Slow, Costly Clinical Trials Drag Down Biomedical Breakthroughs
The biotechnology industry is on the brink of making advances that will amount to secular miracles, from restoring sight to the blind to exorcising cancer through gene-edited immune cells. That progress is threatened, however, by the extremely high, and rapidly increasing, cost of running clinical trials. If we do not bring these costs down, we risk either waiting too long for efficacious drugs or paying for drugs that may not work at all. As the global leader of the biotech industry, the U.S. should lead the way in generating rigorous evidence in clinical trials. To that end, this brief proposes six important reforms:
- Deregulating clinical trial advertising
- Fixing tax disincentives for research participants
- Extending the orphan disease “income exclusion” exemption of $2,000 to all research participants
- More Institutional Review Board (IRB) transparency
- Walking back misguided FDA and congressional diversity mandates
- Affirming FDA’s commitment to 21st-century methods of trial monitoring
Clinical drug trials are very expensive. According to one estimate, for trials conducted in 2015–17, the median cost was $41,413 per patient. But in some trials, that figure is significantly higher. According to George Yancopoulos, cofounder and chief scientific officer of Regeneron, per-patient costs can now reach $500,000—a stark increase from the $10,000 per patient that was typical when he was first starting in the field. As Kenneth Getz, head of the Tufts Center for the Study of Drug Development, put it: “In my 30 years of experience in the clinical trial industry, costs have only moved in one direction: up.”
Some structural factors have increased the costs of Randomized Controlled Trials (RCTs). First, because medicine is increasingly focused on chronic diseases, trial duration has necessarily lengthened. Longer trial durations give valuable information on long-term efficacy and safety and thus should be encouraged. The number of subjects needed for clinical trials has also increased because new medicines compete in a crowded field of incumbents, and it takes more subjects to demonstrate incremental gains versus the standard of care.
In addition to these structural factors already mentioned, Getz believes that high levels of fragmentation among vendors and service providers (e.g., specialty services and technology solutions companies) might play a role. This fragmentation prevents larger and more efficient contract research organizations (CROs), as well as data-management and service providers, from exploiting economies of scale to produce more affordable trials—though a biotech executive with 20+ years of experience attributed that fragmentation to how slow-moving and bureaucratic larger CROs tend to be, which limits any advantages to their size. Perhaps another factor is complacency in the CRO industry, as one blogger has argued.
The Wrong Ways to Lower Costs
One way to reduce these costs is to go through an FDA process known as Accelerated Approval, which imposes less up-front scrutiny on new drugs. This pathway allows for earlier approval based on improvement in some surrogate marker that is thought to predict survival, rather than on a demonstrated improvement in survival itself. Partly as a response to rising costs—and as a way to speed patient access to promising new drugs—pharmaceutical companies have made increasing use of Accelerated Approval, especially for oncology drugs.
For drugs approved in this way, however, confirmatory studies actually demonstrating survival are supposed to be conducted after approval. But as Dr. Vinay Prasad aptly documents in Malignant, which focuses on cancer treatment, such confirmatory studies are only occasionally performed. The result has been a proliferation of questionably efficacious oncology drugs over the last 25 years. The Accelerated Approval pathway should not be eliminated: it has had some impressive successes. One review found that 75% of the non-oncology drugs that received Accelerated Approval between 1992 to 2017 would eventually obtain regular approval, and the pathway sped drug approval by about 4.5 years.
In some cases, FDA has even allowed for the use of these intermediate endpoints—previously suited only for Accelerated Approval—in the regular approval process. For example, everolimus—which can be combined with standard-of-care treatment for advanced breast cancer—gained approval based on an improvement in “progression-free survival,” which refers to the period of time a patient is free of signs of disease progression (usually measured with imaging or blood tests). But trials have not detected any increase in overall survival.
