Research Data Disclosure in the Digital Age

We live in a digital age of rapidly evolving means and opportunities to collect data. The amount of data available grows exponentially with new technology, and meaningful use of individual-level health data holds the promise of supplementing clinical care and improving health outcomes. As researchers explore ways to use research participant data to promote individual and population health, investigators and participants must acknowledge the changes that arise when conducting research in this Digital Age. In a recent Ethics Review, we described our research team’s experience with developing and implementing ecological momentary assessment (EMA) to enhance adverse event monitoring following the return of amyloid-β (Aβ) imaging results to individuals with mild cognitive impairment (MCI) [1]. Preliminary findings showed that participants actively engaged in almost-daily phone calls for two weeks after Aβ imaging result disclosure, and that use of EMA may be considered in older adults as a way to engage and capture real-time data to provide safeguards in research. In response to our Review, Nebeker et al. [2] presented an “intermediate research model”, proposing integration of the traditional research model within the current digital “Internet of Things” era. Their response highlighted the potential for both researchers and research participants to access and learn from the data collected and introduced “MyTerms”, a concept whereby participants can tailor the return of their individual-level data to the terms and conditions that align with their own wishes. Based on the National Institutes of Health guidelines, Nebeker et al. [2] compare and contrast the ethical challenges of this intermediate research model with those posed by the traditional research model. Our study team’s use of EMA for adverse event monitoring is an example of how data can be collected via a low-tech platform, while still capturing meaningful, time-sensitive participant data. Thank you for the opportunity to expand on the response by Nebeker et al. [2]. Our response focuses on research participant receipt of individual-level research data and further ethical considerations.

“MyTerms” is an innovative approach for promoting informed, prospective decisions about the type of research results that one may wish to have returned to them, as an individual participant in the study. “MyTerms” has the potential to foster a sense of empowerment among prospective research participants at the time of consent. Specifically, potential participants may feel empowered by the apparent transparency and control over their participation that such a model would afford. This may, in turn, promote a culture of reciprocity wherein both researchers and participants may both benefit from aspects of the research process. It is also crucial to frame future research questions that consider and define factors that deliberately examine feasibility and clinical impact. A few factors to consider include: 1) patient-level factors (e.g., interest in the information to be returned, possible changes in insurance coverage with new knowledge), 2) provider-level factors (e.g., time, cost, clinical utility, disease management considerations), 3) funder-level factors (e.g., cost, return on investment, actionable results), and 4) community- or societal-level factors (e.g., culture shift in healthcare expectations). The “MyTerms” approach prioritizes customization in data exchange with participants which will require additional staff time and resources. There is also a need for data management across the study—study development to the period following study completion presents potential obstacles. For example, changes in participant preferences and study modifications during implementation will require additional consent, and in some cases, preferences may not be honored depending on the study findings or study modifications. Similarly, the digital data lifespan necessitates long-term data management to ensure data security, and this requires time and resources that may exceed the study duration and funding.

As investigators begin or continue to share data, there are also concerns with re-identification of participant-level characteristics within datasets when merged with publicly-available data sets. Datasets may share a common element connecting datasets together or predictive analytics may be used to make informed assumptions regarding the identity of a research participant. As open science becomes the gold standard by which to share emerging and ongoing research, the ability to combine “non-identifiable” datasets will increase, and researchers must continue to protect confidentially of individual participants. The ability to identify individuals within individual-level “de-identified” data will need to consider possible ways other people (researchers or otherwise) may be able to deduce or identify individual participants with study data when merged with other available datasets. Protection of human subjects is imperative if ethical research is to be upheld, and although we cannot predict the ways in which data can be used in the future, it is the researcher’s responsibility to consider the potential implications by discipline and domain. Currently, there are differences in how individual-level study participant data are shared and made available to individuals outside of the direct research team. For example, National Institutes of Health Public Data Repositories require that publicly-funded study data be made publicly available for reuse, unless restrictions or limitations on data submission and requirements are necessary [3]. In addition to formal repositories, individual-level data can be compiled or pooled across studies, but this requires considerable data management across studies and adequate protection of human subjects. Data collected outside of traditional research settings does not currently require the same oversight and protections as those collected in research settings. This creates a gap between research- and non-research collected data protections and sharing that impacts access and use. One group addressing the maintenance of privacy across all data-collecting and sharing settings is the Future of Privacy Forum (FPF, http://www.fpf.org), funded by the National Science Foundation and Alfred P. Sloan Foundation. They seek to advance the responsible use of data and uphold privacy by bringing academic ideas to non-academic settings [4]. FPF interdisciplinary working groups created a central repository for “privacy-related guidance documents, reports, codes of conduct, and other resources that can help ...   navigate complex issues and implement initiatives in privacy-protective ways.” [4] Further, this work also calls upon policymakers to take the lead in presenting and supporting responsible use of data across public and private domains.

For most, improving patient outcomes is the ultimate goal of clinical research. The patient is the cornerstone of clinical research, without whom we could not pilot, test, and translate health advances into everyday care and practice. Reviewers asked, “if it is perhaps contrary to this principle to not consider participants as deserving of return of study information.” This should be true for both the data collected that participants can “see” (e.g., weight displayed on a scale) as well as those data that may require further explanation or even counseling to understand and emotionally process (e.g., results from a positron emission tomography [PET] scan). Not every stage of research tackles a clinical paradigm change. The purpose may be to describe or create the building blocks by which an effective solution will adopted by providers and patients in the future. Depending on the research study, it is likely that not every unit of data is appropriate or useful for a participant to receive. As presented by Nebeker et al. [2], there must be thoughtful consideration on the part of the researchers, institutional review boards (IRBs), and participants as to what can (or should) foreseeably be shared with and without counseling or clarification. In the case of our study examining participant receipt of Aβ PET imaging results, it was essential that there was counseling provided by trained study staff, supplemented by adverse event monitoring through EMA, as well as knowledge of a participant-known healthcare provider if follow-up was needed.

In conclusion, we appreciate how Nebeker et al. [2] extended our work in new directions, and we look forward to hearing about “MyTerms” in the future. As digital health data collection and sharing grow within research, ethical obligations to participants must remain a priority to researchers, IRBs, funding agencies, and consumers. Ethical data management and consideration of new data analytic approaches and technology require investments and oversight across the research lifespan.