in clinical practice. The Committee recalled that the indication in the marketing authorisation for dexamethasone intravitreal implant was for the treatment of adult patients with visual impairment due to DMO who are pseudophakic or whose DMO is considered insufficiently responsive to, or unsuitable for, non‑corticosteroid therapy. It noted that this was narrower than the general population with DMO which was enrolled in the clinical trials and specified in the final NICE scope, which was finalised before the marketing authorisation was granted. The Committee was mindful that it could only make recommendations within the marketing authorisation (see section 2.1). After considering comparators within the context of the marketing authorisation (see sections 4.4 and 4.5), it concluded that the 4 potentially eligible populations were:

  • people with a pseudophakic lens with central retinal thickness (CRT) of 400 micrometres or more

  • people with a pseudophakic lens with CRT less than 400 micrometres

  • people who do not have a pseudophakic lens and with DMO that does not respond to non‑corticosteroid treatment, or for whom such treatment is unsuitable

  • people with a pseudophakic lens and with DMO that does not respond to non‑corticosteroid treatment, or for whom such treatment is unsuitable.

4.3 The Committee considered the clinical pathway for people with DMO in relation to the marketing authorisation for dexamethasone intravitreal implant. It heard from the clinical experts that treatment options vary according to CRT, whether their disease is unresponsive to non‑corticosteroid treatment and whether such treatment is unsuitable for them, and whether a person has a pseudophakic lens.

4.4 The Committee discussed the clinical pathway for people based on CRT levels. For people with a CRT of 400 micrometres or more, the clinical experts stated that ranibizumab is given as recommended in NICE technology appraisal guidance on ranibizumab for treating diabetic macular oedema. For people who have a CRT of less than 400 micrometres, the Committee noted that ranibizumab is not recommended in NICE technology appraisal guidance on ranibizumab for treating diabetic macular oedema and it heard from clinical experts that the prescribing options in clinical practice are laser photocoagulation and bevacizumab (outside its marketing authorisation). The Committee heard that bevacizumab was used in some centres and therefore took it into account in its decision‑making.

4.5 The Committee discussed the clinical pathway for people with DMO that is insufficiently responsive to non‑corticosteroid therapy (such as ranibizumab, bevacizumab and laser photocoagulation). The Committee noted that for people who have a pseudophakic lens and chronic disease, fluocinolone acetonide intravitreal implant is recommended in NICE technology appraisal guidance on fluocinolone acetonide intravitreal implant for treating chronic diabetic macular oedema after an inadequate response to prior therapy. It heard from the clinical experts that they perceived that fluocinolone acetonide intravitreal implant had been used with caution in the UK because some of its side effects can be difficult to reverse, given its long duration of action. The clinical experts noted that it is given every 36 months, meaning that if intraocular pressure becomes elevated (a class effect of intravitreal corticosteroid treatment), it can be difficult to lower and surgical intervention may be needed. The Committee heard from the clinical experts that people with DMO that is insufficiently responsive to non‑corticosteroid therapy who do not have a pseudophakic lens are monitored but do not have active treatment (watch‑and‑wait).

4.6 The Committee heard from the clinical experts that there are no clinical criteria for determining whether a treatment is unsuitable for a person with DMO. However, they suggested that treatment with an anti‑VEGF agent (for example, ranibizumab) is likely to be unsuitable for people who cannot attend monthly appointments, people who have had a recent cardiovascular event or stroke, and people who have a phobia of needles.

4.7 The Committee considered the most relevant comparators based on the final NICE scope and what it had heard from the clinical experts. The Committee concluded that, based on current practice, the relevant comparators for dexamethasone intravitreal implant in the 4 subpopulations of people with DMO (see section 4.2) are as follows:

  • ranibizumab in people with a pseudophakic lens with a CRT of 400 micrometres or more

  • laser photocoagulation or bevacizumab in people with a pseudophakic lens with a CRT less than 400 micrometres

  • fluocinolone acetonide intravitreal implant in people with a pseudophakic lens and whose chronic DMO does not respond to non‑corticosteroid treatment

  • watch‑and‑wait for people who do not have a pseudophakic lens and whose DMO does not respond to non‑corticosteroid treatment, or for whom such treatment is unsuitable.