Reducing the “drug lag”—which economists have long criticized FDA for—represents an important success. But FDA should not allow Accelerated Approval–standards to be used in the regular process. And it should consistently enforce post-approval requirements on drugs accepted through Accelerated Approval, so that useless drugs are more quickly recognized and their approval rescinded. The recently passed Food and Drug Omnibus Reform Act of 2022 (FDORA)  does move in this direction, giving FDA more ways to encourage sponsors to conduct confirmatory studies.
Historically, another way for pharmaceutical companies to conduct large and (ideally) fast trials has been to outsource them to other countries. From 2005 to 2012, registered trials doubled in low-income countries and quadrupled in lower-middle-income countries. But there is a significant problem with this approach. New drugs have to show benefits compared not simply with a placebo but with the current standard of care for the condition in question—but the current standard of care in other countries may not be the same as it is in the United States. An international trial, then, may show that a new cancer drug is better than, say, China’s standard of care, but we wouldn’t know whether it outperforms the U.S. standard of care.
Good Trials Help Everyone
In addition to weeding out ineffective drugs, high regulatory standards can also produce unexpected benefits. In 2008, for example, FDA issued guidance that required long-term follow-up on cardiovascular outcomes for anti-diabetic drugs. The required follow-up studies on two classes of these drugs—GLP-1 agonists and SGLT-2—helped reveal the myriad benefits that they may have in other diseases, including heart failure, kidney disease, and obesity. This promises great benefit to public health because GLP-1 agonists have proved very safe, while many current diet pills (such as rimonabant) have historically had a mixed track record on safety. Instead of making it even easier for ineffective drugs to get past regulators or continuing to neglect required post-marketing studies, we should make it easier to generate clinical evidence.
The chaos of the early days of the Covid pandemic spurred new innovations in cheap and effective clinical trials. Among the most successful was the UK RECOVERY trial. Professor Martin Landray, a co–chief investigator, summarized the trial’s impact in conversation:
At a cost of $20 million for 48,000 patients, the RECOVERY trial cost about $500 per patient. This platform trial has produced answers for 10 different therapeutics, four of which reduce the risk of death for the sickest patients with Covid-19 and six of which have no meaningful benefit. Thus, that is about $50 per patient per answer.
Compared with the median clinical trial per-patient cost of about $40,000, RECOVERY was a bargain. It led to new discoveries about very old drugs like dexamethasone, which was estimated to have saved more than 630,000 lives worldwide, by cutting mortality among Covid-19 patients requiring oxygen therapy and/or ventilation. RECOVERY also found that tocilizumab (a drug usually used for rheumatoid arthritis) further reduced the risk of death in similarly sick Covid-19 patients with elevated levels of a marker of inflammation (C-reactive protein). Tocilizumab was subsequently approved by more than 30 regulatory agencies around the world. The trial also found that REGEN-COV2 (a monoclonal antibody combination targeted at the viral spike protein) saved lives in Covid-19 patients who lacked antibodies to Covid (termed “seronegative”), leading to emergency-use approval for the drug in the U.S. and beyond.
High-quality and affordable trials were also carried out in the United States. Professors David Boulware and Carolyn T. Bramante and the COVID-OUT team ran a decentralized clinical trial that tested three generic drugs (metformin, ivermectin, and low-dose fluvoxamine) for their ability to prevent death or emergency-room visits with Covid (later analysis of trial data also found a possible benefit of metformin in preventing long Covid). In conversation, Bramante estimated that their per-person trial costs netted out to less than $1,950 per person, while the original hydroxychloroquine arm (with a much shorter follow-up period) cost a little less than $200 per patient, per Dr. David Boulware. Few, if any, biomarkers or expensive lab tests had to be collected, and much of the workflow to collect outcome information was automated. These important but frugal trials show that an order-of-magnitude cost reduction in trials is eminently possible.
