TABLE 3

Effects of the Intervention on Prevalence of Any (Versus None) Non-core Food and Drink Consumption and Television Viewing

% Consumers/ ViewersEffects of the InterventionEffects of the Intervention Accounting for Covariates
ControlInterventionOR (95% CI)aPOR (95% CI)bP
Midintervention (mean child age 9 [1.1] mo)
 Noncore drink intake13.87.10.48 (0.24–0.95).030.46 (0.23–0.93).03
 Sweet snack intake29.026.30.78 (0.48–1.27).320.73 (0.46–1.17).19
 Savory snack intake22.320.10.81 (0.52–1.25).330.78 (0.51–1.22).28
 Television viewing72.968.80.81 (0.53–1.24).330.79 (0.53–1.18).25
Postintervention (mean child age 19.8 [2.2] mo)
 Noncore drink intake27.322.50.81 (0.51–1.30).380.81 (0.51–1.28).37
 Sweet snack intake76.370.70.69 (0.43–1.10).120.69 (0.43–1.10).12
 Savoury snack intake52.058.61.25 (0.87–1.81).231.24 (0.86–1.79).25
 Television viewing80.780.51.01 (0.61–1.65).970.91 (0.56–1.46).69
  • a Marginal logistic regression models, estimated using generalized estimating equations with information sandwich (“robust”) standard errors, were fitted to compare binary outcomes between the trial arms, taking account of clustering. An exchangeable correlation structure was specified for these analyses.

  • b Marginal logistic regression models, estimated using generalized estimating equations with information sandwich (“robust”) standard errors, were fitted to compare binary outcomes between the trial arms, taking account of clustering. An exchangeable correlation structure was specified for these analyses. The dietary intake models adjusted for mothers’ education level and child’s overall energy intake. The television-viewing model adjusted for mothers’ education level and child’s age.