Hospitals are great institutions: they treat people who are sick and send them on their way. One of the arts of medicine is to know when they are ready to go.

Not in bronchiolitis of infancy. We have a number to tell us: usually considered 93 or 94. The oxygen saturation (SpO₂) at which normoxia has been regained and if the infant is feeding satisfactorily (which they usually are at that stage) then they are homeward bound. Safe.

The problem is that oxygen saturation levels are frequently slow to recover in bronchiolitis, with supplemental oxygen treatment correspondingly prolonged.1 Parents balance the concern that their child needs oxygen with a frustration that their bouncing baby is attached by tubing to a wall in a ward. Hospitals are clogged each winter by the ubiquitous peak of infants with slowly dwindling oxygen requirements while recovering from bronchiolitis. How much more sensible paediatric service organisation could be, without this peak.

So, providing medical care at home to infants recovering from bronchiolitis, when that care is solely the provision of supplemental oxygen, seems sensible.

Two recent studies have begun to explore this, providing supplemental oxygen at home to infants with bronchiolitis who have presented to hospital. In both studies infants considered high risk (prematurity, apnoea, pre-existing respiratory/cardiac disease) were excluded.

In the first study, infants with bronchiolitis were randomly assigned home oxygen from the emergency department (ED) of a US tertiary hospital.2 Infants, aged 2–24 months, with clinical bronchiolitis and an arrival SpO₂ of 87% or less were assessed for stability following 8 h of ED observation. Those considered stable (adequate feeding, no significant respiratory distress) and still requiring oxygen (SpO₂ ⩾90% with ⩽1 l/minute oxygen) were discharged …