Clinical effectiveness

4.8 The Committee considered the evidence presented by the company on the clinical effectiveness of dexamethasone intravitreal implant. It noted that the main sources of evidence were the MEAD randomised controlled trials, which compared dexamethasone intravitreal implant with sham procedure in a general population of people with DMO (that is, a broader population than that covered by the marketing authorisation). It noted that the company had provided subgroup analyses for people with a pseudophakic lens, but not for people whose condition did not respond to non‑corticosteroid treatment, or for whom it was not suitable. The Committee agreed that dexamethasone intravitreal implant resulted in a greater average change in mean best corrected visual acuity (BCVA) than sham procedure for the general DMO population (dexamethasone 3.5 letters versus sham 2.0 letters, p=0.023) and for the subgroup of people with a pseudophakic lens (dexamethasone 6.5 letters versus sham 1.7 letters, p<0.001). The Committee concluded that dexamethasone intravitreal implant is a clinically effective treatment for DMO compared with sham procedure.

4.9 The Committee considered the high discontinuation rates in the dexamethasone and sham procedure arms in the MEAD trials. It heard from the clinical experts that the duration of the trials was 3 years which was longer than many other clinical trials in DMO populations; this may have resulted in more people dropping out of the trials. It heard from the company that many people in the sham procedure arm withdrew from the trials because their vision did not improve. Furthermore, if patients in the sham procedure arm needed additional treatment for their vision, the treatment was offered and then the patients were excluded from the trial. Therefore, the people in the sham arm of the MEAD trials are unlikely to be a true representation of the people who have watch‑and‑wait in clinical practice, because in practice people would be less likely to expect an improvement and to seek treatment changes. The Evidence Review Group (ERG) noted that people also dropped out of the dexamethasone arm of the trials but to a lesser extent. The Committee accepted that people remaining in the sham arm of the MEAD trials may not be representative of people who would have dexamethasone intravitreal implant in clinical practice and concluded that the MEAD trials might have underestimated the benefit of dexamethasone intravitreal implant compared with sham procedure.

4.10 The Committee considered the evidence for dexamethasone intravitreal implant and its comparators that was derived from the company's network meta‑analysis. It noted that evidence from direct comparisons was not available for all comparators, so a network meta‑analysis had been carried out by the company. The Committee noted the ERG's concerns that there were high levels of clinical and statistical heterogeneity associated with the network meta‑analysis and that there were wide 95% credible intervals around a large number of the relative risk estimates. It further noted that the MEAD trials informed the comparison of dexamethasone intravitreal implant and watch‑and‑wait and questioned whether a network meta‑analysis was necessary when direct trial data were available. It noted that the relative effects in the network meta‑analysis for dexamethasone intravitreal implant compared with sham were calculated using data only from the MEAD trials. The Committee concluded that although the company's network‑meta‑analysis was associated with uncertainty, it was acceptable to inform its decision‑making except for the comparison of dexamethasone intravitreal implant with sham (watch‑and‑wait). For this comparison, it concluded that its deliberations should focus on the data from the MEAD trials because these head‑to‑head results were more robust than those of the network meta‑analysis (which were based only on MEAD trial data).

4.11 The Committee considered the evidence on adverse events associated with dexamethasone intravitreal implant. It noted that the overall frequency of adverse events in the MEAD trials was acceptable and that there were fewer ocular adverse events in people with a pseudophakic lens compared with the general DMO population. It acknowledged that this was likely to be because cataracts were included as ocular adverse events, and people who have had their lens replaced with a pseudophakic lens cannot develop cataracts. The Committee concluded that dexamethasone intravitreal implant had an acceptable adverse event profile in people with DMO.

Cost effectiveness

4.12 The Committee considered the cost‑effectiveness analyses presented by the company and the critiques, corrections and exploratory analyses performed by the ERG. The Committee noted the following in the company's original submission and analyses supplied in response to clarification:

  • For people with DMO that has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is unsuitable, the company presented a comparison of dexamethasone intravitreal implant with watch‑and‑wait in its base case and with fluocinolone acetonide intravitreal implant as a scenario analysis.

  • For people with a pseudophakic lens, the company presented a comparison of dexamethasone intravitreal implant with ranibizumab in its base case and with laser photocoagulation, bevacizumab and watch‑and‑wait as a scenario analysis.

  • The ERG presented the same comparisons as the company in its exploratory analyses.