HHS in the Mirror
A 2014 report commissioned by the U.S. Department of Health and Human Services (HHS) on the barriers to clinical trials is a useful starting point in understanding the needed reforms. The pharmaceutical industry itself imposes some of the most important barriers to cheaper trials, such as increasing protocol complexity or attempts to wring out more data per participant through extensive and expensive bloodwork and imaging. This impulse is understandable, given how valuable this kind of data can be. But a 2022 report by the Tufts Center for the Study of Drug Development found, not surprisingly, a positive correlation between the number of endpoints measured in a protocol and trial costs. There may be no regulatory fix for overcollection of data, though it might be gently discouraged if regulators would stop demanding a lot of irrelevant preclinical minutiae. However, the HHS report also identifies several regulatory reforms that could be pursued, such as transitioning away from source-data verification, clarifying overlapping regulation, and weighing the risks and benefits of overcompliance with extant privacy laws.
Regulatory Overlap and Ambiguity
One theme highlighted in the 2014 report is death by a thousand overlapping bureaucratic cuts. For instance, the report notes that some oncology research is governed by up to four federal agencies: “FDA ... the Office for Human Research Protections (OHRP), the National Institutes of Health (NIH), and the Office for Civil Rights (OCR),” which makes compliance with regulation unnecessarily burdensome and confusing. In other cases, ambiguous guidance from FDA makes companies overly cautious with trial design. In both cases, forcing federal agencies to come to a clear consensus in a timely fashion would resolve these issues.
Cures Act Brings Some Positive Change
The 2016 21st Century Cures Act directed FDA to harmonize regulations on human-subject research and use more “real-world evidence” (a catch-all term that can refer to large-scale observational data or to pragmatic clinical trials). The latest FDA rulemaking on human-subject regulation, pursuant to Congress’s instructions, was proposed in September 2022. In short, the proposed rule would bring FDA closer in line with the principles that have been adopted by 20 other government agencies (known as the Common Rule). It also exempts research that has progressed to the data-analysis stage from continued Institutional Review Board (IRB) review—a form of oversight that is normally required for any research that could be (very broadly) construed as human experimentation, including research that merely follows subjects as they undergo routine clinical care. The proposed rule also more forcefully demands that research conducted across multiple sites use a single IRB. Overall, the proposal is a tardy, but welcome, attempt to address the regulatory ambiguity identified in the 2014 HHS report.
Unfortunately, in other respects, the proposed rule may further slow research. The proposal would require that any trials involving “vulnerable” subjects, including the “economically or educationally disadvantaged,” provide “additional safeguards” to protect the rights and welfare of these subjects. It would also require that in selecting the members of IRBs that regularly review research involving such subjects, “consideration shall be given to the inclusion of one or more individuals who are knowledgeable about and experienced in working with these categories of subjects.” In practice, this could require anything from additional paperwork to additional training for staff to administrative meetings dedicated solely to these topics. It is too early, and the language is too vague, to tell. While the previous version of this rule called for efforts to increase gender diversity on IRBs, those efforts had to be “nondiscriminatory” and not made on “the basis of gender”—a requirement that the new rule would remove. It is too early to tell what effect these changes will have, but as Carl Schneider has argued at length, piling on protections for “vulnerable populations,” given the extreme safety of research, provides little value, while slowing research in those populations.
The more recent Food and Drug Omnibus Reform Act of 2022 is a mixed bag. It requires FDA to issue guidance on the use of digital technologies in trials and adaptive trial designs, which would help streamline them. However, it also continues to push an evidence-free diversity mandate, instructing FDA to require “diversity action plans” for new drugs and devices and to convene a public workshop on trial diversity. It also imposes reporting requirements so that Congress can grade FDA on its diversity efforts—all of which is a further drain on agency time and resources.
