The high cost of many new anticancer medicines significantly impedes breakthrough discoveries from reaching patients. A commonly heard refrain is that high prices are necessary to compensate for the high costs of research and development (R&D). Yet, there are promising policy proposals aimed at improving affordability without compromising innovation. In seeking new policy solutions, we argue for a shift away from entrenched opinion toward an evidence-based discourse that is grounded in experiments and real-world pilot studies. We offer a novel perspective and practical recommendations on how empirical evidence could and should be gathered to inform evidence-based policy interventions that lead to sustainable medicine prices in oncology.
The high cost of innovative anticancer treatments is increasingly recognized as a global challenge. During 2020, global spending on anticancer medicines increased by 14.3%, reaching $164 billion. By 2025, spending is projected to reach $269 billion (1). This growth is, to a large extent, driven by the extremely high prices of innovative anticancer treatments. Worldwide, patients' treatment success and financial well-being suffer from the severe consequences of the direct and indirect costs of oncological treatments. Financial toxicity in the United States is mostly driven by high co-payments and low insurance coverage (2), but it also exists in countries with a comprehensive social health care system, such as Germany (3). Ensuring that lab discoveries translate into innovative medicines that reach as many qualifying patients as possible in a timely manner is a daunting moral challenge for the biomedical community and for society at large.
It is presumed that the high level of investment in clinical research in oncology is driven by high expected returns. Hence, the pharmaceutical industry warns that lower prices will lead to reduced investments and to fewer innovative treatments reaching the market. However, there remains a gap, or rather a missing association, between treatment costs and clinical benefit for anticancer medicines, which suggests unresolved economic inefficiencies. These inefficiencies add to the uncertainty about the strength of the relationship between therapeutic pricing and investment in research and development (R&D; ref. 4).
There is an active discussion on possible solutions to create sustainable prices that combine affordability and well-targeted incentives for R&D. In a literature review, we identified promising policy options in the pricing environment, such as joint procurement, two-part tariffs, and transparency; in the R&D environment, such as orphan drug reform, public research, and public clinical trials; and in the intellectual property system, such as de-linkage (see Supplementary Table S1 for a description; ref. 5). However, the evidence underlying these proposals is extremely thin. Moreover, the effectiveness and feasibility of these policies are perceived as highly dependent on levels of international collaboration and industry compliance (6).
Likewise, the political discussion appears to be stuck—opposing fronts and arguments are often mired in opinions or theoretical models. There is insufficient availability of (causal) empirical evidence on which fact-based policy can be built (5). The obvious challenges in trying out new policies in practice start a vicious cycle that perpetuates delays in the accumulation of evidence. In this commentary, we argue that there are options for moving from opinion-based to evidence-based policy making. We discuss how selected examples of promising policy options could benefit from controlled studies similar to laboratory, field, and natural experiments (see Fig. 1). These can provide data-driven insights and empirically conclusive evidence in support of policy change. When gathering experimental evidence is not possible, we elaborate on the usability of policy pilots guided by strong economic principles.
Empirical Evidence for Advantages of Transparency
Although transparency laws have been introduced in many U.S. states, secret discounts of list prices together with confidentiality clauses continue to hide the real prices paid by European countries. The debate is stuck between supporters of more transparency who argue that it will improve accessibility through lower prices and opponents who argue that it will eliminate targeted discounts to poorer countries and could even reduce R&D investments. Such a debate is hard to settle with a purely theoretical approach. Outcomes such as R&D investments and net prices will depend on the acceptability of transparent price differences and ultimately on the distribution of R&D costs across countries and payers. As this is an expression of a value judgment and a function of solidarity, it is intrinsically an empirical question and one that can be investigated with laboratory experiments.
Laboratory (behavioral) experiments are increasingly used by social scientists and policy makers to preassess the viability of interventions whose outcomes depend upon human decision-making. Typically performed on student populations, they share many of the advantages of clinical in vivo experiments. These are, most notably, full control over the environment, relatively low cost, and randomized allocation to experimental conditions to allow causal implications to be drawn. A crucial question is how findings from this environment generalize to the real world (7). In many cases, laboratory results in economics are predictive of real-world outcomes (7, 8). Also within health care, laboratory experiments are gaining importance. Examples include research on payment schemes, for example, pay for performance, fee for service, and capitation (9–11). Laboratory findings obtained via best practices, such as the inclusion of monetary incentives, have proven to generalize to the health care field. For instance, a series of articles to reduce care cost demonstrate that the tendency to overtreat patients in the laboratory is replicated by dentists in the real world (12, 13). Although laboratory experiments are often predictive of real-world outcomes, they should ideally be followed up by additional evidence, preferably from field experiments or, when this is unfeasible, policy pilots.
