Strategies which aim to positively impact on weight, physical activity, diet and sedentary behaviours in children from zero to five years. A systematic review of the literature

Authors


Karen Campbell, Centre for Physical Activity and Nutrition Research, School of Exercise and Nutrition Sciences, Deakin University, 221 Burwood Highway, Burwood, 3125, Australia. E-mail: kcamp@deakin.edu.au

Summary

Preventing the development of obesity in children is an international health priority. To assess the effectiveness of interventions designed to prevent obesity, promote healthy eating and/or physical activity and/or to reduce sedentary behaviours in 0–5-year-old children, a systematic review of the literature was performed. Literature searches were limited to articles published between January 1995 and June 2006, printed in English and sampling children aged 0–5-years. Searches excluded literature concerned with breastfeeding, eating disorders, and interventions which were school-based or concerned with obesity treatment. Two reviewers independently extracted data and assessed study strengths and weaknesses. Nine included studies were grouped based on the settings in which they were delivered. Most studies involved multi-approach interventions, were conducted in the USA and varied in study designs and quality. All showed some level of effectiveness on at least one obesity-behaviour in young children. These studies support, at a range of levels, the premise that parents are receptive to and capable of some behavioural changes that may promote healthy weight in their young children. The small quantity of research heralds the need, particularly given the potential for early intervention to have long-lasting impacts on individual and population health, to build in a substantial way upon this evidence base.

Introduction

Preventing the development of obesity in children is an international health priority (1). Current estimates suggest that the prevalence of overweight and obesity is rapidly increasing around the world (2). Contemporary data highlight that these increases begin in early childhood, with the prevalence of overweight increasing in developed and most developing countries over recent decades (3). Overweight and obesity are recognized to have numerous negative impacts on children’s health and wellness during childhood and through to adult life (4,5). In addition, obese children tend to become obese adults and treatment is difficult and costly (2). Cochrane reviews have established that opportunities for prevention are poorly understood and yet of paramount importance (6). Given this, there is an urgent need for research regarding the capacity to influence, in early life, the development of children’s obesity-promoting behaviours (7).

Recent evidence highlights that obesity-promoting dietary habits, such as high consumption of energy-dense foods and fluids, previously documented in children (7,8), are also evident in infants and toddlers (9). Further, the nutritional quality of diet is known to decline during childhood (10). These early dietary patterns couple with evidence of high levels of sedentariness in early childhood. For example, 17% of 0–11-month and 48% of 12–23-month-old children in the USA watch more than the recommended 2 h of television per day, and this proportion of children increases throughout childhood (11). Very little physical activity trend data during infancy and early childhood is available; however, it appears that physical activity levels fall throughout childhood (12).

In addition to population trend data, there is evidence of tracking of children’s dietary (13), sedentary and physical inactivity behaviours (14) from childhood to adolescence and adulthood, which appear to impact on adult health (15). Thus, the obesity-promoting behaviours learned and supported during these early years establish lifestyle behaviours that may track throughout the lifespan. Given this, it is reasonable to posit that early childhood provides a unique and circumscribed opportunity within which to establish lifestyle behaviours that will promote health and minimize the risk of the development of fatness.

While there have been recent reviews of interventions designed to prevent obesity and obesity-promoting behaviours in older children (16–18), there have been no reviews which have focused specifically on obesity prevention interventions in younger children. Given the importance of the early years (birth to five) for the development of health-promoting behaviours, we have undertaken a systematic review of the literature in this area. The main objective of the review was to assess the effectiveness of interventions designed to prevent obesity, promote healthy eating and/or physical activity or reduce sedentary behaviours in 0–5-year-old children.

Methods

Search strategy

The search strategy used for this review involved several phases. In the first instance, we identified reviews of childhood obesity prevention and/or interventions that targeted obesity-promoting behaviours (i.e. diet, physical activity, sedentary behaviours). Hand searching of these reviews facilitated the identification of key researchers and programmes of work in this area. Publications arising from these programmes of work were identified and retrieved for inclusion in the review. Second, we contacted key informants, to assess any new/emerging literature that would be published within the time frames in which we conducted the review. Finally, systematic searches of Medline, PsychINFO, Academic Search Premier, Communication and CINAHL databases were conducted using the search strategy outlined in Fig. 1. Literature searches were limited to articles published between January 1995 and June 2006, printed in English, and sampling children aged 0–5 years. All abstracts generated from these searches were screened by one author for potential inclusion suitability. Twenty-three abstracts were identified as of interest and full copies of these articles were retrieved and assessed by both authors. Of these, nine met our inclusion criteria (noted above) and have been included in the review. Exclusion criteria sought to exclude articles that focused on breastfeeding, eating disorders were school-based or obesity treatment interventions.

Figure 1.

Search strategy.

