In a May article in Health Affairs, we proposed that the federal government consider using an existing law to negotiate or compel lower drug prices in the United States for certain important drugs with excessive prices. We’ve been gratified at the interest the proposal has generated on Health Affairs Blog and elsewhere. We believe it illuminates a significant opportunity to improve access to important medicines, and also to more efficiently allocate our health care dollars.

A recent commentary on this blog by Henry Grabowski, however, expresses concern about the impact of our proposal on future drug innovation. As we will explain, we believe that the impact on innovation in fact would be minimal, and the results for society highly positive.

As we described in our original article, 28 U.S.C. Section 1498 permits the federal government to “use” patents that currently cover new drugs or other patented inventions. Patents are intended to create incentives for private investment in innovation, but in the current marketplace, they also allow pharmaceutical manufacturers to command exorbitant prices because they preclude generic competition. The price differentials between the US markets and other settings can be staggering — for example, a 12-week course of sofosbuvir (Sovaldi) to treat hepatitis C infection has a list price of $84,000 in the US, but can be obtained for under $500 dollars from generic suppliers in India.

The “government patent use” law that we discussed in our paper allows the federal government to produce or import low-cost versions of patented medicines for government programs, as long as it also awards patent holders “reasonable and entire compensation.” We likened the provision to the government’s power of eminent domain over land. In the case of drugs, government patent use allows the government to take steps that help the entire population, such as combatting a major public health epidemic like hepatitis C infection, in the face of a private entity with property rights that block such steps (unless the government pays the price set by the patent holder).

Key to our proposal is the method we described to set compensation. We suggest that the government offer patent holding companies compensation “keyed to the amount invested in the relevant drug, adjusted for the risk of failure and to permit companies to earn reasonable or average profits.” We believe that this approach would provide royalty awards that are sufficient to induce the development of new medicines, because companies could be assured of compensation for their investment, while also producing more efficient prices.

In his blog response, Grabowski first argues that the US government may not fulfill its part of the deal, and would take advantage of companies that have already sunk their research and development (R&D) costs. This concern is easily allayed: In this context, patent holders have recourse in the federal courts, which are required and empowered by law to assess “reasonable” compensation. The government patent use approach, perhaps distinct among the manifold existing proposals to address overpricing of drugs, draws strength both from the impartiality and the fact-finding abilities of courts.

But Grabowski’s central objection is that any reduction in profit that pharmaceutical manufacturers might face through the Section 1498 process, as opposed to the maximum amount they can extract in the current system, would harm innovation, possibly dramatically, because of the pharmaceutical industry’s high risks, long development timelines, and skewed distribution of returns (that is, a few entities account for most of the sales and profits).

It is precisely to address these issues that our proposal recommends not only that companies be compensated for their R&D, but also compensated for risk, and afforded reasonable profits. To preserve ex ante incentives, we must consider the likely yield of investments in R&D from the perspective of a company: If a drug has only a 50 percent chance of success, companies will only invest $1 if there is a chance of a $2 return.

Ideally, courts or agencies would gather tailored information about risk from companies, but to give a sense of the importance of the multiplier, a company that invested $500 million with a 1 in 10 chance of success (one estimate of the success rate for drugs entering Phase 1 trials) could earn up to $5 billion in compensation. We also suggest that companies be afforded a reasonable profit on top of that. (In these ways, our proposal differs substantially from the one criticized in the simulation analysis Grabowski refers to, which assumed far less generous “break-even pricing” for the entire top decile of drugs.)

The industry is indeed characterized by skewed distributions and many failures, but because risk will be compensated for, companies should nonetheless make investments in promising prospects. We should expect some error in calculations of reasonable compensation, but can account for those, too, with additional rewards to offset the possibility for initial underestimates. Indeed, investment of this sort will be more likely if there is certainty about the returns that can legitimately be expected, which is why we urge compensation parameters for government patent use to be clearly established in policy or judicial doctrine. It is folly to suggest that there is no uncertainty today that dampens investment: there is widespread public disdain for the high cost of many drugs, and far more austere proposals will undoubtedly get a hearing in the coming years.

Two final points are worth making. First, the approach to compensation under this statute should focus on the cost of R&D, and not on the cost-effectiveness of drugs. As we note in our paper, “cost-effectiveness tools are not empirically based on assessments of how much governments are able or willing to spend on health,” and can sometimes recommend the use of interventions that are fiscally unsustainable and could be obtained at a much lower price, without sacrificing incentives for R&D.

Grabowski’s contribution to the debate only counts costs on one side of the ledger. One cannot draw conclusions about the final balance of harms and benefits of any approach to innovation policy by focusing exclusively on the costs to industry. Our approach should be to consider the costs and benefits to society, and to focus on the net effect of any costs with respect to R&D and gains with respect to access. Substantial data available today confirm that pharmaceutical manufacturers’ current drug pricing practices have led to rationing. But rather than take this issue on, Grabowski’s response effectively ignores it.

In sum, even if our proposal generated lower profits in some cases for industry, the net gains for the population could be quite positive. If lower-cost drugs were available and we worked to eliminate hepatitis C in the United States in a few years, the social benefits—including to those not infected in the future—would undoubtedly be vast. We continue to believe that they dwarf any possible risks to innovation.

In the end, Grabowski’s comments repeat points that industry trade groups or insiders have expressed before. There is nothing that reflects an attempt to account for the grim realities imposed on patients by the existing regime. We think a new approach is needed. The vast majority of Americans agree.

Authors’ Note

Professor Kapczynski is funded by the Laura and John Arnold Foundation. Dr. Kesselheim is a Greenwall Faculty Scholar in Bioethics and is also funded by the Harvard Program in Therapeutic Science, the Laura and John Arnold Foundation, and the Engelberg Foundation.