Preventing the Misuse of Breakthrough Health Research

Introduction

Rapid advances in biotechnology hold promise for treating diseases once deemed incurable. However, breakthrough technologies can be a double-edged sword, with both benevolent and malicious applications. As biotechnology becomes more accessible at an accelerating rate, it often outpaces our ability to establish comprehensive guidelines that safeguard us from the associated risks. For ARPA-H to champion breakthrough health research, it must not only pioneer disruptive solutions, but also ensure accessibility and anticipate the potential secondary impacts on society, especially in terms of biosecurity risks. 

We appreciate the opportunity to provide comments on ARPA-H’s Ethical, Legal, and Social Implications (ELSI) Initiative and particularly commend agency leadership for introducing the two new ARPA-(H)eilmeier Questions (HQs), which address vital components of a comprehensive ELSI framework:

• HQ9: To ensure equitable access to all people, how will cost, accessibility, and user experience be addressed?
• HQ10: How might this program be misperceived or misused (and how can we prevent that from happening)?

Focusing our response on these two novel questions, we provide the following recommendations: 

Rethink health innovation and implementation to ensure equitable access

ARPA-H’s ELSI initiative aims to develop frameworks that enhance healthcare equity, echoing HQ9’s focus on accessibility and affordability. Given that many existing HHS agencies focus on health R&D, we believe the value-add of ARPA-H will be largest if it targets neglected research areas and transition pathways. While agencies like BARDA excel in developing new FDA-approved therapeutics, ARPA-H’s mission extends further. It has a clear mandate to develop platform technologies, computational tools, or predictive models that may not lead to an FDA-approved therapeutic but address unmet needs in health, placing special emphasis on the customer experience.

Furthermore, certain health technologies, like vaccines, often see a shortfall in substantial private-sector investment due to the sometimes unpredictable demand for such products. Some innovations fostered by ARPA-H will require market-shaping mechanisms or public-private partnerships to progress from research and development to widespread adoption.

We applaud ARPA-H for launching PATIO (the Project Accelerator Transition Innovation Office), dedicating an entire office to technological transition and commercialization. The Customer Experience hub in Dallas, TX, further ensures that ARPA-H creates technologies that are both accessible and affordable for all. Building on these initiatives, ARPA-H could take the following measures:

Recommendation 1: Focus on neglected research areas and transition pathways 

• ARPA-H should leverage its mandate to create new capabilities, like platform technologies or computational tools. While these may not result in FDA-approved treatments, they fill a gap within the HHS R&D ecosystem and can synergize with efforts from agencies like BARDA or the NIH, as well as private industry. 
• Due to an uncertain return on investment, the private sector may hesitate to invest in technologies that are intended as public goods. ARPA-H should play a pivotal role in advancing these technologies by addressing R&D hurdles and spearheading proof-of-concept prototypes — for instance, developing miniaturized sensors for measuring pathogens or pollutants in the air, which could empower individuals to monitor the air they breathe in real time.
• We support ARPA-H’s emphasis on the customer experience, and its close collaboration with the FDA and the Centers for Medicare & Medicaid Services. Given these relationships, we are hopeful that ARPA-H can help change how the federal government interfaces with healthcare recipients. Innovative strategies, like introducing new reimbursement frameworks or enhancing collaborations with entities like the Department of Veterans Affairs (VA), can ensure that health access remains broad-based, and isn’t confined by factors like geographical location or socio-economic status. 

Recommendation 2: Shape market incentives through innovative “pull” mechanisms

• The success of Operation Warp Speed stands as a shining example of public-private collaboration. For innovations that face market failures, ARPA-H should similarly seek to explore innovative pull mechanisms like Advanced Market Commitments (AMCs), prizes, or milestone-based payments. Investigating different pull mechanisms ensures participation across the entire spectrum of research and transition partners: While large companies might lean on AMCs, milestone-based payments could attract smaller companies, and prizes could incentivize academic research or software development.
• More broadly, ARPA-H should explore the diverse landscape of innovation financing mechanisms. IFP hosts the Metascience Working Group, which focuses on projects that align with ARPA-H’s objectives. We currently partner with the National Science Foundation to explore different ways to fund high-risk, high-reward proposals, and are happy to provide further recommendations informed by our research to ARPA-H.

Anticipate and mitigate downside risk

ARPA-H is willing to take significant technical risks, acknowledging that cutting-edge technologies can fail due to unexpected obstacles. However, the risk landscape extends beyond technical risks, and includes misperceptions, accidents, or misuse (HQ10). We agree with ARPA-H that those who spearhead novel technologies and capabilities will be the first to encounter these risks – and possess the opportunity to address them.

The potential for misuse is of particular concern due to the dual-use nature of many of the technologies ARPA-H will develop. Recent advances in biotechnology and life science research may enable ARPA-H to achieve its lofty goals, but they carry inherent misuse risks, especially in the realm of biosecurity.

Both states and terrorist groups have weaponized biology. In 1979, the clandestine Soviet biological weapons program caused an accidental anthrax leak, killing at least 66 people. And in 1992, the Japanese cult Aum Shinrikyo sent 40 people to collect Ebola virus during an outbreak in the Democratic Republic of the Congo, aiming to use the virus for terrorist attacks. 

Aum Shinrikyo’s attempt to collect Ebola thankfully did not succeed. Yet today, an increasing number of individuals have the skills and tools to engineer pathogens like Ebola, corona– or influenza viruses from scratch, exposing us to the risk of deliberate misuse (biosecurity risks) and accidents (biosafety risks). Our protective measures sometimes lag behind: online platforms provide step-by-step protocols and the synthetic DNA required for creating pathogens in the lab. Further, not all synthetic DNA orders currently undergo screening for potential hazards, such as sequences matching harmful pathogens. 

