Blissett 2016

Methods	Study design: Randomised controlled trial Funding: “Funded by the Feeding For Life Foundation (grant reference number 11‐1170). ”
Participants	Description: Children aged 2 to 4 years and their principle caregiver (parent) N (Randomised): 120 parent‐child dyads Age: Child (mean): Prompting no modelling = 27 months, Prompting and modelling = 29 months, Modelling ‘control’ group = 31 months Mothers (mean): Prompting no modelling = 34 years, Prompting and modelling = 26 years, Modelling ‘control’ group = 35 years % Female: Child: 45% Parent: 98% SES and ethnicity: Not specified Inclusion/exclusion criteria: “Inclusion criteria for children included the absence of known food allergies or disorders affecting eating, current or recent major illness or diagnosed intellectual disabilities.” Recruitment: “Caregivers and their children were recruited through the Children and Child Laboratory database, which contains information on families in which caregivers have indicated an interest in research participation at the University of Birmingham.” Recruitment rate: Unknown Region: UK
Interventions	Number of experimental conditions: 3 Number of participants (analysed): Prompting no modelling = 35 dyads Prompting and modelling = 37 dyads Modelling ‘control’ group = 27 dyads Description of intervention: Prompting no modelling: “Caregivers were asked to use physical prompts to eat the novel fruit (NF) (including passing the food to the child, moving the food towards the child, holding the NF up to the child’s face, encouraging the child to touch the NF).” Prompting and modelling: As well as using physical prompts as in PNM, “The caregivers assigned to this condition were also asked to try the NF themselves.” Modelling ‘control’ group: “Caregivers in this condition were not given any information about prompting, but were simply asked to taste the NF themselves.” Duration: 1 day Number of contacts: 1 Setting: Lab Modality: Face‐to‐face Interventionist: Parents Integrity: Prompting no modelling: “Of an original sample of fifty, fifteen were classed as non‐compliant: ten caregivers failed to prompt a minimum of three times, and five caregivers were removed from the group because they ate the NF. This left a sample of thirty‐five parents who physically prompted but did not model eating the fruit.” Prompting and modelling: “Of an original sample of forty‐three dyads, six were non‐compliant because the parent failed to prompt three times or more, leaving a sample of thirty‐seven parents who prompted and modelled eating the fruit.” Modelling ‘control’ group: “There were twenty‐seven dyads in this condition, in which the parent modelled eating of the fruit; all were compliant with this request.” Date of study: Unknown Description of control: N/A
Outcomes	Outcome relating to children's fruit and vegetable consumption: Consumption of novel fruit (grams): “All meal items were weighed on scientific scales before and after consumption.” “Owing to differences in weights of the different NF offered, it was not possible to compare conditions based on simple weight of consumption. Therefore, we calculated consumption of the NF based on the percentage consumed of the whole portion offered.” 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: < 1 day Length of follow‐up post‐intervention: Same day Subgroup analyses: None Loss to follow‐up: Prompting no modelling: 30% Prompting and modelling: 14% Modelling ‘control’ group: No loss to follow‐up Analysis: Unknown if sample size calculation was performed
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The random sequence generation procedure is unclear. The authors indicate that block randomisation was used to allocate to groups in blocks of 10 participants with conditions changing each week, allocated in order of recruitment
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	Low risk	Fruit intake is an objective measure of child’s fruit intake and unlikely to be influenced by performance bias
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Fruit intake All meals were weighed on scientific scales before and after consumption therefore at low risk of detection bias
Incomplete outcome data (attrition bias) All outcomes	High risk	Used a per‐protocol analysis rather than an intention‐to‐treat analysis and therefore at high risk of attrition bias
Selective reporting (reporting bias)	Unclear risk	There is no study protocol therefore it is unclear if there was selective outcome reporting
Other bias	Unclear risk	There was a significant difference in children’s ages and child’s age was controlled for in analyses. Therefore the risk of other bias is unclear