    The Committee noted that the company's new evidence submitted in response to the appraisal consultation document compared dexamethasone intravitreal implant with watch‑and‑wait in people with DMO that has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is unsuitable. The Committee concluded that it had been presented with cost‑effectiveness estimates for dexamethasone intravitreal implant in all necessary subpopulations to inform its decision‑making.

4.13 The Committee considered the cost of residential care in the company's economic model. The ERG noted that the cost of residential care was overestimated which caused the cost of severe vision loss (BCVA <35 letters) to be overestimated (see section 3.82). This was a key driver in the model for people with DMO that has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is unsuitable. The Committee noted that the company used the cost of local authority care in its model and heard from the ERG that the unit cost of private residential care should have been used in the model (see section 3.82). The Committee concluded that the costs of severe vision loss had been overestimated in the company's original calculations.

4.14 The Committee considered the utility values used in the company's model. The Committee noted that the utility values were based on trial data and spanned a relatively narrow range. The Committee discussed that it was possible that the utility values may have underestimated the disability resulting from the worst health state. It acknowledged that the company had not provided any evidence that the utility values used in the model were a good fit to the data nor how the fit would compare with more complex models that allowed interaction between eyes. However, the Committee noted that published utility values have a number of limitations; in particular, they are not specific to people with DMO (Brown 1999, Brown et al. 2000, Czoski‑Murray et al. 2009). It heard from the ERG that data from the MEAD trials are likely to be more relevant to people with DMO than the published utility values. The Committee acknowledged that the company's approach to inclusion of utility values in the model had some limitations, but so too did the published utility values available. On balance, the Committee agreed that the company's utility values were suitable to inform its decision‑making despite these limitations. However, it also concluded that neither approach was ideal and that both had shortcomings that inhibited the accurate estimation of the cost effectiveness of dexamethasone intravitreal implant for DMO.

4.15 The Committee considered other elements of the company's model. It acknowledged that the ERG had concerns about several factors that could have biased the results. These included modelling transitions for each eye independently, 'normalising' the transition probabilities in the model to sum them to 1 and assuming that the relative effect of dexamethasone intravitreal implant compared with sham procedure was stable for 3 years:

  • The Committee noted that modelling the transitions for each eye independently was a more realistic approach than that used in previous appraisals of eye conditions, which sometimes modelled the vision in only 1 eye. It further noted that the company's approach could result in the worse‑seeing eye (WSE) having a better BCVA than the better‑seeing eye (BSE).

  • The Committee heard that the company's approach to 'normalising' the transition probabilities in the model so that they summed to 1 meant that the relative risks used in the model were different from those provided by the network meta‑analysis. The Committee also discussed the assumption in the company's model that movement between health states was restricted to 1 move up or down per cycle. The Committee noted the analyses carried out by the company in response to clarification showed that not restricting the movement between health states had a large impact on the cost‑effectiveness results for the comparison of dexamethasone intravitreal implant with watch‑and‑wait. The Committee heard from the ERG that restricting movements to up or down 1 health state per cycle did not reflect trial evidence from the MEAD trials.

  • The Committee also noted that the assumption that the relative effect of dexamethasone intravitreal implant compared with sham was stable for 3 years was not observed in the MEAD trials, which showed that the relative effect was not stable beyond 12 months.

    The Committee concluded that these assumptions reflected neither the natural course of the disease nor the observed clinical trial data, and that this increased the uncertainty of the results of the model.

4.16 The Appraisal Committee considered whether it should take into account the consequences of the PPRS 2014, and in particular the PPRS Payment Mechanism, when appraising dexamethasone intravitreal implant. The Appraisal Committee noted NICE's position statement in this regard, and accepted the conclusion 'that the 2014 PPRS Payment Mechanism should not, as a matter of course, be regarded as a relevant consideration in its assessment of the cost effectiveness of branded medicines'. The Committee heard nothing to suggest that there is any basis for taking a different view with regard to the relevance of the PPRS to this appraisal of dexamethasone intravitreal implant. It therefore concluded that the PPRS Payment Mechanism was irrelevant for the consideration of cost effectiveness of dexamethasone intravitreal implant.

