Skip to main content
. 2020 May 12;17:60. doi: 10.1186/s12966-020-00963-2

Table 2.

Association between soft drink intake and body weight change in the health workers cohort study, 2004–2010 (N = 1268)

Characteristics Model 1
β-coefficient (95% CI)
Model 2
β-coefficient (95% CI)
Soft drink (servings/day) 0.03 (−0.64, 0.70) − 0.02 (− 0.84, 0.79)
Time (years) 0.46 (0.23, 0.68) 0.45 (0.22, 0.68)
Soft drink x time 0.10 (0.00, 0.19) 0.11 (−0.02, 0.23)
Physical activity
 Low (referent)
 High 0.42 (−0.19, 1.03) 0.34 (−0.40, 1.08)
Physical activity x time −0.11 (− 0.23, 0.01) −0.10 (− 0.24, 0.05)
Soft drink x physical activity 0.18 (− 0.84, 1.20)
Sex x time 0.08 (−0.03, 0.18) 0.07 (− 0.03, 0.18)
Baseline age x time −0.02 (− 0.02, − 0.02) −0.02 (− 0.02, − 0.02)
Soft drink x physical activity x time − 0.03 (− 0.27, 0.21)
AIC test 12,149.59 12,153.29
BIC test 12,301.39 12,316.77

Model 1: Individual-level fixed effects model of two-way interaction terms (soft drink and time, physical activity and time, sex and time, and baseline age and time), adjusted for education, chronic diseases, smoking status, TV viewing time per week, total daily sleep, alcohol intake and food groups: red meat, total dairy, fruits, vegetables, nuts, yogurt, white bread, tortillas and orange juice. Model 2: Individual-level fixed effects model of three-way interaction terms (soft drink, time and physical activity), included two-way interaction terms (soft drink and time, physical activity and time, physical activity and soft drink, sex and time, and baseline age and time) adjusted by the same set of covariates of model 1. Age and sex were centered at the baseline mean in both models. CI confidence interval, AIC Akaike information criterion, BIC Bayesian information criterion