'Non-therapeutic research' in children, i.e. research that will not directly benefit the patients involved, is indispensable for improving the quality of paediatric care. However, as children cannot give consent, such research is strictly limited. According to the Dutch Medical Research Involving Human Subjects Act (WMO) non-therapeutic studies with children are only allowed if the risks are negligible and the burdens are minimal. A Dutch expert committee ('committee Doek') investigating whether the WMO should be made less rigid, recently recommended discouraging the use of the distinction between 'therapeutic' and 'non-therapeutic' research and maintaining the absolute limit regarding risks and burdens only for observational studies in children younger than 12 years. This paper discusses these recommendations and argues that in principle, the risks and burdens of all research that has no direct benefit for the subjects should be kept to a minimum.