For the growing number of multi-indication medicines, challenges arise in assessing and paying for value across medicine indications with a uniform price (see our previous report and further publications for more context on the topic).

In recognition of this challenge, and as part of the Voluntary Scheme for Branded Medicines Pricing, Access, and Growth (VPAG) 2024, DHSC is updating the Commercial Framework to be more explicit about the availability of commercial flexibilities, including for indication-specific pricing (referred to here as indication-based pricing [IBP]).

During the consultation process for phase 1 Woods and colleagues published a paper funded by the NIHR Policy Research Programme (“the EEPRU paper”). The EEPRU paper estimated the overall population health effects of uniform pricing, indication-based pricing, and alternative commercial arrangements for new pharmaceuticals in the UK NHS. They found that enhanced pricing flexibilities lead to better access and health benefits from multi-indication medicines, but – counterintuitively – poorer population health overall because of higher expenditure.

Given the important timing and topic, and its explicit relevance to this live and evolving policy question, we summarise the EEPRU paper’s analysis in this article and offer a short critique, outlining the main limitations and considerations for its interpretation.

Overall, the EEPRU paper’s analysis supports the role of pricing flexibilities in improving incentives, access, and health benefits from innovation. Beyond this, the finding of population health loss reflects their assumption that NICE is using the wrong value-for-money threshold in its decision-making, and therefore that all positive NICE decisions lead to a health loss. We argue that this does not reflect reality and should therefore not be considered informative for the design of the NHS Commercial Framework.

What does the EEPRU paper aim to achieve and what is its main finding?

The report details a hypothetical model comparing the long-term overall population health effects of three pricing policies:

• Uniform Pricing (UP) whereby the same price applies to all indications;
• ‘Pure Indication-Based Pricing’ (IBP) whereby there is a separate price for each indication; and
• Commercial Flexibility (CF) whereby UP applies as standard but IBP is allowed for indications that would not be launched under UP.

An approval norm of £30,0000 per QALY is assumed for UP and IBP scenarios, whereas an approval norm of £20,000 per QALY is assumed for indications requiring CF.

(Side-note: the threshold for CF reflects Principle 4 of the proposed NHS commercial framework: “Confidential complex commercial arrangements are expected to be considered only for products which represent value at or below the lower end of the standard NICE threshold or other applicable thresholds”. It should be noted that the application of a £30,000 per QALY approval norm applied to UP and IBP mischaracterises NICE’s actual policy, which is to use a range from £20,000 – £30,000 per QALY, where – above a most plausible ICER of £20,000 per QALY gained – judgements take into account specific factors such as uncertainty and whether health benefits are captured comprehensively. The application of a £30,000 per QALY approval norm in the EPPRU paper analysis could therefore overestimate expenditure.)

The report estimates (and compares) the population health effects of reimbursing a new multi-indication health technology under the three policies. Two analyses are performed: static and dynamic, where the latter considers incentives for future innovation. The authors choose to focus on the static (no innovation effects) scenario in the headline result.

The authors find that IBP and CF improve access to new medicines compared to UP but raise expenditure and lead to population health losses due to the health opportunity costs of the additional expenditure. They conclude that – for IBP to improve population health – it would need to be accompanied by a reduction in the approval norm (i.e. the cost-effectiveness threshold used to judge value for money) for all indications.

Below, we outline our main points for consideration.

The report finds that IBP and CF improve access to – and the health benefits derived from – multi-indication medicines.

The authors simulate the effects of each policy on the availability of a hypothetical medicine that has a sequence of three indications with a certain value profile. Much like a modelling exercise we presented in our most recent report on this topic, the analysis demonstrates that some indications will not be launched under UP, for example where doing so would require a reduction in price across all indications that would reduce revenue overall (Cole, Neri and Cookson, 2021). IBP and CF thereby increase the health benefits derived from new medicines by enabling access to new treatment indications, i.e. new groups of patients that can benefit from an existing medicine (a “follow-on” indication). This represents a hugely important and growing route to new treatment options for many patients, particularly in oncology, immunology, and rare diseases, among others.

Adopting policies that maximise access to cost-effective treatment options should be considered a priority, in our view. The EEPRU paper corroborates that IBP and CF could support this.

Some readers may wonder why the story does not end here, which is where micro-economic theory takes us, along with the practical examples they (and we) have set out.

This is because the authors go on to conclude that IBP and CF lead to reduced overall population health.

Here’s why….

The finding of population health loss is not uniquely relevant to IBP, and mainly reflects the authors’ belief that the NICE cost-effectiveness threshold is too high.

