. 2023 Dec 5;330(21):2084–2095. doi: 10.1001/jama.2023.22114

Table 2. Outcome Measures.

	Mean (SD)							Effect size: difference in 12-mo changes (95% CI)^a (n = 458)	P value^a
	Normative values	Early adenotonsillectomy (n = 231)			Watchful waiting (n = 227)
	Normative values	Baseline	12 mo	Change from baseline to 12 mo	Baseline	12 mo	Change from baseline to 12 mo
Coprimary outcomes
Caregiver BRIEF GEC T score^b	50 (10)	55.3 (12.2) [n = 230]	52.1 (11.3) [n = 196]	−3.1 (9.4) [n = 195]	56.0 (12.5)	53.7 (11.2) [n = 196]	−1.9 (8.6) [n = 196]	−0.96 (−2.66 to 0.74)	.27
Go/No-go sustained attention^c^,^d	NA	2.0 (1.1) [n = 229]	2.2 (1.1) [n = 184]	0.2 (1.2) [n = 182]	2.1 (1.0) [n = 222]	2.3 (1.0) [n = 184]	0.1 (1.2) [n = 182]	0.05 (−0.18 to 0.27) [n = 455]	.68
Secondary outcomes								Effect size: difference in 12-mo changes (FDR-adjusted CI)^d	FDR-adjusted P value^d
Pegboard dexterity (average)^e	NA	32.5 (14.6) [n = 227]	27.4 (9.1) [n = 187]	−5.3 (7.4) [n = 183]	32.8 (11.9)	26.7 (6.4) [n = 187]	−5.9 (8.0) [n = 187]	0.76 (−0.92 to 2.43)	.37
Caregiver-reported CBCL^f	50 (10)
Total problems T score		53.0 (11.0) [n = 227]	48.4 (10.8) [n = 186]	−4.5 (9.0) [n = 183]	53.3 (11.3) [n = 224]	51.6 (10.9) [n = 182]	−1.4 (7.5) [n = 182]	−3.09 (−4.90 to −1.28) [n = 454]	<.001
Externalizing problems T score		51.1 (10.8) [n = 227]	48.1 (10.3) [n = 186]	−3.2 (8.5) [n = 183]	51.2 (11.7) [n = 224]	49.6 (11.2) [n = 182]	−1.6 (8.2) [n = 182]	−1.54 (−3.34 to 0.26) [n = 454]	.09
Internalizing problems T score		51.8 (11.2) [n = 227]	47.8 (10.9) [n = 186]	−3.8 (10.0) [n = 183]	52.1 (11.3) [n = 224]	51.1 (11.0) [n = 182]	−0.7 (8.3) [n = 182]	−3.05 (−5.07 to −1.04) [n = 454]	.003
Attentional problems T score		57.5 (8.2) [n = 227]	55.2 (6.6) [n = 186]	−2.3 (7.1) [n = 183]	57.3 (7.8) [n = 224]	56.0 (6.7) [n = 182]	−1.1 (5.8) [n = 182]	−1.19 (−2.55 to 0.17) [n = 454]	.09
PSQ-SRBD^g	0.2 (0.1)	0.5 (0.2) [n = 229]	0.2 (0.2) [n = 189]	−0.2 (0.2) [n = 187]	0.5 (0.2)	0.4 (0.2) [n = 193]	−0.1 (0.2) [n = 193]	−0.16 (−0.20 to −0.12)	<.001
mESS^h	5.4 (3.7)	6.9 (4.7) [n = 227]	5.0 (5.3) [n = 188]	−1.8 (4.9) [n = 184]	6.9 (4.6)	6.2 (5.1) [n = 193]	−0.7 (4.4) [n = 193]	−1.18 (−2.15 to −0.21)	.01
OSA-18ⁱ	NA	51.2 (15.7) [n = 229]	35.6 (13.9) [n = 188]	−15.8 (14.4) [n = 186]	52.7 (17.4)	46.5 (17.3) [n = 193]	−6.0 (14.6) [n = 193]	−9.75 (−12.84 to −6.65)	<.001
Caregiver-reported PedsQL^j
Total score		75.9 (13.2) [n = 229]	78.4 (16.0) [n = 189]	2.1 (14.9) [n = 187]	77.7 (12.8) [n = 226]	75.0 (15.9) [n = 193]	−2.6 (15.0) [n = 193]	4.76 (1.44 to 8.09) [n = 457]	.005
Physical score		79.5 (19.1) [n = 229]	81.1 (21.9) [n = 189]	0.7 (23.5) [n = 187]	82.1 (16.3) [n = 226]	76.4 (23.2) [n = 193]	−5.3 (24.3) [n = 193]	6.53 (1.29 to 11.78) [n = 457]	.01
Psychosocial score		73.9 (13.6) [n = 229]	77.0 (14.8) [n = 189]	2.8 (14.2) [n = 187]	75.3 (14.1) [n = 226]	74.3 (14.5) [n = 193]	−1.1 (13.3) [n = 193]	3.88 (0.89 to 6.88) [n = 457]	.01
BMI (percentile)	50 (25-75)^k	65.0 (30.0)	70.4 (27.4) [n = 187]	5.1 (14.3) [n = 187]	62.0 (32.1)	66.2 (31.4) [n = 188]	3.2 (13.3) [n = 188]	1.86 (−0.88 to 4.60)	.18
Systolic blood pressure (percentile)	50 (25-75)^k	63.6 (23.0) [n = 218]	58.6 (27.9) [n = 180]	−4.5 (28.9) [n = 173]	57.6 (27.0) [n = 216]	61.6 (26.0) [n = 184]	4.8 (31.4) [n = 177]	−9.02 (−15.49 to −2.54) [n = 451]	.006
Diastolic blood pressure (percentile)	50 (25-75)^k	55.5 (20.5) [n = 218]	51.5 (22.2) [n = 179]	−4.9 (22.5) [n = 172]	52.5 (21.5) [n = 216]	54.6 (22.1) [n = 184]	2.2 (20.8) [n = 177]	−6.52 (−11.59 to −1.45) [n = 451]	.01
Heart rate, /min	NA	80.8 (9.1)	78.8 (8.9) [n = 154]	−2.0 (7.7) [n = 154]	81.2 (9.5)	79.7 (9.5) [n = 152]	−1.8 (7.9) [n = 152]	−0.23 (−2.13 to 1.67)	.82
AHI^l		Baseline, No. (%)	12 mo, No./total (%)		Baseline, No. (%)	12 mo, No./total (%)		Effect size: risk difference at 12 mo, % (FDR-adjusted CI) ^d ^, ^m (n = 458)
≥3	NA	2 (0.9)	2/154 (1.3)		1 (0.4)	20/152 (13.2)		−11.2 (−17.5 to −4.9) [n = 306]	<.001
≥5	NA	0	0/154		0	11/152 (7.2)		−7.1 (−11.8 to −2.5) [n = 306]	.002

