FDA Breakthrough Therapy Designation Reduced Late-Stage Drug Development Time

Abstract

The Food and Drug Administration’s (FDA’s) breakthrough therapy designation (BTD) program was created to increase patient access to safe and effective therapies by supporting the efficient clinical development of qualifying, clinically meaningful therapies. Using a new data set of key development milestones for drugs approved between 2006 and 2020, including both BTD drugs and a set of comparator drugs identified by FDA experts, we estimated the BTD program’s impact on time spent in late-stage clinical development, measured as the elapsed time between a drug’s end-of-Phase-II meeting with regulators and its approval for marketing. Our analysis suggests that the BTD program lowers late-stage clinical development time by 30 percent. Our findings provide insight into future regulatory and innovation policies aimed at driving efficiency in medical product development to ensure timely patient access to the most clinically meaningful therapies.

In the United States, the efficient development and regulation of new therapies is a central health care policy issue. Newly approved drugs and biologics (hence-forth “drugs”) drive improvements in health outcomes, but they also place substantial pressure on public and household budgets.¹ The high costs of new, on-patent drugs are partly driven by the research and development investment necessary to bring these products to market.

In the US, drug developers typically support new drug approval applications with clinical (efficacy and safety) data from at least two Phase III or “pivotal” trials. These late-stage drug development trials are expensive and time intensive relative to earlier clinical development activities: The average cost of a single Phase III trial ranges from $12 million to $53 million compared with $1 million to $7 million for a Phase I trial and $7 million to $20 million for a Phase II trial.² Further, the estimated length of time to complete these Phase III trials is four years.³ Reducing times to market can have significant financial implications; estimates from analysis of the priority review voucher program indicate that an extra year of marketing time could be worth up to $322 million to a drug developer with a top-selling product.⁴

In an effort to drive efficiencies in drug development and expedite patients’ access to important new therapies, policy makers have implemented regulatory interventions aimed at accelerating drug development timelines, and as a consequence, this may reduce the financial and opportunity costs of drug development. One key type of policy focuses on identifying and facilitating mechanisms for shortening the time needed to develop and approve a new drug. (For an overview of FDA programs for expediting drug development and review, as well as a simple drug development timeline, see online appendix exhibits A1 and A2.)⁵

We studied a relatively new regulatory mechanism aimed at increasing efficiencies in the drug development process: the Food and Drug Administration’s (FDA’s) breakthrough therapy designation (BTD) program.⁶ This program was created by the FDA Safety and Innovation Act of 2012 to expedite the development and review of drugs for serious and life-threatening conditions with unmet needs.⁷ Per the FDA, a drug receiving BTD is eligible for the following: intensive guidance on an efficient drug development program, beginning as early as Phase I clinical trials; FDA organizational commitment involving senior managers; and all fast-track designation features (primarily including the ability to submit an approval package using a “rolling review” of materials).⁷

Although the BTD program spans two medical product centers within the FDA, we focused specifically on the program as administered by the Center for Drug Evaluation and Research (CDER), where the majority of BTD requests are reviewed. (For simplicity, hereafter we use “FDA” and “CDER” interchangeably.)

The BTD program is a meaningful program for assessing the ability of regulatory programs to affect drug development time, for three key reasons. First, the program focuses on expediting the development of the most clinically meaningful drugs. The bar to receive a BTD is higher than for other designation programs at the FDA. The BTD requires that companies present “preliminary clinical evidence [that] indicates that the drug may demonstrate substantial improvement over available therapy on a clinically significant endpoint.”⁷ The requirement to present clinical evidence of efficacy means that only drugs with the potential to drive significant improvements in patient care are included in the program—that is, precisely the types of therapies whose development policy makers, patients, and clinicians desire to expedite.

