The case for clinical trial abundance

A supply-side reform agenda for one of our most urgent problems

This piece was prepared for the Clinical Trials Abundance Initiative, a project by the Institute for Progress (IFP) with support from Renaissance Philanthropy, and presented at a policy workshop in Washington, DC on October 7th, 2024.

Willy Chertman, M.D., is a Biotechnology Fellow at Institute for Progress and an Adjunct Fellow at the Manhattan Institute. He was formerly a Biosecurity Fellow in the office of Senator Todd Young (R-IN).

Ruxandra Tesloianu is a PhD student in Genomics at the Sanger Institute who writes about innovation and other topics. She has a degree in biochemistry from Oxford University.

---

In 2014, the first SGLT2 inhibitor, dapagliflozin, was approved for patients with Type II diabetes for the control of blood glucose levels. Since then, the list of indications for this class of molecules has expanded to include kidney disease (2021) and heart failure (2020) in non-diabetics, both important sources of morbidity and mortality. Despite being an already developed drug with safety data, these repurposings took seven and six years, respectively.

As in the case of dapaglifozin, so in many others. Even after a drug is discovered, its development faces many challenges — particularly the practical hurdles of conducting clinical trials. These studies are not only expensive, constituting an appreciable fraction of drug development costs,¹ but also difficult to execute. They often stretch for years, struggling to find and retain enough participants. They frequently fail to reach completion. There is also an invisible cost — promising studies never launch because researchers or companies are deterred by these practical barriers. That means many medical questions remain unanswered not because of scientific limitations, but due to the logistics and costs of running trials. Because randomized controlled trials (RCTs) play a key role in the drug development pipeline and help generate high-quality evidence in other contexts, making RCTs cheaper, faster, and easier to carry out would have large spillover benefits for biomedical innovation and healthcare resource allocation.²

Some clinical trials are unavoidably expensive and difficult: for example, trials with new biologics have to be manufactured in small batches, administered with physician supervision to watch for unknown adverse events, and involve collection and analysis of many biological samples per patient. But many trials need not be expensive or difficult. When they involve treatments that are routinely administered, they should be integrated into routine clinical care as much as possible.

The need to make drug development more efficient has become increasingly pressing. US healthcare spending growth is predicted to reach nearly 20% of GDP by 2032 and exceed GDP growth itself for structural reasons, like an aging society. Meanwhile, given high medication prices and little political appetite to cut Medicare spending, there is mounting pressure to reduce drug development costs. In the face of these cross-pressures, the best policy approach is a supply-side innovation agenda, aimed at lowering the costs of trials.

We live in an era of converging biotech and AI innovations. Our abilities to read (via sequencing), engineer (using tools like CRISPR/Cas9) and analyze biology (with AI) have all reached important milestones. If these innovations are able to accelerate preclinical R&D, RCTs will become an even more critical bottleneck in bringing new treatments to patients. Without speeding up and reducing the costs of RCTs, breakthrough scientific discoveries risk becoming backlogged in laboratories and testing phases while patients wait longer than necessary for treatment. This potential bottleneck has not gone unnoticed by leaders in AI, including the CEO of Anthropic, Dario Amodei.³

We have several reasons to be optimistic about our ability to cut clinical trial costs and timelines. One proof-of-concept is the RECOVERY trial, which cost about 1/80th of a traditional RCT and likely saved hundreds of thousands of lives by demonstrating the efficacy of steroids for COVID-19. RECOVERY showed the enormous cost and time savings possible if trials are kept tightly focused on important questions and trial enrollment/organization is made as easy as possible. We can also look at historic examples of large trials (e.g., the polio vaccine field trials) that ran on time and answered important questions, by avoiding cumbersome and unnecessary administrative delays. More abstractly, the vast majority of patients are never enrolled in trials, implying a large pool of patients that could theoretically be enrolled in trials.⁴

Many stakeholders agree on the urgency of the problem, often framed as clinical trial modernization. These include high-level regulators like FDA Commissioner Robert Califf and former Deputy Commissioner Janet Woodcock, academic groups like Duke Margolis, pharmaceutical executives, and even longtime industry critics like Vinay Prasad.⁵ Reducing the cost and difficulty of generating high-quality medical evidence is a rare area where most experts agree on the goals.

Despite this shared concern, substantive policy change is lacking. Some attempts in this direction have been made in recent years. For example, in an attempt to prod industry into using a cheaper form of data monitoring, the FDA has signaled acceptance of risk-based monitoring since 2013, and also recently issued guidance making Form 1572 (which documents all trial investigators) slightly less burdensome. Congress has also directed FDA to speed uptake of digital health technologies like wearables in trials, though progress there has been limited. Yet these attempts at change have been scattered and not part of a comprehensive push. As a result, progress has been minimal relative to the scale of the challenge.

