Creating a Culture of Ethics in Biomedical Big Data: Adapting ‘Guidelines for Professional Practice’ to Promote Ethical Use and Research Practice

Abstract

Two scientific domains that are crucial in “Biomedical Big Data”, computing and statistics, do not typically require “training in the responsible conduct of research” or research ethics. While “responsible conduct of research” (RCR) comprises interactions with subjects (human and non-human), it also involves interactions with other scientists, the scientific community, the public, and in some contexts, research funders. Historically, the development or emergence of disciplines and professions tend to involve a semi-simultaneous emergence of professional norms and/or codes of conduct. However, Biomedical Big Data is not emerging as a single discipline or profession, and engages practitioners from many diverse backgrounds. Moreover, the place of the data analyst or the computer scientist developing analytic algorithms seems to be too granular to be considered specifically within the activities that comprise “responsible research and innovation” (RRI). Current legal and policy-level considerations of Biomedical Big Data and RRI are implicitly assuming that scientists carrying out the research and achieving the innovations are exercising their scientific freedom – i.e., conducting research – responsibly. The assumption is that all scientists are trained to conduct research responsibly. In the United States, federal agencies funding research require that training in RCR be included – some of the time. Because the vast majority of research that was federally funded has not included Biomedical Big Data, RCR training paradigms have emerged over the past 20 years in US institutions that are not particularly relevant for Big Data. While it would be efficient to utilize such established, well-known, easily-documented RCR training programs, this chapter discusses how and why this is less likely to support the development of professional norms that are relevant for Biomedical Big Data. This chapter will describe an alternative approach that can support ongoing reflection on professional obligations, which can be used in a wide range of ethical, legal, and social implications (ELSI), including those that have not yet been identified. This may be the greatest strength of this alternative approach for preparing practitioners for Biomedical Big Data, because the ability to apply prior learning in ethics to previously unseen problems is especially critical in the current era of dynamic and massive data accumulation. To support the development of normative ethical practices among practitioners in Biomedical Big Data, this chapter reviews the guidelines for professional practice from three statistical associations (American Statistical Association; Royal Statistics Society; International Statistics Institute) and from the Association of Computing Machinery. These can be leveraged to ensure that, in their work with Biomedical Big Data, participants know and understand the ethical, legal, and social implications of that work. Formal integration of these (or other relevant) guidelines into the preparation for practice with data (big and small) can help in dealing with ethical challenges currently arising with Big Data in biomedical research; moreover, this integration can also help deal with challenges that have not yet arisen. These outcomes, which are consistent with recent calls for the institutionalization of reflection and reasoning around ELSI across scientific disciplines, in Europe, are only possible as long as the integration effort does not follow a currently-dominant paradigm for training in RCR. Preparing scientists to engage competently in conversations around ethical issues in Biomedical Big Data requires purposeful, discipline-relevant, and developmental training that can come from, and support, a culture of ethical biomedical research and practice with Big Data.