Second, the administration of the BTD program is a major priority for regulators. The institutional commitment to “breakthrough” drugs is conducted at the highest levels of CDER leadership. According to the agency, “FDA will seek to ensure that a sponsor of a product designated as a breakthrough therapy receives timely advice and interactive communications to help the [drug developer] design and conduct a drug development program as efficiently as possible.”⁷ For example, the CDER Medical Policy Council (which includes senior leadership such as the center director and the director of the Office of New Drugs) is consulted on complex or novel drug development issues. Administration also entails direct involvement by staff-level leadership and, in particular, senior review and project management staff, who ensure efficiency in FDA-firm interactions and throughout the FDA’s extensive collaborative, cross-disciplinary review process. Although guidance and involvement by senior-level staff are not unique to BTD drugs, the commitment to ensuring efficiency in these processes is a central pillar of the BTD program.

Third, the activities of the BTD program are targeted at streamlining the most lengthy and costly stage of drug development: the later stages of clinical trials. Drug developers generally apply for this designation before beginning late-stage drug development, and, as a result, the program’s main effects are most likely to be observed between the end of Phase II trials and the submission of a marketing application. One potential outcome of the program is that the BTD program may allow drug developers to design and conduct more efficient pivotal clinical trials because of earlier FDA engagement.⁸ However, there are multiple additional avenues through which the BTD may affect late-stage drug development. FDA officials describe these benefits as including “ongoing consultation regarding trial conduct and discussion of needed trial modifications, discussions on the approach to analyses of trial end points, enhanced guidance…on manufacturing to assure earlier readiness for production [and] faster Agency responses to emerging issues.”⁹

Multiple studies have investigated the impact of FDA expedited programs, including the BTD program, and found both that novel drugs have increasingly used these programs and that participation in them is correlated with shorter development time.^10–14 There have also been descriptive analyses of the BTD program; these studies find shorter development time for BTD drugs, with a median development time (from initiation of investigatory trials to approval) reported as 4.9 years.^15–17 Only one prior study used a causal inference methodology to assess the impact of the BTD program on development time and found that it shortened clinical development time by nearly 25 percent, without any evidence of an increase in reported adverse events.¹⁸

This study built on the previous literature by conducting a causal inference analysis using a recognized measure of late-stage development time that has not previously been studied in the context of the BTD program, as well as a novel control group identified by FDA experts.

Study Data And Methods

To analyze the impact of the BTD program on drug development time, we conducted a multi-variate regression analysis to plausibly isolate the program’s effect on late-stage drug development time, as described below.¹⁹

Data Sources

We considered all new molecular entities (NMEs) first approved by the CDER during 2006–20. NMEs are novelsmall-molecule drugs and biological products that have not previously been approved in the United States.

Data on NME approvals and their regulatory characteristics were obtained from the CDER’s public data set, Compilation of CDER New Molecular Entity (NME) Drug and New Biologic Approvals.²⁰ Regulatory characteristics of drugs in the analysis sample included BTD status, approval date, whether the NME was a small-molecule drug or a biologic, whether the NME used any other FDA expedited program (for example, accelerated approval, fast track, or priority review), and whether the NME was designated as an orphan drug.

We excluded from our analysis novel diagnostics (n = 25), such as contrast agents used in radiology procedures, and contraceptives (n = 3). These products have atypical development programs that are not representative of the larger (therapeutic) drug development landscape.

We classified the indication for each NME in our sample into one of fourteen mutually exclusive disease categories, using the World Health Organization’s Anatomical Therapeutic Classification system.²¹

Finally, to proxy for resources and regulatory experience that a drug developer can draw on, we constructed a variable indicating whether the developer was a publicly listed or privately held company.

IDENTIFYING A NOVEL CONTROL GROUP

To identify the impact of the BTD program, we first defined a group of control BTDs—that is, drugs from before the implementation of the BTD program in 2012 that possess characteristics that would have qualified them for a BTD.

The control BTDs were manually assigned by two physician experts who are former members of the CDER’s Medical Policy Council (that is, experts in the BTD determination process). Each expert received a list of NMEs approved between 2006 and 2012 whose applications for approval were submitted before the initiation of the BTD program in mid-2012. The two experts studied the clinical review documents from the approval packages of each NME and determined whether they believed that the clinical data would have supported the NME’s designation as a break-through therapy, had the program been available at the time.