We recently conducted a laboratory experiment to gauge the effect of transparency. In a randomized cross-country study, we evaluated the effect of increased price and R&D cost transparency on affordability and accessibility to treatment (14). In total, 400 participants from the Netherlands, Germany, Poland, and Spain participated in the experiment and were assigned to negotiate on behalf of a single-country payer or a pharmaceutical company. Participants were randomized into one of three experimental arms. In the first arm, the control arm, prices and R&D cost were confidential (full secrecy). In the second arm, negotiated prices were made transparent (price transparency). In the third arm, both prices and verifiable R&D costs were made transparent (full transparency).
We found that patient access was the highest in full transparency, where prices reduced substantially (−26%) without negatively affecting R&D investments. In contrast, price transparency had no significant effect on average prices or number of patients treated, although it did reduce R&D investments significantly (−17%). Here, we saw some indications for price increases for low-income countries, but no evidence of their exclusion as low-income countries continued to negotiate lower prices compared with their richer neighbors.
Another important finding is that the possibility for the company to share the risk of high R&D costs was crucial in maintaining high investments in R&D despite lower overall prices. Prices responded most strongly to varying levels of R&D costs in full transparency, whereas their response was completely absent in price transparency. This remark is relevant because, to our initial surprise, R&D investments did not suffer in full transparency where prices diminished, while we found a reduction in R&D investments in price transparency where prices continued to be high.
These findings contribute to the policy discussion in three ways: (i) to reduce concerns that transparency will eliminate targeted discounts to low-income countries entirely; (ii) to discriminate between interventions and specifically, to give priority to a policy that combines price and R&D costs transparency; and (iii) to highlight that truthful reporting of R&D costs offers an opportunity for risk sharing, which is essential to encourage R&D investments. In our original article, we discuss ways in which this can be encouraged (14). Overall, the experiment shows that a laboratory approach can contribute empirical data to the discussion.
Experimentation to Inform Joint Procurement Policies
Another policy that receives strong support by stakeholders and can be usefully investigated in the laboratory is joint procurement, that is, the consolidation of multiple payers into a single, and therefore stronger, bargaining position. Consolidated negotiations and purchasing are an interesting option within a nation, as well as in cross-country/cross-state collaborations. Joint procurement has the potential to increase the contractual power of buyers and leads to lower prices. Open questions remain as to whether this will be true empirically, what the size of the effect will be, and whether incentives to invest in R&D will be altered. The discussion also revolves around which type of joint procurement is more beneficial. Some suggest the grouping of entities with similar purchasing power. Others propose that even payers of heterogeneous financial capacity should join forces and separately discuss an equitable and fair distribution of R&D costs to maintain high investments. It is largely an empirical question whether the coordination over R&D cost distribution is easier to achieve in the current system of bilateral negotiations or with a joint procurement model with internal assessments and redistributions among buyers, and how this depends on the composition of the procurement initiatives. Other important questions are which players will decide to join forces and whether there will be any buyer excluded from purchasing coalitions.
So far, we seldom see objective evaluations of these consolidations. Joint procurement has been used by the European Union (EU) for COVID-19 vaccines and medication, but this initiative was not empirically evaluated and crucially lacks the comparison with a counterfactual identical scenario where vaccines are not procured jointly. There exists some nonexperimental evidence of the positive effect of joint procurement in the real world (15). In Jordan for example, the first tender for off-patent antibiotics, anti-HIV, and antituberculosis agents has led to 2.4% to 8.9% of savings. Centralized purchasing of hepatitis C medicines was estimated to reduce prices by 90% in Colombia (16). A cross-country laboratory experiment that compares different forms of joint procurements to the existing system of bilateral negotiations, all else equal, could provide useful complementary empirical support for decision-making.
Orphan Drug Reform: Exploiting Naturally Occurring Randomness
The legal framework for orphan designations is a policy area that can benefit from natural experiments, another form of controlled empirical evidence. In both Europe and the United States, orphan drug legislation incentivizes the development of medicines for patients with rare diseases. It does so by improving profit margins via various benefits such as additional market exclusivity, tax credits, reduced administrative cost, and grants for clinical trials. An obvious question is to what extent these benefits are helpful in boosting R&D investments and how they affect the successful development and quick introduction of new medicines. This is especially relevant for oncology, where advances in personalized medicine through biomarkers and targeted therapy have changed clinical treatment and R&D most prominently. This shift has led to smaller subgroups of patients per tumor type, which challenges the disease-driven designation of the orphan drug system. Hence, evaluating the effectiveness of orphan drug regulation is useful for the field of oncology. Importantly, this evaluation has possible wider implications in that it can shed light on the responsiveness of R&D investments to variations in profit margins.
