An Approach to Consolidating Pediatric Hospital Beds During the COVID-19 Surge

We used data from the 2018 American Hospital Association Annual Survey Database²  to quantify the availability of pediatric hospital beds within the continental United States. Ratios of beds per 1000 children (<18 years old) were calculated by using population data from the US Census.³  We excluded psychiatric, chemical-dependence, and neonatal beds. We report descriptive statistics and/or numbers for states, hospital service areas, and hospital referral regions.⁴  The distance from the centroid of counties to hospitals was estimated by using the Haversine formula.⁵  All analyses were conducted by using Jupyter notebooks and Python 3.7. Interactive visualizations were produced with Plotly (available at https://bit.ly/ped_beds_covid).

From a universe of 6218 institutions, 3547 general hospitals with functioning emergency departments were identified. Of these, 3524 were located within the continental United States, and 1193 maintained a total of 30 798 pediatric beds. Only 52 institutions were characterized as exclusively pediatric (9904 beds, 32.2% of total US pediatric capacity), but 319 were sufficiently specialized to require PICUs. These were distributed more widely and housed 23 085 beds, or 74.9% of total US pediatric inpatient capacity.

The distribution of all pediatric beds and of those in hospitals with PICUs is shown in Fig 1. The statewide fraction of pediatric beds contained in hospitals with PICUs ranged from 100% in Washington, District of Columbia, to 0% in Wyoming (median: 72.1%; interquartile range [IQR]: 61.0%–82.4%). The median distances from county centroids to the nearest pediatric bed increased from 24.4 miles (IQR: 13.7–37.2) under baseline conditions to 52.3 miles (IQR: 31.9–77.7 miles) with diversion to PICU hospitals.

Populational distance estimates based on county census data are presented in Fig 2.

Notably, 93.2% of hospital service areas and 39.8% of hospital referral regions did not contain a PICU. In the few regions with multiple PICUs (eg, Baltimore, Boston, Chicago, El Paso, Fort Lauderdale, Houston, Indianapolis, and Los Angeles), additional consolidation could be undertaken with little impact on travel distance.

Decades of pediatric hospital care consolidation has left a system that is highly dependent on a subset of specialized centers.^6,7  In a pandemic that preferentially spares children,⁸  this provides an opportunity for those centers to assume the entire burden of pediatric inpatient care. Here, we show that roughly three-quarters of all pediatric beds are currently maintained in centers with PICUs, so the remaining capacity could theoretically be accommodated if a corresponding volume of nonemergent caseload is deferred. Because, from previous estimates, it is suggested that at least 1 in 5 admissions to children’s hospitals can be considered elective and that admissions to general hospitals tend to have shorter lengths of stay,⁹  such accommodation may be achievable. However, as acknowledged by the Children’s Hospital Association,¹  the degree to which this can be practically implemented will depend on many factors, including local practice characteristics, operating guidelines, transport systems, and the surge in patients with coronavirus disease 2019. Additionally, in our analysis we suggest that the immediate cost of such a strategy is, on average, a doubling of travel distances for families seeking urgent or emergent pediatric care. Additional costs from delayed care and financial dislocations will depend on the duration of the disruption and the extent to which elective pediatric medical and surgical care can be deferred. General pediatricians must be prepared to monitor and manage patients whose care has been delayed and to continue this through an extended period of backlog management. When the pandemic has passed, hospital administrators must be encouraged to reopen community pediatric services amid the related financial pressures.¹⁰ 

IQR

interquartile range