. 2019 Jan 15;16(2):232. doi: 10.3390/ijerph16020232

Table 2.

Cardiorespiratory fitness and obesity outcome measures among children from Port Elizabeth, South Africa, at baseline (February 2015) and after a multidimensional physical activity intervention at the 16-month endline survey (May 2016). Values are unadjusted means (standard deviations) unless specified otherwise and estimated effects of the physical activity intervention on the mean changes in the respective outcome measures between baseline and endline, adjusted for the respective baseline value of sex, age, HAZ, SES index, Hb, soil-transmitted helminth (A. lumbricoides and T. trichiura) and intestinal protozoa (Cryptosporidium spp. and G. intestinalis) infection.

Variables	Schools with Physical Activity Intervention		Schools without Physical Activity Intervention		Intervention Effect ^a
	Baseline	Endline	Baseline	Endline	Estimate ^b (95% CI)	p-Value	ICC ^c
	(n = 300)	(n = 264)	(n = 446)	(n = 255)	Estimate ^b (95% CI)	p-Value	ICC ^c
Cardiorespiratory fitness
Shuttle run (laps)	35.6	34.5	36.8	35.3	−0.56 (−4.67 to 3.56)	0.79	0.04
Shuttle run (laps)	(17.0)	(17.9)	(17.4)	(18.7)	−0.56 (−4.67 to 3.56)	0.79	0.04
VO₂max ^d (mL × kg⁻¹ × min⁻¹)	45.8	43.5	46.3	44	−0.14 (−1.17 to 0.88)	0.78	0.03
VO₂max ^d (mL × kg⁻¹ × min⁻¹)	(4.1)	(4.7)	(4.3)	(4.8)	−0.14 (−1.17 to 0.88)	0.78	0.03
Obesity
BMIZ ^e	−0.1	−0.1	0	0.2	−0.17 (−0.24 to −0.09)	<0.001	<0.01
BMIZ ^e	(1.2)	(1.3)	(1.2)	(1.3)	−0.17 (−0.24 to −0.09)	<0.001	<0.01
Skinfolds ^f (mm)	9.0	9.6	9	10.1	−1.06 (−1.83 to −0.29)	0.007	0.02
Skinfolds ^f (mm)	(4.5)	(4.6)	(4.4)	(5.9)	−1.06 (−1.83 to −0.29)	0.007	0.02
Mean of self-reported physical activity ^g	9.1	9.0	7.8	9.9	−1.08 (−2.36 to 0.18)	0.09	0.04
Mean of self-reported physical activity ^g	(3.7)	(3.1)	(3.8)	(3.4)	−1.08 (−2.36 to 0.18)	0.09	0.04

^a Schoolchildren from the intervention group accomplished a multidimensional physical activity intervention programme between baseline and endline, as described in Figure 2; ^b Estimate of the physical activity intervention effect on the change in the respective outcome measure from baseline to endline, with 95%-confidence interval, P-value and ICC. The underlying linear mixed models included binary factor variables for all three intervention programmes (i.e., the physical activity intervention, the health education and the nutrition education programme) along with the baseline values of age, sex, height-for-age Z-score (HAZ), haemoglobin, socioeconomic (SES) index, protozoa- and soil-transmitted helminth (STH) infection status and random effects for classes and schools; ^c Proportion of unexplained variance attributable to clustering within schools and classes (intraclass correlation coefficient; ICC); ^d Using age-adjusted test protocol from Léger et al. [20]; ^e Sex-adjusted BMI-for-age Z-score (BMIZ); ^f Average of six measurements (triceps and subscapular three times each); ^g Score generated based on self-reported physical activity in the personal free time over the past 7 days and intense exercises outside structured school hours (range: from 1 to 14; 14 being the most active).