Ethics in Biomedical Research – A Refresher Course

On 9 March, 2023, the Centre for Bioethics and Humanities held a webinar on Ethics in Biomedical Research. A/Prof Dr Teguh Haryo Sasongko (in the photo above), the Acting Deputy Director of the Institute for Research, Development, and Innovations at the International Medical University (IMU) in Kuala Lumpur, presented on ‘The Principles of Participant Protection in Biomedical Research involving Human Subjects’. He emphasized the importance of informed consent, confidentiality, and the protection of vulnerable populations in biomedical research, as well as prioritising the safety and well-being of human subjects. Meanwhile, Prof Patrice Francois Boursier and Dr Wong Siaw Ming, both of whom are affiliated with the Centre for Digital Health and Health Informatics at the IMU in Kuala Lumpur gave a talk on ‘Big Data Analytics and its Ethical Challenges’ in the context of biomedical research.

For the first session, A/Prof Dr Teguh highlighted the history of human rights violations in biomedical research, such as the Nazi experiments, the Tuskegee Syphilis Study, and the Plutonium Experiment, and emphasized the importance of preventing such violations from occurring again. To guide ethical conduct in biomedical research involving human subjects, researchers and institutions have developed a set of principles, including respect for persons, beneficence, non-maleficence, and justice.

A/Prof Dr Teguh also provided several concepts to guide ethical decision making in biomedical research, including the universalizability test, impartiality test, and interpersonal justifiability test. In Malaysia, various guidelines address different types of biomedical and health research involving human subject participants, based on the principles of respect for persons, beneficence, non-maleficence, and justice. These guidelines require that researchers obtain informed consent from participants, minimise risks, and ensure that participants receive appropriate care.

At the international level, the World Medical Association has developed the Declaration of Helsinki, a set of ethical principles for medical research involving human subjects.

The Declaration requires that researchers obtain informed consent from participants, minimise risks, and ensure that the benefits of the research are distributed fairly. It also requires that researchers conduct research in a way that respects the culture and dignity of the participants.

Other guidelines are available at this website: https://www.sidcer-fercap.org/pages/guidelines.html.

In conclusion, A/P Dr Teguh emphasized the importance of participant protection in biomedical research involving human subjects and the application of these principles to minimise the risk of human rights violations. Ethical decision-making concepts and national and international guidelines can provide further guidance in ensuring the ethical conduct of biomedical research involving human subjects.

As for the second session, both Prof Patrice and Dr Wong highlighted the 5Vs of big data – Volume, Velocity, Variety, Veracity, and Value – and discussed the different types of big data sources in healthcare. They also explained how big data analytics is used in biomedical research, such as diagnosing diseases, selecting treatment options, and performing predictive analytics for disease prevention.

They identified the primary difference between big data analytics and everyday informatics as its complexity in how data is stored, accessed, and processed, as well as the machine learning component that enables predictive and prescriptive capabilities. These characteristics pose new challenges to the 5 principles in data ethics: intent, outcome, privacy, ownership, and transparency. The main ethical challenge lies in the nature of big data as a secondary data source, as it is typically captured as and when it is generated and set aside to be used later. This poses challenges for transparency and consent.

Both speakers also discussed potential solutions to these challenges, such as broad consent with the support of a trusted party to ensure proper use of the data in the future. They also highlighted different techniques to address the distributed nature of big data architecture, such as data anonymization, de-identification, pseudonymization, and differential privacy.

However, the speakers noted that some ethical issues, such as poor data quality, unrepresentative data, and inherited biases, are more difficult to tackle and could lead to the development of biased algorithms if used as training data. The black-box nature of machine learning algorithms also requires human interventions to detect the deliberate efforts of manipulating the algorithm’s behavior for malicious purposes.

Hence, they suggested a more comprehensive approach to addressing these ethical challenges, such as implementing a corporate-wide data ethics framework that cultivates ethical data practices and oversees end-to-end data collection and usage throughout the organisation on an ongoing basis. They also suggested reviewing the regulatory framework to incorporate data ethics considerations into the current Data Privacy Acts and imposing penalties on the responsible and/or accountable parties for misconduct.

Overall, the second session provided valuable insights into the ethical challenges surrounding big data analytics in biomedical research and highlighted the need for a comprehensive approach to address these challenges.