Procurement regimes and medical innovation

Jeffrey Clemens, Parker Rogers 10 March 2020



Healthcare spending has risen substantially since 1960. In the US, spending has risen from 5% to nearly 18% of GDP, while across the UK, France, Germany, and Japan, it has risen, on average, from just under 4% to just under 11% of GDP. Research has long connected this growth in cost to medical innovation (Smith et al. 2009, Cutler 2004). 

Expensive new medical technologies cause dilemmas around health system costs, access to care, and equity (Chandra and Skinner 2012, Shepard et al. 2019). Countries vary in how they resolve these dilemmas. In the US, for example, treatments including proton beam therapy, the cancer drug Avastin, and the hepatitis C drug Sovaldi are used regularly. By contrast, in the UK the National Health Service (NHS) has strongly limited their coverage (Hawkes 2015, Boseley 2016, Limb 2019). This reflects the NHS's assessment that these treatments are too costly to finance without severe restrictions. 

But why has medical innovation brought cost-increasing enhancements to quality, rather than cost-reducing advances in productivity? And can this be influenced by policy, or is it driven by immutable medical science? 

In our recent work (Clemens and Rogers 2020) we show that fixed-price procurement arrangements can lead medical inventors to increase the emphasis on reducing costs. This is particularly true when prices are low enough that innovation to reduce cost is essential to make sales profitable. We stress that this does not translate directly into a policy recommendation; medical innovation on the frontiers of both cost and quality can have remarkably high value to society.

Using wartime prosthetic device patents to study medical innovation

Our analysis focuses on two striking episodes in the history of prosthetic device innovation: the US Civil War and WWI. These episodes were associated with large increases in demand after battlefield amputations. Sources place the number of surviving amputee veterans at around 35,000 for the US Civil War, and at around 300,000 globally for WWI.

There was an important difference between public procurement models during these two historical episodes. During and after the Civil War, amputees received a fixed allowance to acquire artificial limbs from approved manufacturers. The allowance was lower than pre-war prices for artificial limbs, and manufacturers were prohibited from 'balance-billing' amputee veterans to collect additional charges. 

The programme was thus relatively stingy to manufacturers, and created incentives for them to keep their costs down. WWI, by contrast, is associated with 'cost-plus' procurement arrangements and a shift towards government-led provision of medical devices.

Existing data on patents and clinical trials do not tell us about the detailed economic attributes on which inventors have focused. Analysing how incentives shape the extent to which inventors have emphasised quality versus cost thus required us to generate new data. We used machine learning tools to construct a novel dataset based on the texts of historical patent documents. We constructed variables that describe whether patents have emphasised features including a prosthetic device’s comfort, its appearance, its materials, and the simplicity of its production. 

The quantity of prosthetic device patents

In our empirical analysis, we compare patenting of prosthetic device innovations (our treatment class) to patenting in other medical and mechanical technology classes (our control classes). More specifically, we compare pre-war patenting across our treatment and control classes to patenting during and after the wars. 

Patenting for prosthetic devices rose substantially for several years during both the Civil War and WWI, with increases beginning two or three years into either conflict (Figure 1). At their peaks, the increases in prosthetic device patenting relative to patenting in other medical and mechanical technology classes exceeded 100 log points. 

Figure 1 Patents awarded for prosthetic devices and all other medical devices during the Civil War and WWI 

Source: Clemens and Rogers (2020) using patent data developed by Berkes (2018). 
Notes: The dashed vertical lines are drawn to indicate the years during which prosthetic device patenting rates were elevated. These years are marked in later figures.

Inventors’ emphases on cost and quality 

The demand shock associated with the Civil War generated substantial increases in effort to reduce the cost of producing prosthetic devices. During the Civil War, the average prevalence of the variables we created to capture improvements to the production process temporarily doubled in prosthetic device patents, but was flat in our control groups. This increase in cost-oriented innovation was plausibly driven by the design of the US government's procurement programme. A far more modest increase in production-process innovation occurred during the WWI period. Figure 2 illustrates these differences.

Figure 2 Emphasis on traits by year of production for prosthetic device patents during the Civil War and WWI 

Source: Clemens and Rogers (2020) constructed from the texts of historical patent documents.

Prosthetic device patents contained an increased emphasis on traits connected to mass production during both wars. That is, wartime patents suggest shifts away from bespoke prosthetic limbs. This common shift in emphases is consistent with economies of scale in the supply chain. More specifically, our 'adjustability' variable captures the tendency of prosthetic limb inventors to create limbs that could be mass-produced and subsequently adjusted to the amputee’s height or stump size. Inventors increased their emphasis on this trait across both wars (Figure 3).

Figure 3 Emphasis on adjustability by year of production for prosthetic device patents during the Civil War and WWI

Source: Clemens and Rogers (2020) constructed from the texts of historical patent documents.

But the prosthetic device patents of the Civil War and WWI diverged with respect to product quality. Civil War-era prosthetic device patents exhibited a substantial increase in emphasis on comfort. By contrast, WWI-era prosthetic device patents de-emphasised comfort and exhibited an increase in emphasis on appearance (Figure 4).

These differences can be plausibly, though not definitively, linked to a WWI-era shift in choice away from veterans and towards medical professionals. This shift was accompanied by a heightened emphasis (by both the government and medical professionals) on the re-employment, strenuous rehabilitation, and social reintegration of amputee veterans. 

