Re-evaluating cost effectiveness of universal meningitis vaccination (Bexsero) in England: modelling studyBMJ 2014; 349 doi: https://doi.org/10.1136/bmj.g5725 (Published 09 October 2014) Cite this as: BMJ 2014;349:g5725
All rapid responses
Re: Re-evaluating cost effectiveness of universal meningitis vaccination (Bexsero) in England: modelling study
1. Would the authors please tell us whether reducing the age range from 0-99 down to say, 0-80 be not more reasonable? I assume that the Chancellor will be happy if oldies like me were not a charge on HIS purse.
2. Ms Gina Johnson ( 6 Aug 2015) sought urgent guidance on post-vaccine pyrexia. Six months later, there is still a deafening silence from the manufacturers and from the DoH. Do the manufacturers and the government consider their silence to be " in public interest"?
3. Dr Rappouli et al ( Novartis, 22 Octobef 2014) say that " cost effectiveness has its place but so does common sense". Cost-effectiveness is a matter of pounds, shillings and pences ( old money) while common sense is often a matter of individual or group judgement. " The man on the Clapham omnibus" - says......
4. Those in direct DoH employment or its QUANGOs might benefit from not ruffling the feathers of the upper echelons - howsoever important it might be, for the public that the feathers be ruffled. Can some ethicists contribute to this topic? There was a time when the JCVI was believed to be utterly impartial. NICE, I believe is now either ignored or gagged ( witness the saga of the nurse staffing.) Is the JCVI still completely independent of the government and of the industry?
Competing interests: 1. Ignorance of mathematical modelling 2. I wonder if direct financial interest is easily defined. After all, academic research carried out in one sphere can ease one's ascent in to lucrative career in to commercial world. 3. A belief that failure of manufacturers and the JCVI to respond to legitimate questions on this forum is plain wrong.
The new UK Meningitis B infant vaccination campaign will begin in September. Public Health England guidance (1) tells us that 39% of eight week old babies will develop a fever after this vaccination, even with the (controversial) use of paracetamol prophylaxis. NICE guidance CG160 (2) states that a baby aged under three months who develops a fever of 38 degrees or above should be classed as a "red light" and admitted to hospital for investigations. Should we admit 39% of healthy babies after their first vaccinations? Or should we use our common sense, and risk an indefensible claim for overriding national guidance if the baby is in the early stages of a serious illness?
If the Department of Health wants to reduce avoidable hospital admissions, it needs to clarify this issue very quickly.
1. Public Health England (2015). Immunisation against meningococcal B disease for infants aged from two months. https://www.gov.uk/government/uploads/system/uploads/attachment_data/fil...
2. NICE (2013). Feverish illness in children: assessment and initial management in children younger than five years. https://www.nice.org.uk/guidance/cg160
Competing interests: No competing interests
REPLY re ‘Re-evaluating cost effectiveness of universal meningitis vaccination (Bexsero) in England’ responses
We thank Prof Black, Dr Rappuoli et al. and Glennie et al. for their interest in our work and for raising interesting issues, to which we would like to add some comments.
Models have considerable value in helping to understand the potential impact of different interventions on public health and which factors are critical to the success of a programme and the potential cost-effectiveness of alternate strategies. We do not believe we have adopted a set of singularly vaccine unfavourable conditions in our base case model. Of note, a team from the University of Warwick was commissioned by the Department of Health on behalf of the JCVI to undertake an independent review of the input parameters and modelling. They found “no major defects with the model” and recommended changes (mostly around quality of life estimates) which would have led to the vaccine appearing less cost-effective than in our base-case analysis.
In a situation where resources for health are not endless models can be used to inform priority setting. We agree with Prof Black and Rappuoli et al. that mathematical models and cost-effectiveness analysis should be viewed as tools to aid decision making rather than “gate keepers”, indeed we state in the discussion that we recognise models are a simplification of the real world setting, that there remains substantial uncertainty in the model parameters and that given this value judgements need to be made.
We conducted our analyses in line with the guidance from the Working Group on Uncertainty in Vaccine Evaluation and Procurement used by the JCVI. We did not eschew peer reviewed data, but did seek to use the latest available evidence, which included both data from the published literature as well as evidence submitted as part of the consultation response to the interim JCVI statement. As a result some of the data included comes from as yet unpublished sources, similarly to the Novartis published model on the subject. A summary of parameter values with sources is provided as an appendix to our paper.
