Skip Navigation

Rheumatology 2007 46(11):1729-1735; doi:10.1093/rheumatology/kem221
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Bravo Vergel, Y.
Right arrow Articles by Sculpher, M. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bravo Vergel, Y.
Right arrow Articles by Sculpher, M. J.
Related Collections
Right arrow Spondylarthropathies
Right arrow Pharmacology
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2007. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

The cost-effectiveness of etanercept and infliximab for the treatment of patients with psoriatic arthritis

Y. Bravo Vergel, N. S. Hawkins, K. Claxton, C. Asseburg, S. Palmer, N. Woolacott1, I. N. Bruce2 and M. J. Sculpher

Centre for Health Economics, 1Centre for Reviews and Dissemination, University of York, YO10 5DD York and 2ARC Epidemiology Unit, University of Manchester, M13 9PT Manchester, England, UK

Correspondence to: Y. Bravo Vergel, Centre for Health Economics, University of York, Heslington, YO10 5DD York. E-mail: yb3{at}york.ac.uk


   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 Acknowledgements
 References
 
Objective. Tumour necrosis factor (TNF) antagonists have been shown to improve the outcomes in patients with rheumatoid arthritis (RA) and psoriatic arthritis (PsA). We assess the cost-effectiveness of two TNF antagonists and so-called ‘palliative care’ for the treatment of active PsA from the perspective of the UK National Health Service (NHS).

Methods. Bayesian statistical methods were used to synthesize evidence from three Phase III trials, identified through a systematic review, and estimate the relative efficacy of etanercept, infliximab and palliative care. A probabilistic decision analytic model was then used to compare these treatments after the failure of at least two conventional disease-modifying anti-rheumatic drugs (DMARDs), following the British Society for Rheumatology (BSR) guidelines for use. The primary outcome measure, quality-adjusted life years (QALYs), was derived from utility values estimated as a function of disability measured by the Health Assessment Questionnaire (HAQ). The deterioration experienced in HAQ at treatment withdrawal (rebound) was incorporated using alternative scenarios to represent best- and worst-case assumptions. The model was extended beyond the trial duration to a 10-yr and lifetime horizon, using available evidence and expert opinion-based assumptions on disease progression. Resource utilization was based on literature, national databases and expert opinion. Prices were obtained from routine NHS sources and published literature.

Results. At a 10-yr time horizon, the incremental cost-effectiveness ratio (ICER) for etanercept compared with palliative care was £26 361 per QALY gained for the best-case rebound scenario, which increased to £30 628 for the worst-case. The ICERs for infliximab compared with etanercept were £165 363 and £205 345 per QALY, respectively. These findings are mainly explained by the fact that infliximab has higher acquisition and administration costs without substantially superior effectiveness compared with etanercept. Results were sensitive to estimates of rebound assumptions at withdrawal and the time horizon.

Conclusions. Only results for etanercept remained within the range of cost-effectiveness estimates considered to represent value for money in the NHS by the National Institute for Health and Clinical Excellence. Further research appears most valuable in relation to the short-term effectiveness, utility parameters and assumptions regarding the effect of rebound.

KEY WORDS: Cost-effectiveness, Etanercept, Infliximab, Psoriatic arthritis, Bayesian evidence synthesis


   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 Acknowledgements
 References
 
Psoriatic arthritis (PsA) is an inflammatory arthritis closely associated with psoriasis [1, 2], which can result in significant health-related quality of life (HRQL) impairment as a result of joint deformity and psychosocial disability [2, 3]. It is a chronic and progressive disorder, ranging from mild synovitis to severe progressive erosive arthropathy [4], with a significant percentage of patients, especially those with polyarticular disease, developing joint damage and deformities [5, 6]. Effective treatment for PsA needs to consider both skin and joint disease, especially if both are affected significantly. Most patients will require therapy with disease-modifying anti-rheumatic drugs (DMARDs) in order to control disease progression, but their use is largely derived from analogy from rheumatoid arthritis (RA) management as well as a limited clinical trial evidence base in PsA itself [3]. Such traditional DMARDs are associated with several problems, including a slow onset of action and the requirement for close patient monitoring due to associated toxicities [7–11].

Newer strategies for the treatment of PsA and the most common forms of rheumatic diseases [12] have focused on blockade of key inflammatory cytokines, in particular, tumour necrosis factor (TNF), which has been implicated in the pathogenesis of both psoriasis and PsA [13, 14]. Etanercept (Enbrel®) and infliximab (Remicade®) are TNF antagonists able to alter the immune function and in particular to inhibit inflammatory response [3, 15]. Both have been licenced in Europe for the treatment of active and progressive PsA in adults not responding adequately to conventional therapy. Based on existing Phase III randomized control trials [16–18], they claim to combine greater efficacy with less toxicity, but their acquisition costs are substantially higher than conventional DMARD therapy. It remains unclear whether a potential reduction of disease progression in responders could offset other possible significant downstream costs associated with the disabling nature of the condition. Cost-effectiveness evaluations are particularly pertinent in this kind of context, and although the cost-effectiveness of TNF antagonists is well-established in patients with RA [19–22], to date there is limited evidence of their cost-effectiveness for the treatment of PsA [23].

