Hide table of contents

This piece makes the argument that eligible effective altruists should consider participating in Oxford’s COVID reinfection challenge study

For some relevant background information, Josh is the President and Julia is the COO at 1Day Sooner, a nonprofit that advocates for people who want to participate in high-risk, high-reward medical studies, which was born in the COVID-19 pandemic to represent potential COVID-19 challenge volunteers. Both of us would sign up for a COVID-19 challenge study if we were to be eligible. (We live in the US and so far these studies have only allowed participants already in the UK to register.)

The Oxford Covid-19 Reinfection Study has a high expected value for contributions to infectious disease research, vaccination policy, and as a precedent for conducting challenge studies during future pandemics. For those eligible, (healthy UK residents aged 18-30 who had COVID-19,) we think participation could be considered an effective altruist decision. 

That said, this is less trodden territory than effective altruist decisions like earning to give or career selection. One of our key goals with this post is to solicit input from the EA community both on the case we offer below and also the broader question of how EAs should think about volunteering as research participants more broadly. 


Table of contents

  1. Intro/summary
  2. Benefits
    1. Direct Study Benefits
    2. Indirect study benefits
    3. Direct benefits of an Individual Participating
    4. Indirect benefits of an Individual Participating
  3. Costs
    1. Risk of Complication and Risk of Fatality
    2. Risk of Long-Term Effects
    3. Discomfort and Inconvenience
  4. Final Thoughts
  5. Conclusion

1. Intro

In April 2021, The University of Oxford announced it would conduct a human challenge study where volunteers who have previously had COVID-19 are re-exposed to the SARS-CoV-2 virus in a controlled manner (University of Oxford, 2021). To participate, you must be aged 18-30, be in excellent health, have previously tested positive for COVID-19 infection that did not cause you to be admitted to hospital for treatment, be registered with a GP practice in the UK, and be willing to travel to Oxford. (Both vaccinated and unvaccinated people are eligible to participate). The study design requires that you reside in a quarantine unit at the University of Oxford for at least 17 days, and then attend nine follow-up visits over the following 12 months. Participants will receive compensation of at least £4,995. (Oxford Vaccine Group, 2021).

Oxford’s study design involves a safety and dose finding group of 8 participants, and a later safety and dose confirmation group of up to 20 or possibly 40 participants. As of the time of writing, the study is under way and fourteen participants have completed their time in quarantine. The trial team is experiencing difficulties with recruitment and requested our assistance. However we have been kept at arm’s distance from the Oxford (and Imperial) studies and have not reviewed their protocol, so we have fewer details about the study than we’d like.


2. Benefits

a. Direct Study Benefits

Learning about diseases is valuable, and studies can generate a variety of benefits which are hard to predict beforehand. There are many unanswered questions about infection-induced immunity against SARS-CoV-2 that a reinfection study can help us investigate. For example, we don’t know how long protection after COVID-19 infection lasts, or why some people can get infected a second time. Challenge studies with previously infected participants can help answer these questions. 

Scientific Rationale and Capabilities for SARS-CoV-2 Human Challenge Studies (Rapeport et al., 2021)


Potential Impacts

Practically Possible Only with Human Challenge Studies

Substantially Accelerated or Scientifically Improved by Human Challenge Studies

Innate immune, T-cell, and humoral responses that act during early SARS-CoV-2 infection  Immune responses start before any symptoms develop and are critical for protection but cannot be studied in natural infection (since patients are recognized only after symptom onset).



Durability of natural and vaccine-induced immunityIn previously infected or vaccinated people, challenging at predefined time points clearly shows how long protection lasts, clarifying the need and timing for boosters or other interventions.



Susceptibility to reinfectionBy testing reinfection with SARS-CoV-2 in seropositive persons, researchers can immediately establish whether previous infection is protective.*



Host defense factors that mediate protection and viral clearanceMarkers that differ between infected and uninfected participants can be used to calibrate and develop diagnostics, predictors, and new vaccines. Definition of correlates of protection in field studies is limited by confounding by uncontrolled viral and other factors.*



*It may be possible to partially infer or determine answers to these questions using other study designs, but only if community transmission is ongoing.

While studying immune responses in cell culture, animal models and passive observation of natural infection are all helpful, a human challenge model uniquely allows researchers to study the immune response in humans up close from the moment of exposure. It is hard to predict ahead of time what concrete social benefits will accrue from improving our basic understanding of natural immunity to SARS CoV-2, but we offer some illustrative examples below.