The case fatality rate of 5-10% for meningococcal disease quoted in our introduction relates to laboratory confirmed cases. For the age-stratified incidence and case fatality rate of disease in the model we used Hospital Episode Statistics data from 2005/06-2011/12 in order to additionally capture cases confirmed on clinical grounds alone. Whilst using HES data may mean it is possible a small number of deaths from meningococcal disease that occur outside of hospital are missing from our analyses the principal reason for the case fatality rate we used being lower than 5-10% is due to the many more cases we assume in our model compared to laboratory reports.
We did not weight QALYs in our model in line with NICE guidance that states equity weightings should not be applied. Additionally, whilst we agree with Glennie et al. that there is evidence that societal preference is to not weight QALYs equally, we are not aware of any UK data that could be used for preference weighting of QALYs in our models. The MOSAIC study, a case control study of MenB survivors over a three year period in England, offers the only matched controlled study of disease burden after survival of serogroup B meningococcal disease, which may partly explain the lower quality of life loss estimate from MOSAIC compared to previous studies.[6-8] The fact that the EQ-5D as a measure of health-related quality of life is the preference in adults, but was not designed for children and thus may underestimate the loss due to meningococcal disease was recognised by the JCVI when they specified a quality of life adjustment factor (QAF) of x3 for use in the models.
The JCVI process and that of NICE for assessing cost-effectiveness is analogous insomuch that extra considerations can be included by JCVI through the use of QAF or cost adjustment factors (CAF) and extra considerations can be included by NICE by increasing the cost-effective threshold. The JCVI code of practice states that through using quality of life and cost adjustment factors where necessary the “sort of additional considerations that according to NICE guidance permit an Appraisal Committee to recommend a technology that exceeds the £20,000 threshold (up to around £30,000) should effectively have been allowed for”. We also considered scenarios without a QAF applied (table 6) which for a 2,3,4+12 month strategy and £75 per vaccine dose resulted in a cost/QALY of £ 365 300. Extending this and taking the JCVI recommended strategy of vaccination at 2, 4 +12 months without the addition of the QAF and considering various thresholds (as may be considered by NICE) the cost effective vaccine price was £3, £6, and £12 assuming thresholds of £20,000, £30,000 and up to £50,000 respectively.
Uncertainty around the model parameters is a key issue in the consideration of vaccination with Bexsero against all serogroup meningococcal disease. Here we considered a great number of scenario analyses to explore the impact of this uncertainty and presented results using both 3.5% and 1.5% discount rates. We are explicit about the model not being probabilistic and our statements about cost-effectiveness are based on the incremental cost-effectiveness of vaccination judged against a threshold. Dr Rappuoli and colleagues are correct that there is no provision in the model for technological improvements and thus excludes potentially higher costs in prosthetics, though it is difficult to see how such as yet unknown improvements in patient care or available vaccines could be incorporated into such models.
1 Joint Committee on Vaccination and Immunisation. Minute of the meeting on Tuesday 11 and Wednesday 12 February 2014. https://app.box.com/s/iddfb4ppwkmtjusir2tc/1/2199012147/18992168807/1
2 Huels J, Clements KM, McGarry LJ, Hill GJ, Wassil J, Kessabi S. Modelled evaluation of multi-component meningococcal vaccine (Bexsero®) for the prevention of invasive meningococcal disease in infants and adolescents in the UK. Epidemiol Infect 2014; 142: 2000-12.
3 Ladhani SN, Flood JS, Ramsay ME, Campbell H, Gray SJ, Kaczmarski EB, et al. Invasive meningococcal disease in England and Wales: implications for the introduction of new vaccines. Vaccine 2012; 30: 3710-6.
4 National Institute for Health and Clinical Excellence. Methods for the development of NICE public health guidance (third edition). 2012. http://publications.nice.org.uk/methods-for-the-development-of-nice-publ...
5 Viner RM, Booy R, Johnson H, Edmunds WJ, Hudson L, Bedford H, et al. Outcomes of invasive meningococcal serogroup B disease in children and adolescents (MOSAIC): a case-control study. Lancet Neurol 2012; 11: 774-83.
6 Ruedin HJ, Ess S, Zimmermann HP, Szucs T. Invasive meningococcal and pneumococcal disease in Switzerland: cost-utility analysis of different vaccine strategies. Vaccine 2003; 21: 4145-52.
