Design: cross-sectional appraisal of hospital guidelines using AGREE II, with concordance against Surviving Sepsis Campaign 2020
Population: 14 hospital-specific paediatric sepsis guidelines from 13 countries across five continents
Primary outcome: AGREE II quality rating, consistency across 12 treatment recommendations, concordance with the reference standard
Effect: all 14 rated low quality; concordance 42–100%; consistency 93% for time to antibiotics, 43% for steroid indications

The SENTINEL International study collected the hospital-specific paediatric sepsis guidelines actually in use at its participating sites — 14 documents from 13 countries across five continents — and appraised them with the AGREE II instrument, then checked each against the Surviving Sepsis Campaign 2020 recommendations.

All 14 were rated low methodological quality overall. The weakest domains were rigour of development, scoring between 0 and 50%, and editorial independence, 0 to 58%. Concordance with the reference standard nevertheless ranged from 42% to 100%, and consistency across the 12 key treatment recommendations was highest for initial laboratory investigations, time to antibiotics and fluid bolus volume, all at 93%. It was lowest for steroid indications, at 43%, with fluid resuscitation composition also varying.

The honest reading is reassuring about the first hour and uncomfortable about everything after it. The steps that determine survival in the first sixty minutes are broadly the same wherever you work. What differs is what happens next — which fluid, and when to give steroids — and those are precisely the decisions a junior doctor makes by following the local document. If your unit's guideline has no dated development record and no declared conflicts, that is the AGREE II finding in concrete form, and it is fixable without rewriting the clinical content.

Check when your unit's sepsis guideline was last reviewed and by whom
Look specifically at what it says about fluid composition and steroid timing — that is where guidelines diverged
Do not assume a local document reflects current international recommendations because it is in use
Keep the time-critical steps unchanged: investigations, antibiotics and bolus volume were consistent everywhere
Record the development process and conflicts of interest at the next revision — those were the lowest-scoring domains

Why it matters

The document a trainee follows at three in the morning was almost certainly written without the rigour anyone assumes it had.

The statistics, in plain English

AGREE II scores a guideline's process, not whether its advice is correct — which is why every document here could score low and still concord fully with the reference standard. Read the concordance range of 42% to 100% as the spread across hospitals rather than an average anyone experiences: some units are working from a document that disagrees with the international standard on more than half of the twelve recommendations. The 43% consistency on steroids reflects genuine uncertainty in the evidence, not simply sloppy writing.