The determinations of both experts were then compared. More than 80 percent of expert determinations were identical across evaluators. The drugs with different determinations were discussed in a meeting with one of this study’s investigators. If consensus could not be reached between the experts, or if there was uncertainty around the determination, the decision was discussed with additional subject-matter experts within the CDER to provide clarification. Final determinations were made for all of these NMEs (whose applications for approval were submitted before the initiation of the program), resulting in the classification of 38 control BTDs and 155 control non-BTDs.

Our resulting data set consisted of four distinct groups of drugs: control BTDs approved between 2006 and 2012 (NMEs classified as BTD-type by FDA experts), true BTDs approved between 2012 and 2020 (NMEs that received a BTD), control non-BTDs approved between 2006 and 2012 (NMEs classified as non-BTD-type by FDA experts), and true non-BTDs approved between 2012 and 2020 (NMEs that did not receive a BTD).

MEASURING TIME SPENT IN LATE-STAGE DEVELOPMENT

Our primary outcome of interest was the measurement of late-stage development time. We defined this as the elapsed time, in days, between a central development milestone meeting, referred to as the end-of-Phase-II meeting, between the FDA and the drug developer and the date on which the FDA approved the NME.

The end-of-Phase-II meeting typically takes place before the initiation of the pivotal trials and is used to discuss the design and conduct of these trials. Based on FDA definitions, we defined pivotal trials as the final clinical trials or, infrequently, the single clinical trial, used to support the approval of an NME through the provision of safety and efficacy data.²² The end-of-Phase-II meeting can include discussions of eligibility criteria (inclusion or exclusion) that define the study population, trial endpoints, dosing, data and safety monitoring plans, and other statistical and clinical facets of the planned pivotal trials. The goal of these meetings is to give the FDA an opportunity to formally comment on critical aspects of the drug developer’s planned trials and discuss and reach consensus on any differences in preferred approaches.

To identify the date associated with the end-of-Phase-II meeting for the NMEs in our analysis sample, we relied on two sources of data. Where possible, the meeting date was collected from publicly available review documents associated with the drug’s approval, which were retrieved via the Drugs@FDA website.²³ Dates were typically found within the “Summary of Presubmission/Submission Regulatory Activity” section of the clinical review. When multiple end-of-Phase-II meetings were reported, we used the date of the earliest meeting.

There were three notable categories of exceptions to this end-of-Phase-II meeting identification method. First, in cases where the meeting date was not available from public sources (n = 185), one of the authors used an internal FDA database to identify meeting dates, where possible. Dates that were coded via this method, as well as instances where no date was identified, are considered commercial confidential information by the FDA and cannot be made public.

Second, although the end-of-Phase-II meeting typically marks the last meeting between regulators and drug developers before the initiation of the pivotal trials, in some cases, pivotal trials may be conducted in an earlier phase. When this happens, the final meeting before pivotal trials begin is the end-of-Phase-I meeting, rather than the end-of-Phase-II meeting. In such cases, we used the end-of-Phase-I meeting date as a surrogate for the (nonexistent) end-of-Phase-II meeting date. For simplicity, hereafter we use “end-of-Phase-II” to describe all pre–pivotal trial meetings.

RELATIONSHIP BETWEEN BTD STATUS AND LATE-STAGE DEVELOPMENT TIME

In our regression models, we examined whether drugs that were part of the BTD program experienced shorter late-stage development time compared with our novel control group of drugs that may have met the BTD qualification criteria before the establishment of the BTD program itself.

Our sample included drugs that did not meet the BTD qualification criteria in both the pre- and post-BTD implementation periods, enabling us to control for overall trends in drug development, FDA review times, and other factors that are not associated with the establishment of the BTD program but that might affect clinical development time. The distribution of our dependent variable was right-skewed, so we took the natural log-transformation of the variable and estimated ordinary least squares models with robust standard errors. We controlled for and modeled several potential cofounders, including the other FDA programs that are also intended to expedite the drug development or review process. Detailed variable descriptions and the according regression specifications are in appendix exhibit A3.⁵ To further motivate our analysis, appendix exhibit A4 presents a case study of two drugs in our sample with the same approved indication, only one of which received a BTD.⁵

INFORMAL IMPACT ESTIMATE

To illustrate the financial implications of our results, we performed an informal calculation to estimate the impact of shorter late-stage development time. For this exercise, we adopted a typical late-stage development time of four years (consistent with our data) and a range of discount rates of 7–15 percent to account for the opportunity costs of capital investments.^24,25 This calculation is formally presented in appendix exhibit A5.⁵

LIMITATIONS

There were several limitations to our study. First, given that there have been technological shifts in the science of drug development over time, it is possible that the drugs we used as controls in the earlier period of our study were not ideal comparators to later-approved drugs in the analysis sample. Indeed, this is also a limitation for using our findings to predict the future: Because of both technological shifts and the types of therapies coming to market, the effect of the BTD program may change in the years ahead.