7 Erickson L, De Wals P. Complications and sequelae of meningococcal disease in Quebec, Canada, 1990-1994. Clin Infect Dis 1998; 26: 1159-64.
8 Buysse CM, Raat H, Hazelzet JA, Hulst JM, Cransberg K, Hop WC, et al. Long-term health status in childhood survivors of meningococcal septic shock. Arch Pediatr Adolesc Med 2008; 162: 1036-41.
9 Joint Committee on Vaccination and Immunistion Code of Practice June 2013. https://www.gov.uk/government/uploads/system/uploads/attachment_data/fil...
Competing interests: HC, MH, and WJE received support from the Department of Health for the manuscript BMJ 2014;349:g5725
The publication showing the parameters underlying the recent JCVI decision is very welcome. It provides comparison with the JCVI’s previous interim negative opinion and the original study. It shows that there are a range of possible cost effective options.
However Meningitis Research Foundation (MRF) remains concerned that this published model undervalues the impact of meningococcal B infection. As we have previously argued, the MOSAIC study which underlies several of the parameters used in this model underestimates health impact and cost in several areas. In particular, it delivers a snapshot of serogroup B disease outcome and no single study can possibly cover all time periods: MOSAIC studied survivors approximately 3 years after their acute illness. This current model corrects for the lack of information on short-term impacts, a deficiency of the previous model, by applying the results of a PHE study that found a substantial impact of the disease on quality of life (QoL) for all children affected by MenB in the year following the acute illness.
However, they were unable to correct for late onset health impacts or costs to the NHS that develop after the first 3 years. This omission of late onset effects is likely to result in significant underestimation of the long-term impact of MD.
In the MOSAIC study, 60% of participants were assessed aged 3-5 years, and only around 8% of participants were older than 11 years of age:
• Cognitive impacts are most likely to be evident after a child reaches school age: there is a growing body of evidence that children who survive bacterial meningitis and meningococcal septicaemia are more likely to struggle at school[7-10] due to cognitive impairment, deficits in executive function and consequent behavioural difficulties. These problems often become more acute when children reach secondary school, where there are greater demands on pupils’ ability to organise and prioritise. These problems are likely to have an impact on QoL and NHS/PSS costs, especially for children consulting their GPS, CAHMS and neuropsychologists.
• Institutionalisation and major interventions for children with major cognitive and behavioural problems are rare but have very high NHS/PSS costs. These tend to occur in late childhood or adolescence, after the child has grown physically too big for a parent to control, or in middle age after parents have died or become too frail to care for them.
Growth plate damage rarely becomes apparent until more than 3 years after the acute illness. Consequent surgery, attachment of external fixators, and physiotherapy are costly and have a significant impact on QoL.
We are not aware of any study that could provide detailed data for such impacts to be included in the model, although we have approached the JCVI, offering to set up a focus group of people affected to fill this gap. However, as the authors acknowledge, no model can perfectly represent the real world setting.
We are also concerned that the framework the JCVI and modellers have to use is inherently unfair to interventions which prevent uncommon but severe illnesses in children:
• The EQ-5D, the tool for assessing quality of life impact used by NICE and the JCVI, is insensitive to QoL impacts in young children. It consists simply of 5 questions, in which the respondent rates whether the person affected has problems walking about, problems washing and dressing, problems doing their usual activities, pain or discomfort, anxiety or depression. For a child who is too young to walk about/wash and dress/ have ‘usual’ activities/display anxiety or depression, these questions are largely irrelevant. This means it is very difficult for a parent to answer meaningfully about a child of theirs who has survived meningococcal disease, especially now that the incidence of MD is weighted more towards the youngest age groups.
• Costs and benefits are discounted at 3.5% per year, which has the effect of disadvantaging children because much higher importance is placed on immediate health gains compared with those that are sustained far into the future.
• There is strong evidence that the public prefer prevention over cures, and would rather prevent death or severe disability in a few than mild illness among the many, but there is no allowance for this in the framework.
The current study applies a quality of life adjustment factor (QAF) of 3 to address the problem of underestimation of quality of life impacts and public preference for interventions for severe diseases as well as the innovative nature of the MenB vaccine Bexsero
The authors consider that applying the QAF up-front is equivalent to considering technologies that exceed a £20,000 – 30,000/QALY threshold as NICE does.
