Altai Z, et al. - Researchers quantitatively examined the probability of self-inflicted humeral fracture in non-ambulant infants secondary to rolling, by using personalised computed tomography (CT)-based finite element models. They used three whole-body post-mortem CT observations of children at the age of rolling. They computed the mechanical moment required by each infant to perform a rolling manoeuvre. By applying this value to the finite element model, they simulated spontaneous rolling from the prone to the supine position. They found substantially lower maximum predicted strains when compared with the elastic limit of the bone. Findings challenged the plausibility of self-inflicted humeral fracture secondary to rolling in non-ambulant infants, and it was suggested that without the assistance of an external force, this mechanism alone is unlikely to cause a humeral fracture.