Where a child lives predicted intensive care admission for respiratory illness

Design

retrospective national cohort study linking intensive care admissions to area-level exposures

Population

41,859 respiratory admissions in 30,442 children under 16 to English paediatric intensive care units, 2014–2023

Primary outcome

incidence of intensive care admission for respiratory illness, and in-unit mortality

Effect

most deprived areas adjusted IRR 1.7 (95% CI 1.4–2.2); highest smoke exposure adjusted IRR 1.3 (1.1–1.5)

This retrospective cohort covered every paediatric intensive care admission for respiratory illness in England over a decade: 41,859 admissions in 30,442 children under 16, between January 2014 and December 2023, linked to area-level deprivation, ethnicity and neighbourhood smoking prevalence.

Admission incidence was higher in children of Asian ethnicity (incidence rate ratio 1.5, 95% CI 1.3 to 1.7) and Black ethnicity (1.7, 1.4 to 1.9) than in White children, and higher in the most deprived areas than the least (adjusted IRR 1.7, 1.4 to 2.2). Children living where neighbourhood cigarette smoke exposure was highest had an adjusted incidence rate ratio of 1.3 (1.1 to 1.5). The disparity extended to outcome: in-unit mortality was higher for children of Asian ethnicity (odds ratio 2.1, 1.8 to 2.5) and for those from the most deprived areas (adjusted OR 1.6, 1.2 to 2.0).

The smoking finding is area-level, not household-level, which makes it weaker as a causal claim and more interesting as a policy one — it is the ambient exposure of a neighbourhood rather than a parent's habit. For clinicians the usable part is the outcome disparity. A child from a deprived area arriving in intensive care is more likely to die there, and that is not explained by the admission threshold. It argues for lower thresholds for review and earlier escalation in exactly the families who present later.

• Ask about household smoking at every respiratory presentation and offer cessation referral to the adult
• Lower your threshold for review and escalation in children from deprived areas — their outcomes were worse, not just their admission rates
• Record ethnicity accurately; these disparities are invisible without it
• Do not read the area-level smoking result as a statement about a particular household
• Make safety-netting advice concrete and written, not verbal

Why it matters

The mortality difference persisted after admission, which means it is not explained by who gets referred.

Don't overread it

Neighbourhood smoking is an area-level measure — it cannot establish that second-hand smoke in a particular home caused a particular admission.

The statistics, in plain English

Incidence rate ratios compare how often admissions happen per child-year between groups, so 1.7 for the most deprived areas means roughly 70% more admissions, not 70% more risk for an individual child. The mortality odds ratio of 2.1 for Asian ethnicity is unadjusted, while the deprivation figure of 1.6 is adjusted — they are not directly comparable, and the unadjusted one will partly reflect deprivation itself. Ethnicity here is a marker for a set of social and environmental exposures, not a biological variable.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.