Machine learning approaches can improve the quality and efficiency of clinical decision-making processes. At the same time, experiences with big data and machine learning in other areas of society show that there is reason to worry about the neutrality and fairness of ML-based systems. Concern arises particularly in respect of the impact of ML-based systems on vulnerable minority groups, especially groups that have suffered from inequality and discrimination in the past. From a legal perspective, the attention to algorithmic discrimination has been particularly strong in the contexts of the criminal justice system, employment, and online, personalized advertisement. To protect people from unfair treatment, a fundamental right to ‘non-discrimination’ is recognized in international human rights conventions, in the European Convention of Human Rights, in EU law, and in the national laws and constitutions of many countries. As is typical for laws implementing fundamental rights, non-discrimination laws are broadly formulated and require contextual interpretation before they can be operationalized in a certain decision context. The PhD project explores the definition of discrimination in EU law in the specific context of clinical decision-making, and then asks how EU law can facilitate effective control with algorithmic discrimination in ML-based decision support systems. The latter question is viewed in light of prior scholarly critique of the (lack of) effectiveness in enforcement mechanisms available under EU non-discrimination law.