Potential Social Benefits from Improved Scientific Understanding: In places with a constrained supply of COVID-19 vaccines, the level of protection received from a previous infection can inform important policy decisions. like whether people who have previously been infected should receive one vs two vaccine doses, and whether they should receive additional shots. Follow-on studies could clarify the more effective timing post-infection for administering boosters and help determine whether fractional doses make sense for seropositive individuals. The more effectively limited vaccines are distributed, the more deaths and hospitalisations can be prevented. 

Human challenge studies have contributed vital scientific knowledge to advance vaccine development in recent decades. Challenge data can help reduce uncertainty in the early stages of vaccine development by allowing for the optimal allocation of resources toward the most promising vaccine candidates, reducing the costs of clinical development, and encouraging investment in larger-scale trials (Rohrig et al., 2021). The Oxford reinfection study may eventually help with the development of better COVID-19 vaccines and treatments. A reinfection model could become a safer and more broadly used COVID challenge model with which to test future antivirals and next-generation vaccines. Other lessons learned from the study may be more difficult to predict - the volume of data collected in challenge studies means that it’s quite possible we’ll learn something unexpected. 


b. Indirect Study Benefits

Completing a successful COVID-19 reinfection study could increase the utility of conducting additional challenge studies for other coronaviruses, including as part of efforts to develop a universal coronavirus vaccine. Developing a universal coronavirus vaccine would mitigate the risk of viral mutations, both for COVID-19 and in potential future coronavirus pandemics. They could also eliminate a significant percentage of cases of the common cold, which while generally not lethal, have significant social and economic costs. 

Additionally, a successful reinfection study will make it easier to plan the use of challenge studies for future pandemics ahead of time. Success with study recruitment would help 1Day Sooner’s reputation and our ability to promote strong research in the future, which may benefit effective altruist goals (such as developing effective vaccines against hepatitis c, tuberculosis, malaria, and group A strep) more broadly.  


c. Direct Benefits of an Individual Participating

We propose two standards by which participating in this study could be considered an effective altruist act. 

  1. Shapley Value: Given that the supply of study participants is scarce, participants could in some sense each be credited with a percentage of the eventual value of the study. There are multiple plausible strategies for estimating that percentage, such as giving each participant equal weight, or estimating the odds that the individual’s participation is the tipping point needed for results to be reported. To give a rough sketch: if each participant is given equal weight, then each individual participating in a 35 person study contributes 2.8% of the value. By this logic, if social gains yielded by the study as a whole were greater than one 2 millionth of past disease burden to date, given that total COVID disease burden to date has been very roughly 70 million disability-adjusted life years, participating in a reinfection study would be expected to save a year of someone’s life.
  2. Qualitative Standard: If a project is determined to meet EA standards by making a comparatively high contribution to an important cause area, and an individual can play a meaningful role in the project, then it could be considered an act of effective altruism to play that role. The Oxford reinfection study has the potential to make a significant contribution to public health and pandemic preparedness, and an individual can play a meaningful role by enrolling. Accordingly, participating in the study could be an EA act.


d. Indirect Benefits of An Individual Participating

It’s possible that one person’s participation will inspire others to participate in high-reward medical studies. If a participant tells their eligible friends about their participation and the reasons behind it, it could cause them to consider participating as well. A handful of people volunteering with 1Day Sooner mentioned friendly word-of-mouth as their introduction to challenge studies. Participants in the Imperial challenge study have found opportunities to share their motivations and experiences with the public in major news outlets (Fraser-Urquhart, 2021, Strasburg, 2021), and Oxford participants who chose to do the same could reach even more potential participants. (Volunteers should feel free to reach out to 1DaySooner for assistance with media engagement, should they wish to.)


3. Costs

a. Risk of Symptomatic Infection, Severe COVID-19, or Fatality

Symptomatic Infection

The goal of the study is to establish the lowest dose of virus required to cause re-infection (a positive throat and nose swab) but with the aim of causing minimal or no symptoms. The study design begins with a very low dose of the virus, increasing the dose until it reliably causes re-infection as identified by viral shedding from the nose or throat. However, it is still possible that participants will experience systems. Common symptoms of COVID-19 include fever, headache, fatigue, dry cough, loss or change in taste and/or smell, sore throat, diarrhoea, and body aches (McShane, 2021).

