Since the COVID-19 pandemic began, the search for an effective treatment has been fraught. In the United States, public attention has focused on the anti-malarial drug hydroxychloroquine, in large part because of President Trump’s endorsement of the drug. After he amplified a small study in France that suggested the drug could be an effective treatment, prescription sales of hydroxychloroquine skyrocketed. Well-developed, randomized clinical trials have since found that hydroxychloroquine is not an effective treatment for COVID-19, but this has done little to reduce interest in the drug—especially in the White House, where Trump and his aides most recently latched on to a less rigorous, observational study that purported to demonstrate the drug’s effectiveness.
The debate over hydroxychloroquine has become deeply politicized, which has obscured more nuanced debates within the scientific community over what constitutes actionable evidence. With social media platforms acting as key arbiters in the circulation of health information, these nuances are particularly important for the major platforms to understand. By framing decisions around removing content related to hydroxychloroquine as a choice between harmful medical misinformation vs. science, platforms may be closing off space for inquiry and debate and incentivizing the consumption of more medical misinformation instead of less. We advocate an approach that gives voice to experts and allows dialogue on different approaches to medical uncertainty.
Medicine in a time of uncertainty
Our collective reactions to hydroxychloroquine highlight the underlying tensions about when and how to act in times of uncertainty and demonstrate how people weigh health risks and prioritize action vs. inaction differently. It is critical to understand those differences because the COVID-19 pandemic will continue to force politicians, researchers, and the public to make health decisions based on unsettled science. While the hydroxychloroquine story is sometimes viewed as a battle between legitimate scientific information vs. dangerous misinformation, this fails to consider debates within medicine about when new evidence reaches the level of “actionability.”
In medicine, there is a long-standing impulse to prescribe a range of treatments for the sake of doing something, even in cases where medical evidence is fairly clear that those treatments are ineffective. Doctors still place stents in stable patients with cardiovascular disease, despite mounting evidence against this practice. By taking action, patients and medical providers can feel a sense of control over health risks. In the face of a new disease, the impulse to act can be even more pronounced, and high-profile medical officials have argued in favor of quick action against COVID-19. In April, a petition started by Philippe Douste-Blazy, France’s 2017 candidate to lead the World Health Organization, and signed by other prominent medical figures, argued that hydroxychloroquine should be made widely available to patients pending further evidence of effectiveness.
This inclination toward action—with Trump’s encouragement—led the Food and Drug Administration to grant emergency approval in late March to use the drug as a treatment for COVID-19. It is difficult to speculate about Trump’s intentions when he advocated the use of hydroxychloroquine. It is additionally an open question as to whether the FDA was pressured into taking this action because of politicization of the agency, or whether this decision was made after careful consideration of the benefits and risks of the drug. But as evidence from randomized studies showed that patients with COVID-19 did not benefit from the drug, the FDA rescinded that approval on June 15 and thereafter published a report summarizing safety issues it had catalogued.
The FDA’s approach mirrors that of some medical experts, who favor action in the face of uncertainty. Didier Raoult, the French physician whose preliminary evidence provided the justification for the drug to be given widely to patients, did so because he argued the potential benefit to patients fighting a relatively unknown disease was great enough to outweigh the risks posed by recommending a drug on the basis of study with a small sample size. A 2014 report from the World Health Organization similarly argued that it was ethical to give Ebola patients drugs that were still untested in humans.
In recent years, this debate within the medical community has turned into a political one, with conservative lawmakers working to cut FDA regulations and lower evidentiary standards for new drugs. Proponents of this approach argue that patients should be the ones choosing whether to start using a new drug “at their own risk.” In 2018, Trump signed a federal “Right to Try” act giving terminally ill patients access to experimental drugs before they are approved by the FDA (although critics argued that these drugs were already made available to patients when requested). While “Right to Try” laws have recently been championed by conservative lawmakers, this argument does not fall cleanly along ideological lines. In the 1980s, it was progressive HIV/AIDS activists pushing the FDA to allow the use of unproven treatments.
Thus, current debates that position science on the political left, and personal liberty on the political right, are oversimplified. Medical professionals often offer a range of treatment options to patients, and personal choice is an important part of clinical practice. Furthermore, even experts disagree: This is why patients with rare or serious illnesses are encouraged to seek second and third opinions from different specialists. As information about COVID-19 continues to trickle to the public, the risk of describing the debate over how to treat the disease as one of science vs. liberty and truth vs. falsity may weaken the public’s willingness to abide by simple public health protocols.
The ‘do no harm’ approach
The medical community is divided on the question of when to prescribe experimental, unproven drugs. Many health researchers prioritize inaction until the medical evidence stabilizes and emphasize that the use of unproven drugs can also cause unintended negative effects. In the case of hydroxychloroquine, these researchers stress that the evidence is not sufficiently strong to promote its use to treat COVID-19 and that the drug could pose a serious danger to patients, for example by causing heart arrhythmia.