The framework applied in the EEPRU paper leans strongly on the controversial assumption that England’s health system opportunity cost is £15,000 per QALY, whereas NICE’s approval norm (cost-effectiveness threshold used for decision-making) is assumed to be £30,000 per QALY (as we have already noted, in reality this is the upper-end of a threshold range). This is based on previous work by the same research group, which attempted to calculate the marginal productivity of NHS spend between 2003 and 2012 (Lomas, Martin and Claxton, 2019). Such attempts to estimate a “supply-side cost-effectiveness threshold” are extremely complex and have had limited impact on policy due to their limitations (see this article if you are interested in reading more about those).

The use of that research to inform the EEPRU policy analysis has strong implications: by assuming that the opportunity cost of spend is well below the approval norm this implies that all new health technologies adopted at the current approval norm (i.e. all technologies approved by NICE) reduce overall population health. Comparisons of the three policies are therefore reduced to comparing the relative population health loss between them, compared with potential gains associated with increased access to new medicines.

There are plenty of reasons to believe that £15,000 per QALY is not a reliable estimate of the opportunity cost of spending.

In fact, in an analysis of this estimate, Zamora and Towse (2023) identify several sets of plausible structural assumptions that would place the threshold estimates from the study within the current NICE range of £20,000–£30,000.

Furthermore, there is an underlying assumption that all QALYs are equal, when in fact NICE places a higher weight on several treatment characteristics (e.g. the severity modifier or the higher HST threshold used for ultra-rare conditions). Within section 7 of the EEPRU paper the authors discuss the possible impact of higher approval norms for certain treatment indications, but only to illustrate a wider gap between the approval norm and supposed opportunity cost, and the consequently “larger health losses”; there is no considered reflection or discussion that any of these additional weights may be appropriate (as suggested by the significant policy debate and discussion that have led to them).

Finding the “right” level for the approval threshold has been (and always will be) an area of considerable interest and debate.  But we contend that policy analysis, as presented in this EEPRU paper, which is fundamentally based on the notion that NICE decisions do more harm than good, should not underpin the strategy of a government and health service that supports NICE and its decisions.

The EEPRU paper’s main criticism of IBP and CF is that they would increase health expenditure. But this investment is primarily incurred in providing access to cost-effective therapies in extra indications.

Differential reimbursement to reflect differential value, however implemented, is a form of price discrimination. In economic terms, price discrimination increases social welfare if it increases the total volume consumed. The EEPRU paper accepts that total volume increases in the IBP context. This means that IBP has a positive impact on social welfare; the extra spend on these extra people treated is an inevitable consequence, but this represents good value to the NHS (and should also be set within the broader context of the VPAG which places a cap on total medicines expenditure).

While increasing social welfare (or “surplus”) sounds like a laudable aim and outcome, in practice we may care about how that “surplus” is distributed, namely between the producer (the pharmaceutical manufacturer) and the consumer (represented by the UK’s third-party payer: the NHS). If prices are set up to their maximum value-based price, then the manufacturer (economically speaking) retains the full “value” of that surplus until patent expiry. The EEPRU analysis deems this to be a downside of IBP: in a UP world, manufacturers may accept a price below the value-based price for some indications, meaning the NHS would be getting an extra “good deal” (i.e. positive consumer surplus) for that indication. Under IBP they hypothesise that those good deals (i.e. paying below what is already a cost-effective price) would disappear, with all indication-level prices set to their maximum value-based price.

It should be noted that the EEPRU paper argues that this would not apply to the CF scenario, where IBP only applies for indications otherwise lost due to a UP (therefore maintaining some indication prices under their value-based level). Therefore, the extra requirement within CF for new indications to represent value “at or below” the lower end of the standard cost-effectiveness threshold (£20,000 per QALY) would represent a double-win for the NHS at the expense of industry revenues.

It should also be noted that the assumed sequence of indication launches matters. While the EEPRU analysis demonstrates some alternative scenarios in section 6, there is no discussion of how this impacts access (and to what types of value profiles) and how a benefit of IBP is that it effectively removes the incentive to strategically sequence indication launches, permitting development to be driven by the science and by patient need.

Assumptions around expenditure impact are simplistic and ignore the role of brand competition

By assuming value-based prices for each indication remain fixed along the whole patent term authors ignore any effect of on-patent competition in the distribution of surplus. Furthermore, under IBP, the authors fail to take into consideration the potential benefits of IBP and CF in the form of increased competition: by effectively permitting access to, and incentivising R&D in further indications/sub-populations, more treatment options and therefore higher competition at the indication-level is likely to put downward pressure on prices within the patent term, leading to a natural transfer of surplus from producer to consumer (to read more about the economics of this issue in general, see this report (Cole et al., 2018)). Competition in markets with high epidemiological interest (e.g. Hep C) has been shown to have this effect (Lindgren et al., 2022; Roediger et al., 2019; Berdud et al., 2018).