Success requires a comprehensive supply-side reform agenda — similar to what advocates have outlined for housing and energy infrastructure. Some solutions parallel these fields, like cutting administrative red tape and reducing veto points. Others are unique to clinical trials, such as modernizing outdated systems like the U.S. National Death Index. Articulating and carrying out such an agenda will be a big undertaking, but we think the ideas below are a good start.

On that basis, we organized a workshop bringing together clinical trialists, academics, industry professionals, regulators, and policymakers to present and vet eight concrete policy proposals. These are summarized and linked below, with another forthcoming.

Beyond these specifics, many of our memos follow the guiding question: "What would a permanent, US-scale RECOVERY trial look like and accomplish?" With dramatically cheaper trials, we would more quickly sift through poorly evidenced clinical practice. New therapies would cost less to test in humans, and we would have answers and innovation sooner. Beyond speeding up the approval of new drugs, cheaper and faster trials would also allow more kinds of questions to be asked. When a large trial costs $100 million to carry out, some questions simply don’t get asked. Outside of medicine, environmental interventions like indoor far-UV need pilot studies to advance, but are bottlenecked by high costs. Other memos target specific high-leverage bottlenecks, like human challenge trial regulation, or the public health value of greater FDA transparency.

This series of memos will introduce each policy idea in detail:

1. Evidence for the People: Facilitating Community-Based Trials, Martin Landray — To make clinical trials more accessible, federal regulators should embrace international best practices for trial conduct. They should replace lengthy consent forms with concise documents patients can actually understand.

2. Democratized Clinical Trials: Reducing Regulatory Burden and Ensuring Coverage, Willy Chertman — Clinical trial access is held back by excessive paperwork and patchy Medicaid coverage. Streamlining documentation requirements and fixing coverage gaps would help bring research beyond academic medical centers.

3. Replacing Source Data Verification and Review with Risk-Based Monitoring in Clinical Trials, Michael Lingzhi Li — Traditional clinical trial monitoring practices waste resources on unnecessary source-data verification. The FDA and NIH should incentivize (and consider mandating) adoption of more efficient risk-based approaches that focus attention where it matters most.

4. Improving ClinicalTrials.gov, Mark Webb — ClinicalTrials.gov's clunky search interface blocks patients and physicians from finding potentially life-saving trials. Simple updates to search functionality could accelerate trial recruitment and encourage better sponsor engagement with the website.

5. Unblocking Human Challenge Trials for Faster Progress, Alastair Fraser-Urquhart, Josh Morrison — Human challenge trials can provide early “in-human” data, but to be more broadly utilized they require revisions of burdensome FDA manufacturing requirements and clear regulatory guidance.

6. Fair Pay: Legalizing Comprehensive Compensation in Clinical Trials, Jake Eberts and Allison Foss — To speed up medical research, regulators should allow competitive trial compensation, and Congress should ensure that participants do not lose eligibility for welfare benefits.

7. Building More Efficient Clinical Trial Infrastructure for Rare Diseases, Frank David — The NIH and FDA should build shared infrastructure for rare disease platform trials, allowing multiple companies to test treatments more efficiently by sharing control groups and quickly swapping out ineffective therapies.

8. Improving FDA Transparency for Public Health, Stuart Buck — Congress should require the FDA to publish redacted versions of complete response letters, allowing public health researchers and industry to learn more from each new drug application. To build capacity for redaction at scale, the FDA should procure LLMs for internal use and develop benchmarks for AI-assisted redaction.

9. Improving Regulator Capacity Through Artificial Intelligence, Enlli Lewis — Augmenting FDA reviewers with AI tools to improve productivity could speed drug approvals without lowering standards. To do so, the FDA should bring AI expertise in-house, develop internal FDA reviewer benchmarks, and engage FDA reviewers in the adoption process.

Acknowledgement: Special thanks to all the authors of the policy memos, the attendees at the October 7th workshop, Trevan Locke, Nitya Sridhar, Milos Milijkovic, Randall Lutter, Michael Sklar and Clint Hermes for helpful comments throughout, and to Renaissance Philanthropy for support.

1. The best methodology for calculating drug development costs is a matter of controversy, but clinical trials may account for about 60% of those costs.

2. Partial exceptions include the use of the Animal Rule in public health emergencies and the use of historical controls in certain rare diseases. But the vast majority of drugs fall outside these categories.

3. In a recent essay detailing how AI could improve the world, he writes that: “actual clinical trials involve a lot of bureaucracy and regulatory requirements that add unnecessary additional time and delay progress.”

4. More specifically: one reason for high trial costs might be that companies are competing for a limited number of patients, but this explanation is implausible in most scenarios.

5. In their book Ending Medical Reversal, Prasad and Cifu state “would medicine be better if a larger proportion of patients were enrolled in clinical trials… This change would have to be accompanied by the reaction of a robust clinical-trials enterprise able to design and conduct them, at low cost, with fewer barriers across diverse settings. For most clinical questions, those for which we do not really know the answer, there should be no more barriers to additional trials.”