While it doesn’t get the same attention as more aggressive uses of untested drugs, this “Do No Harm” approach seems to be gaining more followers in medicine. Organizations dedicated to reducing overuse of unproven treatments, such as Choosing Wisely, have grown precipitously over the past ten years. The proponents of inaction in the face of uncertain evidence place high value on the process of evidence-based medicine and well-designed, randomized, clinical trials. They are skeptical of observational evidence in medicine, citing many cases where treatments or procedures seem to improve outcomes and are later proven ineffective. They argue that high-quality, but often costly and time-consuming, scientific studies are necessary before recommending a treatment because they help combat social and cognitive biases in what we believe to be true.
In their view, hydroxychloroquine and other potential COVID-19 treatments will remain unproven until these high standards of medical evidence have been met, and thus should not be recommended to patients except in specific circumstances.
The challenge for Silicon Valley
Given this spectrum, how can social media platforms appropriately manage information on potential COVID-19 treatments? Since the pandemic began, social platforms have been lauded for their proactive approach to removing potentially harmful medical misinformation. But as social media platforms like Twitter have started to recognize, content removal is an inadequate tool for managing health information in a time when the “truth” about COVID-19 is being actively debated.
Hydroxychloroquine was a relatively obscure treatment option until a misleading tweet caught the eye of Silicon Valley elites and mainstream conservative media. The idea was amplified on Fox News, and within three weeks, the president was urging Americans to take the drug. The Strategic National Stockpile now has 63 million doses of an ineffective drug. While Twitter has since removed the original tweet, the company has demonstrated a haphazard approach to mediating hydroxychloroquine information—removing some tweets and keeping others.
If social media platforms are mediators of health information, they (in consultation with the broader public and established experts) need to develop better systems for adding context to health information that both reduces confusion and allows for open debate. Similarly, journalists have characterized the hydroxychloroquine debate as a story of misinformation, but this isn’t the right terminology to deal with the complexity of the relationship between medical evidence and action.
Social platforms are creating new taxonomies to organize and vet information that often ignore historically grounded debates around the production of knowledge. When platforms simply try to litigate what is true and false, they obscure the contingency of knowledge production, and this is especially problematic in a situation where new research around COVID-19 is constantly in flux. It would be more helpful to build a system where the public could engage and learn about the process of producing medical facts.
It’s not just that we keep learning new information about COVID-19: The uncertainty that we collectively experience in our current moment has to do with the processes and structure of research itself. Research is conducted by diverse bodies and individuals, and many groups tackle small portions of the problem from different academic disciplines and medical specializations. One piece of new information may seem relevant but may be debunked in light of a different study. Research is a piecemeal process, and it is impossible to know beforehand what research will end up being critical to the discovery of a vaccine, a pharmaceutical cure, or an understanding of the disease that becomes pivotal.
How, then, can social platforms and traditional news outlets protect the public while also respecting appropriate debate about emerging treatment options? The first step is to recognize that debates about COVID-19 treatments are imbued with both science and values. We often hear that “the science is being politicized” or that individuals are not “listening to the science.” But scientists themselves still disagree, and this framing minimizes the social processes that are an important part of science and medicine. The way people calculate and manage health risks is influenced by their own values and lived experiences, so the debate about the actionability of uncertain evidence requires input from a range of social groups.
Rather than viewing this process as a battle between science vs. misinformation, we should instead focus our attention on defining the values we want to prioritize as we continue to sort through fast-moving and unstable evidence in a pandemic. Do we want to err on the side of doing something in the face of uncertain science, or reduce the potential for unintended negative effects? What types of evidence do we need, collectively, before we can agree that a treatment is effective or ineffective? Even within this debate, the choices are context specific: A particular disease and the way it presents may push a patient toward action. And some of the values that shape medicine itself are up for debate: In moments of crisis, when supplies and ER beds are short, triaging of patients can prioritize those who are young, white, rich, educated, and already in good health. Without critically examining the values and biases that shape these processes, we risk undermining trust in science and medicine, increasing the likelihood of medical misinformation
How can social platforms navigate this messy epistemological ecosystem that impacts every member of society during a pandemic? The imperative to get this right is especially high as we anticipate a second wave of COVID-19 in the United States. This is not an easy process and isn’t something social platforms can do on their own. It requires bringing experts to the table when making content removal decisions and assembling a truly multi-stakeholder framework—something far more complex than the Oversight Board that Facebook has created to mediate content removal decisions. This is especially true as the major public health bodies, the Centers for Disease Control and the World Health Organization, have either been slow to publish guidelines around COVID-19 or have advocated science that is deeply contested by researchers and academics. Since science is global, and many of the institutions producing cutting-edge research on COVID-19 are outside the United States, an international body of scientists, researchers, and public health experts would need to contribute to this process. Civil society groups and those who represent the diverse needs of vulnerable or minoritized communities would also need to be included.
This process will require input from diverse communities and public health institutions, and transparency about how content moderators are setting the parameters of what counts as settled or unsettled science. But COVID-19 could serve as a moment of public engagement with science that would build, rather than undermine, healthy conversation about how to act on uncertain medical evidence. This seems particularly important in a moment when we are all patients, researchers, and decision-makers about our own health choices in the absence of a properly coordinated federal response to the virus.
Smitha Khorana is a journalist, academic, and former medical student, and is currently the newsroom outreach lead at Data & Society.
Kellie Owens is a researcher at Data & Society.
Facebook and Twitter provide financial support to the Brookings Institution, a nonprofit organization devoted to rigorous, independent, in-depth public policy research.