The aim of this study was to first estimate the relative efficacy of etanercept, infliximab and the so-called ‘palliative care’ for the treatment of active PsA. Then, we evaluate the long-term cost-effectiveness of TNF antagonists compared with palliative care using a probabilistic decision analytical model [24, 25], according to their licenced indications and from a UK National Health Service (NHS) perspective. All the relevant clinical evidence was combined using Bayesian evidence synthesis methods, which allowed for the estimation of the relative efficacy of etanercept and infliximab despite the absence of head-to-head trials. This project was commissioned by the National Coordinating Centre of Health Technology Assessment (NCCHTA) on behalf of the National Institute for Health and Clinical Excellence (NICE) [26].


   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 Acknowledgements
 References
 
Overview
A probabilistic decision analytical model was constructed in order to compare the cost-effectiveness of etanercept, infliximab and palliative care for the treatment of active PsA, in the context of their licenced indications [27, 28] and the British Society for Rheumatology (BSR) guidelines for the use of anti-TNF-{alpha} drugs for the treatment of PsA [29]. A systematic review was conducted to identify relevant clinical and cost-effectiveness studies [26], and Bayesian evidence synthesis methods were used to synthesize the effectiveness evidence [16–18].

The measure of benefits used for the economic evaluation were quality-adjusted life years (QALYs). The structure of the decision analytic model and its underlying assumptions were developed in consultation with a group of clinical experts. A health care service perspective was adopted, with costs and benefits discounted at a 6 and 1.5% annual rate [30], respectively. The price year was 2004/05. Full details of the Bayesian evidence synthesis and the model are available elsewhere [26].

Model structure
A cohort model in the form of a modified-decision tree was constructed. A simplified version of the model structure is shown in Fig. 1. The evidence base which is used to populate the model is discussed separately in the following sections.


Figure 1
View larger version (15K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 1. Simplified version of the model structure.

 
Short-term trial data is used to model the response of patients to TNF antagonists. Initial response to treatment at 3 months (measured by the Psoriatic Arthritis Response criteria, PsARC) and functional status (measured by the HAQ) are combined as efficacy outcomes, using Bayesian evidence synthesis methods. Beyond this period, the underlying Markov structure [31] allows the extrapolation of results up to 10 yrs and for a lifetime horizon (assumed equivalent to 40 yrs) using annual cycles. Given the lack of long-term efficacy and safety data on the use of TNF antagonists and the limited experience on the administration of these drugs for PsA and RA, and in conjunction with expert advice, a 10-yr horizon was applied in the base-case analysis.

For those who respond, there is an ongoing risk of withdrawal at any time period. Initial or later treatment failures were assumed to move on to palliative care, with TNF antagonists being the ‘end of the line’ in terms of active interventions and palliative care assumed to be equivalent to placebo. The progressive nature of the condition was modelled using a natural history progression rate, measured in terms of HAQ. At withdrawal, patients experience some deterioration in HAQ (rebound). In the absence of clinical trial data to characterize this effect, the model considers two alternative rebound scenarios, as illustrated in Fig. 2.


Figure 2
View larger version (14K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 2. Illustration of alternative rebound scenarios.

 
In the rebound to natural history progression, the HAQ after treatment failure returns back to the level and subsequent trajectory it would have been if the patient had not initiated treatment (i.e. TNF antagonists only provide symptomatic relief). In the rebound equal to gain, the patient's disability in terms of HAQ deteriorates by the same amount it improved when first responded to treatment (i.e. TNF antagonists can re-set the curve and delay disease progression). Those patients not responding to TNF antagonists at 3 months or withdrawing from active treatment at any time period in the model also experience deterioration in HAQ in line with the natural progression estimates.

Those patients who respond to TNF antagonists experience an initial gain in HAQ, based on results from the evidence synthesis. In addition to this initial improvement, these patients were also assumed to experience a slower disease progression. In the absence of data, it was assumed that TNF antagonists can halt disease progression in responders. Patients are at risk of all-cause mortality at every time period in the model.

