Resource Allocation in COVID-19 Research: Which Trials? Which Patients?

During the COVID-19 pandemic, ethicists and others have worked to allocate scarce resources in the clinical setting, including intensive care beds, ventilators, and personal protective equipment, in equitable and consistent ways. The sudden increase in COVID-19 research has made clear that the allocation of research resources is also an area that needs attention. Previous allocation work in the clinical setting is particularly helpful in informing allocation in the research setting.

Several scarce resources are required for COVID-19 research. These include participants, facilities, funding to complete the research, and interventions being tested. In what follows, we will focus on participants.

First, COVID-19 patients are unevenly distributed in space and time. Due to a variety of factors, some parts of the world experience surges of cases while others have few. Areas such as Wuhan were initially overwhelmed with patients (and therefore potential research participants) when trials began, yet have had to stop trials as the first wave of cases dissipated. Narrow inclusion criteria in some trials, such as focusing on the most severely affected patients, also limits available research subjects for COVID-19 research on hospitalized patients. Lastly, there may be a mismatch between institutions with a high number of cases and those with the resources to conduct research. Many of the institutions with the most COVID-19 cases are community hospitals with less research infrastructure than larger academic institutions.

Others have previously considered the implications of using the traditional frameworks (egalitarian, utilitarian, prioritarian) as a basis for allocation in a research rather than clinical setting (Frye 2018; Jecker et al. 2018). However, this effort focused on the allocation of a scarce intervention in a clinical setting, rather than the additional complications of scarce funding and participants, as is the case in current COVID-19 research. But there are other similarities between the cases of clinical and research allocation related to COVID-19. Like in the clinical resources, COVID-19 trials risk arbitrary allocation practices. Without additional oversight and consideration, participants might be allocated to whichever investigator has the best connections to the hospital or to whichever research team approaches them first. More studies may be approved than there are available participants, making accrual all but impossible for some trials. Participants in a region may be divided among similarly underpowered trials at many institutions.

One important way in which the participant allocation problem is different from other resource allocation problems is that people are not merely research resources. While it is true that participants are a crucial component of research, they are not resources akin to funding, lab space, or ventilators. Any conversation about how best to manage the number of possible participants given a research agenda should keep this squarely in mind.

An additional difference between participant allocation and other resource allocation is in the overall response to scarcity. In the case of clinical resource allocation, one possibility is to increase the supply of the scarce resource. In the case of clinical resource allocation for COVID-19, this might mean increasing the production of ventilators or personal protective equipment. However, in the case of research participants, purposefully “producing” more creates unique ethical concerns. While researchers are unlikely to wish for additional ill patients, challenge trials (purposeful infection of volunteers to test vaccine efficacy) come closest to replicating the move of alleviating scarcity by increasing participant supply in the research context. Others have discussed the ethical complexity of challenge trials and proposed frameworks for their ethical implementation (Shah et al. 2020).

On the other hand, researchers might be justified in increasing available participants by broadening the study question to a wider population. Some authors argue for testing drugs on less severely affected patients since people with viral infections in general seem to respond better to prophylactic treatment or treatment at early stages of infection (Malley and Lipsitch 2020). This course of action would significantly widen the inclusion criteria and mitigate the shortage of research subjects. Such research interventions, which aim to prevent severe illness and reduce the risk of chronic complications are meaningful research aims (Milleson 2018). We acknowledge that this decreases efforts toward research on late-stage, life-saving interventions. We do not advocate for broadening inclusion criteria just for the sake of being able to recruit sufficient numbers of subjects, given the importance of resisting pandemic research exceptionalism (London and Kimmelman 2020), but this might be one additional consideration when deciding on inclusion criteria in the first place.

Another important difference between clinical resource allocation and research resource allocation is the source and nature of the obligation to allocate resources fairly. In the clinical context, the obligation stems from the clinicians’ duties to their patients and the ethical dilemmas are raised by the need to balance these duties with broader social responsibilities. In the research context, while researchers have obligations to select research participants/patients so as to distribute potential risks and benefits fairly, it is not on the presumption that the research intervention is a proven good (Emanuel et al. 2000). We are considering “resources” from the point of view of the researcher, in terms of what researchers need to conduct their studies. While there is an argument access to research itself must be considered a resource and fairly adjudicated, especially in promising late-stage trials, it is beyond the scope of this paper.

Similar to the clinical case of COVID-19 resource allocation, the stakes for allocation in the case of research allocation are very high. At the time of writing this commentary, more than 100,000 people in the United States have died from COVID-19 related causes, making knowledge about possible interventions very valuable.

While the cases of allocation in the clinical setting and the research setting are different, we believe there are some lessons from one setting that can be usefully applied to the other. First, in the clinical setting, the creation of committees to determine access to scarce resources was widely shared in policies across the country (Antommaria et al. 2020). Such committees remove the (often morally distressing) burden of allocation from individual practitioners and allow for consistency in allocation across an institution.

In the case of clinical allocation of resources related to COVID-19, there is a substantial concern that structures and policies for the allocation of clinical COVID-19 resources unfairly disadvantage particular marginalized groups (disabled people, people of color) both because of on-going structural inequalities and particular allocation mechanisms, such as the use of Quality Adjusted Life Years in calculations of benefit (Savin and Guidry-Grimes 2020). Some of these issues persist in the case of research allocation, but in a different manner. Members of marginalized groups might reasonably be less willing to participate in COVID-19 research because of a lack of trust in medical institutions and/or barriers to access to trials (Smirnoff et al. 2018). Those seeking to provide guidance in resource allocation in the research setting might learn from these failures in clinical allocation policy by proactively addressing them.

One example from clinical resource allocation policy that can be followed by research institutions is to constitute a research triage committee, with ethics involvement, to oversee and adjudicate between proposed trials at institutions. These committees can also design processes for implementation of trials conducted simultaneously. For example, if an individual institution hosts more than one trial for the same intervention, the committee can facilitate a process by which research teams trade off recruitment days: Trial A initially recruits patients which arrive on Monday, Trial B initially recruits patients who arrive on Tuesday, etc.

Such a committee can also facilitate communication and collaboration between trialists in order to avoid the creation of multiple underpowered inconclusive competing studies instead of multicenter trials, and to assess the effects of any changes to inclusion and exclusion criteria on recruitment to all trials.

When an institution has multiple clinical trials on the same intervention, patients must be informed about all trials for which they are eligible. Informing patients of all of their options remains crucial even though it may result in uneven recruitment across trials. For example, when all of their options are made clear, patients may choose to avoid participation in trials which feature a placebo arm, even if the intervention being tested is promising and the trial methodologically sound, resulting in uneven recruitment across all trials at an institution.

The COVID-19 pandemic has clarified that research participants can be a scarce resource in need of allocation strategies to promote fair and meaningful data gathering. Ethicists and policy makers have an opportunity to learn from issues in clinical resource allocation when planning for research resource allocation.