4 Consideration of the evidence
The Appraisal Committee reviewed the data available on the clinical and cost effectiveness of vedolizumab, having considered evidence on the nature of moderately to severely active ulcerative colitis and the value placed on the benefits of vedolizumab by people with the condition, those who represent them, and clinical experts. It also took into account the effective use of NHS resources.
4.1 The Committee heard from patient experts about their experience of ulcerative colitis and the available treatments. The patient experts explained that moderately to severely active ulcerative colitis had a major effect on their life. One expert explained that the disease had been so severe that it had left them housebound, unable to work and often hospitalised. Corticosteroid treatment had resulted in osteoporosis and liver problems. Treatment with vedolizumab had resulted in complete remission from the disease and had given them their life back, including the ability to work. Both patient experts explained that people who do not improve on current treatments are desperate for alternative treatment options. The patient experts further highlighted that many people who have ulcerative colitis are teenagers and younger adults. Unmanaged ulcerative colitis can affect their ability to study, find work, socialise and find a partner, which has a major effect on their quality of life. The patient experts also commented on surgery for treating ulcerative colitis. They stated that the effect of surgery on fertility, its irreversibility, its risks and the potential for a life‑long impact on lifestyle meant that it was a very unattractive option for many people. The clinical experts agreed and commented that some people prefer to put up with severe symptoms to avoid surgery. The patient experts also acknowledged that, despite their concerns, they may need surgery in the future once their drug treatment options were exhausted, but they would want to delay it for as long as possible. The Committee concluded that a drug treatment that improves or brings the disease into remission would have a major effect on quality of life, and that avoiding surgery was important to people with ulcerative colitis.
4.2 The Committee considered the current drug treatment of moderately to severely active ulcerative colitis. The clinical experts stated that the aim of drug treatment is to induce remission, control symptoms, and to decrease the risk of cancer associated with long‑term active ulcerative colitis. The Committee understood that conventional therapy included treatment with aminosalicylates, thiopurines, such as azathioprine and mercaptopurine, and corticosteroids. The Committee was aware that some people may also be offered a TNF‑alpha inhibitor if conventional therapy failed. If 1 or more TNF‑alpha inhibitors failed, the treatment options were limited to conventional therapy or surgery. The clinical experts stated that methotrexate, ciclosporin and tacrolimus were rarely used for these patients because of their adverse effects. They also stated that, after other medical therapies have failed, people are often treated with long‑term high‑dose corticosteroids. This is undesirable because of the many potentially serious adverse effects associated with high‑dose corticosteroid therapy.
4.3 The Committee considered when people may choose to have surgery in the treatment pathway for ulcerative colitis. It was aware that there are 2 types of surgery for ulcerative colitis: ileostomy, in which the small intestine is diverted out through a hole (stoma) in the abdomen with an external bag, and ileo‑anal pouch surgery, in which part of the small intestine is used to create an internal pouch, which allows a person to pass stools through the anus. It heard from the patient and clinical experts that surgery relieves symptoms by removing parts of the colon and rectum that are chronically affected, but is associated with problems. For people who have an ileostomy, these include adjusting lifestyle to manage a stoma and its negative effect on body image and self‑esteem. For people who have ileo‑anal pouch surgery, problems include needing to wake up in the night to use the toilet, and potential complications are pouch leakage, overnight incontinence and pouchitis. It heard from the clinical experts that surgery is typically used when drug treatment for moderately to severely active ulcerative colitis has failed. The clinical experts explained that ileo‑anal pouch surgery for ulcerative colitis usually needs 2 to 3 major operations, and that some people may need further operations if they have complications or problems following surgery. The Committee heard that complications following surgery can result in devastating effects for a few people. The Committee also heard from patient experts that, because of the irreversibility and potential for complications with surgery, it is seen as a last step in the management of the condition when no other options are available. Moreover, the Committee heard that there would also be a risk of infertility, and recalled that many people who have ulcerative colitis are teenagers and younger adults. The clinical experts explained that clinicians would be reluctant to recommend surgery unless all drug treatment options had failed. The Committee concluded that surgery was considered a final treatment option for treating chronic moderately to severely active ulcerative colitis.
