Table 2 Quality criteria that were used to assess the six studies
Study designRandomisationAllocation biasBaseline differencesAttrition rateBlindingVerification of outcome assessment
Reynolds et al. 20001128 elementary schools were randomly assigned to the interventionAllocation method not statedNo significant differences in age, gender, ethnicity, marital status, education level or SES were foundNot statedNot statedFood records, food frequency questionnaire
Friel et al. 199912No randomisationAllocation method not statedAt baseline, mean (SD) scores for behaviour attributes [25.4 (3.6)] were lower in the intervention group than in the control group [26.5 (3.3), p = 0.01]64% completed pre and post testsNot performedValidated visual instrument for food frequency questionnaire
Reinaerts et al. 200713No randomisationAllocation method not statedAt baseline, significantly more “non-native” children in group 2 (distribution) (50%) than in control group (14%, p<0.001); no significant differences in age or sex were found54% completed follow-up measurementsNot performedChildren’s fruit and vegetable consumption was measured by a food frequency questionnaire and a pre-structured food recall that the children’s parents completed at home
Winkleby et al. 19971424 community classes were randomised to one of two curriculum interventionsAllocation method not statedNo significant differences in age, gender or marital status were found82% completed follow-up measurementsNot statedSociodemographic and nutrition knowledge variables were self-reported; dietary intake data were collected using a validated food frequency questionnaire
Calfas et al. 200215Participants from four primary health care providers were randomly assigned to the interventionAllocation method not statedNo significant differences in age, SES, gender, BMI, ethnicity, marital status, employment or education status were found86% retention rateNot statedBlock simplified fat screener, food records and behavioural item scale
Kristal et al. 199916Participants from three clinical centres were randomly assigned to the interventionAllocation method not statedNo significant differences in age, gender, SES, education level or ethnicity were foundNot statedNot performedFood frequency questionnaires, fat- related dietary habits questionnaires