Some modelling assumptions are unclear and may overestimate off-patent access under uniform pricing

There are several aspects of the modelling exercise presented in the EEPRU paper that are unclear or unexplained. For example, the selection of a 100-year time horizon seems inappropriate given the rate of technological change, and that the discount rate reduces to insignificance nearly all of the effects within a much shorter timeframe anyway. More fundamentally, the authors assume that, in a UP scenario, access to indications not launched would become available to the UK population immediately following loss of exclusivity. In practice, regulatory approval may not have even been sought, meaning access after patent expiry would be limited or off-label, which could also act as a barrier for generics and biosimilars. Accounting for all these benefits from generic/biosimilar competition in the UP scenario (and for a 100-year time horizon) is unrealistic.

The modelling contains naïve assumptions about industry costs and decision-making

In the EEPRU analysis, the authors assume indications will be launched in the UK if the incremental price exceeds the incremental cost of production and supply. In reality, indication launches within a country can be costly, with significant investment needed in regulatory and HTA submission and compliance, manufacturing and distribution adjustments, marketing and commercialisation, and any ongoing data collection requirements or commitments. Therefore, the EEPRU paper’s characterisation that R&D costs are sunk, and should therefore not be considered, fails to adequately appreciate the role of price in investment and country-level commercialisation decisions.

When innovation effects are accounted for in the EEPRU analysis (the “dynamic”  scenario) both IBP and CF are shown to improve the overall population health compared to UP at current approval norms.

In the dynamic scenario, authors incorporate the effects of reward level on the production of future innovation (i.e. acknowledging the impact of pricing on whether the indication is even investigated and brought to market at all). In this scenario, they find that IBP and CF can improve overall population health. This happens under some circumstances for IBP, and all circumstances for CF. When the influence of realistic market features are introduced into their modelling (e.g.  higher approval norms for some indications, high-cost comparators for HTA etc.) the authors show that IBP can also increase overall population health, even with the use of a £15,000 per QALY opportunity cost estimate.

The authors choose to focus their conclusions on the “static” scenario, disregarding the impact of UK pricing policy on global incentives to develop medicines.

They thereby suggest that UK policy should be designed on the basis that the UK is irrelevant and inconsequential for the global life sciences industry and innovation: “Focusing on static effects is appropriate if UK pricing policy (as a small driver of global revenue) is not expected to influence drug R&D decision making […].” (Woods et al., pg.9). This position implicitly assumes that the UK optimal pricing policy is to free-ride on other countries’ contributions to fund the global R&D (a position which – if adopted by everyone – would lead to an innovation drought). We do not believe that NSHE / DHSC should embrace a policy based on this principle: first on moral grounds, and second because the UK does matter (research by colleagues on the international influence of NICE demonstrates this).

We cannot assume we operate in a static world when discussing pricing policy for biomedical innovation, and we cannot accept that UK optimal pricing policy should be defined on that basis. We therefore believe the focus for the pricing policy analysis must be the dynamic scenario.

The assumptions used by the authors to explore the “dynamic” impact of IBP and CF are over-simplified and are essentially a rejection of value-based pricing.

Authors introduce the concept of dynamically efficient approval norms. A brief definition of this concept is provided to the reader “approval norm that would maximise long-term health outcomes accounting for the effects of payment level on innovation”. However, the core element of the definition comes out of the blue and is not explained.

The authors impose an approval norm adjustment, based on previous work, which offers “just enough” value (according their calculations) to the innovator to keep them innovating, with the rest designated to consumer surplus (the payer) (Woods et al., 2024).  Through those sums they propose reducing approval norms for UP from £30k per QALY to £11,500 – £15,0000 per QALY and to £9,000 – 11,000 per QALY for IBP. This is essentially a rejection of the well-established and theoretically supported model of intellectual property protection and value-based pricing (Danzon, Towse and Mestre-Ferrandiz, 2015; Bell et al., 2023; Garrison et al., 2019; Danzon, 2018). Pricing policies based on these principles would de-link price and value, and could undermine investment in high-value innovation.