4.4 The Committee explored how decisions to continue treatment with vedolizumab would be made in clinical practice. The Committee was aware that the marketing authorisation for vedolizumab states that continued therapy should be carefully considered if no evidence of therapeutic benefit is observed by week 10. By this time people would have had 3 doses. The clinical experts stated that, in clinical practice, clinicians may assess response at week 14 (before the person has had their fourth dose of vedolizumab). The Committee asked whether a person in remission on vedolizumab would remain on vedolizumab indefinitely. It was aware that there were no trial data about whether remission would be maintained in people stopping vedolizumab. The clinical experts stated that, in clinical practice, clinicians would not want to keep a patient on treatment if they no longer needed it. Patient experts also stated that people would want to avoid the potential side effects of treatment if the treatment was no longer necessary, but were concerned about symptoms recurring if treatment was stopped. The Committee heard from the clinical experts that, in practice, treatment with TNF‑alpha inhibitors may be discontinued if deep remission had been achieved (that is, no clinical symptoms and no inflammation in the colon, (assessed using endoscopy or measuring faecal calprotectin). The same approach was likely to be taken with vedolizumab. The clinical experts stated that people who had a disease flare‑up after stopping treatment with a TNF‑alpha inhibitor may be offered re‑treatment with the same or another TNF‑alpha inhibitor. They explained that the chance of going into remission with re‑treatment depended on whether antibodies had developed against the TNF‑alpha inhibitor. The clinical experts stated that data suggested that rates of antibody formation to vedolizumab were low. Therefore, if a person stopped vedolizumab after their ulcerative colitis entered remission and then had a disease flare‑up, re‑treatment with vedolizumab may be a possibility. The Committee concluded that people who had remission on vedolizumab may stop treatment, but this would not stop people having re‑treatment with vedolizumab if they had a subsequent relapse.
4.5 The Committee considered the generalisability of the population in GEMINI I to the population who would have vedolizumab in clinical practice in England. It understood that GEMINI I was an international study and 2 of the centres were in the UK. It was aware that there were differences in the study entry criteria between the USA and other centres. These differences related to which previous treatments had failed and the use of immunosuppressants during the study (see section 3.12). The Committee heard from the clinical experts that the population included in the trial broadly reflected the population who would be treated with vedolizumab in England. It also heard that differences in immunosuppressant use between trial centres were unlikely to affect the trial's generalisability to clinical practice in England. The Committee concluded that the clinical efficacy results from GEMINI I were generalisable to clinical practice, but that there was uncertainty about whether the proportion of people who had previous TNF‑alpha inhibitor treatment in GEMINI I would be the same as in the population considered for vedolizumab treatment in England.
4.6 The Committee discussed the efficacy estimates for vedolizumab from GEMINI I. The Committee noted that in GEMINI I people had vedolizumab at weeks 0 and 2 and response was assessed at week 6, but the marketing authorisation for vedolizumab states that people should have 3 doses before response is assessed at week 10. The company clarified that the European Medicines Agency made this recommendation based on data from the continued follow‑up of people who had not had a response at week 6 in GEMINI I. It confirmed that the recommendation had been made after the company compiled its submission for NICE (see section 3.6). The clinical experts stated that a 10‑week assessment reflected when response to induction with vedolizumab would be expected to peak, based on the trial data. The Committee noted that in the trial 47% of people had a response to vedolizumab by week 6 and of those who did not have a response, 32% had done so by week 10, and 39% by week 14. The Committee agreed that the trial data, which did not include people whose disease responded after week 6, may underestimate the number of people expected to have a response to induction treatment in clinical practice, in which response is assessed later than 6 weeks. It further commented that data on the effect of vedolizumab maintenance treatment had only been presented for people whose disease responded by week 6 rather than the total population eligible to continue maintenance treatment with vedolizumab in clinical practice. The Committee concluded that, although the efficacy of vedolizumab had been shown in GEMINI I, it may have underestimated the proportion of people who would have a response to induction treatment in clinical practice, and that data on the outcome for those who responded after 6 weeks were not available from the trial.
4.7 The Committee understood that vedolizumab has a marketing authorisation for use in people in whom conventional therapy with or without a TNF‑alpha inhibitor has failed. It noted that, in GEMINI I, 48% of people had previously had a TNF‑alpha inhibitor and that the company had presented the results for 2 subgroups: people who had not had TNF‑alpha inhibitor treatment before, and people in whom TNF‑alpha inhibitors had failed. The Committee noted that in both subgroups vedolizumab was associated with a higher rate of clinical response and remission than placebo. This was consistent with the results for the whole intention‑to‑treat population. The Committee was aware that GEMINI I was not powered to test for a statistically significant difference in the treatment effect of vedolizumab between subgroups. However, the Committee noted that, in the subgroup of people who had a TNF‑alpha inhibitor, there were numerically lower rates of remission and response in both vedolizumab and placebo arms compared with the subgroup of people who had not had a TNF‑alpha inhibitor. The Committee heard from the company and the clinical experts that people in whom TNF‑alpha treatment had failed could be considered to have ulcerative colitis that is more difficult to treat. The Committee agreed that it was useful to consider the 2 subgroups as separate populations because they may differ in their likelihood of having a response or going into remission with drug treatment in clinical practice. The Committee concluded that, based on the data from GEMINI I, vedolizumab was clinically effective in the whole population, and in both subgroups, compared with conventional therapy.