Abbreviations: AHI, apnea-hypopnea index; BMI, body mass index; BRIEF, Behavior Rating Inventory of Executive Function; CBCL, Child Behavior Checklist; FDR, false discovery rate; GEC, global executive composite; mESS, modified Epworth Sleepiness Scale; NA, not available or not applicable; OSA, obstructive sleep apnea; PedsQL, Pediatric Quality of Life Inventory; PSQ-SRBD, Pediatric Sleep Questionnaire–Sleep-Related Breathing Disorder.

^{^a}

The estimated between-group difference in mean change from baseline to 12 months and the corresponding P value are from a linear mixed-effects model with prespecified adjustment for stratification factors (age ≥6 years, overweight/obese, Black/African American) and site effect. The analysis incorporates information from all study participants with at least 1 measurement across the 3 study visits.

^{^b}

The BRIEF GEC section comprises summary measures of behavioral regulation, emotion regulation, and cognitive regulation (BRIEF-2, for children aged 5 to 18 years) or inhibitory self-control, flexibility, and emergent metacognition (BRIEF-P, for preschool-aged children). Caregiver scores ranged from 33 to 102, with higher scores indicating worse functioning. Thirty-five patients were lost to follow-up at 12 months in the early adenotonsillectomy group and 31 in the watchful waiting with supportive care group. All 458 patients contributed information to the difference-in-differences analysis.