People with a pseudophakic lens with CRT of 400 micrometres or more

4.17 The Committee considered the cost effectiveness of dexamethasone intravitreal implant compared with ranibizumab for people who have a pseudophakic lens and a CRT of 400 micrometres or more. It noted that the company's analyses in the pseudophakic population incorporated a range of discounts (from 10% to 50%) applied to the list price of ranibizumab (see sections 3.45 and 3.53). The Committee discussed that in the company's probabilistic base‑case analysis using the list price of ranibizumab, the incremental cost‑effectiveness ratio (ICER) for ranibizumab compared with dexamethasone intravitreal implant was £89,531 per quality‑adjusted life‑years (QALY) gained. It noted that QALY gain with dexamethasone intravitreal implant was lower than that with ranibizumab. The Committee was aware of the actual discount agreed in the patient access scheme for ranibizumab (the level of the discount for this comparator was agreed with the Department of Health and is commercial in confidence – see guide to the processes of technology appraisal), and it agreed that the analyses by the company and the ERG captured this discount. Taking into account the exact discount agreed in the patient access scheme for ranibizumab, the Committee concluded that it did not recommend dexamethasone intravitreal implant because it produced fewer QALYs compared with ranibizumab, and the lower QALY gain was such that it could not justify the marginal difference in costs. Therefore the Committee concluded that dexamethasone intravitreal implant was not a cost‑effective use of NHS resources compared with ranibizumab (the most relevant comparator – see section 4.7) for treating DMO in people with a pseudophakic lens and a central retinal thickness of 400 micrometres or more.

People with a pseudophakic lens with CRT less than 400 micrometres

4.18 The Committee considered the cost effectiveness of dexamethasone intravitreal implant compared with laser photocoagulation therapy in people who have a pseudophakic lens and a CRT less than 400 micrometres. It noted that the company's analyses and the ERG's exploratory analyses in the pseudophakic population (see sections 3.52 and 3.95) showed that dexamethasone intravitreal implant was dominated by laser photocoagulation (that is, laser photocoagulation was less expensive and more effective). The Committee concluded that dexamethasone intravitreal implant was not a cost‑effective use of NHS resources compared with laser photocoagulation for treating DMO in people with a pseudophakic lens and a central retinal thickness less than 400 micrometres.

4.19 The Committee considered the cost effectiveness of dexamethasone intravitreal implant compared with bevacizumab in people who have a pseudophakic lens and a CRT less than 400 micrometres. It noted that the company's analyses and the ERG's exploratory analyses in the pseudophakic population showed that dexamethasone intravitreal implant was dominated by bevacizumab (see sections 3.52 and 3.95). The Committee concluded that dexamethasone intravitreal implant was not a cost‑effective use of NHS resources compared with bevacizumab for treating people with a pseudophakic lens and a central retinal thickness less than 400 micrometres.

People who do not have a pseudophakic lens and with DMO that does not respond to non‑corticosteroid treatment or for whom such treatment is unsuitable

4.20 In its first meeting, the Committee considered the cost effectiveness of dexamethasone intravitreal implant compared with watch‑and‑wait for people who do not have a pseudophakic lens and with DMO that has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is unsuitable based on evidence submitted before consultation. It noted that in the company's base‑case analysis, dexamethasone intravitreal implant dominated watch‑and‑wait in this subgroup. However, the Committee considered the concerns raised by the ERG about the company's model (see sections 3.79 and 3.80), acknowledged the ERG's corrections to the company's model (see section 3.87) and noted the ERG's alternative assumptions (see sections 3.88 and 3.92). It noted that in the ERG's exploratory analyses using data directly from the MEAD trials and with unrestricted moves between health states, the ICER was £1,166,271 per QALY gained for dexamethasone intravitreal implant compared with watch‑and‑wait. It noted that in the ERG's exploratory analyses using data from the network meta‑analysis and with restricted moves between health states, the ICER was £22,049 per QALY gained for dexamethasone intravitreal implant compared with watch‑and‑wait. The Committee was mindful of its earlier conclusion that data from the MEAD trials were more robust than the results of the network meta‑analyses for dexamethasone intravitreal implant compared with watch‑and‑wait (see section 4.10) because of the reliance of the network meta‑analysis on assumptions that did not reflect clinical practice. The Committee noted that, although it preferred using the head‑to‑head MEAD trial data, the ICER of £1,166,271 per QALY gained was likely to be an overestimate because of the high discontinuation rates in MEAD (leading to the sham arm being unrepresentative of patients in clinical practice; see section 4.9), narrow bands of utility values from MEAD (see section 4.14) and the possibility of the WSE having a higher utility value than the BSE in the model (see section 4.15). The Committee acknowledged that although the true value of the ICER was likely to be less than £1,166,271 per QALY, it was extremely unlikely to be within the range normally considered to be a cost‑effective use of NHS resources (up to £20,000–30,000 per QALY gained).

