Skip to main content
. 2021 Feb 22;11:4296. doi: 10.1038/s41598-021-83801-0

Table 2.

Mediation analysis to disentangle the effects of body mass index and smoking in the pathway from education level to health outcomes.

Outcomes Total effect of education Effect after adjusting for BMI Effect after adjusting for smoking Effect after adjusting for both
ORa 95% CI ORb 95% CI %* ORc 95% CI %* ORd 95% CI %*
Mental disorder
Insomnia 0.72 0.68, 0.76 0.75 0.71, 0.80 12 0.74 0.69, 0.79 8 0.77 0.72, 0.82 20
Major depressive disorder 0.78 0.74, 0.83 0.85 0.80, 0.90 35 0.83 0.77, 0.88 25 0.87 0.81, 0.93 44
Suicide attempts 0.56 0.48, 0.67 0.60 0.50, 0.72 12 0.67 0.54, 0.82 31 0.68 0.55, 0.84 33
Cardiovascular disease
Coronary artery disease 0.63 0.58, 0.69 0.67 0.61, 0.73 13 0.72 0.65, 0.79 29 0.74 0.66, 0.82 35
Heart failure 0.51 0.45, 0.59 0.55 0.47, 0.64 11 0.64 0.54, 0.75 34 0.65 0.55, 0.77 36
Total stroke 0.68 0.63, 0.73 0.70 0.64, 0.75 8 0.75 0.69, 0.81 25 0.76 0.69, 0.82 29
Any ischemic stroke 0.66 0.61, 0.72 0.67 0.61, 0.74 4 0.69 0.63, 0.77 11 0.70 0.63, 0.78 14
Large artery stroke 0.51 0.42, 0.63 0.54 0.43, 0.68 8 0.59 0.46, 0.76 22 0.59 0.46, 0.77 22
Small vessel stroke 0.56 0.46, 0.67 0.54 0.44, 0.67 0 0.54 0.43, 0.67 0 0.53 0.42, 0.66 0
Cancer
Breast cancer 0.89 0.83, 0.96 0.92 0.85, 1.00 28 0.93 0.85, 1.01 38 0.94 0.86, 1.03 47
Breast cancer ER− 0.73 0.66, 0.81 0.74 0.66, 0.83 4 0.71 0.63, 0.80 0 0.71 0.63, 0.81 0
Lung cancer 0.49 0.43, 0.56 0.56 0.48, 0.65 19 0.53 0.45, 0.63 11 0.59 0.49, 0.70 26
Other diseases
Rheumatoid arthritis 0.50 0.41, 0.60 0.44 0.36, 0.53 0 0.41 0.34, 0.50 0 0.42 0.34, 0.52 0
Type 2 diabetes 0.52 0.48, 0.57 0.60 0.55, 0.64 22 0.75 0.70, 0.81 56 0.79 0.73, 0.85 64

BMI body mass index, ER estrogen receptor.

*Percentage of the effect of education on the health outcome that is mediated by body mass index, smoking, or both (Formula: (log(OR_total) − log(OR_adjusted))/log(OR_total) × 100)). We replaced the values with zero for those percentage below zero.

aTotal effect without any adjustment.

bAdjusted for the effect of body mass index.

cAdjusted for the effect of smoking (cigarettes per day).

dAdjusted for the effects of both body mass index and smoking behaviours.