Table 1.

Study characteristics of included trials

First author, year published, country	Methods	Participants	Intervention	Outcomes	Conflicts of interest
de Silva-Sanigorski 2011(30), Australia	Study design: Cross-sectional, quasi-experimental design No. of experimental conditions: 1	Service type: Family day care services Socioeconomic characteristics: Intervention: SEIFA = 44·1 Comparison: SEIFA = 41·0 Participants: Intervention: n 1 family day care scheme (28 family day care providers) Control: n 18 family day care schemes (223 family day care providers) Recruitment rate: Not reported	Intervention duration: 4 years Number of contacts: Unclear Description of intervention: Romp & Chomp was a community-based and community-wide programme that sought to change policy, sociocultural and physical aspects of early childhood environments to favour obesity prevention. The intervention activities had a strong focus on community capacity building and developing sustainable changes in areas of policy, sociocultural and physical environments by using a socioecological framework. Project objectives included increasing the capacity of organisations to promote healthy eating and active play, decreasing the consumption of high-sugar drinks, energy-dense snacks and television viewing time and promoting the consumption of water, milk, fruit and vegetables and increasing active play. Implementation strategies: Educational meetings, educational outreach visits, educational materials and others. Unclear who delivered the intervention – ‘a management committee of stakeholder representatives oversaw the implementation of the projects “action” plan Community based intervention’ Theoretical underpinning: Guided by the health promotion principles of Nutbeam(38,39) and the socio-ecological model of health(40,41) Delivery modalities: Unclear Description of control: Not reported – sample matched to the intervention community’s characteristics	Data collection method: The environmental audit used in this evaluation was developed specifically for Romp & Chomp. The audit comprised forty-five questions designed to capture the general characteristics of the setting (e.g. number of children cared for) and factors in the physical, policy, sociocultural and economic environments of the setting that could enhance or inhibit efforts to promote healthy eating and active play for children aged 0–5 years who attend the setting. Validity of measures used: Not reported Child outcomes: Not reported Environmental outcomes: Compared with the comparison sample, intervention children spent less time watching television (coefficient –8·01, 95 % CI –14·90, –1·13, P = 0·03) and using computer/electronic games (coefficient –1·65, 95 % CI –3·12, –0·19, P = 0·03), less time in organised active play (coefficient –32·52, 95 % CI –44·04, –21·01, P < 0·001) and less time in free inside play (coefficient –15·37, 95 % CI –29·40, –1·33, P = 0·03). In the intervention service, there were also significantly more rules related to healthy eating (coefficient –0·42, 95 % CI 0·20, 0·65, P < 0·001), more care provider practices that supported positive meal experiences for children (coefficient 0·19, 95 % CI 0·07, 0·31, P < 0·001), fewer unhealthy food/drink items allowed (coefficient –0·21, 95 % CI –0·41, –0·01, P < 0·05), higher ratings of resources for both nutrition (coefficient 0·63, 95 % CI 0·27, 1·00, P < 0·001) and physical activity (coefficient 0·85, 95 % CI 0·55, 1·16, P < 0·001) and a higher mean rating for the food-related physical environment (coefficient 0·36, 95 % CI 0·14, 0·57, P < 0·001). Intervention costs: Not reported Adverse outcomes: Not reported	Sources of funding: Reported Potential conflicts of interest: Not reported
Trost 2011(37), United States	Study design: Quasi-experimental design with replication in three independent cohorts of family day care services No. of experimental conditions: 1	Service type: Family day care services Socioeconomic characteristics: Not reported Participants: Intervention: Baseline n 196 (pre-assessments) Final follow-up n 199 (post-assessments) Analysed at follow-up n 196 Control: n 297 Recruitment: Not reported Recruitment rate: Not reported	Intervention duration: 12 months Number of contacts: 5 Description of intervention: Healthy Kansas Kids programme – a state-wide initiative focusing on obesity prevention in early childhood settings via a train-the-trainer model. Content included ways to create a healthy lifestyle through nutrition and activities, importance of movement for learning, learning experiences to combine physical activity, food-related activities and children’s books and including children in food preparation Implementation strategies: Educational meetings, educational outreach visits and educational materials delivered by experts in the fields of nutrition and physical activity Theoretical underpinning: Not reported Delivery modalities: Face-to-face Description of control: Not reported – representative sample of registered family day care services operating in the state of Kansas	Data collection method: Nutrition and Physical Activity Self-Assessment for Child self-assessment instrument (NAP SACC-SA). Validity of measures used: NAP SACC-SA scores for nutrition and physical activity were calculated by averaging the responses for the nine nutrition (thirty-six items, Cronbach’s α 0·76) and six physical activity content areas (nineteen items, Cronbach’s α 0·75), respectively. Child outcomes: Not reported Environmental outcomes: Healthy Kansas Kids family day cares exhibited significant improvements in healthy eating (change between 6·9 and 7·1 %) and physical activity scores (change between 15·4 and 19·2 %) (P < 0·05). Within each cohort, pre-intervention scores were not significantly different from the state average, whereas post-intervention scores were significantly higher than the state average. Intervention costs: Not reported Adverse outcomes: Not reported	Sources of funding: Reported Potential conflicts of interest: Nil reported

SEIFA, socioeconomic index for areas.