Half of children admitted with sepsis had been seen in the previous week

Design

retrospective observational cohort using administrative claims data

Population

6,928 paediatric sepsis hospitalisations, ages 0–18, US, 2016–2022; median age 10, 64.3% with complex chronic conditions

Primary outcome

any healthcare encounter in the 7 days before sepsis hospitalisation

Effect

53.5% had a prior encounter (33.7% outpatient, 28.2% ED, 6.2% inpatient); 38.9% of those had an infection-related diagnosis

This retrospective cohort used the Merative MarketScan claims database to identify 6,928 hospitalisations for sepsis in children aged 0 to 18 between 2016 and 2022, requiring 30 days of continuous insurance enrolment beforehand so that prior contacts could be seen. Median age was 10 years and 64.3% had a complex chronic condition.

More than half — 53.5% — had a healthcare encounter in the seven days before admission: 33.7% an outpatient visit, 28.2% an emergency department visit, 6.2% an inpatient stay. Among those with any prior encounter, 38.9% had an infection-related diagnosis documented at it, which works out at roughly one in five of the whole cohort being assessed for infection in the week before being admitted with sepsis.

The finding that stops an easy conclusion is that mechanical ventilation, ICU use and length of stay did not differ between children who had been seen beforehand and those who had not. If earlier contact were systematically identifying sepsis that was then missed, you would expect the pre-seen group to arrive later and sicker. They did not. So this is better read as a description of where the opportunity sits than as evidence that these children were failed: half of paediatric sepsis passes through a clinician's hands in the week before admission, and most of those visits do not look like sepsis at the time.

• Ask about healthcare contacts in the last week when a child presents unwell — a recent visit is a risk marker, not reassurance
• Give explicit return advice at every febrile child discharge, with named deterioration signs
• Flag children with complex chronic conditions, who made up 64.3% of this cohort
• Record the visit that prompted a return, so patterns are visible to the next clinician

Why it matters

The week before admission is a window that exists in half of these children and is not currently used as one.

Don't overread it

This describes where children were seen, not that their sepsis was missed — outcomes were no worse in those with prior contact.

The statistics, in plain English

The absence of difference in ventilation, ICU use and length of stay is the informative negative here — it argues against the interpretation that earlier contact meant missed disease. Claims data record what was billed, not what was clinically apparent, so an 'infection-related diagnosis' may have been a minor upper respiratory code. Employer-insured and Medicaid populations in the US do not map onto Indian presentation patterns, where access and timing differ substantially.

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.