Most children with asthma have their disease easily controlled if low-dose inhaled corticosteroids (ICSs) are regularly and correctly administered. If a child presents with asthma which is apparently resistant to therapy with high-dose ICS and other controllers, then they have problematic severe asthma. However, in light of the UK National Review of Asthma Deaths, definitions of severe asthma based solely on the levels of prescribed treatment are too narrow. A detailed assessment of all such children should be performed. First, the diagnosis of asthma should be confirmed, then co-morbidities assessed. Next, a nurse-led assessment further characterizes the problem, conventionally categorizing the child as either having difficult asthma or severe therapy-resistant asthma. Here, we reassess in particular the interactions between, and management of, these two categories, highlighting that this dichotomous classification may need reconsideration. We use bronchoscopy and an intramuscular steroid injection to determine if the child has steroid-resistant asthma, using a novel, multidomain approach because the adult definition does not apply to around half the children we see. Finally, we highlight some mechanistic data which have emerged from this protocol such as the absence of T-helper 2 (TH2) cytokines even in eosinophilic severe asthma and the potential role of the innate epithelial cytokine IL-33, novel data on lineage negative innate lymphoid cells, which we can measure in induced sputum, and demonstrating that intraepithelial neutrophils are associated with better, not worse asthma outcomes. Severe paediatric asthma is very different from severe asthma in adults, and approaches must not be uncritically extrapolated from adult disease to children.