4.21 In its second meeting, the Committee considered the new evidence submitted by the company in response to the appraisal consultation document and the associated cost‑effectiveness estimates by the company and the ERG. This compared the cost effectiveness for dexamethasone intravitreal implant with watch‑and‑wait in people who do not have a pseudophakic lens and with DMO that has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is unsuitable. The Committee reviewed the plausibility of each of the 4 assumptions that had been altered by the company in addition to those in the ERG's original exploratory base case (that is, which resulted in an ICER of £1,166,271 per QALY gained). Firstly, it accepted the minor change made by the company to modelling the costs of residential care (see section 3.59), noting that this had a minimal effect on the ICER.

4.22 Secondly, the Committee considered the alternative transition matrices implemented by the company. The Committee did not take into account the criticisms from the ERG's written report on the company's new evidence because this did not correctly interpret what happened to patients who discontinued or were censored in the company's economic model (see section 3.99). It acknowledged that the company's new approach reflected the different probabilities of improving or worsening vision depending on the starting health state, which was an advantage over using the natural history transition matrix (see section 3.60). It noted the company's concerns around using natural history data based on diabetic retinopathy (from the Wisconsin Epidemiologic Study of Diabetic Retinopathy) rather than DMO. However, the Committee had concerns over the plausibility of the company's new approach. The Committee was aware of the considerable impact of the company's new transition probabilities on the ICER (see section 3.61) and noted that the incremental QALYs increased from 0.0058 to 0.03746 when this sole change was implemented. It considered that this 6‑fold increase in utility gain was implausible and had likely overestimated the incremental difference in QALYs between dexamethasone intravitreal implant and watch‑and‑wait. It found it unreasonable that the model did not account for any association between worsening of vision and treatment discontinuation because it considered it probable that those patients with deteriorating vision were more likely to discontinue. The Committee believed that assuming that the deterioration was carried forward to the remaining cycles was likely to embed a bias within the model. Therefore, it was not persuaded that adopting the last observed transition before discontinuation to inform the model cycles after discontinuation was plausible. It considered that it would have been preferable to use more of the data before treatment discontinuation. The Committee heard from the clinical experts that the decline in vision according to the natural history of the condition would be expected to follow a curve, and considered that the linear approach applied by the company was not clinically plausible. The Committee consequently agreed that the company's new approach to modelling the transition probabilities was inappropriate. It acknowledged that there was some uncertainty in using the natural history data, but concluded that the company's original transition probability matrices were less inappropriate than those in the company's new evidence submission for using in its decision‑making.

4.23 Thirdly, the Committee considered the utility values used by the company in its base case using the new evidence. It acknowledged that utility values derived from Czoski‑Murray et al. had been accepted in previous appraisals of treatments for DMO but agreed that these values had limitations (see section 4.14). It expressed its preference for utility values derived from clinical trial data, although it considered that the utility values derived from the MEAD trials also had limitations (see section 4.14). It accepted the Czoski‑Murray et al. values but heard they had been incorrectly implemented in the company's economic model. It considered the ERG's correction to the implementation of the utility values to be reasonable, and heard from the company that its base‑case ICER decreased slightly when it used the ERG's implementation method. The Committee concluded that the utility values derived from Czoski‑Murray et al. with the ERG's correction to their implementation were reasonable.

4.24 Fourthly, the Committee considered the clinical continuation rule proposed by the company. It heard from the clinical experts that gaining at least 5 letters would not be an appropriate way of determining response because DMO is a progressive condition and even slowing the rate of deterioration (such as the rate of losing 5 letters) could be seen as clinically beneficial. The Committee also heard that improvement in vision was open to clinical interpretation. It heard that better central vision (for example, so that a patient could see without having to turn their head) would be viewed as an objective measure of improvement, but it would not necessarily be reflected in a gain in letters. Moreover, the Committee heard that the visual acuity tests were not always reliable. The Committee concluded that it was inappropriate to adopt the company's proposed treatment continuation rule and excluded it from its decision‑making.