Severe COVID-19

Severe COVID-19 is uncommon in young healthy adults. The chance of requiring hospital treatment in the 18-30 age group if infected with COVID-19 is estimated to be less than 0.4% (1 in 250 people infected with COVID-19). [Oxford] 1Day Sooner researchers project the hospitalization risk for a seronegative COVID-19 challenge study to be as low as 22.0 per 100,000 [source]. Participants in Group 1 will be given Ronapreve, a cocktail of monoclonal antibodies designed to treat COVID-19 disease if they demonstrate any symptoms related to COVID-19 disease beyond mild symptoms, and will otherwise receive intensive medical care (McShane, 2021).


1Day Sooner researchers project the average odds of an individual death in a seronegative COVID-19 challenge study with participants aged 20-29 to be 8.2 in 100,000 if no effective treatment is used [4]. Given the existence of newer and effective treatments, the risk is plausibly much lower At the same dosing level, risks are expected to be significantly lower for people who have already recovered from COVID-19. For comparison, the chance of death from live kidney donation is about 13/100,000 (Segev et al., 2010).

b. Risk of Long-Term Effects

There is little reliable data available about the risk of long COVID symptoms following reinfection with SARS-CoV-2. This writeup is a good starting point for thinking about the risks of long COVID. The SIREN study looking at COVID-19 re-infection in healthcare workers suggests re-infection may lead to development of COVID-19 disease with fewer symptoms compared to the initial episode. This may indicate that there is a reduced risk for prolonged symptoms following a secondary infection (McShane, 2021).

c. Discomfort, Inconvenience

During the screening process for the study, potential participants undergo intensive medical tests and are asked to share in detail their medical and psychological history. If a participant is selected to participate, they must self-isolate for several days in advance, then participate in a medical quarantine at the University of Oxford for Several weeks. Throughout the day, participants undertake frequent testing including blood draws, nasal swabs, and cognitive tests. These tests may require being woken up early or being required to stay up later than desired. They are not allowed visitors during quarantine and have no in-person interactions other than with the medical staff. It’s not all bad, however: study participants are able to bring with them whatever items they need to occupy themselves (one brought an exercise bike) and volunteers report having access to high-quality coffee and snacks. Additionally, 1Day Sooner offers virtual social opportunities to any  interested Oxford participants.


4. Final Thoughts: The EA Community and Trial Participation

From the beginning, there was strong international interest in COVID-19 challenge study participation - in 2020, over 35,000 people around the world signed up with 1Day Sooner as potentially interested in participating in a COVID challenge study. However, challenge studies are an unfamiliar topic outside of the scientific research community, and the conversation on whether the United States should conduct COVID-19 challenge studies was dominated by bioethicists who viewed participation as an unacceptable ask from researchers rather than an act of service volunteers independently wanted to perform. 

If the EA community were to promote participating in challenge studies as a meaningful act of altruism, and more EAs make it publicly known that they were on standby to participate when eligible, these studies could be filled more quickly and more studies could be conducted. Additionally, EA voices could make valuable contributions to advocacy campaigns in support of more risky studies at the boundary of public and professional approval, such as challenge studies for tuberculosis, which caused 1.4 million deaths in 2019 (WHO, 2020) and hepatitis-C, which caused an estimated 290,000 deaths in 2019 (WHO, 2021).


5. Conclusion

A small number of people enrolling in this study could make a major contribution to research that could lead to many lives saved, through development of better vaccines and treatments for SARS-CoV-2 or coronaviruses more generally, or through the precedent set for future pandemic challenge studies. Being in a challenge study isn’t for everyone, even those who meet the study inclusion criteria. Different people will give different weights to the inconvenience, discomfort, and potential health effects of enrolling and may conclude the costs are excessively high given their personal situation. However, we believe that the benefits of this study outweigh the risks, and that for many eligible people the benefits of participating outweigh the costs.

As mentioned above, we recognize that trial participation involves a different kind of analysis than an individual decision such as a charitable donation, and we hope this piece sparks discussion on how effective altruists should evaluate it.