4.8 The Committee discussed the comparators included in the company's decision problem. The Committee noted that the final scope issued by NICE stated that the comparator for vedolizumab was established clinical management. This may include treatment with aminosalicylates, corticosteroids, thiopurines, calcineurin inhibitors, TNF‑alpha inhibitors and surgery. The Committee noted that in GEMINI I people had conventional therapy, which included aminosalicylates, thiopurines and corticosteroids in both the placebo and vedolizumab arms. It was further aware that in GEMINI I people did not have calcineurin inhibitors or TNF‑alpha inhibitors, and GEMINI I excluded people who were thought to need immediate surgery. The Committee noted that the company did not consider surgery and the calcineurin inhibitors to be comparators for vedolizumab because calcineurin inhibitors would be used for acute rather than chronic treatment, and surgery would be for severe disease that could not be managed with drug treatment. The Committee agreed, based on what it had heard from clinical and patient experts, that calcineurin inhibitors and surgery were not relevant comparators. The Committee concluded that conventional therapy and TNF‑alpha inhibitors were appropriate comparators for the whole population and both of the subgroups considered in this appraisal.
4.9 The Committee considered the network meta‑analyses presented by the company to estimate the relative effectiveness of vedolizumab compared with adalimumab, infliximab and golimumab. It noted that clinical data for infliximab and golimumab were not available for people who had previously had a TNF‑alpha inhibitor. Therefore, for this subgroup a comparison could only be made between vedolizumab and adalimumab. The Committee understood that the company had presented network meta‑analyses for the subgroups rather than the whole population. The Committee noted the Evidence Review Group (ERG)'s concerns that there were differences between the trials included in the meta‑analyses, and the company had presented results from a fixed‑effect model which was less suitable than a random‑effects model in these circumstances. The Committee understood that a network meta‑analysis for the whole population would include data from studies that included people who had, and had not, taken a TNF‑alpha inhibitor, and that these differences in patient characteristics may affect the results. Therefore, the Committee recognised that the relative effectiveness of vedolizumab compared with the TNF‑alpha inhibitors, obtained from a mixed treatment comparison of the whole population, would be subject to considerable uncertainty.
4.10 The Committee discussed the adverse effects of vedolizumab. It noted that the adverse events reported in GEMINI I were similar in the placebo and vedolizumab arms. It also noted that vedolizumab was taken at the same time as systemic immunosuppressants by some people, so adverse events seen in the trial that were associated with immunosuppression in other areas of the body besides the gut may not be associated with vedolizumab. The Committee was aware that cases of progressive multifocal leukoencephalopathy (PML), a fatal condition affecting the brain, have been seen with natalizumab, an antibody that inhibits α4‑integrin. It was aware that, because vedolizumab also inhibits an α4‑integrin, the incidence of PML in people treated with vedolizumab is being closely monitored, although there have been no reports of PML. The Committee heard from clinical experts that natalizumab inhibits α4‑integrin in all tissues of the bodies including the brain. Vedolizumab targets the gut, so the Committee accepted that the risk of PML in people treated with vedolizumab would be low. The Committee concluded that vedolizumab appeared to be safe and well tolerated by patients.
4.11 The Committee noted that the company presented base‑case results for 3 populations:
the whole population for whom vedolizumab has a marketing authorisation
a population who had not had previous treatment with a TNF‑alpha inhibitor
a population in whom treatment with a TNF‑alpha inhibitor had failed.
The Committee agreed that in view of its previous consideration, it would be useful to consider the 2 subgroups as separate populations (see section 4.7) and it should confine its further consideration to whether vedolizumab was cost effective in the 2 subgroups separately, rather than the cost‑effectiveness estimate for the whole population.
4.12 The Committee considered the original base‑case results (using the company's original patient access scheme) for the population who had not had TNF‑alpha inhibitors before. It noted that there was a large difference in the cost‑effectiveness estimates presented by the company and those presented by the ERG.
In the company's base case, the pairwise ICERs for vedolizumab compared with conventional therapy and adalimumab were under £7000 per quality‑adjusted life year (QALY) gained. Vedolizumab was also more effective and less costly than infliximab and golimumab.
In the ERG's exploratory base case, the incremental analyses showed that adalimumab dominated vedolizumab. The pairwise ICER for vedolizumab compared with conventional therapy was over £50,000 per QALY gained.
