There are no studies comparing the cost-effectiveness of biological therapies for plaque psoriasis in children and young people. Furthermore, none of the companies participating in this appraisal have submitted an economic evaluation in this population. Our review of previous NICE TAs of plaque psoriasis in adults was conducted to examine existing decision-analytic models and identify important structural assumptions and highlight key areas of uncertainty and the potential issues associated with generalising evidence from the adult population to a population of children and young people. In this section we summarise the key areas of uncertainty identified in adults in light of potential implications for thede novomodel in children and young people.
Model structure
Although in clinical practice, treatment with biological therapy is expected to be sequential, that is, patients are switched to further lines of biological therapy on failure of the first-line biological therapy, the majority of the TAs did not consider treatment sequencing. Lack of evidence to inform treatment sequencing, especially on the efficacy of the treatments depending on the previous therapies received, appeared to be the main reason for not formally modelling treatment sequences in all but one appraisal (TA368102). Given that there is very limited evidence to support the cost-effectiveness of the sequential use of treatments in adults and that no evidence exists in children and young people (seeChapter 3), any attempt to model treatment sequences in the population of children and young people will be highly uncertain.
Clinical effectiveness evidence
Because of a lack of head-to-head trials comparing the biological treatments with each other, NMA was used to compare the treatments with each other indirectly. There was concern that not all trial populations matched those of the decision problem because of variation in the inclusion criteria, with some trials not explicitly excluding individuals who had not failed non-biological systemic therapy. Placebo or BSC was not defined consistently across the trials, which introduced heterogeneity in placebo response rates. Similar issues were identified in the clinical effectiveness evidence for children and young people (seeChapter 3), with the evidence base even more sparse, with only three RCTs and no common comparator across the trials.Chapter 4describes how NMA was used to expand the evidence base in children and young people by drawing strength from the full network of evidence available for adults, while attempting to account for heterogeneity between trial populations (i.e. children and young people vs. adults) and placebo response rates.
ASSESSMENT OF EXISTING COST-EFFECTIVENESS EVIDENCE
NIHR Journals Library www.journalslibrary.nihr.ac.uk
Long-term response and withdrawal rates
To extrapolate data beyond the clinical trials, previous appraisals in adults have assumed that responders to treatment maintain their PASI response rate over time until treatment withdrawal. The same all-cause withdrawal probability of 20% per annum has been assumed for all biological therapies in the absence of any long-term withdrawal data. Given the paucity of long-term data in children and young people, this parameter will also be uncertain in this population.
Health-related quality of life
Most of the previous TAs in adults used utility values based on an estimate of the relationship between PASI response rates and changes in DLQI scores mapped onto EQ-5D utility values. Although some TAs applied EQ-5D data collected directly in RCTs, this was limited to data collected in the trials sponsored by the companies and no evidence synthesis methods were used to synthesise the utility estimates. The estimates of utility gains from treatment were variable across subgroups of patients defined by baseline DLQI score, with greater gains achieved for individuals with worse baseline HRQoL. The size of the utility gains in previous appraisals was considered to be largely uncertain and it represented a key driver of the cost-effectiveness results. It is expected that utility gains associated with treatment will also be highly uncertain in the population of children and young people because of an absence of EQ-5D data in this population. InChapter 6, a review of HRQoL data in children and young people is reported. Scenario analyses are used to explore the impact of uncertainty on the cost-effectiveness results.
Resource use and costs
The resource use and costs associated with BSC has been one of the key drivers of cost-effectiveness in adult appraisals. In particular, the duration of and costs associated with inpatient hospitalisation stays for individuals who do not respond adequately to treatment have been highly uncertain. Until the publication of CG153,11the resource use and costs associated with BSC in adult TAs were largely informed by assumptions and expert opinion. The two TAs that followed the guideline (TA350101and TA368102)
supplemented it with resource use data from cohort studies of patients treated for psoriasis with biological treatments. However, the patient population from whom the data were collected was likely to reflect a sicker population than that defined by the NICE scope for these appraisals and the uncertainty associated with the estimates was not sufficiently explored. The search described inChapter 5(seeMethods) did not identify any evidence on the resource use and costs involved in BSC in children and young people. The use of evidence in adults supplemented by clinical expert opinion to inform the costs of BSC in children and young people is discussed in depth inChapter 6. Scenario analyses are used to explore the implications for the cost-effectiveness results of uncertainty in the assumptions made about BSC, particularly in relation to hospitalisation LOS.
