This article investigates patterns of pharmaceutical development activity around the creation of the Breakthrough Therapy Designation (BTD), a regulatory award administered by the Food and Drug Administration (FDA) in the United States. The program was created in 2012 in response to a series of early stage clinical trial results in the late 2000s, where the therapies tested showed exceptional gains in efficacy. Although encouraging for patients and sponsors, these cases of “large treatment effects seen early” were problematic because they eroded genuine clinical uncertainty for subsequent testing activities. This erosion could make the late stage phase 3 trials (the most important in the process) ethically contestable on the grounds that control group patients would be significantly less likely to survive than treatment group patients. By instituting the BTD, the FDA introduced a large amount of regulatory flexibility and support for sponsoring firms to mitigate these tensions.

We study the question of whether the program’s creation may have affected industry-wide patterns of drug development activity. The question is motivated by the observation that the program may have indirectly created strong incentives for the industry to pursue the designation. These incentives operate on several important margins of pharmaceutical profitability, including time to market, development costs, and the level of uncertainty faced by developers. On the basis of these considerations, we hypothesize that the creation of the BTD program may have fueled drug innovation specifically aimed at taking advantage of its incentives.

For our empirical analysis, we rely on a comprehensive sample of about forty-five thousand experimental therapies introduced to the clinical trial protocol over the years 2005–2021. This extended sample gives us approximately symmetrical pre and post periods with respect to the program’s 2012 creation. In the dataset, each observation corresponds to an “indication”, that is, a pairing of a drug molecule and a targeted condition, which is the same level at which BTD awards are granted.

For identification, we take advantage of one of the BTD’s main qualifying criteria. Namely, eligible drugs must target a serious or life-threatening disease. Based on this requirement, we assume that the program’s impacts focused on drug innovation targeting conditions that are likely to be considered serious by the FDA. We quantify this likelihood using historical award data from other FDA programs for serious conditions. Specifically, we construct a policy exposure metric that scores targeted conditions in the unit interval. Conditions associated with zero exposure have a minimum likelihood that the FDA will consider them serious or life-threatening; those associated with unit exposure have a maximum likelihood.

Consistent with our primary hypothesis, our main set of empirical results links one higher standard deviation policy of exposure with an increase of 30% in the number of new indications entering the clinical trial process after the program was established. These effects do not reflect the continuation of pre-existing trends and are derived primarily from molecularly targeted drugs, which, given their highly deliberate pharmacology, are more likely to produce substantial efficacy improvements. Consistent with the extended development cycles of the pharmaceutical industry, our estimates reveal that this upswing in new indication introductions exhibited a multiyear lag. It became statistically discernible in 2016 and continued to escalate until 2019, after which it stabilized. We estimate that the program has led to increases of 13% and 19% in the number of indications entering the clinical trials protocol for the respective periods of 2016–2018 and 2019–2021.

Our second set of analyses investigates the potential impact of the program on pharmaceutical productivity, measured by the success rates of development. We consider two potential effects. The first is based on the program’s requirement for applicants to demonstrate substantial improvements over alternatives in early stage trials, potentially incentivizing industry-wide riskier project selection. However, our matched-sample analysis does not reveal any impact on productivity. This result, in line with recent findings on the innovation effects of the Medicare Part D program (Dranove et al., 2022), is consistent with high levels of risk aversion among pharmaceutical developers.

The second productivity effect that we investigate concerns BTD-awarded indications, which benefit from substantial regulatory support and flexibility. Our evidence reveals a significant productivity advantage for these indications, as demonstrated by higher stage advancement rates and successful launches in the US market. Since BTD-awarded indications are a small minority of the population (less than 1%), the effect does not affect industry-wide patterns.

By revealing a surge of new indications entering the development process without a corresponding reduction in productivity, our findings support the prediction of a delayed increase in the number of indications reaching the market. We empirically investigate this prediction by examining patterns of new indication launches in the US market. Our evidence aligns with this prediction, showing a total increase in launches of about 7.4% manifesting toward the end of the sample period, between 2019 and 2021. This figure, which accounts for the approvals derived from the BTD-awarded indications, is not substantially greater compared to the increase of 5.5% observed when these indications are excluded. These results indicate that about three-quarters of the effect of the BTD on commercialization can be attributed to innovation motivated by the aim of accessing the benefits of the designation, even if these are not ultimately accessed. This result is consistent with the view that BTD incentives increase the marginal expected returns to experimentation (i.e., introducing a new indication), with developers acting as Bayesian-updating agents who may choose to continue the development of a non-awarded indication if the signal returned by the experiment is positive enough.

