Van Dijk and colleagues present three cases to illustrate and discuss the relationship between medicalisation and overdiagnosis. In this commentary, I consider each of the case studies in turn, and in doing so emphasise two main points. The first is that it is not possible to assess whether overdiagnosis is occurring based solely on incidence rates: it is necessary also to have data about the benefits and harms that are produced by diagnosis. The second is that much is at stake in discussions of overdiagnosis in particular, and that it is critical that work in this area is conceptually rigorous, well-reasoned, and empirically sound. van Dijk and colleagues remind us that overdiagnosis and medicalisation are not just matters for individual patients and their clinicians: they also concern health systems, and society and citizens more broadly.