21st-Century Trial Monitoring
Another area that could benefit from regulatory reform is FDA’s approach to source-data verification. In 1998, the agency issued guidance that effectively required in-person double-checking of every bit of data submitted by a given site to the sponsor of the trial. In practice, this meant frequent “site visits” where the sponsor of a drug would send employees to various clinical trial sites to verify data. In 2013, FDA followed the European regulatory agencies in backing away from this burdensome approach, issuing new guidance that relaxed requirements. Instead, the agency adopted a new strategy of “risk-based monitoring,” which involves more targeted verification of data. In this approach, data can be remotely monitored for anomalies that require further investigation. It also called for focused in-person site visits when necessary, such as at the initiation of a trial with a complex protocol.
Implementing the risk-based approach, however, has proved more difficult. A recent industry survey showed that, as of 2021, most risk-based monitoring principles had not yet been widely implemented—save for one (remote monitoring), which was widely adopted only during the Covid pandemic. Despite its rapid adoption, remote monitoring led to no discernible rise in non-Covid-related study protocol disruptions. That’s encouraging, but there is much room for continued improvement.
The founders of Lindus Health, a next-generation CRO aimed at accelerating clinical trials, believe that widespread adoption of risk-based monitoring would reduce costs substantially, and they suggest that FDA could take steps to spur the use of the approach. For example, FDA could encourage reviewers to say in the pre-submission process: “This trial’s source-data verification processes are too onerous, and we recommend risk-based monitoring instead.” If done frequently enough, this would reassure CROs that FDA is truly committed to risk-based monitoring. Given the substantial autonomy that FDA reviewers enjoy, top-down guidance is necessary but insufficient. In addition, FDA should pursue cultural change at the reviewer level. This would take the form of updated reviewer education such as workshops instructing how to assess risk-based monitoring practices, as well as a recognition program, akin to the existing FDA “Scientific Achievement Awards,” which would celebrate employees who demonstrate “Excellence in Risk-Based Monitoring.”
Transparency for Smarter Regulation
Another problem highlighted in the HHS report is the difficulty generated by our current privacy regime. Privacy laws, principally the Health Insurance Portability and Accountability Act of 1996 (HIPAA) make IRBs cautious about recontacting patients who have dropped out of studies. While a carefully worded consent form will permit recontact even after withdrawal, not all researchers anticipate this problem. In addition, some IRBs frown upon the centralized collection of patient identifiers that would aid in an attempt to recontact. Moreover, consent forms cannot anticipate all future data uses: a clinical trial designed to answer a specific clinical question, in the fullness of time, may yield other valuable information, such as through reanalysis of biological samples with new methods. This is referred to as “secondary use of data.” A recent stunning example of regulator intransigence on such a question was the FDA preventing early testing for Covid in biological samples originally collected for flu surveillance.
According to legal expert Clint Hermes, many research institutions have by now found ways to comply with privacy regulations while minimizing the burden on researchers. For example, they may use HIPAA authorizations to allow for participant recontact for a specific study or IRB waivers for more complex cases. In addition, HHS has historically allowed researchers within a given institution to review protected health information for recruitment purposes, and privacy regulations do not prevent researchers from contacting patients about enrolling in a trial after viewing their protected health information. However, some institutions without sophisticated regulatory departments lack the necessary expertise to strike this balance. As a consequence, Hermes believes that there is more variability in IRB behavior, ranging from “unnecessary and sometimes irrational burdens ... to plainly unethical and illegal research with rubber stamp IRBs.” This may explain why IRB horror stories persist, even as research continues largely unabated at experienced institutions.
Professor Holly Fernandez Lynch, an IRB expert, has argued that transparency may be a way to systematically improve IRB decision-making. In practice, this would mean requiring IRBs to post decisions and meeting minutes online, in a manner similar to The Ethics Application Repository (TEAR) in New Zealand. Since IRBs are occasionally privy to commercially important information such as the clinical trial protocols of a drug under development, redactions would be permitted, and there could be a delay between the IRB decision and its release to the general public. IRB transparency would have a host of downstream benefits—it would facilitate empirical study of the IRB system and allow IRBs at different institutions to learn from one another. More important, perhaps, is that researchers would be able to tell how the IRB at their institution compares with its peers and push for change.
