Table 2.

Impact of adjustment of occupational status-all cause mortality associations for distal, proximal and cumulative health behaviours by cohort and sex

	1952 cohort (N = 1,444)	1932 cohort (N = 1,550)
N (%) died	120 (8.3)	719 (46.4)
Hazard ratio (95 % confidence interval) for all-cause mortality in respondents with manual versus non-manual occupational status
Age- and sex-adjusteda	1.80 (1.25, 2.58)	1.66 (1.43, 1.93)
Adjusted for 4 baseline behavioursb	1.39 (0.95, 2.02)	1.43 (1.23, 1.66)
% attenuationc	44 %	30 %
Adjusted for 4 most recent behavioursd	1.30 (0.88, 1.92)	1.40 (1.19, 1.63)
% attenuationc	55 %	34 %
Adjusted for 4 cumulativebehaviourse	1.14 (0.77, 1.69)	1.28 (1.10, 1.50)
% attenuationc	77 %	51 %
	1932 men (N = 704)	1932 women (N = 846)
N (%) died	385 (54.7)	334 (39.5)
Hazard ratio (95 % confidence interval) for all-cause mortality in respondents with manual versus non-manual occupational status
Age-adjusteda	1.59 (1.30, 1.95)	1.74 (1.41, 2.16)
Adjusted for 4 baseline behavioursb	1.31 (1.06, 1.62)	1.53 (1.23, 1.91)
% attenuationc	42 %	24 %
Adjusted for 4 most recent behavioursd	1.30 (1.05, 1.62)	1.53 (1.21, 1.92)
% attenuationc	43 %	24 %
Adjusted for 4 cumulativebehaviourse	1.18 (0.95, 1.46)	1.41 (1.12, 1.77)
% attenuationc	65 %	38 %

^aHazard ratio comparing all cause mortality in respondents with manual versus non-manual occupational status

^bHazard ratio additionally adjusted for each of smoking, drinking, physical activity and fruit and vegetable intake at baseline

^c % attenuation = 100 × (HR_unadjusted − HR_adjusted)/(HR_unadjusted – 1)

^dHazard ratio additionally adjusted for each of smoking, drinking, physical activity, and fruit and vegetable intake at last wave or wave immediately preceding death

^eHazard ratio additionally adjusted for mean smoking, drinking, physical activity and fruit and vegetable intake over all waves at which respondent was known to be alive