TABLE 6.

Cross-validation of total energy expenditure predictions from age, weight, height, physical activity, and resting metabolic rate

Model		Prediction coefficients							Cross-validation predictions1
												Regression2
		PA	A	B x age	D x weight	E x height	F x RMR	R2	rc	Bias5	LOA3	Intercept	Slope	rs
			(constant)	(y)	(kg)	(cm)	(kJ/d)			kJ/d (%)	kJ/d (%)	(kJ/d)
Group	Sex	Prediction based on age, weight, and height
A	Females	.4	4834.5	−53.9	33.1	23.4		0.69	0.82	−75 (−2.3)	−1944 to 1793 (−26.5 to 22.0)	1159	0.87	0.82
	Males	.4	4780.3	−58.4	51.9	21.6		0.77	0.86	−12 (−1.3)	−1957 to 1934 (−23.2 to 20.6)	1522	0.85	0.88
B	Females	.4	3934.7	−44.3	41.5	21.2		0.72	0.83	52 (−0.8)	1690 to 1795 (−23.9 to 22.3)	1562	0.80	0.82
	Males	.4	6768.6	−74.2	48.9	19.2		0.74	0.85	29 (−1.0)	−2090 to 2148 (−23.8 to 21.8)	1078	0.89	0.86
		Prediction based on age, weight, height, and RMR
A	Females	.4	−5541.6	−8.9	6.0	42.5	1.13	0.89	0.94	35 (0.5)	−1100 to 1169 (−14.0 to 14.1)	318	0.95	0.93
	Males	.4	−5971.5	−15.3	8.9	44.3	1.12	0.89	0.94	−69 (−1.2)	−1405 to 1267 (−14.8 to 12.4)	802	0.93	0.94
B	Females	.4	−5046.3	−7.5	14.4	40.1	0.96	0.88	0.93	4 (0.5)	−1190 to 1182 (−15.6 to 14.6	495	0.94	0.92
	Males	.4	−5320.0	−19.2	5.4	44.0	1.14	0.89	0.94	74 (0.04)	−1287 to 1436 (−12.6 to 13.0)	64	0.99	0.94

LOA, limit of agreement; PA, physical activity; RMR, resting metabolic rate.

¹TEE predicted in group A using coefficients generated in Group B and vice versa; R²= coefficient of determination.

²Regression = measured TEE versus predicted; r_c = Lin’s concordance correlation coefficient; r_s = Spearman correlation coefficient

³LOA = 1.96 SD limits of agreement.

⁴PA values according to age group and sex (see text and Table 4).

⁵Measured – predicted