Bayesian evidence synthesis
The existing trials [16–18] included placebo as controls but no study has directly compared etanercept and infliximab, so an indirect approach was required. A Bayesian evidence synthesis enabled the comparison of the two anti-TNF-{alpha} therapies with each other based on all available trial evidence using indirect comparison methods [32, 33], as the network of evidence allowed for a treatment comparison through the placebo option. The evidence synthesis consisted of two linked meta-analyses that estimated the respective response rates of etanercept and infliximab on the one hand, and the mean reductions (i.e. improvement) in HAQ score conditional on response to treatment on the other. The placebo effect was also taken into account, and so the response rate and mean reduction in HAQ score of the placebo option was also estimated in the evidence synthesis. This data was then used to adjust the active treatments’ efficacy by accounting for the placebo effect observed in the trials [26].

The absolute change in HAQ conditional on response from the Mease et al. [17] and IMPACT [18] trials was made available at request by the pharmaceutical companies. Data from one unpublished cohort study of PsA patients based at the Academic Unit of Musculoskeletal Disease, University of Leeds [34], was used to inform the change in HAQ score experienced by patients not taking any active treatment (i.e. natural progression).

Between-trial variability was modelled using a random effect baseline for the probability of PsARC response to placebo and natural progression. Etanercept and infliximab's treatment effects on the probability of PsARC response were modelled using a fixed-effect model, additive to the placebo probability of response on the log-odds scale. The treatment-specific change in HAQ score for responders and non-responders was also modelled as a fixed effect, incremental to the natural progression baseline. The evidence synthesis also estimated the change in HAQ for non-responders to either treatment, as we were aware that PsARC does not fully capture treatment success. The mean HAQ change for non-responders to etanercept and infliximab were also incorporated in the model at the initial 3-month period. Table 1 lists the four parameters modelled in the Bayesian evidence synthesis. See Woolacott et al. [26] for further details.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Clinical effectiveness parameters based on the results of the Bayesian evidence synthesis

 
Model inputs
The marginal posterior distributions for the clinical effectiveness parameters of interest are summarized in Table 1. The full posterior distributions, which preserve the information on distributional shape and parameter correlations, were used to populate the cost-effectiveness model.

The probability of long-term failure from 3 to 20 months as reported in Geborek et al. [35] was used to estimate the annual probability of withdrawal for both etanercept and infliximab. Mortality rates were based on standard UK age- and sex-specific life-tables [36] adjusted by the higher mortality risk associated with PsA [19]. There is no evidence of a differential mortality risk between therapies.

Apart from the cost of the TNF antagonists themselves (including drug acquisition, administration and monitoring), the direct costs associated with PsA were estimated as a function of patients’ HAQ score based on a published study on RA [37]. These costs were considered to incorporate the costs associated with palliative care. No additional costs were assigned to the palliative care strategy in order to avoid the potential for double-counting this cost for severe patients [26]. Drug costs were taken from the British National Formulary [38] based on licenced dosages. Other unit costs were obtained from national databases [39, 40] and expert opinion (I. Bruce, personal communication). The relationship between HRQL (in terms of utility) was also estimated as a function of HAQ, based on published results derived from a linear-regression model using individual patient data on HAQ and EQ-5D from a cohort of PsA patients [23].

Results of the analysis are conditional on three specific features of the patient cohort under treatment: the baseline HAQ, age at treatment initiation and the patient's weight, which determined the number of infliximab dosages required. The former were estimated based on the average HAQ score and age from the three trials included in the analysis [16–18], the latter was based on patients’ mean weight in the Infliximab Multinational Psoriatic Arthritis Controlled Trial (IMPACT) [18], which is equivalent to four infliximab vials per infusion. See Table 2 for a summary of the base values of the main input parameters used in the model.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Summary of other parameters associated with the base-case model

 
Analytical methods
The uncertainty in the individual parameters was fully characterized using the full posterior distributions of the clinical effectiveness parameters and the probability distributions summarized in Table 2; probabilistic analysis was conducted using Monte Carlo simulation. The results of the model are presented in two ways. Firstly, mean lifetime costs and QALYs for the three strategies are reported and their cost-effectiveness compared, estimating incremental cost-effectiveness ratios (ICERs) using standard decision rules [41]. Secondly, decision uncertainty is presented in terms of the probability that each intervention is considered cost-effective for a given cost-effectiveness threshold. The model was implemented in Microsoft Excel and the specialist software WinBUGS [42].

A series of alternative scenarios is also presented to assess the implications of other uncertainties in the model. Only those that demonstrated a relevant impact on results are reported here [26]. These include running the model for alternative rebound scenarios (a best-case scenario if rebound is equal to gain and a worst-case scenario for rebound equal to natural history progression), and for a lifetime horizon; using a 3.5% annual discount rate for both costs and QALYs and three infliximab vials per infusion. Only results for males are presented here, as differences by gender were demonstrated to be minor [26].