4.25 Taking the above issues into account (see sections 4.21–4.24), the Committee considered that the true value of the ICER would be greater than the ERG's new exploratory base‑case ICER of £127,645 per QALY gained (because the ICER would increase if the clinical continuation rule was omitted from the economic model). It noted that this far exceeded the range normally considered to be a cost‑effective use of NHS resources (up to £20,000–30,000 per QALY gained). The Committee concluded that dexamethasone intravitreal implant was not a cost‑effective use of NHS resources compared with watch‑and‑wait for treating DMO in people who do not have a pseudophakic lens and have DMO that has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is unsuitable. Accordingly, the Committee did not recommend dexamethasone intravitreal implant for this group.

People with a pseudophakic lens and with DMO that does not respond to non‑corticosteroid treatment or for whom such treatment is unsuitable

4.26 The Committee considered the cost effectiveness of dexamethasone intravitreal implant compared with fluocinolone acetonide intravitreal implant for people who have a pseudophakic lens and DMO that has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is unsuitable. It observed that the company's ICER for fluocinolone acetonide intravitreal implant compared with dexamethasone intravitreal implant in people with DMO that has not responded adequately to non‑corticosteroid therapy was £24,591 per QALY gained. It noted that the company's and the ERG's analyses incorporated a range of discounts (from 10% to 50%) applied to the list price of fluocinolone acetonide intravitreal implant (see sections 3.51 and 3.97). The Committee was aware of the actual discount agreed in the patient access scheme for fluocinolone acetonide intravitreal implant (the level of the discount for this comparator was agreed with the Department of Health and is commercial in confidence – see guide to the processes of technology appraisal) and it agreed that the analyses undertaken by both the company and the ERG captured this discount. The Committee noted that when the exact discount agreed in the patient access scheme for fluocinolone acetonide intravitreal implant was taken into account, there was little difference in the total costs and QALYs of fluocinolone acetonide intravitreal implant and dexamethasone intravitreal implant. Therefore, it considered that the cost effectiveness of dexamethasone intravitreal implant is likely to be similar to that of fluocinolone acetonide intravitreal implant. The Committee considered that, on balance, dexamethasone intravitreal implant would provide an alternative treatment option to fluocinolone acetonide intravitreal implant. The Committee concluded that dexamethasone intravitreal implant was a cost‑effective use of NHS resources in this group and so recommended it as a treatment option for DMO in people who have a pseudophakic lens and whose DMO has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is not suitable.

Summary of Appraisal Committee's key conclusions

TA349

Appraisal title: Dexamethasone intravitreal implant for treating diabetic macular oedema

Section

Key conclusion

Dexamethasone intravitreal implant is recommended as an option for treating diabetic macular oedema (DMO) only if:

  • the implant is to be used in an eye with an intraocular (pseudophakic) lens and

  • the diabetic macular oedema does not respond to non‑corticosteroid treatment, or such treatment is unsuitable.

Because it considered that the cost effectiveness of dexamethasone intravitreal implant was likely to be similar to fluocinolone acetonide intravitreal implant when the patient access scheme for fluocinolone intravitreal implant is applied, the Committee recommended dexamethasone intravitreal implant as a treatment option for treating DMO in people who have a pseudophakic lens and whose DMO has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is not suitable.

The Committee noted that dexamethasone intravitreal implant produced fewer QALYs compared with ranibizumab in people with a pseudophakic lens and a central retinal thickness of 400 micrometres or more. When the ranibizumab patient access scheme was incorporated, the lower QALY gain was such that it could not justify the marginal difference in costs and so it concluded that dexamethasone intravitreal implant was not a cost‑effective use of NHS resources in this patient group.

The Committee noted that dexamethasone intravitreal implant was dominated by laser photocoagulation and bevacizumab in people with a pseudophakic lens and a central retinal thickness less than 400 micrometres (that is, dexamethasone intravitreal implant was less effective and more costly) and concluded that it was not a cost‑effective use of NHS resources in this patient group.

The Committee considered the ICERs for dexamethasone intravitreal implant compared with watch‑and‑wait in people who do not have a pseudophakic lens and who have diabetic macular oedema that has not responded adequately to non‑corticosteroid therapy, or for whom such treatment is unsuitable. It considered that the true value of the ICER would be greater than the Evidence Review Group's (ERG's) new exploratory base‑case ICER of £127,645 per QALY gained. The Committee concluded that dexamethasone intravitreal implant was not a cost‑effective use of NHS resources compared with watch‑and‑wait in this group, and accordingly did not recommend dexamethasone intravitreal implant.