Figure 4 User traits of patents, comfort and appearance, for prosthetic device patents during The Civil War and WWI 

Source: Clemens and Rogers (2020) constructed from the texts of historical patent documents.

Implications and conclusions

It is important to understand how healthcare policy shapes medical innovation, for example in the context of debates over US legislation to reduce drug prices. In this context, the Congressional Budget Office attempted to estimate the effects such legislation might have on future rates of innovation. Drawing on a set of papers that estimate the relationship between market size and broad-based flows of pharmaceutical innovation (Acemoglu and Linn 2004, Blume-Kohout and Sood 2013, Dubois et al. 2015), the Congressional Budget Office estimated that the imposition of lower prices would at least moderately reduce future drug development (Congressional Budget Office 2019). 

Most research on the effects of markets on medical innovation has focused on the pharmaceutical sector. Other well-known papers in this area include research by Finkelstein (2004) and by Budish et al. (2015). Our paper, by contrast, attempts to understand what drives the development of new pieces of medical equipment. Another such paper (Clemens 2013) analyses patenting in the wake of the US Medicare program’s enactment, and additional papers have focused on estimating the effects of regulatory uncertainty on incentives per se (Stern 2017).

To the best of our knowledge, ours is the first analysis of the effects of procurement models on the extent to which medical innovations emphasise production costs versus quality. We show that the design of incentives can have substantial effects on these margins. Specifically, we show that fixed-price procurement arrangements can induce far greater effort to reduce production costs than cost-plus procurement arrangements. This raises questions about regulatory regimes, payment systems, and the possibility of 'missing innovation'.

The incentives created by payment systems and regulatory approval processes may inadvertently quash invention along various margins. For many cancer drugs, for example, the US Food and Drug Administration (FDA) has historically required new drug applications to demonstrate improvements in life expectancy. Budish et al. (2015) have highlighted how these approval standards, alongside fixed patent terms, may inadvertently blunt incentives for long-term research on early-stage cancers. 

The FDA’s emphases may also result in 'missing innovations' to reduce side effects, to otherwise improve the quality of life, or to reduce cost. More broadly, we should be wary of whether approval processes and payment systems generate too much emphasis on safety and quantifiable dimensions of quality ( life expectancy, for example). This may come at the expense of innovation to reduce cost, or to improve a more expansive definition of a patient’s quality of life (such as reducing side effects). 


Acemoglu, D, and J Linn (2004), “Market Size in Innovation: Theory and Evidence from the Pharmaceutical Industry”, Quarterly Journal of Economics 119(3): 1049-1090

Blume-Kohout, M E, and N Sood (2013), “Market Size and Innovation: Effects of Medicare Part D on Pharmaceutical Research and Development”, Journal of Public Economics 97: 327–336. 

Berkes, E (2018), "Comprehensive Universe of US Patents (CUSP): Data and Facts", unpublished working paper and data set. 

Boseley, S (2016), “NHS ’abandoning’ thousands by rationing hepatitis C drugs”, The Guardian, 27 July. 

Budish, E, B N Roin, and H Williams (2015), “Do Firms Underinvest in Long-Term Research? Evidence from Cancer Clinical Trials”, American Economic Review 105(7): 2044–85.

Chandra, A, and J Skinner (2012), “Technology Growth and Expenditure Growth in Health Care”, Journal of Economic Literature 50(3): 645–80.

Clemens, J (2013), “The Effect of US Health Insurance Expansions on Medical Innovation”, NBER working paper 19761.

Clemens, J, and P Rogers (2020), “Demand Shocks, Procurement Policies, and the Nature of Medical Innovation: Evidence from Wartime Prosthetic Device Patents”, NBER working paper 26679.

Cutler, D (2004), Your Money or Your Life: Strong Medicine for America’s Health Care System, Oxford University Press.

Dubois, P, O De Mouzon, F Scott-Morton, and P Seabright (2015), “Market Size and Pharmaceutical Innovation”, RAND Journal of Economics 46(4): 844–871.

Finkelstein, A (2004), “Static and Dynamic Effects of Health Policy: Evidence from the Vaccine Industry”, Quarterly Journal of Economics 119(2): 527–564.

Hawkes, N (2015), “NHS England Drops 16 Medicines from Cancer Drugs Fund”, British Medical Journal 35: h4803.

Limb, M (2019), “How NHS Investment in Proton Beam Therapy is Coming to Fruition”, British Medical Journal 364: l313.

Shepard, M, K Baicker, and J S Skinner (2019), “Does One Medicare Fit All? The Economics of Uniform Health Insurance Benefits”, NBER working paper 26472.

Smith, S, J P Newhouse, and M S Freeland (2009), “Income, Insurance, and Technology: Why Does Health Spending Outpace Economic Growth?”, Health Affairs 28(5): 1276–1284.

Stern, A D (2017), “Innovation under Regulatory Uncertainty: Evidence from Medical Technology”, Journal of Public Economics 145: 181–200.



Topics:  Health economics

Tags:  procurement, innovation, Medical, prosthetic, healthcare

Associate Professor in the Department of Economics, University of California at San Diego

PhD student in Economics, University of California, San Diego


CEPR Policy Research