In addition, well-meaning scientists may actively seek to enhance dangerous viruses in the lab. For instance, in an effort to better predict potential H5N1 influenza pandemics, scientists previously engineered the extremely lethal but poorly transmissible H5N1 avian influenza virus to improve its mammal-to-mammal transmission. However, these scientists may overlook the potential dangers of publishing their findings, which could inspire malicious actors to engineer enhanced pandemic-capable pathogens and misuse them. While there are ongoing efforts by the U.S. government to develop a revised oversight policy for research with certain pathogens and toxins in order to mitigate risks from dual-use research, much work toward a comprehensive framework remains.

Next to intentional bioweapon attacks, accidents also pose risks, and laboratory-acquired infections are surprisingly common. Experts have estimated that there is at least a 0.2% chance of a laboratory-acquired infection per BSL-3 laboratory year. This number might seem low for an individual laboratory, but in light of more than 1,300 BSL-3 labs registered with the Select Agents Program in the U.S. alone, the cumulative risk is notable. Furthermore, even pathogens with seemingly low case fatality rates can wreak global havoc. Despite SARS-CoV-2 having a reported case fatality rate of about 1% in the U.S., it led to over one million deaths. 

Taken together, this implies that even though misuse events or accidents tied to ARPA-H innovations might be rare, a single incident, like the release of a self-spreading pathogen, could have catastrophic global consequences. It’s essential to prevent these extreme risks upfront.   

Encouragingly, the U.S. government has already made commendable strides in establishing risk mitigation practices. For instance, the National Science and Technology Council released guidance on research security programs at federal funding agencies. At ARPA-H, such a program could encompass elements like biosecurity, cybersecurity, foreign travel security, and research security training for relevant personnel. More specifically, ARPA-H could take the following actions:

Recommendation 3: Establish a Research and Technology Protection Office

• Similar to IARPA, ARPA-H could establish a Research and Technology Protection (RTP) Office to proactively identify and mitigate security risks. We recommend this RTP Office to serve as ARPA-H’s primary liaison for research security, overseeing elements like biosecurity, cybersecurity, foreign travel security, or export controls. It could offer relevant training to agency staff and support Program Managers (PMs) throughout a project’s lifespan and beyond. To promote transparency, the creation of an accessible website could be beneficial in providing relevant information to the general public. 

Recommendation 4: Prioritize biosecurity measures and foster collaboration with security agencies

• ARPA-H should place special emphasis on biosecurity by equipping PMs with the necessary training and support from subject-matter experts throughout the course of a project. For example, ARPA-H could support the formation of advisory ELSI panels tailored to different programs, drawing inspiration from DARPA’s Safe Genes program. 
• The RTP Office could foster collaboration with national security agencies. As a civilian agency, ARPA-H may be more exposed to espionage or cyberattacks than its defense-oriented counterpart, DARPA. Anticipating and mitigating biosecurity risks will be particularly important as advances in biotechnology increasingly intersect with national security concerns. 

Recommendation 5: Anticipate unintended consequences by investing in ELSI forecasting capability

• When choosing and developing new programs, PMs should take security risks into account. For some programs, the associated risks will be obvious, and PMs’ excellent subject-matter expertise makes them well-equipped to identify these. Yet, compared to other research agencies, ARPA-H’s use-inspired approach to innovation is more technology-agnostic. As a result, some programs might yield unexpected technological breakthroughs, which makes it harder to identify all possible risks upfront. Furthermore, invention is often driven by the combination of different technologies. Anticipating every conceivable application of a novel technology, along with its associated risks, is vastly more complex than for a technology viewed in isolation.
• ARPA-H should continuously reassess security risks throughout a project’s duration, particularly considering how its research outputs might interact with existing technologies, and the ensuing potential for misuse or unintended consequences. A specific strategy that has seen considerable progress in recent years is technology forecasting. By employing probabilistic methods, technological developments and their ensuing ELSI implications can be better predicted. ARPA-H should seek to invest in ELSI forecasting capabilities and embed them within its risk assessment framework, for instance, by creating an internal forecasting team.

Recommendation 6: Develop an expanded guidance framework

• To pursue technological innovation in alignment with ARPA-H’s ELSI vision, PMs could proactively steer the relative timing of technologies, a strategy sometimes referred to as “differential technological development.” For instance, PMs could aim to develop risk-decreasing technologies before risk-increasing ones, like experts are currently developing post-quantum cryptography before building actual quantum computers due to the risk they might pose to information security. Another approach could be “safety by design,” for example, modifying engineered microorganisms in a way that restricts their survival to specific environments.
• To aid PMs in their decision-making process, ARPA-H should extend its hidden questions behind the Heilmeier Questions to include concrete risk mitigation strategies. A similar approach was pioneered by former IARPA director Jason Matheny, who created an internal questionnaire for PMs making new research proposals. Building on these questions, ARPA-H could develop an expanded guidance framework to complement its “hidden questions,” such as:
  • How can we prevent, mitigate, and monitor for accidents?
  • How can we evaluate misuse risks throughout a project’s life cycle in order to mitigate them preemptively?
  • How can we assess the secondary impacts of a program’s research outputs? Relatedly, how can we integrate technology forecasting into our risk assessment process?  
  • Can we use one or more of the following strategies to prevent misuse and unintended consequences?
    • Can we substitute this technology with a lower-risk alternative? 
    • Can we advance a risk-decreasing technology before a risk-increasing one? 
    • Can we integrate safety mechanisms?

We applaud ARPA-H’s dedication to developing a holistic ELSI framework that considers the secondary impact of its innovations.