Table 2:

Primary outcomes

Outcome	Had at least one during study				Within Group Change		Between Group Difference
	At least one during study, N (%)	Risk difference (95% CI)	Risk ratio (95% CI)	P	Estimate (95% CI), ORa	P	Estimate (95% CI), RORb	P
Severe asthma exacerbations
Control	5 (25)	0.05 (−0.21, 0.31)	1.25 (0.39, 3.99)	1.00	0.61 (0.23, 1.61)	0.32
Intervention	4 (20)				0.31 (0.03, 2.76)	0.29	0.51 (0.05, 5.68)	0.58
Moderate asthma exacerbations
Control	13 (65)	0.05 (−0.25, 0.35)	1.08 (0.67, 1.75)	0.74	0.50 (0.27, 0.91)	0.02
Intervention	12 (60)				0.23 (0.06, 0.86)	0.03	0.47 (0.11, 1.95)	0.30

Outcome	Mean (SE)					Within Group Change		Between Group Difference
	Exit	Δ Exit-Rand	P	Unadjusted β (SE)	P	Estimate (95% CI)	P	Estimate (95% CI)	P
Asthma Control Test (ACT)
Control (n=13)	21.1 (3.8)	0.15 (2.5)	0.04			0.17 (-1.17, 1.51)	0.80
Intervention (n=13)	23.2 (1.6)	2.00 (2.2)		1.85 (0.93)	0.06	2.02 (0.88, 3.15)	0.001	1.85 (0.09, 3.61)	0.04
Childhood Asthma Control Test (CACT)
Control (n=7)	22.9 (4.3)	0.29 (3.5)	0.94			−0.31 (−1.87, 1.25)	0.70
Intervention (n=7c)	25.2 (1.9)	1.67 (4.4)		1.38 (2.18)	0.54	3.96 (-1.27, 9.20)	0.14	4.27 (-1.19, 9.73)	0.13
Pediatric Asthma Quality of Life Questionnaire (PAQLQ)
Control	6.20 (0.9)	0.25 (0.7)	0.48			0.25 (−0.06, 0.55)	0.11
Intervention	6.65 (0.4)	0.42 (0.6)		0.18 (0.21)	0.41	0.54 (0.20, 0.88)	0.002	0.30 (−0.16, 0.75)	0.20

Notes: For asthma exacerbations and asthma control outcomes which were obtained at multiple visits, n=39 for exit and Δ Exit-Rand analyses and n=40 for analyses within and between groups. For PAQLQ which was obtained at 2 visits, all analyses n=39. P value of risk difference and risk ratio obtained from Fisher Exact Test or X², where appropriate. P value for Δ Exit-Rand obtained from two-sided Wilcoxon rank sum. Models were not adjusted for other variables. Within group change obtained from generalized linear models adjusting for correlation between values including interaction term (between group difference). For asthma exacerbations (both severe and moderate), estimates obtained are from visit 2 ˗7 (no randomization values) and are adjusted for visit 2 values. Boldface indicates statistical significance (P<0.05).

^aWithin-group changes for binary outcomes are quantified by odds ratios (OR); within a given group, OR=λ indicates that the odds of outcome at study exit are λ times the odds of outcome at baseline.

^bBetween-group differences are quantified by ratios of odds ratios (ROR); a ROR=γ indicates that the OR for intervention participants is γ times the OR for controls.

^cN=6 for exit and △ Exit-Rand, n=7 for analyses within and between groups.