The authors’ original study used a quality of life utility loss for survivors of 0.2 which was based on three published studies[14-16]. This new analysis uses a utility loss for survivors of 0.222 obtained by multiplying the utility loss figure provided by one of the authors of the MOSAIC study (which is difficult to dispute because it is unpublished) by the QAF.
The application of the QAF has thus had the net result of a minimal increase in health gain and brings with it a rigid threshold of £20,000/QALY compared to the more flexible 30,000/QALY used in the original study.
We agree with their conclusion that “models can be valuable tools to gain a greater understanding of the potential impact of an intervention but there is inherent uncertainty associated with such modelling and value judgments need to be made”. In light of all the concerns described above, we feel that to offset these deficiencies the more vaccine- favourable scenarios the model provides should be seriously considered.
MRF would like to see a rapid and positive conclusion to the current negotiations between the Department of Health and the manufacturer, which have been underway for three months, to secure a cost effective price for an infant immunisation programme. The vaccine was licensed in January 2013, and recommended by JCVI for an infant programme in March this year. We are now approaching a second winter peak in cases of a vaccine preventable disease which will see more lives lost.
1. Christensen, H., et al., Re-evaluating cost effectiveness of universal meningitis vaccination (Bexsero) in England: modelling study. BMJ, 2014. 349: p. g5725.
2. JCVI. JCVI interim position statement on use of Bexsero meningococcal B vaccine in the UK. 2013 [cited 2013 August]; Available from: https://www.gov.uk/government/uploads/system/uploads/attachment_data/fil....
3. Christensen, H., et al., Introducing vaccination against serogroup B meningococcal disease: an economic and mathematical modelling study of potential impact. Vaccine, 2013. 31(23): p. 2638-46.
4. Meningitis Research Foundation. Response to the JCVI Interim position statement on the use of Bexsero meningococcal B vaccine in the UK. 2013; Available from: http://www.meningitis.org/news-media/our-response-to-the-jcvi-72756.
5. Viner, R.M., et al., Outcomes of invasive meningococcal serogroup B disease in children and adolescents (MOSAIC): a case-control study. Lancet Neurol, 2012. 11(9): p. 774-83.
6. Iain Kennedy and Shamez Ladhani. Abstract E17 Quality of life changes in children diagnosed with group B meningococcal disease in months following illness. 2013; Available from: www.meningitis.org/conference2013.
7. Stevens, J.P., et al., Long term outcome of neonatal meningitis. Arch Dis Child Fetal Neonatal Ed, 2003. 88(3): p. F179-84.
8. Anderson, V., et al., Cognitive and executive function 12 years after childhood bacterial meningitis: effect of acute neurologic complications and age of onset. J Pediatr Psychol, 2004. 29(2): p. 67-81.
9. Bedford, H., et al., Meningitis in infancy in England and Wales: follow up at age 5 years. BMJ, 2001. 323(7312): p. 533-6.
10. Halket, S., et al., Long term follow up after meningitis in infancy: behaviour of teenagers. Arch Dis Child, 2003. 88(5): p. 395-8.
11. Als, L.C., et al., Neuropsychologic function three to six months following admission to the PICU with meningoencephalitis, sepsis, and other disorders: a prospective study of school-aged children. Crit Care Med, 2013. 41(4): p. 1094-103.
12. Luyten, G.P., Goos, P., Kessels, R., Beutels, P.,. Prevention, cure and public preferences for funding health care: a D-efficient discrete choice experiment. in International Health Economics Association, 9th World Congress. 2013. Sydney, Australia.
13. Joint Committee on Vaccination and Immunisation. JCVI Minute of the meeting on Tuesday 11 and Wednesday 12 February 2014. 2014; Available from: https://www.gov.uk/government/groups/joint-committee-on-vaccination-and-....
14. Ruedin, H.J., et al., Invasive meningococcal and pneumococcal disease in Switzerland: cost-utility analysis of different vaccine strategies. Vaccine, 2003. 21(27-30): p. 4145-52.
15. Erickson, L. and P. De Wals, Complications and sequelae of meningococcal disease in Quebec, Canada, 1990-1994. Clin Infect Dis, 1998. 26(5): p. 1159-64.
16. Buysse, C.M., et al., Long-term Health Status in Childhood Survivors of Meningococcal Septic Shock. Arch Pediatr Adolesc Med, 2008. 162(11).
