Harnack 2012

Methods	Study design: Randomised controlled trial – cross‐over Funding: "Funded by a grant from the Robert Wood Johnson Foundation Healthy Eating Research program."
Participants	Description: Preschool‐aged children attending a Head Start centre in Minneapolis, Minnesota, USA N (Randomised): 57 children Age: 2 to 3 years = 51% 4 to 5 years = 49% % Female: Not specified SES and ethnicity: Child: Non‐Hispanic African‐American = 76%, Hispanic or Latina/Latino = 6%, Multi‐racial = 13%, American Indian = 4%, Non‐Hispanic White = 2% Parent education: Less than high school = 9%, High school graduate = 42%, Some college = 49% Inclusion/exclusion criteria: Not specified Recruitment: “Children in three preschool classrooms were recruited. A consent form and letter explaining the study was sent to parents.” Recruitment rate: 98% (57/58) Region: Minneapolis, Minnesota (USA)
Interventions	Number of experimental conditions: 3 Number of participants (analysed): Overall = 53 Description of intervention: Fruit and vegetable first: “During the fruit and vegetable first experimental weeks all fruits and non‐starchy vegetables on the lunch menu were served traditional family style five minutes in advance of other menu items. Children were allowed to begin eating the fruit and vegetable items served first, with the remaining menu items (e.g. milk, entrée, side dishes) placed on the tables for traditional family style meal service five minutes following distribution of the first course. All other usual meal service practices remained the same during the fruit and vegetable first experimental condition.” Provider portioned: “During the provider portioned experimental condition, a plate was prepared for each child that contained a specific quantity of each menu item.” Duration: “Each condition was implemented for two one‐week periods over the six week period, for a total of two weeks per condition” Number of contacts: Unclear, each day of the 6‐week period (dependent on how many days children attend) Setting: Preschool Modality: Face‐to‐face Interventionist: Classroom teachers Integrity: No information provided Date of study: Unknown Description of control: Usual ‘control’ meal service: “ "During each day of the control weeks, the usual traditional family style meal service approach to serving lunch meals at the center was followed. During usual lunch meals at the center children are seated around tables, and each food item on the menu is passed around the table from child to child in serving bowls for self‐service.”
Outcomes	Outcome relating to children's fruit and vegetable consumption: Child’s consumption of fruit and vegetable serves (1 cup equivalents). Study staff trained and certified in conducting lunch observations recorded food intake on a meal observation form. “The lunch observation data were entered into Nutrition Data System for Research (NDSR), a dietary analysis software program.” Outcome relating to absolute costs/cost effectiveness of interventions: Not reported Outcome relating to reported adverse events: Not reported Length of follow‐up from baseline: 6 weeks Length of follow‐up post‐intervention: Immediately Subgroup analyses: None Loss to follow‐up: Overall = 7% Analysis: Unknown if sample size calculation was performed
Notes	Sensitivity analysis ‐ primary outcome: Primary outcome not stated, fruit and vegetable intake is the only outcome
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomly allocated to experimental group but the random sequence generation procedure is not described
Allocation concealment (selection bias)	Unclear risk	There is no information provided about allocation concealment and therefore it is unclear if allocation was concealed
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Intake: There is no mention if children were blinded and so it is unclear how this may impact children’s vegetable intake
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Intake: Observers made visual estimations of food amounts to determine the amount taken but it is unclear if observers were blinded to condition. Food amounts may not be accurately estimated by observers
Incomplete outcome data (attrition bias) All outcomes	Low risk	3/57 (93%) completed the study and therefore the risk of attrition bias is low
Selective reporting (reporting bias)	Unclear risk	There is no study protocol therefore it is unclear if there was selective outcome reporting
Other bias	Low risk	Contamination, baseline imbalance, & other bias that could threaten the internal validity are unlikely to be an issue