Results

Summary of included studies

The nine included studies were delivered through a variety of settings (family/home, group, primary care, pre-school/childcare and mixed settings). Most studies involved multi-approach interventions and were conducted in the USA. Many studies targeted socio-economically at-risk families through well-established health service infrastructures. While the designs of studies varied substantially and all had some methodological weaknesses, all showed some level of effectiveness on at least one obesity-promoting behaviour in young children.

Description of key studies

Included studies have been grouped by the settings in which they were conducted (family/home, group, primary care, pre-school/childcare and mixed settings), and are reported below. Further details regarding these studies are presented in Tables 1–5.

Table 1.  Family-/home-based interventions
StudySampleInterventionOutcomesEvaluation
  1. BMI, body mass index; n/a, not available; NS, not significant; RCT, randomized control trial; SES, socio-economic status.

Harvey-Berino & Rourke (19) USA & CanadaAge: 9 months−3 years.
No.: 20 control, 23 intervention.
Retention: 93%, n = 40 (two intervention families lost to follow-up and one intervention family with incomplete follow-up data).
Demographics: Native American families, average of 13 years’ maternal education.
Inclusion criteria: (a) child aged between 9 months and 3 years old; (b) child was walking; (c) mother BMI over 25 kg m−2; (d) mother agreed to keep all treatment appointments.
Focus: physical activity and healthy eating.
Study design: RCT.
Data collections: baseline (week 0) and post-intervention (week 16).
Control condition: 16 weekly home visits by indigenous peer educator. Parenting support programme focusing on general parenting skills.
Intervention: 16 weekly home visits by indigenous peer educator. Parenting support programme focusing on how improved parenting skills could facilitate the development of appropriate eating and physical activity behaviours in the children.
Anthropometry: weight-for-height z-scores decreased in the intervention group and increased in the control group (−0.27 vs. 0.31, P = 0.06). No difference in proportions of children exceeding the 85th or 95th percentile. No difference in maternal BMI.
Dietary intake: total energy intake decreased for intervention but not control group children (−39.2 vs. 6.8 kcal kg−1 per day, P = 0.06). No difference for children’s fat intake or for maternal dietary intake.
Physical activity/sedentary behaviour: no differences observed for child or maternal physical activity.
Other relevant outcomes: mothers in the intervention group reduced their use of restrictive child feeding practices (−0.22 vs. 0.08, P < 0.05). No other differences were observed.
General comments: showed the home-visiting model to be a feasible means of delivering interventions. Evidence of effectiveness for modifying relative weight and caloric intake in children but appeared to have no impact on children’s fat intake or physical activity, or on maternal measures. Focused on Native American families.
Strengths: good compliance (all but two mothers completed all 16 sessions), good outcome measures, intervention delivered by single peer educator.
Weaknesses: small sample, no follow-up beyond the end of the intervention.
Worobey et al. (20) USAAge: mean 27 months.
No.: 87.
Retention: 69%, n = 60 with complete information.
Demographics: mainly Hispanic families (74%).
Inclusion criteria: (a) child <72 months; (b) income ≤250% of federal poverty level; (c) at least one of the at risk circumstances (child identified for health risk, identified as high risk prenatally, or psychosocial reasons).
Focus: healthy eating.
Study design: pre-/post-test intervention without control. Assessment of the nutrition education component of a broader child health programme (Prevention-Oriented System for Child Health).
Data collections: baseline and approximately 8 months later.
Control condition: none. Intervention: individualized home visits (weekly, monthly or bimonthly dependent on needs) to deliver child nutrition education to parents. Topics focused on increasing iron intake, vitamin supplementation, increasing water and juice consumption, reducing high-fat snacks and increasing healthy snacks.
Anthropometry: n/a.
Dietary intake: children’s total caloric intake decreased by approximately 17% (1698– 1411 kcal, P < 0.05). No change in macronutrient proportions. Vitamin supplement use increased (60–87%, P < 0.01).
Physical activity/sedentary behaviour: n/a.
Other relevant outcomes: none.
General comments: showed individualized child nutrition education for parents has the ability to decrease children’s caloric intake, but unclear whether a more healthy diet was consumed post-intervention. Primarily designed to reduce lead exposure so many of the messages/outcomes may not be relevant to other populations. Focused on socio-economically disadvantaged families.
Strengths: no specific strengths.
Weaknesses: no control group. Single 24-h dietary recall. Outcomes for key messages not reported. Unclear whether single or multiple interventionists used. Individually tailored so topics covered and visit frequency was not consistent. Other aspects of the child health programme may have impacted the outcomes. Appears to be no longer-term follow-up.
Wardle et al. (21) UKAge: 34–38 months.
No.: 45 control, 48 information intervention, 50 exposure intervention.
Retention: 98%, n = 140 (one control and two exposure group families withdrew).
Demographics: mainly Caucasian families (74%), 68% primary carers had been engaged in full-time education beyond 20 years of age.
Inclusion criteria: none stated.
Focus: increasing liking and consumption of vegetables.
Study design: RCT.
Data collections: baseline (0 week) and post-intervention (2 weeks).
Control condition: wait list.
Intervention: Information – parents were informed about the ‘5-a-day’ fruit and vegetable recommendations and given a leaflet with advice and suggestions for increasing children’s fruit and vegetable consumption. They were invited to ask questions about healthy eating. Exposure – parents instructed to offer their child a taste of their target vegetable (low-preference) every day for 14 consecutive days. Parents provided with techniques to encourage tasting and told importance of not offering rewards for consumption. Daily ‘vegetable diary’ provided for parents to record their experiences and for children to record their liking for the vegetable using stickers.
Anthropometry: n/a.
Dietary intake: liking of target vegetable increased more in exposure than control (P < 0.01) or information (P < 0.001) groups. Preference ranking of target vegetable increased more in exposure than control (NS) or information (P < 0.05) groups. At follow-up, 65% of exposure group children rated target vegetable as liked and almost 30% as most liked (preferred). Voluntary consumption of the target vegetable at the taste test increased significantly only in the exposure group (47–77% of children).
Physical activity/sedentary behaviour: n/a.
Other relevant outcomes: 17/20 interviewed parents thought the exposure advice was useful. 11/20 had used the method with other foods. 7/10 exposure group parents interviewed thought there had been a lasting effect on their child’s liking of the target vegetable. Duration of the exposure method was criticized by interviewed parents.
General comments: this parent-delivered intervention demonstrated improved liking, preference and consumption of a target vegetable by repeated exposure.  The sample was predominantly high-SES families which may limit its generalizability. Strengths: a relatively simple parent-delivered intervention that could be widely implemented for low cost.
Weaknesses: there was poor compliance in the exposure group with 29% completing less than 10/14 tastings. There was no quantitative follow-up beyond the 2-week intervention period to assess sustainability of impact. The use of the ‘vegetable diary’ may have constituted a reward and hence impacted the results.
Table 2.  Group-based interventions
StudySampleInterventionOutcomesEvaluation
  1. n/a, not available; PA, physical activity.