Competing interests: I receive fees for speaking at conferences and taking part in ad hoc advisory boards for manufacturers of meningococcal vaccines. All such fees are donated to Meningitis Research Foundation. Meningitis Research Foundation is a charity that is solely concerned with promoting and funding research into early recognition, best treatment, best aftercare and best possible prevention of meningitis and associated infections. Over 95% of the charity's income is from public donations. Meningitis Research Foundation has received sponsorship for conferences and meetings, and unrestricted educational grants for its charitable activity from vaccine manufacturers.
The introduction of cost-effectiveness analysis has been one in which the UK has led the world, and as a measure to allow for equitable rationing of health budgets, it is clear that cost-effectiveness modelling has merit. However it needs to be remembered that such models are tools and not scientific facts. This is particularly pertinent in the case of vaccination for rare diseases. The uncertainty inherent in parameters such as epidemiology, effectiveness of treatments over time, long term outcomes in survivors and the impact of herd protection mean that models should be viewed with caution and be one of a number of tools, rather than the only mechanism for decision making.
In the case of Bexsero, the uncertainty and sensitivity of the assumptions are reflected in the fact that Christensen’s paper is the third such iteration of this model. These changes in these models have led to different conclusions at consecutive Joint Committee on Vaccination and Immunisation (JCVI) assessments.
When modelling in an uncertain environment, parameter input data becomes key. Rightly there has been a push towards greater transparency concerning pharmaceutical company sponsored trial data and publications. The BMJ has been at the forefront of this effort (BMJ Open Data Campaign). It is disappointing therefore that the Christensen paper has eschewed peer reviewed data for critical inputs such as sequelae rate and disutility, and case fatality rate, preferring instead to use personal communications from the authors (evident in the appendix to the publication). This approach, if originating from industry, would most certainly be challenged; surely the same should apply across other disciplines, including the authors of the Christensen paper.
A closer examination of sequelae detailed in the paper, reveals that the modelling does not capture the true cost of surviving MenB. It is notable that the average disutility for survivors was reduced by two thirds from the first published version of the JCVI model (Christensen 2013). This reduction followed a personal communication from a single study and is based on a methodology (EQ-5D) that is not fit for assessing the health state of young children (Wille et al, 2010) – The JCVI did introduce a so-called QALY adjustment factor bringing the disutility back to the level of the 2013 model. Unfortunately, it is positioned as a panacea covering all of the uncertainties and special circumstances surrounding Bexsero and MenB, quelling any hope for judgement to be applied as is customarily allowed in NICE analyses.
Another example of where the true cost of MenB is being underestimated concerns amputations, for which the NHS provision is currently for basic prostheses only. The existence of an NHS England fund to provide for higher grade prostheses for ex-military amputees, allowing for a more active life, points to the fact that these basic limbs are not suitable for young, otherwise healthy amputees just setting out in life (NHSE 2013). There is no provision in modelling for technological improvements and thus excludes higher cost in prosthetics over the 100 year period. The approach is akin to patients today being satisfied with 1914 model prosthesis, which is obviously absurd. A significant number of impacted families, supported by the meningitis charities, dedicate considerable time and effort on fundraising to support the economic burden it puts on these families. Similarly the cost of loss of executive function in meningitis is not captured well, despite data showing significant numbers of survivors have IQ<85 and late presentations of executive function loss leading to educational and behavioural issues that limit their long term potential (Anderson 2004). A further example is the average case fatality rate used in this model of <4%, which differs dramatically from the range 5 -10%, quoted in the introduction to the paper and also by the JCVI – yet no explanation is provided.
Among a general issue with lack of transparency it is especially surprising then that the JCVI have chosen to abandon the established NICE methodology of adapting a cost-effectiveness threshold of £20,000-30,000+ per QALY gained to allow for factors such as vulnerability of the population considered, innovation etc. in order to use a novel and as yet unclear methodology from the Working Group on Uncertainty in Vaccine Evaluation and Procurement – without conducting any consultation. This lack of clarity makes it difficult to understand the factors considered by the JCVI when assessing a vaccine.
