Table 2.

Associations Between Age at Onset of Independent Walking and Hip Shape Mode Outcomes in the MRC National Survey of Health and Development

Mode	Group	Model	Regression coefficient	95% CI		P	Sex interaction P
HM1	Combined	1	−0.036	−0.058	−0.014	<0.01	0.40
		2	−0.032	−0.055	−0.010	<0.01	0.55
HM2	Men	1	−0.014	−0.047	0.019	0.4	0.09
	Women		0.022	−0.008	0.052	0.16
	Men	2	−0.001	−0.035	0.032	0.95	0.09
	Women		0.036	0.005	0.066	0.02
HM3	Combined	1	0.016	−0.005	0.038	0.14	0.42
		2	0.019	−0.003	0.041	0.09	0.57
HM4	Men	1	0.042	0.008	0.076	0.01	0.01
	Women		−0.012	−0.040	0.016	0.39
	Men	2	0.040	0.005	0.075	0.023	0.01
	Women		−0.006	−0.034	0.023	0.69
HM5	Combined	1	−0.003	−0.025	0.020	0.82	0.53
		2	0.006	−0.017	0.029	0.63	0.56
HM6	Combined	1	0.02	−0.002	0.042	0.07	0.31
		2	0.02	−0.002	0.042	0.08	0.42
HM7	Combined	1	−0.032	−0.054	−0.009	<0.01	0.99
		2	−0.027	−0.050	−0.004	0.02	0.96
HM8	Combined	1	−0.005	−0.027	0.017	0.68	0.96
		2	−0.008	−0.031	0.015	0.50	0.98
HM9	Combined	1	−0.001	−0.023	0.021	0.95	0.33
		2	−0.005	−0.028	0.017	0.65	0.28
HM10	Combined	1	−0.01	−0.031	0.011	0.34	0.89
		2	−0.005	−0.027	0.016	0.63	0.7

Regression coefficients are the difference in mean HM scores per 1 month increase in walking age. Where sex interactions were evident (P for interaction < 0.1), sex‐specific associations are presented. Model 1 adjusted for sex (if men and women are combined) and CRF location. Model 2: model 1 + birthweight + father's occupational class + adult occupational class + height + appendicular fat mass + appendicular lean mass. Only results from basic and fully adjusted models are presented for brevity. When each set of covariates were adjusted for in turn, there was no evidence that any one specific set of factors was responsible for the attenuations observed between the models shown here.