The Committee recognised that there was a very large difference in the base‑case ICERs presented by the company and the ERG, and explored the following model assumptions to understand this difference:
1‑year stopping rule (see section 4.13)
utility values (see section 4.14)
frequency of surgery and its costs (see section 4.15)
costs of post‑surgery care (see section 4.16).
4.13 The Committee considered the 1‑year stopping rule for biological treatments (vedolizumab, adalimumab, infliximab or golimumab). It noted that the company's model assumed that people would have biological treatments for a maximum of 1 year, after which they would switch to conventional therapy. However, biological treatments continued throughout the first year even if there was a loss of response. The Committee noted that the ERG had considered that biological treatments would continue until there was a loss of response. The Committee was aware that the clinical experts had stated that there was no established stopping rule for vedolizumab or the TNF‑alpha inhibitors when used to treat ulcerative colitis. However, people with confirmed remission may stop treatment if it is likely that their remission will be maintained without continued treatment. The Committee noted that if people stayed on biological treatments for 3 years rather than 1 (as modelled by the company in a scenario analysis) this increased the ICER for vedolizumab compared with drug treatment in the population who had not had TNF‑alpha inhibitors before. Similarly the ERG's scenario, in which people were assumed to continue treatment with biological treatments until loss of response, resulted in an increase in the ICER. The Committee considered that staying on vedolizumab for longer periods of time reduced the likelihood that vedolizumab would be cost effective in the population who had not had a TNF‑alpha inhibitor before. The Committee concluded that a similar stopping rule to that recommended in NICE's technology appraisal guidance on infliximab (review) and adalimumab for the treatment of Crohn's disease was appropriate and was likely to reflect how clinicians would prescribe vedolizumab in clinical practice.
4.14 The Committee discussed the utility values used in the company's model and the ERG's exploratory analysis, noting that utility values had large effects on the ICERs. It noted that the company had used EQ‑5D data from GEMINI I to estimate the utility associated with ulcerative colitis, but had used data from a different source to estimate the utility in the surgery and post‑surgery health states. The Committee was aware that this resulted in people with post‑surgery remission having a worse quality of life than people with moderately to severely active ulcerative colitis, which the ERG considered to be implausible. The Committee noted that all the other sources of utility values presented by the company and the ERG showed that the utility following surgery was higher than with moderately to severely active ulcerative colitis. The Committee was aware that the ERG had used utility estimates from Woehl et al. in their exploratory base case and they had also presented alternative utility estimates from Swinburn et al. The Committee considered that because these estimates were obtained from abstracts rather than full published studies and included small numbers of participants, there was uncertainty about their generalisability to clinical practice. The Committee noted that an important difference between the Woehl et al. and Swinburn et al. estimates was that the post‑surgery utility from Woehl et al. was estimated to be similar to having mild disease; whereas in Swinburn et al. it was associated with a lower utility, closer to that of moderately to severely active ulcerative colitis. The Committee heard from the clinical experts that even without complications, surgery had a substantial effect on people's lives. The patient experts added that, although they had not had surgery, the anticipated effect on their life was sufficient for them to delay it for as long as possible. The Committee agreed that quality of life may be improved after surgery compared with having moderately to severely active ulcerative colitis, but the magnitude of the difference was unclear. The Committee agreed that the Woehl et al. estimates may not fully capture the lifelong effect of surgery on a person's quality of life and that the Swinburn et al. estimates could be regarded as equally valid. The Committee therefore considered the cost‑effectiveness estimates based on both sets of utility values.
4.15 The Committee discussed the number of surgical procedures a person was assumed to have in the model, and the associated costs. The Committee noted that when the company's model was run over 10 years people would have 4 operations, and over a lifetime time horizon up to 19 operations. The Committee heard that the ERG considered that the total number of operations, and therefore the costs, had been overestimated. The Committee heard from the clinical experts that surgery for ulcerative colitis was normally carried out in 3 stages in separate operations, and a person could have further surgery if there were complications or further problems. The Committee understood from the ERG that it considered that the costs of surgery from Buchanan, which were used by the company, represented the total cost of multiple operations. The Committee heard from the clinical experts that costs reported by Buchanan only accounted for the cost of 1 operation. The Committee agreed that if this were the case, the ERG's exploratory base case would have underestimated the cost of surgery. However, the company's assumptions overestimated the costs of surgery because of the number of operations included. The Committee concluded that the total costs of surgery in the company's base case were too high and those in the ERG exploratory base case were too low.