If the approach set forth by Christensen et al truly reflects the outlook of cost-effectiveness analyses future innovation will be impeded, contrary to government rhetoric for the UK to become a global leader in life sciences. A vaccine such as Bexsero can take upwards of 20 years to come to market, and without recognition of this investment commercial organisations will struggle to convince shareholders of the value of continued investment in areas where the science is challenging, trial programmes are difficult and expensive and health economics subject to uncertainty and discrimination. It is worth noting that nearly a third of the clinical development of Bexsero took place in the UK, while since licensure less than 3% of Bexsero doses have ended up in the UK.
Therefore Government – much like the industry – must be willing to keep the long-term public health benefits in mind and align the risks and rewards accordingly, otherwise all investments and innovation will cease – as has been the case for antibiotics. It is sobering to read Frank Dobson’s account in this journal earlier this year (Dobson 2014) of what it took to get the UK MenC vaccination programme launched in 1999. The situation was not dissimilar to the current. “A few years after it was introduced, David Salisbury sent me a graph showing deaths flatlining between zero and two a year. He said I should be glad.”
Rino Rappuoli1, James Wassil2, Lars Bonefeld3, and John Porter3
1.Novartis Vaccines and Diagnostics, Srl., Siena, Italy
2.Novartis Vaccines and Diagnostics, Inc., Cambridge, MA, U.S.A.
3.Novartis Vaccines and Diagnostics Limited, Frimley, Surrey, United Kingdo
Dobson, F. BMJ. [0959-8146] 2014 vol:348
H. Christensen et al. Vaccine 31 (2013) 2638– 2646
Wille et al Quality of Life Research August 2010, Volume 19, Issue 6, pp 875-886
NHSE 2013: http://www.england.nhs.uk/ourwork/commissioning/armed-forces/veterans-pr...
Anderson et al 2004 Journal of Pediatric Psychology 29(2) pp. 67–81,
Competing interests: All authors are full-time employees at Novartis Vaccines.
The study by Christensen et al. highlights inherent difficulties of making decisions based solely upon cost-effectiveness models when uncertainty cannot be easily quantified. While uncertainty around parameter estimates is common when conducting appraisals of a medicine, I would argue that vaccines for rare diseases are particularly prone to uncertainty around clinical efficacy, impact on carriage and epidemiology as these are difficult or impossible to accurately assess in short-term randomised trials that are available prior to vaccine introduction and use.
Indeed, as pointed out in a recent letter discussing the economics of vaccination programs (Stephens et al. 2014), it is not appropriate to view cost-effectiveness analyses as static, because key variables can change dramatically over time. Stephens et al. refer to the example the PCV7 pneumococcal vaccine where the estimate of cost-effectiveness was unfavourable ($80,000 per life year saved). None the less, the vaccine was introduced in the US and it was subsequently shown that the vaccine was highly cost effective with an actual cost-effectiveness of $7500 per life year saved due to the fact that models did not include the strong herd immunity benefit associated with the vaccine. Similarly in the UK, many of the herd immunity benefits of the introduction of the meningitis C vaccine were identified long after the cost-effectiveness analysis had been conducted (Ramsay et al. 2003; Mooney et al. 2004). Although both of these examples highlight the difficulty of estimating and including herd immunity effects prior to the actual large scale use of a vaccine, the same concern regarding the lack of available data and true impact could also be applied to vaccine coverage, duration of protection, etc. A recent case in point is Ebola for which ‘cost-effectiveness’ conclusions would look vastly different today than they did a few months ago. As I have argued in a recent review (Black 2013) the inherent uncertainty in CE models and their historical lack of predictive value in accurately assessing vaccines before their introduction should temper our willingness to use these analyses as rigid “gate keepers” for public health interventions but rather they should be viewed as one of many criteria to be considered in adopting a new technology. In fact in the US, the Institute of Medicine has advocated this approach and developed a Smart Vaccines assessment tool to facilitate multifactorial assessment. (Institute of Medicine 2014)
The JCVI utilised the output of the model by Christensen et al. and scenario analyses to inform the decision-making process for Bexsero. The Meningitis B (MenB) model presented by the authors appears to have significant parameter uncertainty. It appears that only one parameter (carriage) is informed by a systematic literature review and whereas many other key inputs are informed by personal communications (five of which are from one of the study authors). This would seem to be inherently problematic. Another potential area of difficulty is using standard health-related quality of life measurement tools for young children, for whom they may not accurately reflect utility (e.g. few two-year olds are able to dress themselves, which is one of the assessment questions). Furthermore the observed case fatality rate resulting from the 2014 study is 3.8% (variable by age), which is markedly lower than the range 5 10%, reported in the introduction and by the JCVI (JCVI 2014). Although there may be valid reasons for these assumptions, the rationale is unclear from the publication.