Fraser-Urquhart, A. (2021, June 24). I had the coronavirus dripped into my nose. On purpose. For science. Washington Post. https://www.washingtonpost.com/outlook/2021/06/24/covid-challenge-trial-participant/

Manheim, D., Wiȩcek, W., Schmit, V., Morrison, J., & 1Day Sooner Research Team. (2021). Exploring Risks of Human Challenge Trials For COVID-19. Retrieved October 15, 2021, from https://onlinelibrary.wiley.com/doi/10.1111/risa.13726

McShane, H. (2021). A SARS-CoV-2 challenge study in previously infected healthy adults. Retrieved October 21, 2021, from https://web.archive.org/web/20211021144714/https://trials.ovg.ox.ac.uk/trials/sites/default/files/trials_attachments/COVCHIM01_%20PIS%20v6.0_27.08.21_Clean.pdf

Oxford Vaccine Group. (2021, April 29). COVID-19 Challenge Study (COV-CHIM 01). Oxford Vaccine Group: All Recruiting Trials. Retrieved October 01, 2021, from https://trials.ovg.ox.ac.uk/trials/covid-19-challenge-study-cov-chim-01

Rapeport, G., Smith, E., Gilbert, A., Catchpole, A., McShane, H., & Chiu, C. (2021). SARS-CoV-2 Human Challenge Studies — Establishing the Model during an Evolving Pandemic. New England Journal of Medicine, 961-964. https://www.nejm.org/doi/full/10.1056/NEJMp2106970

Rohrig, A., Morrison, J., Pugh, J., Savulescu, J., & McShane, H. (2021). Exploring the Ethics of Tuberculosis Human Challenge Models. PhilPapers. Retrieved October 23, 2021, from https://philpapers.org/rec/ROHETE

Segev, D., Muzaale, A., & Caffo, B. (2010). Perioperative Mortality and Long-term Survival Following Live Kidney Donation. JAMA, 303(10), 959-966. https://jamanetwork.com/journals/jama/fullarticle/185508

Strasburg, J. (2021, September 4). Researchers Infect Volunteers With Coronavirus, Hoping to Conquer Covid-19. Wall Street Journal. https://www.wsj.com/articles/researchers-infect-volunteers-with-coronavirus-hoping-to-conquer-covid-19-11630747801

University of Oxford. (2021, April 19). Human challenge trial launches to study immune response to COVID-19. University of Oxford News. Retrieved October 1, 2021, from https://www.ox.ac.uk/news/2021-04-19-human-challenge-trial-launches-study-immune-response-covid-19

WHO. (2020). Tuberculosis. WHO Fact Sheets. Retrieved October 1, 2021, from https://web.archive.org/web/20211001210826/https://www.who.int/news-room/fact-sheets/detail/tuberculosis

WHO. (2021). Hepatitis C. WHO Fact Sheets. Retrieved October 21, 2021, from https://web.archive.org/web/20211021010035/https://www.who.int/news-room/fact-sheets/detail/hepatitis-c

Sorted by Click to highlight new comments since:

Hi Julia and welcome to the forum! I've been following One Day Sooner for quite a while and I've enjoyed my interactions with your colleagues.

I would find this post easier to read/evaluate if there are bottom-line fast estimates of what the cost-effectiveness looks like, either point estimates or 90% credible intervals.

I understand that a lot of this is highly uncertain and you don't have precise numbers, but some approximate numbers will still help! Even something short like

Assuming you're a healthy 30-yo male, we expect the costs to be about ~X hours of your time (including amortized mortality risk) and the expected benefits to be ~Y lives saved in expectation, with very high error bars

can be helpful, as this will help us prioritize between whether taking time off to do this is better or worse than earning to give, or the direct impacts of our current jobs. 

Pedantic, but I'm somewhat uncomfortable with the rhetoric of whether EAs "should" sign up for this (as in, they have an obligation to do so, which they are failing to live up to if they don't), given the personal risks involved. (I think it's reasonable to have a discussion on the object-level question of whether signing up scores well by EA lights – I'm not objecting to that – though I don't personally have a formed opinion on this question either way.)

It interpreted this post as suggesting that participation was supererogatory. It's true that some ethical systems (like classical consequentialism) don't include supererogatory acts, but that seems like a flaw with those systems. Certainly when I typically tell someone they 'should' get a new credit card, or eat more spinach, I don't intend to imply this is mandatory. Nor does this suggestion seem that different from other requests we make of EAs - the additional risk to a healthy young person is probably a much smaller sacrifice than 10% of one's lifetime earnings!

the additional risk to a healthy young person is probably a much smaller sacrifice than 10% of one's lifetime earnings

FWIW, I'm also against people saying "EAs should give at least 10% of their income to charity" – this makes people who don't want to make that sort of commitment feel unwelcome, and my sense is that rhetoric along those lines has hurt movement growth.

Thanks for the comments! Just wanted to quickly say that Larks's interpretation of our intention was correct: we view participation in this study is superogatory (and are really making the argument that this study might be in the range of actions that are considered effective altruist).

Curated and popular this week
Relevant opportunities