Skip to main content
. 2021 Nov 18;326(22):2326–2328. doi: 10.1001/jama.2021.18377

Table 2. Frequency and Duration of Mental Health Boarding by Hospital Characteristics.

Hospital characteristics Frequency of boarding Duration of boarding
No. of youths boarding daily, median (IQR) P valuea Adjusted relative risk (95% CI)b,c Duration of boarding, median (IQR), h P valuea Adjusted linear coefficient (95% CI)b,d
Hospital type
Freestanding children’s hospital (n = 38) 6 (4-9) .004 1 [Reference] 28 (24-48) .003 [Reference]
Children’s hospital within larger institution (n = 35) 4 (2-7) 0.50 (0.41-0.61) 72 (48-96) 23.76 (4.40 to 43.12)
General community hospital (n = 12) 3 (2-4) 0.26 (0.18-0.38) 42 (24-72) −3.50 (−30.61 to 23.61)
Licensed pediatric psychiatric beds
Yes (n = 31) 6 (4-8) .01 1.68 (1.40-2.02) 38 (24-75) .41 8.50 (−9.87 to 26.87)
No (n = 54) 4 (2-6) 1 [Reference] 48 (24-72) [Reference]
Hospital location
Urban (n = 55) 4 (3-8) .29 1 [Reference] 48 (24-81) .52 [Reference]
Suburban/rural (n = 30) 4 (3-6) 0.96 (0.77-1.20) 48 (24-72) −0.22 (−18.63 to 18.19)
Hospital geographic region
Midwest (n = 21) 4 (3-7) .34 1 [Reference] 26 (21-54) .001 [Reference]
Northeast (n = 24) 5 (4-12) 3.22 (2.44-4.25) 72 (60-96) 40.09 (14.91 to 66.27)
West (n = 19) 4 (3-5) 1.08 (0.80-1.46) 44 (26-66) 14.64 (−10.44 to 39.73)
South (n = 21) 6 (2-6) 1.34 (1.02-1.78) 48 (24-48) 10.55 (−13.56 to 34.66)
a

Differences in distributions evaluated using Kruskal-Wallis equality-of-population rank tests.

b

Models adjusted for hospital type, presence of licensed psychiatric beds, hospital location, and geographic region.

c

Relative risks calculated using Poisson regression.

d

Linear coefficients calculated using general linear regression.