Second, for some drugs without an end-of-Phase-II meeting date, we used the end-of-Phase-I meeting date, potentially introducing a concern that our dependent variable was measured inconsistently. However, for these drugs, the end-of-Phase-I meeting date signals the beginning of the pivotal trials. Thus, we believe that within the context of this analysis, this date captured an equivalent point on the drug development timeline: the initiation of late-stage drug development. We note that our regression controlled for both orphan drug designation and accelerated approval status, which are the only types of drugs that may have been coded with an end-of-Phase-I meeting date. In addition, as a robustness check, we ran a regression that removed both types of drugs and found that the results were consistent with those in our preferred specification.

Finally, the experts’ identification of control BTD drugs was necessarily conducted ex post and may have been influenced by the drug’s observed development process (rather than the other way around). These concerns were mitigated by the fact that the experts’ selection criteria for identifying control BTD drugs were based on the drugs’ clinical characteristics. For transparency, we publicly report the list of control and true BTD drugs for future researchers.¹⁹

Study Results

Exhibit 1 presents summary statistics and compares control and true BTD drugs with non-BTD drugs. The final data set included 434 NMEs with an identified end-of-Phase-II meeting date. These 434 NMEs were allocated to the four previously defined groups: control BTD drugs, true BTD drugs, control non-BTD drugs, and true non-BTD drugs.

EXHIBIT 1.

Characteristics of control and true breakthrough therapy designation (BTD) drugs versus control and true non-BTD drugs, 2006–20

	BTD drugs		Non-BTD drugs
Characteristics	Control	True	Control	True
No. of drugs	27	85	129	193
Small moleculea (%)	61	63	84	77
Any FDA expedited programa (%)	89	100	36	50
No. of FDA expedited programsa	1.8	2.0	0.6	0.8
Boxed warninga (%)	50	27	46	29
Orphan drug designationa (%)	39	73	24	29
Publicly listed firma (%)	89	93	75	87
Cancer druga (%)	28	57	25	27
Time measures (average days)b
EOP2 to submission	1,224	1,065	1,317	1,522
EOP2 to approval	1,523	1,298	1,851	1,880
Submission to approval	299	233	534	358

SOURCE Authors’ analysis of publicly available and confidential Food and Drug Administration (FDA) data, 2006–20. NOTES This table shows drug characteristics for the sample of 434 drugs used in the analysis. Detailed variable descriptions are in appendix exhibit A3 (see note 5 in text). All variables were measured at the drug level. “Control” drugs are defined in the text.

^aTo maintain data confidentiality, the indicated drug characteristics were obtained from a sample of drugs with publicly available data for the end-of-Phase-II (EOP2) meeting date. Statistical analyses from this sample show no significant differences when compared with the complete sample.

^bCalculated from the complete sample.

A comparison of means showed that average late-stage development time for control BTD versus non-BTD drugs was similar before the establishment of the BTD program but diverged there-after, an indication of the program’s possible impact (appendix exhibit A6).⁵

BTD STATUS AND LATE-STAGE DEVELOPMENT TIME

Exhibit 2 presents the regression results and demonstrates that true BTD status significantly reduced late-stage development time. The regression estimates from model 1 show that BTD status was associated with a statistically significant (p< 0:001) 32 percent decline in late-stage development time. (The regression results with untransformed coefficients are in appendix exhibit A7.)⁵ The effects remained relatively consistent as additional controls for drug (for example, small molecule and approved indication, as measured by Anatomical Therapeutic Classification codes), firm, and regulatory (for example, orphan drug designation status and inclusion in other FDA expedited programs) features were added (models 2 and 3). In the regression specification that was fully saturated with controls, model 3, true BTD status was associated with a statistically significant (p< 0:001) 30 percent decline in late-stage development time.