   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 Acknowledgements
 References
 
Clinical effectiveness
Table 1 shows the marginal posterior distributions for the clinical effectiveness parameters. The mean posterior probability of responding to therapy measured in terms of PsARC criteria was estimated to be 0.725 (S.D. 0.089) for etanercept and 0.839 (S.D. 0.075) for infliximab. In other words, the evidence synthesis suggested that the two treatments were very similar with respect to initial response rates at 3-month therapy, although infliximab had a slightly higher probability of response.

The evidence synthesis results showed that responders to etanercept and infliximab experienced an improvement measured in terms of HAQ scores, compared with placebo. Incremental to the natural progression baseline HAQ of 0.016 (S.D. 0.007), responders to etanercept treatment experienced a reduction in HAQ of –0.718 (S.D. 0.054), and responders to infliximab treatment of –0.667 (S.D. 0.090). Both these HAQ changes were remarkably different from the placebo effect, estimated as a reduction of –0.280 (S.D. 0.057) in HAQ change.

Base case cost-effectiveness results
Table 3 presents results for a 10-yr time horizon for both rebound scenarios. For the best-case rebound scenario, the mean number of QALYs per average patient was 4.51 for etanercept, 4.64 for infliximab and 3.25 for palliative care. The total mean costs per patient were £44 111 for etanercept, compared with £64 274 for infliximab and £10 718 for palliative care. As expected, for the worst-case rebound scenario, both etanercept and infliximab showed higher costs and fewer benefits: the mean number of QALYs per average patient was 4.36 for etanercept, 4.46 for infliximab and 3.26 for palliative care. The total mean costs per patient were £44 169 for etanercept, compared with £64 418 for infliximab and £10 679 for palliative care. Differences in palliative care results are only due to probabilistic variation.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Cost-effectiveness results

 
The resulting ICER for etanercept compared with palliative care was £26 361 per QALY gained for the rebound equal to gain scenario (best-case) which increased to £30 628 if rebound after treatment failure is assumed to be natural history progression (worst-case). Results for etanercept would be in the cost-effectiveness range of £20 000–£40 000 per QALY, which has been considered to provide value for money in the NHS [43].

Infliximab shows a very high ICER compared with etanercept, which varies between £165 363 per QALY gained for the best-case scenario to £205 345 for the worst-case. These results are explained by the fact that infliximab has higher acquisition and administration costs [estimated as a half-day case per infusion in a rheumatology department based on expert opinion (I. Bruce, personal communication) compared with self-injection costs for etanercept] with only a marginal improvement in effectiveness (a QALY gain over etanercept of 0.12 over 10 yrs for the best-case scenario and of 0.09 for the worst-case).

Etanercept and palliative care showed the highest probability of being cost-effective for a range of thresholds. Etanercept has the highest probability of being cost-effective for a threshold of £30 000–£40 000 for the rebound equal to gain scenario (0.69 and 0.93, respectively) and at a £40 000 threshold for the rebound to natural history scenario (0.88). At lower levels of the threshold, palliative care showed the highest probability of being cost-effective. Even for a £40 000 threshold, infliximab has little likelihood of being cost-effective: 0.009 and 0.005 for the best-case and worst-case scenario, respectively.

Alternative scenarios
Table 4 presents results for a lifetime horizon for both rebound scenarios. As expected, the longer the time horizon the greater the cumulated costs but also the higher the mean total QALYs. This results in a reduction in the ICERs for both technologies, independently of the rebound assumption. The ICER for etanercept for a lifetime horizon was £16 891 per QALY gained for the best-case scenario which increased to £27 805 for the worst-case. Etanercept showed the highest probability of being cost-effective for a threshold of £30 000–40 000 for both rebound scenarios. At the lower threshold of £20 000, palliative care showed the highest probability of being cost-effective (0.26 for rebound equal to gain and 0.96 for rebound to natural history). Results for infliximab provided ICERs which were still above the range considered cost-effective according to the conventional thresholds [43], with an ICER ranging between £84 473 per QALY gained for the best-case scenario to £168 753 for the worst-case. The highest probability of infliximab being cost-effective was 0.16 for a £40 000 threshold for the best-case scenario.


View this table:
[in this window]
[in a new window]

  		TABLE 4. Results of a scenario analysis to assess the sensitivity of model results to alternative assumptions

 
Further scenario analyses were conducted to assess the implications of using different annual discount rates and average number of vials per infusion. Using the alternative discount rate of 3.5% results in higher costs and lower QALYs for all options, and a slightly higher ICER for etanercept and infliximab. Using the alternative assumption of three infliximab vials results in a lower ICER for infliximab whilst etanercept still shows the highest probability of being cost-effective for a threshold of £30 000–40 000. Differences by rebound scenario and for a lifetime horizon showed the same trends and did not affect the ranking of treatments, so only results for the base-case 10-yr horizon are presented.

