Big Data, Machine Learning, and Personalization in Health Systems: Ethical Issues and Emerging Trade-Offs

The use of big data and machine learning has been discussed in an expanding literature, detailing concerns on ethical issues and societal implications. In this paper we focus on big data and machine learning in the context of health systems and with the specific purpose of personalization. Whilst personalization is considered very promising in this context, by focusing on concrete uses of personalized models for glucose monitoring and anomaly detection we identify issues that emerge with personalized models and show that personalization is not necessarily nor always a positive development. We argue that there is a new problem of trade-offs between the expected benefits of personalization and new and exacerbated issues - results that have serious implications for strategies of mitigation and ethical concerns on big data and machine learning.