DISCUSSION NOTES BY CANAY ÖZDEN-SCHILLING:
Our first virtual event after the pandemic’s onset explored issues around the ethics of data collection during COVID-19. We were fortunate to have the discussion led by Dr. Stuart Ray, a professor of medicine and vice chair of medicine for data integrity at JHMI, and Dr. Alexis Walker, an anthropologist and postdoctoral fellow at the Berman Institute of Bioethics, who is doing research on the ethical, legal, and social implications (ELSI) of genomics. In his introductory comments, Ray recapped a number of COVID-related developments he has been involved in as a physician and researcher. The medical school commands a vast dataset including the data of any person who has had their COVID test within the institution. The urgent task is to make that data available under stringent measures to researchers who are able to make discoveries. In addition, the sequencing of the genomes of people infected with COVID-19 rapidly ramped up, using metadata (travel history, time of diagnosis) shared, once again, in urgency but under rules aimed to protect privacy.
Our main reading for the event was one recently co-authored by Walker on the ethics of identifying the information of individuals infected with infectious diseases (along with this one and this one that provided broader context for data collection for public health purposes). In her introductory comments, Walker charted the general territory of the vast world of research on ethics in genomics, much of which is focused on privacy. There are certain specific public health issues that are unique to infectious diseases, characterized by the urgency to act swiftly for public health purposes while protecting the privacy of the individuals surveilled. Walker noted that much of the social science funded by NIH is in the ELSI of genomics. Human genomics can potentially tell us something about one’s susceptibility to specific infectious diseases, in addition to the behavioral side of the issue. Walker concluded her remarks by noting that genetics and genomics data should not be necessarily considered more sensitive than the behavioral kind. Ray followed up with a note on why the current moment is different from the SARS outbreak (2002-2003). During that event, infected people were numerable and importantly, there had not been a single transmission from an asymptomatic individual, which is markedly different from the COVID-19 case, especially in terms of how we might control this pandemic. SARS was dominated by super-spreaders in a way that allowed researchers to focus on those limited areas of dissemination, whereas for COVID-19 we need larger aggregations of data.
Jeremy Greene kicked off the discussion with a general question on what data is being asked to do in the COVID-19 moment and how COVID-19 might change how we think about data. In a moment when urgency may be said to eclipse other reasonable public concerns, academic journals have tended to decline non-COVID-19 related work, while publishing some COVID-19 work that was at the pre-peer review stage, while retracting other, immaturely published articles. We see in this moment, Greene noted, a convergence of the multiple literatures on ethics, infectious diseases, and data. Ray noted that we overall accept there is a loss of privacy and a reorientation of the institutions we trust at this moment. Veena Das offered to re-anchor our discussion in a way that would be inclusive of broader international contexts. There is a certainty to the way epidemiologists are speaking, Das offered, noting the Imperial College report that admitted to recommending feasible policy without necessarily a consideration for ethics. Can we also bring into a discussion of privacy the uneven impacts of social inequality? Gail Geller, a co-author of Dr. Walker, noted a discussion of “epistemic humility” from a recent conference she attended, which denotes the importance of scientists and policy-makers acknowledging what they know and, perhaps more importantly, what they don’t know. But this humility, as Ray pointed out, is hampered by the fact that news anchors gravitate toward scientists who express certainty. Yulia Frumer drew attention to the issue of re-identification that is discussed in our three readings. Carolyn Sufrin called attention to different levels of privacy as emanating from her research on COVID-19 in prisons. Ray also agreed that prisons’ disproportionately non-white population that are surveilled in a for-profit context complicate what a discussion of privacy might mean.
A number of participants wanted to draw attention to the America-centrism of the privacy discussions in the COVID-19 debates. Heba Islam posed the question of whether privacy was an end in itself, especially considering places where the consequence of privacy violation can be relatively more severe. Naveeda Khan followed up on the question of whether the concept is robust enough to be moved across contexts. Veena Das drew attention to the modes of surveillance that already exist in a way that should compel us to think about privacy, not as generalized anxiety, but as a context-specific problem, bringing in the example of India’ recent Supreme Court decision on whether life is a fundamental right or a right accorded to individuals by the state. Khan and Das both dwelled on the failures of data collection, when hugely consequential privacy violations are made in the service of acquiring data without a clear purpose.
Ray responded by noting that context-specific social inequalities might color our notions of privacy—the consequence of privacy violation in the US context, for example, is not being able to guarantee that someone will remain insurable after a data breach. We also had a brief discussion on the spectrum of government versus corporate surveillance, with Ray noting that corporate surveillance continues to be suspect without the right structures of accountability in place. We concluded the event with more questions on what technological fixes and epidemiological models could accomplish for us, to be asked again at future COVID-19 events.