4.16 The Committee considered the costs of stoma care for people who had an ileostomy. The ERG had stated that it was unclear whether the company had included the costs of stoma care in its model. It had therefore presented a scenario analysis in which costs of stoma care were included. This resulted in vedolizumab being less costly than all other options. The Committee considered that it was appropriate to include the costs of stoma care in the model, but noted that the ERG's estimate of the costs over a 6‑month period (£315) may be a low estimate. The Committee had heard that for the ongoing multiple technology appraisal of infliximab, adalimumab and golimumab for treating ulcerative colitis after conventional therapy has failed, post‑surgery care costs may be £1000–3000 per year. The Committee concluded that the costs of stoma care should be included in the model, but that these may have been underestimated by the ERG. The Committee noted that increasing the cost of stoma care was likely to improve the cost effectiveness of vedolizumab.
4.17 The Committee further considered the ERG's exploratory base case (using the company's original patient access scheme) for the population who had not had TNF‑alpha inhibitors before. The ERG estimated that adalimumab was associated with 0.02 more QALYs than vedolizumab and cost £12,574 less (adalimumab dominated vedolizumab). The pairwise ICER for vedolizumab compared with conventional therapy was £53,000 per QALY gained. However, in the company's submission the pairwise ICERs with adalimumab and conventional therapy were £7000 and £5000 per QALY gained respectively. The Committee was aware that a revised patient access scheme had been approved by the Department of Health and noted that incorporating this would result in lower ICERs. The Committee noted that the ERG's analysis assumed continued treatment with biological treatments until loss of response, only 1 surgical procedure, relatively low stoma care costs, Woehl et al. utility values and a lifetime time horizon. The Committee considered that the utility values reported by Swinburn et al. were as plausible as those by Woehl et al. and that the company's assumption that people stopped treatment at 1 year was not unreasonable. The Committee understood that if the ERG's exploratory base case was adjusted by applying the Swinburn rather than the Woehl utility values, and assuming a 1‑year stopping rule, the ICER for vedolizumab was either less than £20,000 per QALY gained relative to its comparators or it was a dominant treatment option (that is, more effective and less costly than its comparators). The Committee concluded that, taking into account the uncertainty of the utility values and the costs of surgery and post‑surgery care, vedolizumab for people who had not had TNF‑alpha inhibitors before was likely to be a cost‑effective use of NHS resources if a stopping rule was applied and if vedolizumab was provided to the NHS at the price agreed in the patient access scheme.
4.18 The Committee discussed the revised cost‑effectiveness estimates submitted by the company for people in whom treatment with a TNF‑alpha inhibitor had failed. The Committee was aware that these analyses included the revised patient access scheme, a 10‑week response assessment time and also included some amendments to the model suggested by the ERG (see section 3.41). The Committee considered that these changes were improvements to the company's previous analyses. The Committee noted that, depending on the source of utility values, the ICERs for vedolizumab compared with conventional therapy were £27,500 (Swinburn et al.), £31,900 (Woehl et al.) and £37,000 (base‑case utilities) per QALY gained. The Committee noted comments received during consultation that it may be possible to identify a subgroup of patients who would be likely to respond to vedolizumab despite failure of a TNF‑alpha inhibitor, but heard from the clinical expert that this was unlikely to be achievable in clinical practice. The Committee expressed a preference for the use of published utility values from Woehl et al. or Swinburn et al. to those used in the company's base case and noted that the corresponding ICERs were around the upper limit of the range normally considered to be a cost‑effective use of NHS resources. On balance, the Committee concluded that taking into account the uncertainty of the utility values and the costs of surgery and post‑surgery care, vedolizumab for people in whom TNF‑alpha inhibitors had failed could be considered a cost‑effective use of NHS resources if a stopping rule was applied and if vedolizumab was provided to the NHS at the price agreed in the patient access scheme.
4.19 The Committee considered whether vedolizumab was an innovative treatment. It noted that vedolizumab has a different mechanism of action to other drug treatment options for ulcerative colitis. It further noted that because vedolizumab suppresses immune activity only in the gut, this was a step‑change in the management of ulcerative colitis because other immunosuppressants affect immune activity in the whole body. The Committee noted that the clinical experts had stated that the benefits of targeted immunosuppression with vedolizumab may not have been fully seen in GEMINI I because some people had vedolizumab plus a systemic immunosuppressant. The Committee concluded that vedolizumab is an innovative technology, and that some of its benefits, such as its targeted immunosuppression, might not be fully captured in the model. The Committee noted that the impact of any such benefits could not be quantified with the available data.
4.20 The Committee was aware that surgery for ulcerative colitis may reduce fertility, which may disadvantage people who are yet to have a family. The Committee agreed that drug treatments rather than surgery were the main comparators for vedolizumab and therefore this was not an equalities issue in this appraisal.