ICU and You · Six Impossible Things? · Number 1
Things we call impossible that are already true somewhere — and the sixth, which is yours
“There's no use trying,” she said: “one can't believe impossible things.”
“I daresay you haven't had much practice,” said the Queen. “When I was your age, I always did it for half-an-hour a day. Why, sometimes I've believed as many as six impossible things before breakfast.”
Lewis Carroll, Through the Looking-Glass, 1871 — the White Queen, not the Hatter
This series is for the dreamers. Not idle dreaming — the useful kind. Each edition sets out five things that are commonly described as impossible, and which are nothing of the sort.
That is why the word wears quotation marks and the title ends in a question mark. Every one of the five below is already true, right now, for some children, in some places. None of them waits on a discovery. What stands between them and every child is distribution, money, attention and will — which are human arrangements, and human arrangements can be changed by humans.
Intensive care has a particular place in this. We are where these failures become visible; several of them are made or unmade in our own units; and we are unusually well placed to say so, because nobody can accuse us of not having seen it.
The White Queen managed six. Here are five. The sixth is the whole point, and it belongs to you.
About 25 million children a year develop sepsis. Around 3 million die. Eighty-five per cent of both happen in low- and middle-income countries.
The argument that this is a distribution problem rather than a medical one used to be an assertion. Since 2024 it has been a measurement.
When the international task force derived the Phoenix sepsis criteria, they applied the same diagnostic definition to more than three million encounters across sites in the United States, Bangladesh, China, Colombia and Kenya. Mortality among children meeting the criteria:
7.1 per cent in higher-resource settings.
28.5 per cent in lower-resource settings.
Four times the death rate, with the definition held constant. That difference is not biology. It is oxygen, antibiotics within the hour, a functioning laboratory, someone trained to recognise it, and a bed to put the child in.
Sepsis mortality has already fallen substantially worldwide, which is the encouraging half. The World Health Assembly adopted a resolution on sepsis in 2017 committing member states to prevention, recognition and management. The knowledge is not the missing piece.
Recognition systems are ours. We write the protocols, teach the courses, run the simulations and staff the retrieval phones. And when a protocol developed in a well-resourced unit is exported, we are the people who should be asking whether the system that makes it safe travelled with it — see the second impossible thing, where that question turns out to matter enormously.
Roughly 374,000 children under five died of diarrhoea in 2024. Severe wasting affects about 13.6 million children, and only about a third of them get treated.
In 1978 The Lancet ran an unsigned editorial on oral rehydration solution titled Water with sugar and salt, and called it “potentially the most important medical advance this century”. It was not an overstatement.
Bangladesh proved the point at scale and with almost no technology. Between 1980 and 1990, BRAC sent workers door to door and taught oral rehydration to more than twelve million households. By the 2010s ORS was in use in something like eighty per cent of Bangladeshi homes, and diarrhoea had fallen to around two per cent of child deaths — a reduction of roughly ninety per cent over three decades.
So here is the uncomfortable arithmetic. A course of oral rehydration salts with zinc costs well under a dollar. It is not patented, not scarce, and not difficult. Global coverage of ORS with zinc for children with diarrhoea is about twenty-two per cent.
That is not a medical failure. There is nothing left to discover.
The FEAST trial randomised more than three thousand African children with severe febrile illness and impaired perfusion to a fluid bolus or no bolus. The children who got the bolus died more often — 10.5 to 10.6 per cent against 7.3 per cent at 48 hours.
Be careful what you take from that. Children with severe acute malnutrition and children with gastroenteritis were excluded. There was essentially no mechanical ventilation or inotrope support in that setting. And the secondary analysis found the excess deaths were cardiovascular — arrhythmia and refractory hypotension — not the pulmonary oedema everybody had assumed.
The lesson is not that fluid is dangerous. It is that a protocol exported without the system that makes it safe can kill children. Which is precisely why "just give them our guideline" is not the answer to the first impossible thing either.
We see these children at the end — the ones who arrived too late, and the ones in whom refeeding is now its own hazard. We are also, as a specialty, in the habit of assuming our protocols generalise. FEAST is the standing correction to that habit, and we should be the ones teaching it.
In Australian and New Zealand paediatric intensive care, about 97 to 98 per cent of children now survive. The question has quietly changed from whether they live to what they live with.
This is the one we own outright. Nobody else makes it, and nobody else can fix it.
Paediatric post-intensive care syndrome spans physical, cognitive, emotional and social health, and the paediatric framing adds something the adult version cannot: the outcome is measured against an expected developmental trajectory, and the family is part of the unit being assessed. A child does not return to a previous baseline. A child is supposed to be moving.
Two numbers make it concrete. Delirium is present in about a quarter of children in paediatric intensive care at any given moment, rising to something near forty per cent among those still there at day six — and its contributors include the restraints, the benzodiazepines, the immobility and the broken sleep that we supply.