1.1

4.17–4.19, 4.25, 4.26

Current practice

Clinical need of patients, including the availability of alternative treatments

The Committee heard from the clinical experts that for people with acentral retinal thickness (CRT) of 400 micrometres or greater, ranibizumab is given as recommended in NICE technology appraisal guidance on ranibizumab for treating diabetic macular oedema. For people who have a CRT of less than 400 micrometres the Committee noted that ranibizumab is not recommended in NICE technology appraisal guidance on ranibizumab and it heard from clinical experts that the prescribing options in clinical practice are laser photocoagulation and bevacizumab (outside its marketing authorisation).

The Committee acknowledged that for people who have a pseudophakic lens and chronic disease which has not responded to prior therapy, fluocinolone acetonide intravitreal implant is recommended in NICE technology appraisal guidance on fluocinolone acetonide intravitreal implant for treating chronic diabetic macular oedema after an inadequate response to prior therapy. It heard from the clinical experts that they perceived that fluocinolone intravitreal implant had been used with caution in the UK because some of its side effects can be difficult to reverse, given its long duration of action.

The Committee heard from the clinical experts that people with DMO who do not have a pseudophakic lens are monitored but do not have active treatment (watch‑and‑wait).

The Committee heard from the clinical experts that there are no clinical criteria for determining whether a treatment is unsuitable for a person with DMO. However, they suggested that treatment with an anti‑VEGF agent (for example, ranibizumab) is likely to be unsuitable for people who cannot attend monthly appointments, people who have had a recent cardiovascular event or stroke, or people who have a phobia of needles.

4.4–4.6

The technology

Proposed benefits of the technology

How innovative is the technology in its potential to make a significant and substantial impact on health‑related benefits?

The Committee concluded that patients and clinicians considered dexamethasone intravitreal implant to be a valuable option that could offer longer‑term benefits than anti‑VEGF treatments to some people with DMO, and for people with DMO that is unresponsive to non‑corticosteroid treatment.

The company did not make any claim for innovation.

4.1

What is the position of the treatment in the pathway of care for the condition?

The Committee recalled that the indication in the marketing authorisation for dexamethasone intravitreal implant was for the treatment of adult patients with visual impairment due to diabetic macular oedema who are pseudophakic or who are considered insufficiently responsive to, or unsuitable for, non‑corticosteroid therapy and was mindful that it could only make recommendations within the marketing authorisation. After considering comparators within the context of the marketing authorisation, it concluded that the 4 potentially eligible populations were:

  • People with a pseudophakic lens with CRT of 400 micrometres or more.

  • People with a pseudophakic lens with CRT less than 400 micrometres.

  • People who do not have a pseudophakic lens and with diabetic macular oedema that does not respond to non‑corticosteroid treatment, or for whom such treatment is unsuitable.

  • People with a pseudophakic lens and with diabetic macular oedema that does not respond to non‑corticosteroid treatment, or for whom such treatment is unsuitable.

2.1

4.2

Adverse reactions

The Committee concluded that dexamethasone intravitreal implant had an acceptable adverse event profile in people with DMO.

4.11

Evidence for clinical effectiveness

Availability, nature and quality of evidence

The Committee accepted that the high discontinuation rates in the MEAD trials may have underestimated the benefit of dexamethasone intravitreal implant compared with sham procedure.

The Committee noted that the relative effects in the network meta‑analysis for dexamethasone intravitreal implant compared with sham were calculated using data only from the MEAD trials. It concluded that the trial data from the MEAD trials were more robust than the results of the network meta‑analysis for the comparison of dexamethasone intravitreal implant and sham procedure (watch‑and‑wait).

4.9

4.10

Relevance to general clinical practice in the NHS

The Committee heard from the clinical and patient experts that dexamethasone intravitreal implants last much longer than anti‑VEGF treatments, and that the need for repeat treatment visits with anti‑VEGFs can be problematic.

4.1

Uncertainties generated by the evidence

The Committee accepted that the high discontinuation rates in the MEAD trials may have underestimated the benefit of dexamethasone intravitreal implant compared with sham procedure.

The Committee concluded that the data from the MEAD trials were more robust than the results of the network meta‑analysis for the comparison of dexamethasone intravitreal implant and sham procedure (watch‑and‑wait).

4.9

4.10

Are there any clinically relevant subgroups for which there is evidence of differential effectiveness?

