Discussion

Our analysis identified important differences in three areas central to HTA for prescription drugs: the methodologies used, the way results are interpreted, and the populations or drugs carved out from standard HTA assessment. These choices have important ethical implications related to access to drugs and lead HTA organizations reach different conclusions and make divergent recommendations. The HTA organizations of different countries are each situated in their own cultural, social, and political contexts, so the ethical implications should be seen as value choices reflective of that context.Reference Torbica, Fornaro, Tarricone and Drummond60

One of the most important differences we observed was in the ways HTA organizations calculate costeffectiveness. Across the organizations, there is an underlying consequentialist assumption to maximize population health and well-being with the resources available, but how that is determined and what become priorities depends on the approach taken to cost-effectiveness. For example, if supplementary costs, like productivity, are included, then the analysis is likely to favor drugs for patient groups who work rather than those unable to work due to disability or age. While including productivity may disadvantage the young, reducing the discount rate for long-term effects favors drugs for conditions affecting young people and makes them more likely to be cost-effective. These choices in the methods reflect decisions about whether to include a broader scope of societal economic costs, like lost productivity and upfront payments for long-term benefit, or individual costs and benefits, lost wages and lasting health effects.

We also identified substantial variations in how HTA recommendations are made. For example, HTA organizations vary in how they consider the context of a finding that a drug is not cost-effective. Though QALYs are currently the best tool for comparing health states and health gain from treatment, the incremental cost-effectiveness ratio accounts for absolute QALY gain, not how badly off a patient group is to start. Both a move from 0.2 to 0.3 QALYs and from 0.9 to 1.0 QALYs involve the same QALY gain of 0.1, but one group has worse health than the other. HTA organizations differ in how they seek to meet the needs of the worst-off, an important ethical value. ICER has attempted to do this quantitatively with QALY shortfall reporting, and other organizations include need or severity as a qualitative consideration when making recommendations.

Similar challenges arise with attempts to quantify equity concerns. The HTA organizations we reviewed emphasized concern about not increasing health inequities and identifying equity concerns the HTA may raise.61 However, none of the organizations attempt to redress health inequities by applying weightings to QALYs, with the exception of Chuikyo’s vague 5% discount for “social and ethical perspective” subgroups. Partly this is a methodological challenge, but it is also a broader ethical question about whether the health care system should aim to address a broad set of inequities through drug access for some groups. PBAC has decided to cover access for indigenous groups and not others; CADTH has considered similar measures but worries about unintended consequences.62

Table 3Differences in methods and considerations for recommendations among HTA organizations

This review has important lessons for US policymakers considering setting up an HTA-like system to negotiate drug prices in Medicare. The choices in HTA approaches are inherently value judgments that will affect coverage and access to drugs, and therefore should be reflective of American values. First, the health outcome most used in HTA, the QALY, introduces the potential for variation in cost-effectiveness calculations since utility scores for health conditions vary by country: even in similar countries, like Canada and the US, there are differences in how people rate the utility of health states.Reference Oremus, Tarride, Clayton, Canadian Willingness-to-Pay Study and Raina63 Therefore, a robust US HTA method needs to use fair and representative USspecific utility scores. AHRQ has so far published one such data set that captures a nationally-representative sample of scores.64

Another choice US policymakers will have to face is whether to use a cost-effectiveness threshold or efficiency frontier, two different ways of defining when a drug is good value for money. Notably, HTAs that use thresholds to identify the opportunity costs of introducing a new drug set them too high to accurately capture opportunity costs. In the UK, the actual opportunity cost threshold has been estimated at £13,000/ QALY: if more than £13,000 is spent on producing one QALY, more than one QALY will be lost elsewhere in the health system.Reference Claxton, Martin, Soares, Rice, Spackman, Hinde, Devlin, Smith and Sculpher65 However, NICE uses a threshold of £20-30,000/QALY. ICER uses an upper threshold of $150,000/QALY, but the most generous estimate of an actual US threshold for private plans is $84,000/QALY.66 ICER accounts for this difference by addressing opportunity costs with its budget impact analysis, alerting decision makers when a drug would unsustainably increase health spending.67 By contrast, efficiency frontiers — the approach adopted in Germany and France — are more consistent with what US private insurers do when they negotiate with manufacturers for a better price on a drug. As the US government considered implementing HTA assessments, efficiency frontiers would allow Medicare to identify an appropriate and proportional price for a new drug without having to apply a cut-off, which would help overcome political resistance to cost-controls and limit-setting via cost-effectiveness.

