Many guidelines for prevention and treatment still locate persons in risk classes (e.g. low, moderate, high) on the basis of thresholds placed on a continuous metric for a single criterion (e.g. risk of developing x). These 'traffic light' signals can lead to inferior decisions through their mono-criterial focus and lack of preference-sensitivity to the multiple criteria relevant to the person. It is arguably unethical to communicate to someone that they are at low, moderate, or high risk of x solely on the basis of the unpublished and often unknown preferences of the group that has set the classification thresholds. Any prior classification and labelling will interfere with the individual's balanced processing of information on the performance of all treatment options on their multiple relevant criteria - including treatment side effects and burdens as well as main benefit - and jeopardise meeting the requirements for fully informed and preference-based consent to any subsequent action. Personalised decision support tools based on Multi-Criteria Decision Analysis can help fulfil these objectives, with apomediative (at home) e-decision support especially appealing because of its empowering and resource-saving potential. The individual's absolute risk score is required in these tools since any threshold-based risk classification will interfere with the coherence of the analysis across the multiple criteria.