IRBs are regulated by the Office for Human Research Protection (part of HHS) but are also supervised by FDA when they are involved in drug trials. In the past, regulatory reform in this area has been slow and incremental, so a significant change such as IRB transparency would likely require congressional intervention, whether by amending the underlying statute or by directing the agency to engage in new rulemaking with firm timelines, ideally with funding tied to timeliness.
Beyond the 2014 Report
In addition to the recommendations in the HHS report, three more ideas deserve consideration: (1) relaxing rules on clinical trial advertising; (2) fixing the tax disincentive that discourages low-income people from participating; and (3) modifying diversity mandates. In combination, these changes will speed trial recruitment and cut trial costs.
Advertising and Recruiting—Be Bold
When it comes to recruiting participants, clinical trials remain stuck in 1998, as FDA guidelines strictly regulate the formatting of trial advertisements. Broad guidelines on clinical trial advertisements have led researchers to keep them as boring as possible. Anything that unduly emphasizes payment is suspect: using bold text to highlight compensation, for example, is prohibited. (One paper highlighted an ad that told participants : “Transportation and parking fees will be paid for.” It might be okay to advertise this information, but not in bold text. That is, apparently, just too alluring to participants and thus possibly unethical.) Font size is also problematic: “[Ads] should not emphasize the payment or the amount to be paid, by such means as larger or bold type.”
A natural inclination when faced with national regulation that seems unnecessarily restrictive is to look for international comparisons. However, in contrast to tax policy or criminal law, international variation in clinical trial regulation is minimal. Many high-income countries are bound in some way by the Good Clinical Practice (GCP) guidelines developed by the International Council for Harmonisation (ICH). Regulators in the European Union, Japan, and the U.S. abide by ICH GCP standards, which are supposed to protect trial participants while minimizing the burden of additional compliance work on multinational pharmaceutical companies. These guidelines are not as detailed as FDA guidance and thus do not lay out explicitly which type of advertisements for clinical trials are acceptable. But they do consider advertising part of the subject recruitment process and therefore subject to IRB approval. Overall, it appears that many are as strict as, or stricter than, U.S. guidelines on clinical trial advertising.
FDA should maintain strict standards for truthfulness in clinical trial advertising but not regulate the aesthetic presentation of such information. FDA should strike out the following section of its operative guidance document: “The IRB should review the final copy of printed advertisements to evaluate the relative size of type used and other visual effects . . . [and] should not emphasize the payment or the amount to be paid, by such means as larger or bold type.” Subject recruitment is already a difficult and time-consuming process. The aggregate impact of Australian cigarette packaging–style advertisements for clinical ads represents a significant cost to medical research and to society in the long run.
Don’t Tax Positive Externalities
While trial expenses are tax-deductible for pharmaceutical companies, incentives for research participation—many bioethicists and IRBs are allergic to calling it “payment”—are taxed. Olympians who win prize money (up to $1 million) and volunteer firefighters receiving certain incentives benefit from tax exemption, under the theory that they generate positive externalities for all of us. Ensuring that participation in clinical trial research (an activity with large positive externalities) is properly incentivized is a no-brainer. Fixing this would be simple but would require Congress to make research incentives tax-deductible.
Participants in trials on rare diseases (“orphan diseases”) are already allowed to deduct $2,000 worth of income from those trials from Supplemental Security Income (SSI) and Medicaid annual limits. Extending this proviso to all participants in research trials would be straightforward and may have the additional benefit of improving socioeconomic diversity in trials by removing barriers that especially poor participants face.
