The effect of a sugar sweetened beverage tax among different socioeconomic groups: A systematic review

Abstract

sleeping patterns, physical activity, dietary quality, parenting self-efficacy and feeding practices. Data from these studies were prospectively planned to be combined into an IPD set. The principal analysis method for aggregating the data across trials was a one-step linear modelling approach that included treatment effect as a fixed effect and accommodated the clustering introduced by one trial. Multiple-imputation was used to evaluate the impact of missing data for the primary endpoint. Results: The study cohort of four trials comprises 2196 women/infants. Data for the primary outcome were available for 78% of the cohort. The mean BMI-z score at 18—24 months was 0.67 (95% CI: 0.60—0.74) for the intervention group and 0.80 (95% CI: 0.73—0.87) for the control group. The estimated difference was −0.13 (95% CI: −0.23 to −0.03; p = 0.012). The multiple-imputation analysis estimated the difference at −0.10 (95% CI: −0.20 to 0.00; p = 0.04). Median duration of breastfeeding was significantly longer in the intervention arm (35 weeks [95% CI: 30—37] versus 28 weeks [95% CI: 26—30]; HR = 0.89 [95% CI: 0.8—0.98]; p = 0.022). The proportion of children viewing TV for >1 h/day was significantly lower in the intervention arm (24% versus 35%; OR = 0.63 [95% CI: 0.49—0.81; p = <0.001]). There were no significant differences in overweight/obesity, sleeping patterns or physical activity. Conclusion: Early interventions for the prevention of childhood obesity appear effective in achieving a reduction in BMI z-scores at 18—24 months to a moderate degree, as well as prolonging breast feeding time, and reducing TV viewing time.