Not applicable.

Estimate of the size of the clinical effectiveness including strength of supporting evidence

On the basis of the MEAD trials, the Committee concluded that dexamethasone intravitreal implant is proven to be an effective treatment compared with sham procedure in people with diabetic macular oedema.

4.8

Evidence for cost effectiveness

Availability and nature of evidence

The Committee concluded that it had been presented with cost‑effectiveness estimates for dexamethasone intravitreal implant in all necessary sub‑populations to inform its decision‑making.

4.12

Uncertainties around and plausibility of assumptions and inputs in the economic model

The Committee acknowledged the ERG's concerns about several factors that could have biased the results, including modelling transitions for each eye independently, 'normalising' the transition probabilities in the model to sum them to 1 and assuming that the relative effect of dexamethasone intravitreal implant compared with sham procedure was stable for 3 years. The Committee concluded that these assumptions reflected neither the natural course of the disease nor the observed clinical trial data, and that this increased the uncertainty of the results of the model.

The Committee considered the transition probabilities used for treatment discontinuation or censoring. It was not persuaded that adopting the last observed transition before discontinuation to inform the model cycles after discontinuation was plausible, as assumed in the company's new analyses. The Committee concluded that the company's original transition probability matrices according to disease natural history were less inappropriate for using in its decision‑making than those in the company's new evidence submission.

4.15, 4.22

Incorporation of health‑related quality‑of‑life benefits and utility values

Have any potential significant and substantial health‑related benefits been identified that were not included in the economic model, and how have they been considered?

The Committee noted that the utility values used in the company's model were based on trial data. It acknowledged that the company's approach to inclusion of utility values in the model had some limitations, but so too did the published utility values available. On balance, the Committee concluded that neither approach was ideal, but it agreed that the company's utility values were suitable to inform its decision‑making, despite these limitations. However, it also concluded that neither approach was ideal and that both had shortcomings that inhibited the accurate estimation of the cost effectiveness of dexamethasone intravitreal implant for DMO.

4.14

Are there specific groups of people for whom the technology is particularly cost effective?

Not applicable.

What are the key drivers of cost effectiveness?

For people with diabetic macular oedema that has not responded adequately to non‑corticosteroid therapy or for whom such treatment is unsuitable, the key driver of cost effectiveness in the model is the cost of residential care used for people with severe vision loss.

4.13

Most likely cost‑effectiveness estimate (given as an ICER)

For people with a pseudophakic lens with CRT of 400 micrometres or more, the Committee concluded that if the confidential patient access scheme for ranibizumab was included, it did not recommend dexamethasone intravitreal implant because its lower QALY gain with a marginal difference in costs was not a cost‑effective use of NHS resources compared with ranibizumab.

For people with a pseudophakic lens with CRT less than 400 micrometres, the Committee noted that dexamethasone intravitreal implant was dominated by laser photocoagulation therapy and bevacizumab.

For people without a pseudophakic lens with diabetic macular oedema that is unsuitable for or insufficiently responsive to non‑corticosteroid therapy, the Committee considered that the true value of the ICER would be greater than the ERG's new exploratory base‑case ICER of £127,645 per QALY gained.

For people with a pseudophakic lens with diabetic macular oedema that is unsuitable for or insufficiently responsive to non‑corticosteroid therapy the Committee noted that, when the exact discount agreed in the patient access scheme for fluocinolone acetonide intravitreal implant was taken into account, there was little difference in the total costs and total QALYs of fluocinolone acetonide intravitreal implant and dexamethasone intravitreal implant. Therefore, it considered that the cost effectiveness of dexamethasone intravitreal implant is likely to be similar to fluocinolone acetonide intravitreal implant.

4.17–4.19, 4.25, 4.26

Additional factors taken into account

Patient access schemes (PPRS)

A patient access scheme is in place for 2 of the comparators – ranibizumab and fluocinolone acetonide intravitreal implant. These are taken into account in the Committee's conclusions.

The Committee concluded that the PPRS Payment Mechanism was irrelevant for the consideration of cost effectiveness of dexamethasone intravitreal implant.

4.17, 4.26, 4.27

End‑of‑life considerations

Not applicable.

Equalities considerations and social value judgements

No issues relating to equality considerations were raised in the submissions, or in the Committee meeting.

  • National Institute for Health and Care Excellence (NICE)