Value of information
The aforementioned results indicate that there is uncertainty surrounding the choice between etanercept and palliative care. The expected cost of this uncertainty (the probability of error and the health and resource consequences of error) can be interpreted as the expected value of perfect information (EVPI), since perfect information can eliminate the possibility of making the wrong decision [44–46]. In the context of research prioritization, the EVPI represents the maximum expected value of additional research to inform this decision. The population EVPI for the two rebound scenarios for the base-case 10-yr time horizon is reported in Table 5, and shows that further research to inform this decision could be worthwhile.


View this table:
[in this window]
[in a new window]

  		TABLE 5. Population EVPI and EVPPI for both rebound scenarios

 
The value of information associated with particular uncertain parameters within the model (EVPPI) is also presented in Table 5. The value of information associated with the short-term effectiveness was between £25.39 million and £28.96 million (for the best and worst rebound scenario, respectively), which is substantially higher than for any other element of evidence. The second highest value of information was that associated with the utilities (between £11.92 million and £10.73 million).


   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
Acknowledgements
 References
 
The results show that only results for etanercept remained within the conventional cost-effectiveness range of treatments to be recommended by NICE. These findings are mainly explained by the fact that infliximab showed higher acquisition and administration costs without practically superior effectiveness. The cost of the therapy itself is the main factor contributing to the cost difference between both strategies: for a 10-yr follow-up, the cumulative therapeutic costs as a proportion of cumulative total costs was 79.11% for etanercept compared with 84.04% for infliximab. These proportions were practically identical under both rebound scenarios.

Based on research undertaken during a similar period to our own, Bansback et al. [23] have also demonstrated the cost-effectiveness of etanercept after the failure of two conventional DMARDs, using patient data from one of the same Phase III etanercept trials incorporated in our Bayesian evidence synthesis [17]. Despite differences in the modelling approach and the use of sequences, the result in that study for the base-case analysis at 10-yr horizon for the comparison of etanercept against the least effective of the two strategies (ICER of £28 000 per QALY gained for the combination therapy of methotrexate plus cyclosporine, followed by palliative care after failure) can be compared with those presented here: an ICER of £26 361 for the best-case and £30 628 for the worst-case rebound scenario, respectively.

There are a number of issues that should be kept in mind when interpreting these results. First, the model considers the cost-effectiveness of etanercept and infliximab compared with each other and with palliative care. This is equivalent to assuming that TNF antagonists would be used ‘end of line’ once DMARD therapies have been tried and failed. A strict interpretation of the licences of etanercept and infliximab would suggest that all DMARDs should have been tried prior to their use. In practice, there are only four DMARDs routinely used in PsA (sulphasalazine, methotrexate, cyclosporine and leflunomide), of which only one (leflunomide) is actually licenced for PsA treatment. Even if a strict interpretation of the licences is not applied, it remains unclear how many DMARDs really should be tried; however, the recent BSR guidelines for the use of anti-TNF agents in PsA advise that at least two DMARDs should be used prior to considering anti-TNF therapy [29].

Second, the model was not able to incorporate the potential HRQL impact of TNF antagonists on psoriasis. The extent of psoriasis involvement could not only have an impact on patient's quality of life but also be an important cost driver in the most severely affected patients. However, to date, no validated measure exists that can incorporate both the impact of the arthritic and skin component of the disease on the quality of life of PsA patients. The use of HAQ as the measure of disability that drives both quality of life and direct costs is consistent with the main economic evaluations published on the use of TNF antagonists in rheumatic diseases [19–22].

The last issue relates to the lack of long-term data on the use of TNF antagonists. Potential severe adverse events have not been incorporated in the model so results should be interpreted with caution. For the base-case scenario, short-term effectiveness results were extrapolated up to 10 yrs as a reasonable assumption based on expert clinical advice (I. Bruce, personal communication), but the reality is that there is limited experience on the administration of TNF antagonists both for PsA and other rheumatic conditions, so the number of years a patient can be safely on these drugs is uncertain. In our model, patients still responding at the end of the 10-yr time horizon were 22.7% for those on etanercept and 26.2% for those on infliximab, approximately.

In general, the population EVPI in this area is substantial and suggests that further research may be potentially worthwhile. The EVPI associated with the model parameters suggests that, if further research is commissioned, it will require experimental design and should focus on the short-term effectiveness parameters. However, additional research on quality of life in PsA would also be valuable. This would not necessarily require experimental design and could be conducted relatively quickly and cheaply.