EXHIBIT 2.

Effect of the breakthrough therapy designation (BTD) program on late-stage drug development time, 2006–20

	No. of days between the end-of-Phase-II meeting and NME approval
	Model 1	Model 2	Model 3
BTD status (%)	8.06	2.74	1.71
BTD status × post-BTD implementation (%)	−31.75***	−32.70***	−30.30***
Additional controls (%)
Small molecule	—a	−1.19	−0.40
Boxed warning	—a	4.08	−4.08
Orphan	—a	7.47	7.57
No. of FDA expedited programs	—a	−8.33***	—a
Priority review time	—a	—a	−7.13
Fast track	—a	—a	−3.34
Accelerated approval	—a	—a	−19.99***
Publicly listed firm	—a	−8.15*	−8.06*
Cancer drug	—a	−9.15	−6.85
Controls: approval year	Yes	Yes	Yes
Controls: drug characteristics	No	Yes	Yes
Controls: diseaseb	No	Yes	Yes
No. of observations	434	434	434
R-squared	0.1735	0.2351	0.2446

SOURCE Authors’ analysis of Food and Drug Administration (FDA) data, 2006–20. NOTES This table presents the regression results of ordinary least squares models examining the effect of the BTD program on late-stage development time for drugs first approved between 2006 and 2020. The outcome is the number of days between the end-of-Phase-II meeting and drug approval. The results shown here are based on an exponential transformation of the regression estimates shown in appendix exhibit A3 (see note 5 in text). For example, the coefficient on BTD × post-BTD implementation (which is shown in appendix exhibit A3) is associated with a (exp(B) − 1) × 100 percent change in duration (which is shown in this exhibit). BTD × post-BTD implementation is an indicator variable for a drug that is a true BTD with a marketing application that is submitted to the FDA on or after the start of the BTD program. Detailed variable descriptions are in appendix exhibit A1. NME is new molecular entity.

^aVariables were not included in the model indicated.

^bRefers to controls for the Anatomical Therapeutic Classification code associated with the drug’s first approval.

^*p < 0:10

^***p < 0:001

As expected, participation in additional expedited programs was associated with a decrease in late-stage development time relative to drugs that did not participate in such programs. For example, drugs approved under accelerated approval experienced a 20 percent decline in time spent in late-stage development relative to drugs without accelerated approval. Firm size and regulatory experience (as proxied by whether the firm is publicly listed) were also associated with a decline in late-stage development time, although the effects were less statistically significant, consistent with fewer degrees of freedom (because of multiple controls) in the final regression specifications.

HETEROGENEITY OF BTD PRODUCTS

The regression results presented in appendix exhibit A8⁵ indicate that the BTD program was associated with reduced late-stage development time for each of the four most common diseases; however, the effects varied in magnitude and were all not statistically significant, likely because of the small sample size. The results presented in appendix exhibit A9⁵ indicate that the BTD program may have resulted in larger reductions in drug development time for publicly listed versus privately held drug developers. However, these effects varied in magnitude and were not significant across all groups—again, likely because of the small sample sizes.

ROBUSTNESS

Our results were robust to instead using the time from the end-of-Phase-II meeting to submission of the drug application (appendix exhibit A10), to including only NMEs with publicly available end-of-Phase-II meeting dates (appendix exhibit A11), and to using various alternative estimation samples (appendix exhibit A12).⁵ Alternative estimation samples included excluding NMEs that were approved under accelerated approval or were orphan designated (that is, those that may have had non-traditional or shortened development timelines, including all NMEs using an end-of-Phase-I, rather than end-of-Phase-II, meeting date), and excluding NMEs whose approval packages were first submitted to the FDA in 2011 or 2012 (that is, those approved around the introduction of the BTD program).