The results of the model suggest that the vaccine can be cost-effective in certain scenarios, though exclusively considering a rather low cost-effectiveness threshold (£20,000 per QALY). NICE, whose methodology JCVI follows, recommends considering a range of £20,000 - £30,000 per QALY gained to allow decision makers to take into account other factors of interest (e.g. the degree of uncertainty around cost-effectiveness, ability to capture benefit using health related quality of life, innovation, end of life care, and the non-health objectives of the National Health Service) (NICE 2013). Many of these are relevant to Bexsero – most recently evident when Professor Sir Michael Rawlins awarded the Prix Galien to Bexsero for most Innovative product in the UK. Furthermore, a recent study (Dakin 2013) suggests that in practice, the NICE threshold may be higher than the £20,000 - £30,000 per QALY range reflecting the consideration of other factors in decisions (£39,000 - £44,000 per QALY).
Christensen et al. report that the JCVI considered the evidence using new guidance from the Working Group on Uncertainty in Vaccine Evaluation and Procurement (reported in the JCVI minutes 2014). The authors go onto report that the JCVI process is analogous to the NICE process for taking special circumstances into consideration. Since, to our knowledge, no consultation for the new guidance or official comparison has been conducted, this could be debated. The authors fail to point out that a key consideration of the cost-effectiveness threshold range is uncertainty, which is considered by NICE using a mandatory probabilistic modelling approach. The JCVI adapted the new guidance from the Uncertainty Working group for the first time, which requires a consideration that the vaccine has “no more than a 10% likelihood of not being cost-effective at a threshold of £30,000 per QALY gained” (p36 JCVI 2013). Given the modelling approach, conclusions on cost-effectiveness cannot be made at this stage as no probabilistic analysis was conducted.
Finally, adopting an extreme risk-averse approach to public health decision-making has its own risks and could hamper the adoption of a beneficial intervention. Furthermore, unlike appraisal of therapeutic drugs compared against a well-defined and known standard of care, the downside risk of a wrong decision is unlimited in a vaccination/public health setting because by not introducing a vaccine, the alternative is leaving the public without protection.
Pitman R, Fisman D, Zaric GS, Poostma M, Kretzschmar M, Edmunds J, Brisson M. Dynamic Transmission Modeling: A Report of the ISPOR-SMDM Modeling Good Research Practices Task Force-5. Value in Health. 2012; 15: 828 – 834
Joint Committee for Vaccination and Immunisation. Minutes of the meeting on Tuesday 11 and Wednesday 12 February 2014. 2014.
Stephens DS, Ahmed R, Orenstein WA. Vaccines at what cost? Vaccine. 2014; 32(9): 1029-1030.
Ramsay ME, Andrews NJ, Trotter CL, Kaczmarski EB, Miller E. Herd immunity from meningococcal serogroup C conjugate vaccination in England: database analysis. BMJ. 2003; 326: 365-366.
Mooney JD, Christie P, Robertson C, Clarke SC. The impact of meningococcal serogroup C conjugate vaccine in Scotland. Clinical Infectious Diseases. 2004; 39: 349-56.
Black S, The role of health economic analyses in vaccine decision making. Vaccine. 2013; 31(51): 6046-6049.
National Institute of Health and Care Excellence. Guide to the methods of technology appraisal 2013. April 2013.
Dakin H, Devlin N, Feng Y, Rice N, O’Neill P, Parkin D. The influence of cost-effectiveness and other factors on NICE decisions. OHE Research Paper 12/06r. November 2013.
Institute of Medicine of the National Academies (2012) Ranking Vaccines: A Prioritization Framework - Phase I: Demonstration of Concept and a Software Blueprint. Available at www.nap.edu/smartvaccines. Accessed June 1, 2014.
Institute of Medicine of the National Academies (2013) Ranking Vaccines: A Prioritization Software Tool - Phase II: Prototype of a Decision-Support System. Available at www.nap.edu/smartvaccines. Accessed June 1, 2014.
Joint Committee for Vaccination and Immunisation. Code of Practice. June 2013
Competing interests: Consultant for Novartis, Takeda and WHO; serve on DSMBs for GSK.