Cost-effectiveness results were sensitive to the rebound assumption and time horizon. Alternative structural assumptions, such as rebound effect and HAQ progression whilst responding to treatment, have also demonstrated an important impact on the value of information [48]. The estimates of population EVPPI reported before have not formally integrated uncertainty about the nature of the rebound effect. An indicative sensitivity analysis is possible by assuming equipoise between the rebound equal to gain and rebound back to natural history (i.e. a 50:50 chance). This results in an ICER of £29 348 per QALY gained for etanercept for a 10-yr horizon, that is, halfway to those reported for the alternative rebound scenarios, with infliximab still showing the lowest probability of being cost-effective for a range of thresholds. The relative importance of the value of information results for the different parameters remains approximately the same. However, the short-term effectiveness parameters showed a value of information higher than previously reported, whilst for utilities it was lower (£44.27 million compared with £25.39 million; and £913 556 compared with £11.92 million for a 10-yr horizon, respectively). The population EVPPI for the rebound parameter itself shows a value of £6.89 million. These results, which were based on a 50:50 chance for each rebound scenario, should be interpreted with caution and more formal methods of elicitation could in principle be used to parameterize these uncertainties. However, these results indicate that a potentially high value of information could be associated with the nature of the rebound effect and further research about this uncertainty may well be worthwhile.

To date, this is the first probabilistic decision analytic model examining the cost-effectiveness of both etanercept and infliximab for the treatment of active PsA, according to their licensed indications and from a UK NHS perspective. Within the data and modelling limitations recognized, this study demonstrates the potential cost-effectiveness of etanercept for the treatment of PsA in adults after the failure of DMARD therapies, compared with infliximab and palliative care and for a 10-yr and lifetime horizon. The results of this model, alongside other relevant evidence, was used by NICE in developing their guidance on the use of etanercept and infliximab for PsA in the NHS [49]. The subsequent recommendations clearly reflect the cost-effectiveness of the different treatments: while etanercept received a positive recommendation for the treatment of PsA, infliximab was only recommended for patients intolerant (or contraindicated) to etanercept or who have difficulty with self-administered injections.

Formula


   Acknowledgements
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Acknowledgements
References
 
The authors gratefully acknowledge the contribution of the reviewers from the Centre for Reviews and Dissemination and our other Expert Advisor Dr Robert Chalmers, Consultant Dermatologist. We would also like to thank Prof. Tony Ades for his advice on the statistical analysis. This project was funded by the Health Technology Assessment Programme (project number 04/04/01). The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Department of Health. M.J.S. has also received funding via a Career Award in Public Health funded by the UK NHS Research and Development Programme. The views expressed are those of the authors who are also responsible for any errors.

Disclosure statement: M.J.S. has a minority shareholding in a consultancy company which has worked for Schering Plough but outside the area of psoriatic arthritis. N.S.H. has received fees for consultancy from Schering Plough, the manufacturer of infliximab.


   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Acknowledgements
 References
 

