Big data in health: an ethical framework

Author
Prof. Effy (Eftychia) Vayena
ETH Zurich

Interview with the principal investigators of this NRP75 project.

What was the aim of your project BEHALF?

In the age of data-driven medicine, data generated for multiple purposes – which go beyond health reasons– can provide valuable information about health. The aggregation and analysis of heterogeneous data can result in preventive, diagnostic and therapeutic benefits. Nevertheless, health-related data collection, use, storage, and sharing raise a variety of ethical dilemmas. How to implement informed consent in data-driven biomedical research? How to guarantee research participants’ privacy and protection from dignitary harm? How to ensure a fair distribution of benefits emerging from Big Data research?

Our NRP 75 project BEHALF aimed to

  • Map the ethical and conceptual issues of Big Data in healthcare and biomedical research.
  • Assess whether Big Data issues impact the effectiveness of the existing ethical oversight mechanisms. Which gaps exist in the current oversight mechanisms for uses of Big Data in health research? Particular attention was dedicated to reviewing the strengths and weaknesses of Swiss Research Ethics Committees (RECs).
  • Develop an ethics framework (made of policy recommendations) that could support the work of RECs in Switzerland and abroad by addressing existing regulatory gaps in health research practices involving Big Data.
  • Translate the ethical framework in a concrete toolkit usable by both RECs and researchers.

What were the results?

Our project led to numerous findings which were summarized and published in 13 peer-reviewed open access scientific publications.

By charting the ethical and conceptual issues related to Big Data research in the biomedical sector, our research showed that over the past ten years, the research community became increasingly aware of the potential of Big Data and AI-enabled technologies in the field of health and medicine. Our analysis revealed that privacy is the most debated ethical concern of Big Data and AI-enabled technologies. In fact, researchers and app developers seem to understand the ethical use of Big Data mainly as compliance with existing data protection regulations, ignoring other relevant ethical issues (e.g., research accountability, fair distribution of risks and benefits, individual autonomy and empowerment, and group-level harm mitigation).

Further important findings?

Based on our conceptual and empirical analysis we reflected on the types of reforms needed to improve the ethical evaluation of Big Data projects and the ethical oversight of the digital health tools which rely on Big Data. As anticipated in the original project proposal, in this research phase we organised an expert workshop which aimed to gain feedback on the results of our analyses from international experts, as well as to define RECs reforms and governance recommendations that could be implemented in Switzerland and abroad. The resulting collective paper suggests the need to reform RECs at various levels (i.e., regulatory, procedural, complementary) to enable them to effectively fulfil their role as the key oversight mechanism in biomedical research using Big Data.

In the last phase we translated the ethical recommendations for RECs into something more concrete, an ethics toolkit. The first part is a self-assessment tool designed to help RECs in Switzerland and abroad assess their preparedness and address their weaknesses in the Big Data research context. The second part of the toolkit is an ethical assessment tool in the form of a checklist. The ethics toolkit is available online.

What was a special highlight? What went beyond what was planned?

Using the ethical toolkit and the BEHALF research findings, Swissethics (the umbrella organisation of the Swiss Cantonal Ethics Committees) modified and expanded the templates used by researchers for project submission; specifically, regarding projects involving the further use of data (with and without consent). We consider the integration of our checklist into the templates of Swissethics a major achievement of this project. Our work concretely changed the way in which RECs assess Big Data research, as well as providing additional ethical guidance to researchers in planning and carrying out their projects.

What are the main messages of the project?

  • While privacy and data protection are extremely relevant ethical issues in Big Data research, they should not be considered as ethics panacea. Rather the spectrum of ethical implications should be urgently addressed, going beyond prescriptions of data compliance. While failure to address these ethical challenges could result in harms and distrust both in the technologies and responsible institutions, success might guarantee benefits for individuals as well as the broader society. Our ethics toolkit could help in safeguarding society while adequately ensuring an ethically aligned use of data.
  • Big data research in the biomedical sector and beyond involves a variety of stakeholders, including researchers, developers, the private sector, professional organisations, decisionmakers and the public. These stakeholders may have competing interests and different expectations about how society will look in the future, and which technology tools should be used. To overcome this limitation, public involvement is crucial for orientation towards a common set of values that can inform ethical guidance and technology development.
  • Existing oversight mechanisms for Big Data uses in biomedical research and health technology development must be reformed and strengthened at various levels. For instance, the members of ethical oversight committees should ensure adequate expertise to address emerging technical challenges. Moreover, oversight systems should be adaptable to technological change, including ongoing monitoring and robust assessment of the research project or technology development beyond the initial risk-benefit evaluation.

Does your project have any scientific implications?

As mentioned before, a main scientific implication of this project consisted of the revision of the Swissethics templates for researchers. The findings of this project contributed to the revision and expansion of the Swissethics templates used by researchers for project submissions. Some template sections were expanded, while others were introduced anew, such as sections about biases in the datasets, transparent data processing, strategies to store data securely, risk-benefit assessment, scientific validity of the research, and findings dissemination. The findings of BEHALF also informed the revision of those checklist used by Cantonal RECs to assess whether or not a research protocol is ethically aligned.

It was a goal of this project to clarify relevant ethical issues in Big Data for healthcare and biomedical research from a conceptual and normative perspective. Our checklist – the ethical assessment tool – can guide researchers in conducting more ethically-aligned Big Data research. Using our ethics tool, they can identify how well their project meets the standards of the REC. Thus this tool helps researchers to reflect on and address ethical issues that might emerge in their research. Moreover, it establishes common practices to govern the creation, collection, storage, processing, internal sharing, analysis, dissemination, and re-use of data in biomedical research.

Furthermore, offering clarity about best practice might ultimately benefit the private sector as well. Developers and researchers can use this tool to recognize ethical issues which could cause reputational harm if left unaddressed. In addition, private sector use of this tool would facilitate collaboration with academic research partners. As the health data ecosystem expands, Big Data research will involve partners from many sectors. As public concerns about Big Data use grows, it is important that any Big Data activity in health takes place within a clear and ethically robust framework. Using this toolkit can contribute to ensuring public trust in Big Data activities, and therefore more sustained support.

Does your project have any policy recommendations?

This project highlights gaps in the current regulatory and oversight landscape of Big Data uses for health research and digital health. The following policy recommendations may help to fill these gaps and streamline the processes that govern ethical data uses:

  • Revise and expand the existing ethical standards, going beyond data protection and privacy preserving approaches
  • Translate scattered and general ethical guidance into actionable and streamlined practices
  • Establish minimum international standards for ethically aligned data uses
  • Introduce regulatory, procedural, and complementary reforms to strengthen the role of the existing oversight mechanisms and improve the review process

About the project

Related links