Furthermore, some of the implicit assumptions are unrealistic, for example that the elasticity of innovation is constant, when in reality this will vary by disease and innovation type.  For example, in areas of high unmet need where R&D risk is high, innovation is likely to be much more sensitive to changes in reward. As with the static scenario, the dynamic effects scenario also neglects the impact of competition, implying that the optimal “share” of surplus for innovators can be fixed ex-ante, while ignoring that competition produces an ex-post redistribution of rewards (from producer to consumer) by pushing down prices.

Given our reservations on the underlying assumptions, we do not have confidence in the “dynamically efficient” approval norms proposed by the EEPRU paper, and believe they would insufficiently incentivise future innovation.

Summary and implications

The EEPRU paper introduces and combines some novel elements in a modelling exercise, but there is a need for further careful examination of the analysis, its fundamental assumptions, and the implications of those. In this article, we provide a brief overview of some of the major ones as we see them.

The EEPRU paper supports the usefulness of price flexibilities in enabling greater access to cost-effective treatment opportunities. Beyond this, the results and recommendations rely on the core assumption that current approval norms lead to population health loss. If the reader does not believe that then, by implication, they cannot believe or rely on the reports’ quantitative and policy-related conclusions.

While further research into opportunity costs (“supply-side cost-effectiveness thresholds”) is certainly warranted, that research has yet to change or influence NICE decision-making more generally. It would therefore be inconsistent and distortionary for it to influence the multi-indication medicine policy context. This would put patient populations for whom treatment advances arise more commonly from multi-indication products – such as oncology, immunology and rare disease – at a relative disadvantage. 

While budget impact is often raised as the main concern around introducing pricing flexibilities, beyond our argumentation above this may also have limited relevance in practice due to the expenditure growth cap associated with VPAG in England.

So, let’s create the right environment to foster and encourage valuable and cost-effective innovation, and recognise the potential for multi-indication medicines to deliver that.

References

Bell, E., Berdud, M., Cookson, G. and Besley, S., 2023. Delivering the Triple Win: A Value-Based Approach to Pricing. [online] Available at: https://www.ohe.org/publications/delivering-triple-win-value-based-approach-pricing/ .

Berdud, M., Garau, M., Neri, M., O’Neill, P., Sampson, C. and Towse, A., 2018. R&D, competition and diffusion of innovation in the EU: the case of Hepatitis C. OHE research paper, 18(06).

Cole, A., Neri, M. and Cookson, G., 2021. Payment Models for Multi-indication Therapies.

Cole, A., Towse, A., Lorgelly, P. and Sullivan, R., 2018. Economics of Innovative Payment Models Compared with Single Pricing of Pharmaceuticals. Research Papers. [online] Office of Health Economics. Available at: https://ideas.repec.org/p/ohe/respap/002030.html [Accessed 17 Jul. 2018].

Danzon, P., Towse, A. and Mestre-Ferrandiz, J., 2015. Value-Based Differential Pricing: Efficient Prices for Drugs in a Global Context: VALUE-BASED DIFFERENTIAL PRICING. Health Economics, 24(3), pp.294–301.

Danzon, P.M., 2018. Affordability Challenges to Value-Based Pricing: Mass Diseases, Orphan Diseases, and Cures. Value in Health, 21(3), pp.252–257. 10.1016/j.jval.2017.12.018.

Garrison, L.P., Jackson, T., Paul, D. and Kenston, M., 2019. Value-based pricing for emerging gene therapies: the economic case for a higher cost-effectiveness threshold. Journal of managed care & specialty pharmacy, 25(7), pp.793–799.

Lindgren, P., Löfvendahl, S., Brådvik, G., Weiland, O. and Jönsson, B., 2022. Value appropriation in hepatitis C. The European Journal of Health Economics, 23(6), pp.1059–1070. 10.1007/s10198-021-01409-7.

Lomas, J., Martin, S. and Claxton, K., 2019. Estimating the Marginal Productivity of the English National Health Service From 2003 to 2012. Value in Health, 22(9), pp.995–1002. 10.1016/j.jval.2019.04.1926.

Roediger, A., Wilsdon, T., Haderi, A., Pendleton, K. and Azais, B., 2019. Competition between on-patent medicines in Europe. Health Policy, 123(7), pp.652–660.

Woods, B., Lomas, J., Sculpher, M., Weatherly, H. and Claxton, K., 2024. Achieving dynamic efficiency in pharmaceutical innovation: Identifying the optimal share of value and payments required. Health Economics, 33(4), pp.804–819. 10.1002/hec.4795.

Zamora, B. and Towse, A., 2023. The cost-per-QALY threshold in England: Identifying structural uncertainty in the estimates. Frontiers in Health Services, [online] 2. 10.3389/frhs.2022.936774.