We identified two groups of carve-out conditions with special methods or considerations for HTA. Drugs for cancer have become increasingly expensive and are politically sensitive for HTA organizations, which grapple with questions of how much small gains in life expectancy are worth.68 Different procedures or standards for cancer drugs and ultra-rare diseases, like Germany’s assumption that drugs for rare diseases offer benefit or the UK’s Cancer Drugs Fund and higher threshold for rare diseases, treat some people’s health gains as worth more than others. Such an approach may make sense in the context of a market failure. Treatments for ultra-rare conditions have tended to be very expensive because the small patient populations mean small markets for manufacturers to recoup upfront development costs and make a profit.Reference Sarpatwari and Kesselheim69 This results in higher costs for treatments for ultrarare conditions, so the drugs are less likely to be costeffective. While some HTA organizations consider rare diseases to be a special case, the inverse is never true: lower thresholds are not applied to drugs with a wide market that will impact health system affordability, though modifications to cost-effectiveness analysis to account for this have been proposed.Reference Lomas70 For example, sofosbuvir (Sovaldi), an extremely effective treatment for hepatitis C virus infection, was cost-effective despite a high initial price per course of therapy,71 but in the US its budget impact meant that not all patients could be treated without exceeding the budget of health systems.72 Therefore, some countries, including US Medicaid programs, imposed non-evidence based guidelines to prioritize patients.Reference Barua, Greenwald, Grebely, Dore, Swan, Taylor, Marshall, Pawlotsky, Lazarus, Aghemo, Dore and Grebely73 If context, like a small patient population, is a reason for weighting QALYs as worth more, then one could argue that a similar contextual feature, a large patient population, should also receive special consideration when assessing costeffectiveness. Budget impact analysis also plays an important role in setting an upper limit on how much a health system should spend on a single drug.

It is not clear that special carve-outs would be required in the US for rare diseases or cancer drug innovation. In the US, the 1983 Orphan Drug Act introduced a tax credit to incentivize drug development and granted seven-year market exclusivity to manufacturers of such drugs, which, combined with the high prices charged, make many drugs for ultrarare diseases profitable.74 ICER changed its approach to rare diseases and stopped applying a higher threshold in its HTAs because of the changed contextual features (profitability) and ethical concerns about giving special consideration to rarity.75 Similarly, the US government supports innovation through federally funded research and tax incentives, much of which goes to research on cancer.Reference Nayak, Avorn and Kesselheim76 In the US, with its government and public sector support for research and the development of drugs for cancer or rare diseases, it would magnify the special treatment if HTA also includes carve-outs.

Carve-outs for rare diseases and cancer, consideration and calculations for innovation, terminal conditions, long-term effects, productivity, or equity — these are all social value judgments about what matters most and how to allocate public or private resources for health care. NICE guidance states “an additional QALY should receive the same weight regardless of other characteristics of the people receiving the health benefit.”77 However, each methods choice, like whether to raise a threshold or include productivity, means shifting the calculation of cost-effectiveness and makes additional QALYs receive different weights for different patient groups. The HTA organizations make arguments for why these issues are relevant to consider, and their decisions reflect judgements about economic and social values specific to the country. The US has its own social values, so a US-based HTA process for insurers and federal payers should base its methods on the particular context. There has already been some work exploring people’s values and preferences in designing health insurance and potential limits to benefits; other research has focused on particular questions of age, severity, and productivity in the context of allocating organ transplants.Reference Danis, Ginsburg, Goold, Stahl, Tramontano, Swan and Cohen78 ICER solicited public input on its recent updates to its HTA framework, and other countries can provide examples of how to solicit and incorporate public, social values into HTA.Reference Kreis and Schmidt79 While ICER offers one American approach to HTA, other value frameworks have been proposed, it is not clear the extent to which it reflects social value judgments and whether it should be the defining approach in the US. For a US HTA method to reflect American social values, more research needs to be done into public values on the issues raised in HTA and the choices that need to be made to make recommendations. For Medicare, which enrolls millions of Americans across the country, broad public representation will be important.