And then there is the number that belongs to this fortnight in particular. A Queensland study linked more than five thousand children who survived intensive care before the age of five to their Year 3 NAPLAN results:
14.0 per cent failed to meet the national minimum standard in both numeracy and reading, against 9.0 per cent of matched controls — scoring roughly thirteen points below those controls across domains.
The child we discharged as a survivor turns up in Year 3 as a child who is behind. That is our outcome, appearing in someone else's dataset, years later, where we never look.
This is the weakest evidence base of the five, and pretending otherwise would be the wrong kind of dreaming. The big sedation-protocol trial, RESTORE, was null on its primary outcome. Early mobilisation is feasible and safe with associations rather than demonstrated long-term benefit. Follow-up clinics are widely recommended and essentially unevaluated in children. The often-quoted figure linking care-bundle adherence to lower mortality is an observational association, not a causal one.
So this impossible thing is different in kind from the others. For sepsis and for thirst we know what to do and fail to deliver it. Here we do not yet know what to do — which makes it the one most open to a reader who wants to go and find out.
Measles deaths have fallen 88 per cent since 2000, to about 95,000 a year. Measles cases are now rising.
Fifty years of the Expanded Programme on Immunization is estimated to have saved 154 million lives, the great majority of them infants, and to account for about forty per cent of the global decline in infant mortality over that period. Measles vaccine alone accounts for something like 94 million deaths averted.
Which makes what is happening now the clearest possible demonstration that these things are not ratchets. Around 14 million children received no vaccine at all in a recent year, more than before the pandemic. Global measles coverage sits well below the level needed to stop transmission, and the number of countries with large outbreaks has roughly tripled since 2021.
Nor is this only a problem elsewhere. Australian childhood immunisation coverage is at a five-year low. Coverage at twelve months has fallen more than four percentage points since 2020, coverage at twenty-four months is below ninety per cent, and both are falling faster among Aboriginal and Torres Strait Islander children.
We are the witnesses. In the Texas measles outbreak of 2025, among hospitalised patients with records available, every single one was unvaccinated or of unknown vaccination status; seven per cent needed intensive care.
And malignant pertussis is the starkest teaching case in paediatrics: young infants with hyperleukocytosis, pulmonary hypertension and refractory cardiovascular failure, where one reported cohort lost nine of twenty-three. Children on extracorporeal support, for a disease prevented by an injection given at six weeks of age.
Few people can describe that with more authority than the person who ran the circuit.
In Australia, invasive infection requiring intensive care occurs in Aboriginal and Torres Strait Islander children at about three times the rate of other Australian children, with a mortality rate ratio of 2.65.
Geography should not be a prognostic variable. It is.
Australian mortality from potentially avoidable causes rises steadily with remoteness, reaching more than twice the metropolitan rate in remote areas and higher still in very remote ones. Child mortality at ages nought to four is about twice as high for Aboriginal and Torres Strait Islander children as for other Australian children — and has not changed significantly across the last decade for which figures exist. The 2008 target to halve that gap was not met.
For critical illness specifically, an Australian study across our own intensive care units found invasive infection requiring intensive care at 47.6 per 100,000 Aboriginal and Torres Strait Islander children per year against 15.9 for other children, with invasive Staphylococcus aureus at nearly eight times the rate.
What is genuinely encouraging is that the structural fixes have evidence behind them. Larger paediatric intensive care units have better risk-adjusted survival — for every additional hundred annual admissions the odds of death fall measurably. Specialist paediatric retrieval teams reduce adverse events in transfer. And telemedicine support to non-specialist emergency departments has been shown to cut medication errors in acutely ill children from about one in eight consultations to about one in thirty.
No paediatric tele-critical care study has demonstrated a mortality benefit; what is shown is safer, better care. And the finding that retrieved children do worse than children who present directly is heavily confounded — retrieved children are sicker on arrival, and the studies cannot separate severity from transfer. It would be easy and wrong to say that transfer harms children.
This one is almost entirely ours to build. Retrieval services, the advice line, the conference call, the tele-support to a hospital with no paediatric unit, the training that lets a regional team hold a child safely for six hours — all of that is intensive care reaching outward. The technology already exists. It is a question of who picks up the phone at three in the morning, and whether the system has decided that is somebody's job.
The White Queen managed six before breakfast. I have given you five.
The sixth is yours — and I would genuinely like to know what it is.
What is your sixth impossible thing?
Not a wish. Something you think is described as impossible, or unaffordable, or unrealistic, or simply never mentioned — and which you believe can and should be made real. It might be global or it might be one ward. It might be something about children, or about the people who look after them.
Tell me what it is and why you think it is achievable, and I will publish the best of them — with your name on it or without, as you prefer. If enough of you write in, a future edition of this series will be entirely yours.
The button opens a reply straight to me. If your device blocks it, write to icuandyou@icloud.com instead.