Cause Effect Low+ High+ Beta-coeff.(95% CI) Estimated odds ratio (95% CI)
Age CVD 43.4 65.2 1.54 (1.11,  1.97) 4.65 (3.03,  7.14)
Age Statin use 43.4 65.2 0.84 (0.69, 0.99) 2.31 (1.99,  2.69)
Age Anti-hypertensive 43.4 65.2 1.44 (1.31,  1.57) 4.23 (3.72,  4.82)
Age Family history of CVD 43.4 65.2 -0.31(-0.43, -0.20) 0.73 (0.65,  0.82)
Age Systolic blood pressure 43.4 65.2 10.42 (9.5, 11.34) N/A
Age Sex (men) 43.4 65.2 -0.73 (-0.83,  -0.62) 0.48 (0.43,  0.54)
Ethnic group Diabetes Other Indian 1.64(1.33,  1.95) 5.14 (3.78, 7.00)
    Other Maori 1.03(0.84,1.23) 2.81 (2.31, 3.42)
    Other Pacific 1.86(1.68, 2.04) 6.44(5.39, 7.70)
Ethnic group Smoker Other Indian -0.50 (-0.99,  -0.01) 0.60(0.37,  0.99)
    Other Maori 1.28(1.12,  1.45) 3.60(3.05, 4.25)
    Other Pacific 0.72(0.54,  0.91) 2.06 (1.72, 2.48)
Ethnic group (adj. for age) Family history of CVD Other Indian 0.02(-0.28,  0.32) 1.02(0.75,  1.37)
    Other Maori -0.24(-0.41,  -0.07) 0.79 (0.67,  0.93)
    Other Pacific -1.03 (-1.24, -0.82) 0.36 (0.29,  0.44)
Diabetes (adj. for age) Statin use No Yes 1.94 (1.77,  2.10) 6.94 (5.90,  8.16)
Diabetes (adj. for age) Antihyper-tensive use No Yes 1.68 (1.53,  1.84) 5.38 (4.60,  6.28)
Statin use(adj. for age) Antihyper-tensive use No Yes 1.70 (1.55,  1.86) 5.49 (4.69,  6.42)
Anti-hypertensive (adj. for age) Systolic blood pressure No Yes 7.30(6.28,  8.33) N/A
Smoker (no adj.) CVD No Yes 0.59(0.15,  1.03) 1.80 (1.16,  2.79)
Smoker (no adj.) Total: HDL-cholesterol ratio No Yes 0.51 (0.43, 0.59) N/A
Family history (adj. for ethnic group and age) Statin Use No Yes 0.42 (0.26,  0.58) 1.52 (1.30,  1.79)
Sex (no adj.) Total: HDL-cholesterol ratio Female Male 0.61(0.55,  0.67) N/A

CVD: Cardiovascular disease. HDL: high density lipoprotein. adj.: adjustment. N/A: not applicable.
+for age, comparisons were made at the 84th and 16th centiles to allow comparison with measures of effects from binary variables [21].
Table 1: Lists of causal links andestimated beta-coefficients from linear or logistic regression, adjusted for variables, as indicated in the DAG). Also odds ratios for logistic regression models.