Finally, we also conducted a “placebo test” of the effect of the fast-track program on late-stage development time (appendix exhibit A13).⁵

INFORMAL IMPACT ESTIMATE

We estimated that the BTD-induced reduction in late-stage development time would lower the revenue needed for a drug to achieve profitability by 9–18 percent (appendix exhibit A5).⁵ Equivalently, the reduction in late-stage development time associated with the BTD program would enable investments that were about 22 percent larger than without this reduction.

DISCUSSION

Patients, clinicians, and policy makers all share the goal of efficient access to safe and effective medical products. To this end, the BTD program was created in 2012 to expedite the development of clinically meaningful drugs for serious and life-threatening conditions with unmet needs. We found evidence that the BTD program has been associated with 30 percent faster late-stage development time among NMEs.

Notably, the effects of the BTD were stronger among privately held drug developers, highlighting the additional benefits that the BTD program may have on improving the ability of smaller (or less resourced) drug developers on bringing these drugs to market.

Taken to its logical conclusion, a decrease in late-stage development time may, in turn, lower development costs, stimulate drug developers’ research incentives for future development projects, and ultimately increase patients’ access to safe and effective medicines. We estimated that the BTD-induced reduction in late-stage development time of 30 percent would lower the threshold needed for an NME to achieve profitability by 9–18 percent. These potential benefits were approximated from the perspective of a drug developer; however, the full social benefits of innovations associated with BTD would extend to patients in the form of improvements in quality of life, life extension, and other benefits of novel medical products.²⁶

POLICY IMPLICATIONS

Although we focused on just one regulatory program, there are several ways in which findings from the BTD program may shed light on policies and programs to expedite new therapeutic development in other settings.

First, the BTD experience highlights the importance of designing policies that foster transparency between regulators and drug developers. A key feature of the BTD program is that it formalizes the need for communication between regulators and developers during development, such as intensive regulatory guidance and organizational commitment from senior managers. This stands in contrast to the fast-track program, an expedited regulatory program in which non-clinical evidence is acceptable for designation, that provides nearly all of the same features of the BTD program except this intensive organizational commitment, and that was not associated with declines in late-stage development time over the same time horizon (appendix exhibit A13).⁵ This higher level of FDA guidance thus emerges as the likely contributor to an efficient drug development program, suggesting that drug developers should use regulatory resources and guidance whenever possible.²⁷

Second, in addition to engaging regulators, initiatives aimed at efficient drug development should also continue to engage the primary stakeholders: patients. Patients and care partners can make significant contributions to the drug development process—for example, highlighting appropriate, patient-relevant trial endpoints, which can help reduce the need for redundant or prolonged clinical trials, as well as engaging with the community on new trials and recruitment.²⁸

Third, it is noteworthy that the BTD program was not accompanied by any additional FDA resources such as user fees, nor were new funds appropriated.⁸ Rather, staff-level regulatory resources to serve the program were redirected from existing review programs. However, senior-level regulatory resources may have been diverted from other non–new drug regulatory activities that these senior experts advise on (for example, monitoring manufacturing and drug shortage issues or policy around generic drugs and biosimilars). It is thus possible that the reallocation of effort by these senior regulators had both positive externalities (for example, as they were able to craft uniform policy solutions to frequent new drug issues) and negative externalities (for example, as the senior regulators reallocated time away from other public health–promoting regulatory activities).

Although an analysis that considered all potential externalities was beyond the scope of this study, our findings suggest that regulatory interventions can drive efficiencies, even under resource constraints. However, as a corollary, our results also suggest that a better-resourced regulator, which is equipped to provide relevant feedback at earlier points in the drug development timeline, could also have significant impacts on expediting the development of clinically significant drugs.

Finally, the results of this study may have broad generalizability to newer domestic and international programs that have sought to replicate the BTD program’s goals, such as the European Medicines Agency’s PRIME Program (launched in 2016)²⁹ and the FDA’s break-through devices program (launched in 2018).³⁰

Conclusion

Policy makers and patients have voiced concerns that the high cost and long timelines of drug development may drive high drug prices and impede access to the most clinically meaningful products. The results of this study suggest that regulatory policies that increase transparency between regulators and drug developers are associated with significant declines in time spent in late-stage drug development. This highlights both the potential and the centrality of regulatory policies in expediting drug development, and ultimately providing patients with timely access to lifesaving medicines.