Unfunded Diversity Mandates and “Race Medicine”
Since the late 1990s, FDA has pushed pharmaceutical companies to recruit more minority participants for clinical trials. The reasoning, formalized in 1998 FDA guidance, was that some populations would, for genetic or environmental reasons, respond differently to medications. The best example for this hypothesis was the drug BiDil, approved only for black patients after data from multiple trials showed a benefit in that population but not in white patients. Apart from BiDil, however, no other drugs have been approved with a race-specific indication. Recently, it has become more practical to identify specific genetic variants that influence drug metabolism—an area of study known as pharmacogenomics—which makes relying on self-identified race as a proxy for genotype less defensible.
In recent years, FDA—like many institutions that have quietly bowed to radical identity politics—has moved even further in this direction, issuing guidance that expanded diversity requirements both in 2020 and 2022. Though couched as “non-binding recommendations” to avoid statutory rulemaking requirements, FDA guidance is all but law. In this case, FDA recommends a number of approaches to increase minority representation in trials. Some are sensible, such as “make recruitment events accessible”—although it should be noted that this could also be accomplished through deregulating clinical trial advertising—but others are more controversial and are presented without any evidence for their effectiveness. Another recommendation from FDA’s 2020 guidance verges on a blatant giveaway to diversity training groups: “Provid[e] cultural competency and proficiency training for clinical investigators and research staff . . . decrease biased communication and behavioral practices . . . avoid the use of cultural generalizations and stereotypes in interactions with participants.” Also in 2022, Congress passed FDORA, which, as noted above, requires FDA to collect “diversity action plans” for clinical trials and imposes reporting requirements on FDA.
The call to involve more patient groups in study design imposes more “cooks in the kitchen,” which makes clinical trial design more of a vetocracy. The result will be to impose more costs, in both money and time, on already wasteful and slow trials—and done through FDA fiat.
Unfortunately, even before the most recent diversity guidance, public health authorities have imposed a dangerous diversity orthodoxy. In 2020, NIH director Francis Collins delayed participant accrual in Moderna’s Covid-19 vaccine trial by three weeks to increase minority participation, as Peter Loftus reports in The Messenger. In that case, Collins’s reasoning was especially shoddy—while group differences in susceptibility to infectious diseases (including Covid-19) have been documented (though the causes are debated), he presented no evidence for group differences in response to vaccines. Even if they had been documented, slowing vaccine approval for everyone to ensure a more precise estimate of efficacy in minority populations is unethical under practically any system of morality. Unfortunately, his behavior went largely uncriticized. If the 2022 FDA guidance on diversity goes unrevised, it seems very likely that many more drugs will be similarly delayed as a result.
Future sessions of Congress would ideally rescind the portions of the 2022 FDORA amendment that deal with race: the “diversity action plans” that it requires FDA to collect and FDA reporting requirements. In addition, Congress should instruct FDA to issue new guidance rescinding some of the recommendations that FDA has made in recent draft guidance on clinical trial diversity, and at least return to the 2020 guidance, which was less prescriptive. Long-term, more diversity in clinical trials should be pursued only if the cost-benefit trade-off is worth it, so Congress should instruct FDA to collect estimates from sponsors on the monetary costs of increased diversity in trials, as well as develop estimates of how beneficial more diversity could be.
As the global leader in pharmaceutical R&D and biomedical research, the U.S. should take the lead in fixing a costly and slow clinical trial industry. Congress should directly mandate IRB transparency by statute or instruct FDA to revise 45 C.F.R. 46 to require IRB transparency; extend the $2,000 SSI and Medicaid income exemption from participants in orphan disease trials to all trial participants; grant tax deductibility of research incentives to participants, instead of just companies; and rescind the portions of the 2022 FDORA that deal with race. FDA must issue more guidance on risk-based monitoring and invest in training for reviewers to instruct them in such guidance; develop a Scientific Achievement Award for reviewers who excel at carrying out such guidance; and issue updated guidance on clinical trials that instruct IRBs to desist from regulating the aesthetics (though not the content) of trial advertisements. Doing so would speed the deployment of new drugs and therapeutics without compromising efficacy or safety.