We suggest a natural experiment that can provide causal evidence in this debate. Such an experiment investigates whether the program is useful in stimulating R&D and the successful development and introduction of new medicines. Participation in the experimental orphan drug program is guaranteed to a subset of indications that fall below a cutoff threshold of disease prevalence (e.g., 5 per 10,000 individuals in the EU or fewer than 200,000 patients in the United States). Retrospectively, indications that were just below and/or just above the disease prevalence threshold can be compared with one another in terms of the likelihood to reach the market, the speed of introduction, and total sales. This enables the evaluation of the causal impact of the program. It also offers information on the effect of variation in profit margins, due to orphan designation, on access to medical treatment. The latter is important because of the frequent opposition of policy proposals on the grounds that lower profit margins will reduce investments and thus distort the R&D system.
Two-Part Tariffs: Moving to the Field
With a two-part tariff, payers pay a monthly or yearly subscription fee to get access to a medicine. This lump sum remunerates the company for the R&D costs and also provides a profit margin that is agreeable to the parties. On top of the subscription fee, the buyer pays a per-unit price that covers the production costs but does not lead to extra profits.
Economic theory is unambiguous on the positive effects such a measure can bring. It is a well-known fact that monopolies and oligopolies have incentives to keep prices too high and production levels too low compared with what would be socially desirable. A two-part tariff is theoretically expected to fully solve these inefficiencies. As R&D costs are separately and sufficiently remunerated, prices no longer have to incorporate R&D costs, and all patients can be treated for whom the clinical benefits exceed the relatively small production costs, up to the volume agreed in the contract. In contrast, in the current system there is a significant proportion of patients with cancer who remain untreated even though the expected benefits are larger than the production costs. For instance, in Europe, the lack of access to ipilimumab and abiraterone alone is estimated to have led to a loss of more than 30,000 life-years (17).
The subscription fee can be calibrated such that the incentives for pharmaceutical firms to invest in R&D can be kept at the same level as in the current system. Such calibration should be based on predictions of how many people can benefit from a novel therapy; to ensure access, the remuneration should also be differentiated according to the purchasing power of countries. Two-part tariffs are actually used and found to substantially reduce expenditures of buyers for hepatitis C medicines (18). We expect that a two-part tariff might also work well in the field of oncology in view of the high R&D costs and relatively low production costs, which lead to large economic inefficiencies with a traditional pricing system.
A laboratory experiment on the introduction of a two-part tariff in oncology is, in our view, less straightforward. It may be relatively hard to identify the essential features that a laboratory setting should incorporate for a proper comparison between two-part tariffs and the current pricing system. Ideally, the effectiveness of a two-part tariff would be investigated in a field experiment. In such an experiment, participating countries or payers could commit to a randomized allocation of the price negotiation strategy, with some being assigned to a two-part tariff negotiation process and others to the traditional process. Both groups could then be compared in terms of expenditures and access to the medicine to gauge the causal impact of transitioning to a two-part tariff system. Naturally, it will be very difficult and perhaps impossible to convince the relevant stakeholders to participate in such an experiment. A first step could be testing a two-part pricing system in the field of oncology with a small and well-thought-out policy pilot that ideally involves a group of countries, states, or payers and is followed by careful ex post evaluations.
We contend that there is insufficient empirical evidence underlying policy proposals to increase affordable access to innovative medicines, especially in mono- and oligopolistic settings. Although we recognize that field experimentation on policies to increase the affordability of medicines in oncology is a challenge, we have used experimental economics to provide evidence for policy making. Additionally, we offer a number of recommendations on how empirical evidence could be gathered and provide a foundation to the policy discussions that so far have been highly opinion-based. Evidence that shows that it is possible to balance public and industry interests is urgently needed. Considering the high stakes, as well as the high levels of complexity and uncertainty of pharmaceutical price regulations, more and well-communicated evidence is needed to provide comfort to policy makers in taking actions to reduce expenditure while keeping incentives for R&D investments. This commentary suggests how this can be achieved in practice.
N. Franzen reports grants from a philanthropist (unrestricted) during the conduct of the study. G. Romagnoli reports having recently coauthored a paper on incentivizing mindfulness meditation practices for better mental health. This project, completely outside of the scope of the project under publication, was funded by a Dutch insurance company, Zilveren Kruis Zorgverzekeringen N.V., based on a research partnership between Zilveren Kruis and G. Romagnoli's coauthors on the project at Duke University (the funding was used to hire a research assistant and pay the experimental subjects). G. Romagnoli did not receive financial compensation for this work, and G. Romagnoli's relationship with Zilveren Kruis has since concluded. V.P. Retèl reports grants from a philanthropist (unrestricted) during the conduct of the study, as well as grants from Intuitive B.V. outside the submitted work. W.H. van Harten reports grants from a private sponsorship during the conduct of the study; grants from Organisation of European Cancer Institutes outside the submitted work; and is Chair, Working Group Health Economics, OECI, and CEO Rijnstate General Hospital, Arnhem, the Netherlands. No disclosures were reported by the other authors.