Table 3. Percentage Attenuation of Sex-Adjusted Associations Between Affective Case Accumulation and Mortality by Individual Covariates, Ordered by Total Percentage Attenuation.
Covariate | % Attenuation by No. of times affective symptoms were at case levela | ||
---|---|---|---|
1 | 2 | 3-4 | |
Health conditions | 13.2 | 25.3 | 32.1 |
Lung function | 13.2 | 28.7 | 24.6 |
Physical activity | 11.8 | 25.3 | 23.9 |
Anxiolytic use | 7.9 | 18.4 | 28.4 |
Smoking | 5.3 | 17.2 | 24.6 |
Antidepressant use | 6.6 | 14.9 | 20.1 |
Diet | 6.6 | 12.6 | 16.4 |
Adult social class | 7.9 | 11.5 | 11.2 |
Marital status | 6.6 | 11.5 | 7.5 |
Pulse rate | 6.6 | 4.6 | 12.7 |
Childhood sickness absence | 5.3 | 4.6 | 7.5 |
Educational level | 3.9 | 2.3 | 6.7 |
Social support | 2.6 | 3.4 | 6.0 |
Childhood cleanliness | 2.6 | 5.7 | 3.0 |
Problem drinking | 1.3 | 3.4 | 5.2 |
Adolescent externalizing | 0 | 2.3 | 6.7 |
Body mass index | 1.3 | 3.4 | 3.7 |
Parental abuse | 1.3 | 1.1 | 3.7 |
Parental divorce | 0 | −1.1 | 4.5 |
Childhood social class | 5.3 | 0 | −3.0 |
Stressful life events | 0 | 0 | 0.7 |
Systolic blood pressure | −1.3 | −1.1 | −5.2 |
Fully adjusted (all covariates) | 39.5 | 81.6 | 86.6 |
Percentage attenuation of each covariate was entered separately (ie, not simultaneously) into the sex-adjusted model; since individual covariates are not independent of one another, these figures provide only an approximate indication of the importance of each variable.