^{^c}

The Go/No-go sustained attention d-prime is a signal detection measure that combines a child’s true positive rate on an attention task (correct response to the target stimuli) with their false-alarm rate (incorrect response to the nontarget stimuli). Scores ranged from −0.9 to 4.5, with higher scores indicating greater sustained attention. Forty-seven patients were lost to follow-up at 12 months in the adenotonsillectomy group and 43 in the watchful waiting group. Three patients were excluded from the difference-in-differences analysis.

^{^d}

P values for all secondary outcomes are adjusted so that the set of hypotheses with P values below .05 corresponds to the set of hypotheses that would be rejected under the Benjamini and Hochberg procedure for controlling the FDR at .05.¹⁷ This means that, among all effects considered statistically significant at the .05 level (using the adjusted P values), 5% would be expected to be truly null. Twenty-two P values were included in the adjustment set: all 17 secondary analyses presented in Table 2, as well as analyses of 5 caregiver BRIEF subscale scores (eTable 6 in Supplement 2). The 95% confidence intervals for all 22 of these secondary end points are adjusted for multiple comparisons following the procedure of Benjamini and Yekutieli,¹⁸ with a nominal coverage level of 97%.

^{^e}

The average of National Institutes of Health Toolbox 9-hole pegboard dexterity test times (in seconds) from both the dominant and nondominant hand. Average times ranged from 15.5 seconds to 118 seconds, with higher average times indicating lower manual dexterity. Forty-four patients were lost to follow-up at 12 months in the adenotonsillectomy group and 40 in the watchful waiting group. All 458 patients contributed information to the difference-in-differences analysis.

^{^f}

Scores ranged from 28 to 86 on the CBCL externalizing problems scale, from 29 to 88 on the CBCL internalizing problems scale, from 50 to 97 on the CBCL attentional problems scale, and from 24 to 84 on the CBCL total problems scale (comprising all 3 of the previous scales plus social and thought problems). On each scale, higher scores indicate greater problems. Forty-five patients were lost to follow-up at 12 months in both the adenotonsillectomy group and the watchful waiting group. Four patients were excluded from the difference-in-differences analysis.

^{^g}

Scores range from 0 to 1, with higher scores indicating greater severity. Forty-two patients were lost to follow-up at 12 months in the adenotonsillectomy group and 34 in the watchful waiting group. All 458 patients contributed information to the difference-in-differences analysis.

^{^h}

Scores range from 0 to 24, with higher scores indicating greater sleepiness. Forty-three patients were lost to follow-up at 12 months in the adenotonsillectomy group and 34 in the watchful waiting group. All 458 patients contributed information to the difference-in-differences analysis.

^ⁱ

Scores on the OSA-18 quality of life survey range from 18 to 126, with higher scores indicating a greater negative effect of sleep-disordered breathing on quality of life. Forty-three patients were lost to follow-up at 12 months in the adenotonsillectomy group and 34 in the watchful waiting group. All 458 patients contributed information to the difference-in-differences analysis.

^{^j}

The PedsQL total score comprises performance on 4 subscales: emotional functioning, social functioning, and school functioning (summarized by the psychosocial functioning score) and physical functioning (summarized by the physical functioning score). Scores on all scales range from 0 to 100, with higher scores indicating better quality of life. Forty-two patients were lost to follow-up at 12 months in the adenotonsillectomy group and 34 in the watchful waiting group. One patient was excluded from the difference-in-differences analysis.

^{^k}

Normative values reported as median (IQR).

^{^l}

AHI defined as the average number of apnea or hypopnea (hypopneas with ≥3% oxygen desaturation or arousal) events per hour of sleep, with higher scores indicating more severe obstructive sleep apnea. AHIs were rounded to the nearest tenth. An AHI of 3 or greater at 12 months indicates a progression of disease over the course of the study; an AHI of 5 or greater at 12 months indicates progression to moderate obstructive sleep apnea. Seventy-seven patients were lost to follow-up at 12 months in the adenotonsillectomy group and 75 in the watchful waiting group. One hundred fifty-two patients were excluded from the risk difference analysis.

^{^m}

The between-group difference in prevalence at 12 months was estimated using a linear regression model with robust standard errors, fit to all patients with complete baseline and 12-month information. Point estimates and confidence intervals adjust for baseline apnea-hypopnea index, stratification factors (age ≥6 years, overweight/obese, Black/African American) and site effect. The confidence intervals are also adjusted for multiple comparisons.