Acute severe lower respiratory tract infections (LRTIs) continue to be a major cause of mortality (13%) and morbidity globally [1,2] and in South Africa (SA). [1] Reid et al. [3] showed that pneumonia was the main cause of death (25%) in children aged <5 years in the Metro West health district of the Western Cape Province of SA. [3] In many middle-to low-income countries with resource limitations, the number of child deaths resulting from severe pneumonia (related to hypoxaemia and respiratory failure) remains high despite implementation of international and local guidelines. [4] Additional respiratory support, important in the care of critically ill children, is often unavailable or is perceived as being neither feasible nor safe in resource-limited settings. [4,5] The use of non-invasive ventilation for respiratory support, including nasal continuous positive airway pressure (CPAP) and high-flow nasal cannula (HFNC) therapy, in and outside the paediatric intensive care unit (PICU) is increasing. [5,6] There is limited evidence to support the use of HFNC in children outside the PICU setting and in conditions other than bronchiolitis. [7] The probable mechanisms of action of HFNC include heating and humidification of inspired air, nasopharyngeal carbon dioxide wash-out, reduction in upper airway resistance, and provision of positive distending pressure. [8-12] The current paucity of data on the variable positive distending pressure provided by HFNC raises clinical safety concerns, [7] and many modern HFNC devices therefore now have pressure-limiting valves. However, more robust physiological studies are needed. [7] A pioneering, contextually relevant study from Bangladesh showed that outcomes in terms of 'treatment failure' (defined as two or three of the following characteristics: severe hypoxia, signs of severe clinical distress, or respiratory acidosis on blood gas analysis) were no worse in patients with severe pneumonia who received HFNC than in those who received bubble CPAP, but mortality rates were higher in the low-flow nasal cannula (LFNC) and HFNC groups, indicating a potential risk of harm associated with these modalities. [13] Various observational studies have shown that HFNC reduces respiratory rates and the effort of breathing in infants with bronchiolitis, has very few reported complications, and is well tolerated. [7] Commercial HFNC devices and circuits are expensive. Guaranteeing that a clinically effective and cost-effective form of This open-access article is distributed under Creative Commons licence CC-BY-NC 4.0.