  1. Patel S, Veale D, FitzGerald O, McHugh NJ. Psoriatic arthritis-emerging concepts. Rheumatology (2001) 40:243–6.[Free Full Text]
  2. Ruderman EM. Evaluation and management of psoriatic arthritis: the role of biologic therapy. J Am Acad Dermatol (2003) 49(Suppl):2A:S125–32.
  3. Galadari H, Fuchs B, Lebwohl M. Newly available treatments for psoriatic arthritis and their impact on skin psoriasis. Int J Dermatol (2003) 42:231–7.[CrossRef][Web of Science][Medline]
  4. Pipitone N, Kingsley GH, Manzo A, Scott DL, Pitzalis C. Current concepts and new developments in the treatment of psoriatic arthritis. Rheumatology (2003) 42:1138–48.[Abstract/Free Full Text]
  5. Krueger GG. Clinical features of psoriatic arthritis. Am J Manag C (2002) 8(6 Suppl):S160–70.
  6. Kane D, Stafford L, Bresniham B, FitzGerard O. A prospective, clinical and radiological study of early psoriatic arthritis: an early synovitis clinic experience. Rheumatology (2003) 42:1460–8.[Abstract/Free Full Text]
  7. Dukes M, Aronson J. Meyler's; side effects of drugs: an encyclopedia of adverse reactions and interactions. (2000) 14th. Amsterdam: Elsevier.
  8. British Medical Journal. Clinical Evidence. http://clinicalevidence.bmj.com.
  9. United States Pharmacopoeial Convention. USPDI (2004) 1. Drug Information for the Health Care Professional. Rockville, MD: United States Pharmacopoeial Convention.
  10. Whiting-O’Keefe QE, Fye KH, Sack KD. Methotrexate and histologic hepatic abnormalities: a meta-analysis. Am J Med (1991) 90:711–6.[Web of Science][Medline]
  11. Roenigk HH, Auerbach R, Maibach H, Weinstein G, Lebwohl M. Methotrexate in psoriasis: consensus conference. J Am Acad Dermatol (1998) 38:478–85.[CrossRef][Web of Science][Medline]
  12. Bravo Vergel Y, Torgerson D. Cost-effectiveness of new biologics for rheumatoid arthritis and osteoarthritis. In: Clinical trials in rheumatoid arthritis and osteoarthritis—Reid D, Miller C, Krishan B, eds. (in press) London: Springer-Verlag London Limited.
  13. Parizer DM. Management of moderate to severe plaque psoriasis with biologic therapy. Manag Care (2003) 12:36–44.[Medline]
  14. Gniadecki R, Zachariae C, Calverley M. Trends and developments in the pharmacological treatment of psoriasis. Acta Dermato Venereologica (2002) 82:401–10.[CrossRef][Web of Science][Medline]
  15. Prinz JC. The role of T cells in psoriasis. J Eur Acad Dermatol Venereol (2003) 17:257–70.[CrossRef][Web of Science][Medline]
  16. Mease PJ, Goffe BS, Metz J, VanderStoep A, Finck B, Burge DJ. Etanercept in the treatment of psoriatic arthritis and psoriasis: a randomised trial. Lancet (2000) 356:385–90.[CrossRef][Web of Science][Medline]
  17. Mease PJ, Kivitz AJ, Burch FX, et al. Etanercept treatment of psoriatic arthritis: safety, efficacy, and effect on disease progression. Arthritis Rheum (2004) 50:2264–72.[CrossRef][Web of Science][Medline]
  18. Antoni C, Kavanaugh A, Kirkham B, et al. Sustained benefits of infliximab therapy for dermatologic and articular manifestations of psoriatic arthritis: results from the Infliximab Multinational Psoriatic Arthritis Controlled Trial (IMPACT). Arthritis Rheum (2005) 52:1227–36.[CrossRef][Web of Science][Medline]
  19. Wong JB, Singh G, Kavanaugh A. Estimating the cost-effectiveness of 54 weeks of infliximab for rheumatoid arthritis. Am J Med (2002) 113:400–8.[CrossRef][Web of Science][Medline]
  20. Kobelt G, Jonsson L, Young A, Eberhardt K. The cost-effectiveness of infliximab (remicade) in the treatment of rheumatoid arthritis in Sweden and the United Kingdom based on the ATTRACT study. Rheumatology (2003) 42:326–35.[Abstract/Free Full Text]
  21. Brennan A, Bansback NJ, Reynolds A, Conway P. Modelling the cost-effectiveness of etanercept in adults with rheumatoid arthritis in the UK. Rheumatology (2004) 43:62–72.[Abstract/Free Full Text]
  22. Bansback NJ, Brennan A, Ghatnekar O. Cost-effectiveness of adalimumab in the treatment of patients with moderate to severe rheumatoid arthritis in Sweden. Ann Rheum Dis (2005) 64:995–1002.[Abstract/Free Full Text]
  23. Bansback NJ, Ara R, Barkham N, et al. Estimating the cost and health status consequences of treatment with TNF antagonists in patients with psoriatic arthritis. Rheumatology (2006) 45:1029–38.[Abstract/Free Full Text]
  24. Briggs A, Goeree R, Blackhouse G, O’Brien B. Probabilistic analysis of cost-effectiveness models: choosing between treatment strategies for gastroesophageal reflux disease. Med Decis Making (2002) 33:290–308.
  25. Claxton K, Sculpher M, McCabe C, et al. Probabilistic sensitivity analysis for NICE technology assessment: not an optional extra. Health Econ (2005) 14:339–47.[CrossRef][Web of Science][Medline]
  26. Woolacott N, Bravo Vergel Y, Hawkins N, et al. Etanercept and infliximab for the treatment of psoriatic arthritis. Health Technol Asses (2006) 10. no. 31.
  27. Schering-Plough Ltd. Remicade [Infliximab: summary of product characteristics] (2005) London: Electronic Medicines Compendium.