Each of these areas of uncertainty is considered in more detail in the following chapter as part of the decision-analytic model developed to evaluate the cost-effectiveness of adalimumab, etanercept and ustekinumab in children and young people.
DOI: 10.3310/hta21640 HEALTH TECHNOLOGY ASSESSMENT 2017 VOL. 21 NO. 64
© Queen’s Printer and Controller of HMSO 2017. This work was produced by Duarteet al.under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.
Chapter 6
Independent economic assessment
Introduction
The review of cost-effectiveness evidence in the population of children and young people and the absence of company models highlights the challenges of developing an economic model in this population. The fundamental challenge is the limited clinical evidence base for both short- and long-term outcomes to inform a model. Therefore, any estimation of the cost-effectiveness of biological therapies in children and young people will be subject to a number of uncertainties. These uncertainties cannot be avoided but a clear and transparent approach, that highlights the assumptions entering the economic model, can be pursued to help the decision-maker assess the cost-effectiveness of biological therapies in this population.
Plaque psoriasis is a chronic non-progressive disease that manifests itself in children and young people in a similar manner to that in adults. The main difference between the younger population and adults is the presence of comorbidities in adults (such as high blood pressure, liver impairment and renal impairment), which tend to make adults less well with psoriasis than a younger population. Currently, there is no treatment pathway specific to psoriasis in children and young people in the UK. The management of treatment and approach to care seems to mirror that used in adults. Our clinical advisor, Dr Ruth Murphy, indicated that when there is an absence of evidence it would be reasonable to extrapolate data from the adult population to children and young people. The company submission for ustekinumab also supports this approach for the development of an economic model given that there are few significant differences in the posology or management of chronic plaque psoriasis in children, young people and adults.
The management and treatment of plaque psoriasis depends on the extent and severity of an individual’s disease, local custom and practice. If an individual patient does not respond to or tolerate a particular treatment option, an alternative one is usually tried. This means that treatments are usually‘trialled’on an individual basis until an effective option is found. If an effective treatment is not found, then a patient will receive some form of BSC. This approach to treatment appears to be the same for children and young people and adults, but usually more caution is exercised in the younger population because of the limited availability of licensed treatment options.
The trialling of treatments on treatment failure or intolerance suggests that sequences of treatments could be considered in an cost-effectiveness model, whereby after failure of a first treatment option patients are trialled on a second option and so on, until all options are exhausted. However, this would require additional clinical evidence on the efficacy of the treatments conditional on the previous therapies received. There is very limited evidence to support the cost-effectiveness of the sequential use of treatments in adults and no evidence exists in children and young people (seeChapter 3). Therefore, although the model should ideally explore the sequential use of treatments, any attempt to do so in the population of children and young people would be highly uncertain. Furthermore, the optimum treatment sequence may not be suitable for an individual patient with specific characteristics and when treatment in this population is usually tailored to the child or adolescent because of needle phobia or the presence of psoriatic arthritis. Therefore, an alternative approach may be better whereby the optimum ordering of treatments, in terms of their cost-effectiveness, is established. This can be achieved by comparing each of the alternative treatment options with BSC and then indicating the most cost-effective order in which to give the therapies based on total expected costs and QALYs associated with each treatment option.
The previous York model appears to be the most widely accepted model of chronic plaque psoriasis.97 The five NICE TAs that followed TA103 for the treatment of moderate to severe psoriasis in adults98–102 followed the framework of the York model and these have been accepted by NICE as being relevant to
DOI: 10.3310/hta21640 HEALTH TECHNOLOGY ASSESSMENT 2017 VOL. 21 NO. 64
© Queen’s Printer and Controller of HMSO 2017. This work was produced by Duarteet al.under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.
plaque psoriasis. The main changes that have followed since the advent of the York model have been the availability of new evidence, the methodology for linking efficacy estimates to HRQoL utility values, the parameters used in the model to inform BSC, updated unit costs, time on treatment and the modelling of treatment sequences in the most recent appraisal of apremilast.102It would therefore seem appropriate that the same modelling framework is used for children and young people but with an evidence base informed by outcomes in the younger population. Hence, the structure of our model is very similar to that used in previous TAs in adults and, when evidence is lacking or limited in the population of children and young people, data have been extrapolated from the adult population and supplemented by expert opinion.