Department of Global Health and Social Medicine, King's College London, Bush House, North East Wing, The Strand, London, WC2R 2LS, England.
Department of Educational Research, Lancaster University, Lancaster, England.
Bull World Health Organ. 2020 Apr 1;98(4):239-244. doi: 10.2471/BLT.19.237370. Epub 2019 Jan 17.
There is growing interest in population health research, which uses methods based on artificial intelligence. Such research draws on a range of clinical and non-clinical data to make predictions about health risks, such as identifying epidemics and monitoring disease spread. Much of this research uses data from social media in the public domain or anonymous secondary health data and is therefore exempt from ethics committee scrutiny. While the ethical use and regulation of digital-based research has been discussed, little attention has been given to the ethics governance of such research in higher education institutions in the field of population health. Such governance is essential to how scholars make ethical decisions and provides assurance to the public that researchers are acting ethically. We propose a process of ethics governance for population health research in higher education institutions. The approach takes the form of review after the research has been completed, with particular focus on the role artificial intelligence algorithms play in augmenting decision-making. The first layer of review could be national, open-science repositories for open-source algorithms and affiliated data or information which are developed during research. The second layer would be a sector-specific validation of the research processes and algorithms by a committee of academics and stakeholders with a wide range of expertise across disciplines. The committee could be created as an off-shoot of an already functioning national oversight body or health technology assessment organization. We use case studies of good practice to explore how this process might operate.
人们对基于人工智能方法的人群健康研究越来越感兴趣。此类研究利用各种临床和非临床数据来预测健康风险，例如识别传染病并监测疾病传播。此类研究大多使用公共领域的社交媒体数据或匿名二级健康数据，因此无需伦理委员会审查。虽然已经讨论了数字研究的伦理使用和监管问题，但对于人群健康领域的高等教育机构中此类研究的伦理治理问题却很少关注。此类治理对于学者做出伦理决策至关重要，并向公众保证研究人员的行为符合伦理规范。我们提出了高等教育机构人群健康研究的伦理治理流程。该方法采用研究完成后的审查形式，特别关注人工智能算法在增强决策方面的作用。第一层审查可以是国家层面的开源算法和相关数据或信息的开放科学存储库，这些数据或信息是在研究过程中开发的。第二层审查将由具有广泛跨学科专业知识的学者和利益相关者委员会对研究过程和算法进行特定于部门的验证。该委员会可以作为已经运作的国家监督机构或卫生技术评估组织的分支机构设立。我们使用案例研究来探讨该流程的运作方式。