  28. Wyeth Pharmaceuticals. Enbrel [etanercept: summary of product characteristics] (2004) Electronic Medicines Compendium.
  29. Kyle S, Chandler D, Griffiths C, et al. Guideline for anti-TNF{alpha} therapy in psoriatic arthritis. Rheumatology (2005) 44:390–7.[Free Full Text]
  30. National Institute for Clinical Excellence. Technical guidance for manufacturers and sponsors on making a submission to a Technology Appraisal (2001) London: National Institute for Clinical Excellence.
  31. Briggs A, Sculpher M. An introduction to Markov modelling for economic evaluation. Pharmacoeconomics (1998) 13:397–409.[CrossRef][Web of Science][Medline]
  32. Lu G, Ades AE. Combination of direct and indirect evidence in mixed treatment comparisons. Stat Med (2004) 23:3105–24.[CrossRef][Web of Science][Medline]
  33. Caldwell DM, Ades AE, Higgins JPT. Simultaneous comparison of multiple treatments: combining direct and indirect evidence. Br Med J (2005) 331:897–900.[Free Full Text]
  34. Wyeth Pharmaceuticals. An appraisal submission for the National Institute for Clinical Exellence: etanercept (ENBREL*). (2004) Maidenhead. To establish the clinical and cost effectiveness of the use of etanercept for the treatment of active and progressive psoriatic arthritis in patients who have inadequate response to standard treatment (including disease modifying antirheumatic drug (DMARD) therapy).
  35. Geborek P, Crnkic M, Petersson IF, Saxne T. South Swedish Arthritis Treatment Group. Etanercept, infliximab, and leflunomide in established rheumatoid arthritis: clinical experience using a structured follow up programme in southern Sweden. Ann Rheum Dis (2002) 61:793–8.[Abstract/Free Full Text]
  36. Government Actuary's; Department. Interim life tables 2001–2003. London: Government Actuary's; Department.
  37. Kobelt G, Jonsson L, Lindgren P, Young A, Eberhardt K. Modelling the progression of rheumatoid arthritis: a two-country model to estimate costs and consequences of rheumatoid arthritis. Arthritis Rheum (2002) 46:2310–9.[CrossRef][Web of Science][Medline]
  38. British Medical Association. British National Formulary, No. 46 (2003) London: British Medical Association.
  39. Department of Health. NHS Reference costs (2004.) http://www.dh.gov.uk.
  40. Curtis L, Netten A. Unit costs of health and social care 2004 (PSSRU) (2004.) University of Kent: Kent. http://www.pssru.ac.uk.
  41. Johannesson M, Weinstein S. On the decision rules of cost-effectiveness analysis. J Health Econ (1993) 12:459–67.[CrossRef][Web of Science][Medline]
  42. Spiegelhalter DJ, Thomas A, Best NG, Lunn D. WinBUGS user manual: version 1.4 (2001) Cambridge: MRC Biostatistic Unit.
  43. Raftery J. NICE: faster access to modern treatments? Analysis of guidance on health technologies. Brit Med J (2001) 323:1300–3.[Free Full Text]
  44. Claxton K. The irrelevance of inference: a decision making approach to the stochastic evaluation of health care technologies. J Health Econ (1999) 18:341–64.[CrossRef][Web of Science][Medline]
  45. Claxton K, Ginelly L, Sculpher M, et al. A pilot study on the use of decision theory and value of information analysis as part of the NHS Health Technology Assessment programme. Health Technol Assess (2004) 8:1–103.[Medline]
  46. Claxton K, Sculpher M, Drummond M. A rational framework for decision making by the National Institute for Clinical Excellence. Lancet (2002) 360:711–5.[CrossRef][Web of Science][Medline]
  47. Symmons DPM, Barrett EM, Bankhead CR, et al. The incidence of rheumatoid arthritis in the United Kingdom: results from the Norfolk Arthritis Register. Br J Rheumatol (1994) 33:735–9.[Abstract/Free Full Text]
  48. Bravo Vergel Y, Hawkins N, Asseburg C, Palmer S, Claxton K, Sculpher M. The cost-effectiveness and value of information associated with biologic drugs for the treatment of psoriatic arthritis. In: Poster presented at iHEA (2005) Barcelona. (Available at http://www.york.ac.uk/inst/che/staff/yolanda.htm).
  49. Etanercept and infliximab for the treatment of adults with psoriatic arthritis. In: NICE technology appraisal guidance. (Available at http://guidance.nice.org.uk/TA104/niceguidance/pdf/English).
Submitted 15 January 2007; revised version accepted 20 July 2007.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 J Health Serv Res PolicyHome page
A E Ades, N. J Welton, D. Caldwell, M. Price, A. Goubar, and G. Lu
Multiparameter evidence synthesis in epidemiology and medical decision-making
J Health Serv Res Policy, October 1, 2008; 13(suppl_3): 12 - 22.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Bravo Vergel, Y.
Right arrow Articles by Sculpher, M. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bravo Vergel, Y.
Right arrow Articles by Sculpher, M. J.
Related Collections
Right arrow Spondylarthropathies
Right arrow Pharmacology
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?