In our final analysis, we examine the interpretation of the observed increase in approvals, specifically whether it represents a transitional or a permanent (steady-state) effect. Our evidence suggests that the estimated effect should predominantly be seen as transitional, given that it does not encompass a substantial number of BTD-fueled indications expected to complete their development cycles in the years ahead. Consequently, we anticipate that the steady-state effect could surpass the magnitude of the transitional effect. A back-of-envelope estimate indicates that the 5.5% increase in approvals for non-awarded indications witnessed between 2019 and 2021 may cumulatively increase to 9.2% by 2025.

As a whole, our analysis links the establishment of the BTD program with significant dynamic innovation effects, manifested through increased volumes of indications introduced to the clinical trial process and launched into the market. An distinct aspect of this effect is that a substantial part of it transpired outside of the program (non-awarded indications). These effects appear to have been largely unforeseen given the program’s historical context. In Section 6, we further discuss the unanticipated nature of these effects, as well as other elements that inform welfare evaluations, including issues related to the quality of the spurred innovation.

Our most direct contribution to the literature is related to work that focuses on the impacts of FDA regulations and policies on pharmaceutical innovation. For example, Yin (2008) studied the innovations impacts of the Orphan Drug Designation (ODD) program. A fundamental difference between the BTD and ODD programs is that the latter was deliberately oriented at fostering drug R&D targeted at rare conditions, thus including explicit incentives (e.g., tax breaks). In contrast, the BTD was created as a “escape valve” for potential ethical challenges, and its incentives are implicit. Peltzman (1973) contended that the higher scientific standards for approval introduced through the 1962 Kefauver-Harris amendments had a chilling effect drug R&D. Yin (2009) argued that, to meet the ODD’s rare disease eligibility requirement, sponsors could have directed their efforts toward developing treatments for specific subgroups within non-rare conditions. Olson and Yin (2018) developed evidence consistent with the notion that the FDA’s Pediatric Exclusivity Designation is strategically employed by pharmaceutical developers to prolong the market exclusivity of highly profitable non-pediatric drugs.

Our research also relates to work focusing on how “pull” incentives (Kremer and Glennerster, 2004) shape pharmaceutical R&D. Among others, this literature documents cases in which incentives come from enhanced patent protection (Kyle and McGahan, 2012), compositional changes in patient population (Acemoglu and Linn, 2004, Dubois et al., 2015), consumption-inducing public health policies (Finkelstein, 2004), or insurance expansions (Blume-Kohout and Sood, 2013, Dranove et al., 2022). In all these cases, the pull incentive’s payoff is contingent on the indication’s approval for commercialization. BTD incentives have a somewhat different structure in that they relate to speed-to-market, development costs, as well as the probability of reaching the market.

While empirical analyses have traditionally concentrated on drug R&D when examining how incentives shape medical innovation, several recent studies present compelling evidence that regulatory and demand-driven factors affect innovation outcomes in the healthcare industry in a more general way. This mainly covers innovations related to medical procedures (Dranove et al., 2022a) and medical devices (Clemens and Rogers, 2023, Galasso and Luo, 2017, Galasso and Luo, 2022, Rogers et al., 2022, Stern, 2017). Beyond reflecting a broader scope of operation beyond pharmaceuticals, which account for less than 10% of healthcare spending in the United States,1 these studies highlight types of incentives and outcomes that have received limited attention in research centered on pharmaceuticals. For example, Stern (2017) emphasizes the significance of regulatory uncertainty, while (Galasso and Luo, 2022) underscore the impact of litigation risk. Clemens and Rogers (2023) distinguish between medical innovation aimed at improving quality and innovation aimed at reducing costs. Our contribution to this strand of literature lies in our emphasis on the ethical dimension, a factor that undoubtedly shapes the broader landscape of healthcare care delivery.

Lastly, in the realm of policy discussions, our findings contribute to the systematic evaluation of the BTD program. To date, this evaluation has focused mainly on what happens “within” the program, i.e., how desirable are the outcomes of BTD-awarded indications. The covered issues include the effectiveness of awarded indications (Darrow et al., 2018, Pease et al., 2017, Puthumana et al., 2018, Ross and Redberg, 2015), speed to market (Chandra et al., 2022, Hwang et al., 2018, Shea et al., 2016), user behavior (Kesselheim et al., 2016, Krishnamurti et al., 2015), and the financial and competitive impacts on sponsoring firms and the markets in which they operate (Garfinkel and Hammoudeh, 2021). In contrast, our work highlights the significant spillover effects of the program, that is, those operating “outside” of it.

The rest of the article is organized as follows. In Section 2, we review the institutional background and analyze the nature of the incentives introduced by the BTD. In Section 3, we describe our main data source, key variables, and their variation. In Section 4, we analyze new indication flows and in Section 5, the potential impacts on productivity. A discussion of our main findings is provided in Section 6, and a conclusion is provided in Section 7.