Abstract
This paper poses the question of whether people have a duty to participate in digital epidemiology. While an implied duty to participate has been argued for in relation to biomedical research in general, digital epidemiology involves processing of non-medical, granular and proprietary data types that pose different risks to participants. We first describe traditional justifications for epidemiology that imply a duty to participate for the general public, which take account of the immediacy and plausibility of threats, and the identifiability of data. We then consider how these justifications translate to digital epidemiology, understood as an evolution of traditional epidemiology that includes personal and proprietary digital data alongside formal medical datasets. We consider the risks imposed by re-purposing such data for digital epidemiology and propose eight justificatory conditions that should be met in justifying a duty to participate for specific digital epidemiological studies. The conditions are then applied to three hypothetical cases involving usage of social media data for epidemiological purposes. We conclude with a list of questions to be considered in public negotiations of digital epidemiology, including the application of a duty to participate to third-party data controllers, and the important distinction between moral and legal obligations to participate in research.
Introduction
In traditional public health, a long-standing "social contract" suggests that individuals have a moral duty to participate in research or surveillance to prevent public harm. This duty often justifies the collection of health data without explicit informed consent, particularly during infectious disease outbreaks. However, the emergence of Digital Epidemiology (DE)—which leverages "data exhaust" from social media, search engines, and wearable devices—fundamentally changes the ethical landscape. This paper explores whether this social contract extends to the world of Big Data and proposes a normative framework to determine when a moral duty to participate is ethically justifiable.
The Digital Shift: New Risks and Data Types
Digital epidemiology differs from traditional clinical research in three critical ways that complicate the moral duty to participate:
- Non-medical Origin: Data is often repurposed from contexts not originally intended for health monitoring (e.g., Twitter posts or search queries).
- Granularity and Pervasiveness: DE tracks real-time behaviors and locations, often without the subject's active awareness, and can put groups at risk through profiling even if individuals are not the primary subjects.
- Proprietary Ownership: Data is frequently held by private corporations rather than public health institutions, raising questions about who benefits from the data.
Because of these shifts, the ethical justifications used in the doctor-patient relationship cannot be directly applied to the relationship between a platform user and a public health researcher.
A Framework for Justification: The Eight Conditions
To determine when a moral duty to participate in DE is ethically defensible, the authors propose a normative framework consisting of eight justificatory conditions. These must be met to justify the use of non-medical, proprietary data for public health purposes:
- Strong Public Interest: The research must address significant disease prevention or health promotion goals.
- Plausible Assumption of Benefit: There must be a reasonable expectation that the specific data use will contribute to the stated public health goal.
- Minimal Risk: The risks to participants (e.g., privacy breaches or stigmatization) must not be significantly greater than those of non-participation.
- Negligible Effort: The burden on the data originator should be minimal; participation should not require significant active labor.
- Broad Understanding: While specific consent may be bypassed, subjects should have a general awareness of how their digital footprints might be utilized for research.
- Data Minimization: Researchers must use only the necessary amount of identifiable data required for the study.
- Community Engagement: Researchers must engage with affected communities, especially where there is a risk of group-level harm or stigmatization.
- Harm Mitigation: Strategies for compensation or redress must be established in case of data misuse or unintended harm.
Key Distinctions: Sharing vs. Creating
A vital distinction in this framework is between the duty to share (allowing existing data to be used) and the duty to create (actively generating new data for a study). While a duty to share might be justifiable under the eight conditions, the authors argue that a duty to create—such as requiring citizens to actively post symptoms on social media—is almost never ethically defensible outside of extreme, immediate crises.
The Responsibility of Data Controllers
A significant insight of this analysis is that the moral duty to participate may shift from the individual to the data controllers (e.g., Google, Meta). As gatekeepers of the "public good" inherent in large-scale datasets, these platforms may have a greater obligation to facilitate access for public health purposes than individuals have to provide the data in the first place. This opens a contentious area of international law regarding the social contract of private corporations.
Conclusion
While a moral duty to participate in digital epidemiology can exist, it is highly contingent on the immediacy of the public health threat and the plausibility of the research benefits. This duty remains a moral one; legal coercion should be reserved for "abnormal circumstances" involving overwhelming threats to public safety. Future public negotiations must address the responsibilities of third-party data controllers and the evolving nature of the social contract in the digital age.
Summary updated