Table 2 MAIHDA models for BMI.

From: Multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) to understand how obesity risk varies according to multiple lifestyle behavior recommendations

 

Males

Females

 

Null Model

Main effects model

Main effects model plus confoundersa

Null Model

Main effects Model

Main effects Model plus confoundersa

 

β (95% CI)

β (95% CI)

β (95% CI)

β (95% CI)

β (95% CI)

β (95% CI)

Intercept

28.04 (27.80, 28.29)

29.64 (29.45, 29.84)

29.82 (29.60, 30.05)

26.87 (26.61, 27.12)

28.36 (28.18, 28.53)

28.60 (28.39, 28.81)

Smoking

 Previous

 

–

–

 

–

–

 Current

 

−0.90 (-1.10, -0.71)

−0.93 (−1.13, −0.74)

 

−0.52 (−0.69, −0.34)

−0.50 (−0.67, −0.33)

 Never

 

−1.03 (−1.20, −0.86)

−0.99 (−1.17, −0.82)

 

−0.54 (−0.69, −0.40)

−0.44 (−0.58, −0.30)

Fruit & vegetable

 Not meets

 

–

–

 

–

–

 Meets

 

0.11 (−0.04, 0.26)

0.10 (−0.04, 0.25)

 

−0.32 (−0.45, −0.20)

−0.39 (−0.51, −0.26)

Physical activity

 Not meets

 

–

–

 

–

–

 Meets

 

−1.07 (−1.22, −0.92)

−1.08 (−1.23, −0.93)

 

−1.46 (−1.59, −1.33)

−1.50 (−1.63, −1.37)

Alcohol

 Not meets

 

–

–

 

–

–

 Meets

 

−0.32 (−0.47, −0.17)

−0.31 (−0.47, −0.16)

 

0.19 (0.06, 0.32)

0.17 (0.04, 0.29)

Sleep

 Not meets

 

–

–

 

–

 

 Meets

 

−0.66 (−0.81, −0.51)

−0.64 (−0.79, −0.49)

 

−0.63 (−0.76, −0.50)

−0.56 (−0.69, −0.43)

Level 2 variance

0.70 (0.40, 0.99)

0.05 (0.02, 0.07)

0.05 (0.02, 0.07)

0.76 (0.44, 1.09)

0.03 (0.01, 0.04)

0.02 (0.01, 0.04)

Level 1 variance

15.15 (15.04, 15.26)

15.15 (15.04, 15.26)

15.13 (15.02, 15.24)

19.66 (19.51, 19.82)

19.66 (19.51, 19.82)

19.47 (19.32, 19.62)

 VPC

4.43%

0.33%

0.33%

3.76%

0.15%

0.15%

 PCVb

 

92.80%

92.76%

 

96.01%

96.17%

  1. aConfounders were Age, ethnicity, tdi and employment status.
  2. bAll estimated PCVs are relative to the null model.