I'm Mary Hartman, and we have no conflicts of interest to disclose. To start with some background on this topic, we know that caring for children in the ICU is extremely expensive and resource consuming. But we actually have no national estimates of how many children receive ICU care in this country or for what types of problems. All of our existing estimates of diagnosis prevalence, resource utilization, costs, and outcomes are either outdated or use data sets that are limited to certain geographic regions, types of insurance, health care systems, or types of hospitals. And as the COVID pandemic has made clear, it's really essential to have an accurate national assessment of our current care needs in order to have a good emergency planning, as well as an understanding of trends over time to be able to anticipate our future care and capacity needs. And multiple studies, including many of the ones that we just heard about, have shown that there is higher ICU use and illness severity in the US among multiple different types of patient groups, including African-American and Latino children, children living in poverty and those receiving government insurance, and children with complex medical conditions. We know that in the US, there's a rising prevalence of all of these groups. And so we can hypothesize that with these changing characteristics of the population, that may be associated with increased ICU resource utilization and illness severity, which could all be reflected in changes in our capacity needs, costs, and patient outcomes, which has implications for future health care planning, resource allocation, and care delivery, both before and after hospitalization. We have some evidence that this hypothesis is true. A recently published evaluation of PICU admission trends among many of the large tertiary care children's hospitals in the country found that PICU admissions are rising and that children have increasing rates of medical complexity and resource use. But as we've also mentioned in the previous presentations, standalone children's hospitals are not the only setting in which children receive pediatric ICU care in this country. And so we aimed to conduct a population-based assessment of how ICU admission patterns, use of critical care services, and the characteristics and outcomes of critically ill children are changing over time. So to evaluate this, we conducted a population-based retrospective cohort study using the Health Care Cost and Utilization Project state inpatient databases. These databases contain all inpatient care records from 99% of all discharges within the state and have no exclusions by insurance type or geography. We included states with revenue codes identifying ICU admissions from five years between 2001 and 2019. We ended up with records from 21 states that together represent over a third of the US under-18 population, with most states contributing to at least three or four of the five sample years. We included all children under 18 admitted to a non-neonatal ICU and collected information on patient demographics, hospitalization and outcome data, and facility characteristics. We used data from the US Census to calculate population incidence of admissions and diagnoses and used generalized linear Poisson regression and Kuzik's to assess for trends over time from 2001 to 2019. We then generated age and sex-adjusted national estimates from the 35% of the population studied. So what did we find? There were over two million pediatric hospital admissions over the five included years, of which 275,000 were admitted to a non-neonatal ICU, representing 12.8% of all pediatric admissions. Between 2001 and 2019, the population rate of pediatric hospital admissions dropped by 35%, from 30.1 in 2001 to 19.7 per 1,000 in 2019. But meanwhile, the population rate of ICU admissions decreased only very slightly, from 3.2 to 3.0 ICU admissions per 1,000 children. And this resulted in a 33% increase in the proportion of children in the hospital who received ICU care, rising from 10.6% in 2001 to 15.5% of pediatric hospital admissions in 2019. This chart demonstrates the types of ICUs that children were admitted to. An increasing proportion of ICU admissions were in dedicated PICUs, with PICUs representing 61% of admissions in 2001 and 81% by 2019, while the proportion admitted to general ICUs decreased from 29% to 11% over that time period. Admissions to cardiac, surgical, and trauma burn ICUs remain somewhat similar. A similar trend was observed with increasing regionalization from community hospitals to children's hospitals, with a 66% increase in admission to children's hospitals from 2001 to 2019. ICU patients, shown in dark blue, were increasingly of black or mixed race and Hispanic ethnicity, and were substantially more so compared to the general US population in included states of pediatric patients in those states. And that's shown in light blue there, though these differences did lessen over time. And the proportion of children with public insurance rose by nearly 50% from 2001 to 2019, with more than half of PICU patients using public insurance in 2019 compared to only a third of the general US under 18 population. The proportion of children admitted with pre-existing comorbid conditions rose over the time period to 57% of ICU admissions by 2019, with an even more pronounced increase in the proportion with pre-existing technology to nearly a quarter of ICU admissions by 2019. The most common admitting diagnoses were respiratory conditions and injuries. The proportion of respiratory admissions increased to over a third of all admissions by 2019, with a rise in the population incidence of respiratory admissions by 57% over that time period. And meanwhile, both the proportion of ICU admissions due to injuries and the population-based incidence of injury admissions fell to approximately 60% of 2001 levels by 2019. The proportion of pediatric ICU admissions with multiple organ dysfunction syndrome more than tripled from 6.8% to 21% over that time period. The most common organ failure in early years was neurologic failure, but the proportion of ICU admissions with respiratory failure rose quickly over that time period and became the predominant organ failure from 2010 onward. And associated with this rise in respiratory failure, the proportion of children receiving mechanical ventilation increased from 14.9 to 17.4% of ICU admissions. But despite this increase in complexity and illness severity of ICU admissions, pediatric ICU mortality is actually the lowest it's ever been, with a decrease from 2.5% to 1.8% of all ICU admissions between 2001 and 2019. And this was especially true for children without comorbidities, for whom ICU mortality had fallen to 0.3% by 2016. But importantly, mortality for children with mortalities also declined substantially over that time period. Most notably, however, mortality among children with multiple organ dysfunction syndrome fell by nearly 60% from 2001 to 2019, from 18.1% down to 7.6%. The greatest improvements in survival were among children with three or more organ failures, with really substantial declines in mortality over the past couple of decades among the most severely ill children. But also importantly, by 2019, ICU admissions, sorry, ICU mortality for admissions with single organ failure had fallen to 0.4%, with 88% of ICU deaths occurring among children with MODS. This improved survival of increasingly medically complex children does come at a cost, though. Median hospital length of stay increased over the time period to approximately 4.5 days, with a rising number of long-stay patients. And even after adjusting for inflation, median hospital costs rose by over 80%. And unsurprisingly, the bulk of those costs were spent caring for the most severely ill children, with median costs for patients with MODS or those who required mechanical ventilation nearly tripled the median cost of all ICU admissions at nearly $40,000 each. And finally, after age and sex adjusting the included sample to the general U.S. population, we can estimate that in 2019, there were 239,000 pediatric ICU admissions, including 51,000 children with MODS and 42,000 children receiving mechanical ventilation, with a total annual cost for all non-neonatal ICU admissions for children in the U.S. of $11.6 billion. There were certainly several limitations to this work, including that not all states were represented, as they don't contribute data to HCUP, or they don't include revenue codes that allow identification of ICU care. The diagnoses, comorbid conditions, and organ failures were determined using ICD codes, and changes in prevalence could be due to the change from ICD-9 to ICD-10 coding. And the designation of different types of ICUs could have, to which children were admitted, could have varied from hospital to hospital. However, we certainly can conclude that the prevalence of children requiring or receiving ICU care in the U.S. is rising quickly. ICU care is increasingly regionalized in children's hospitals and dedicated PICUs. And while studies have demonstrated improved outcomes for children treated in dedicated pediatric facilities, regionalization of care may also be associated with lack of timely access to pediatric care for children living in more rural areas. Pediatric ICU patients are increasingly medically complex, more frequently develop an organ failure, and require mechanical ventilation, and have longer hospital stays. And a higher proportion are relying on public insurance, while the costs to those payers are increasing. And so it's essential that pediatric ICU capacity and the healthcare system as a whole adapt to ensure that we both have adequate resources, as well as timely access to care for these children in the future. Thank you very much. Open to any questions. Thank you.

prevalence

ICU care

access to care

medically complex

pediatric ICU capacity