McGarvey et al. (22) USA Fit WICAge: 3 years (0.65 SD).
No.: 151 control, 185 intervention.
Retention: 65% of intervention group; 43% of control group evaluated at 1-year follow-up.
Demographics: parents attending the Special Supplemental Nutrition Program for Women, Infants and Children (WIC) clinics. 59%≤high school education, 54% Hispanic, 17% white, 16% African–American.
Inclusion criteria: attendees at the WIC programme.
Focus: six key dietary and physical activity behaviours (described under ‘intervention’).
Study design: prospective, pre-test post-test, non-randomized control group design.
Data collections: baseline (week 0) and post-intervention (1 year).
Control condition: participation in regular WIC programme, six educational groups and two individual contacts.
Intervention: (a) six educational groups and two individual contacts promoting (1) increase PA, (2) monitor mealtime behaviour, (3) limit household TV viewing, (4) drink water instead of sweetened beverages, (5) consume five fruits or vegetables daily, (6) increase family activities to promote fitness; (b) role modelling by WIC staff; (c) reinforcement by promotion of messages by collaborating organizations in the local community.
Anthropometry: n/a.
Dietary intake: intervention group parents were significantly more likely to report increased frequency of offering their child water rather than sweetened beverages (P = 0.005).
Physical activity/sedentary behaviour: intervention group parents were significantly more likely to report that they were involved in active play with their child (P = 0.009).
Other relevant outcomes: intervention parents were significantly more likely to report observing staff engaging in targeted health behaviours (P = 0.002). They were also significantly more likely to use community activity centres (P = 0.01).
General comments: showed the existing health service delivery model to be a feasible means of delivering interventions. The intervention showed evidence of effectiveness for modifying parent behaviours regarding offering of water over sweetened drinks and taking opportunities to engage in active play, however, did not affect four other targeted behaviours. It focused on socio-economically disadvantaged families.
Strengths: moderately good retention of participants. Intervention delivered in existing infrastructure. Followed over a 1-year period.
Weaknesses: measures used were not validated and involved only subjective self-report. Multiple points of intervention (potentially both a strength and a weakness) means that there is no capacity to understand which aspect (one or more) may have impacted on the behaviour changes. This has implications for the potential cost of interventions.
Table 3.  Primary care-based interventions
StudySampleInterventionOutcomesEvaluation
  1. n/a, not available; RCT, randomized control trial.

Johnson et al. (23) USAAge: n/a.
No.: convenience sample from Special Supplemental Nutrition Program for Women, Infants and Children (WIC). Pre-intervention: 10 204 WIC clients and 205 WIC staff. Post-intervention: 8977 clients and 211 staff.
Retention: n/a.
Demographics: 66% high school graduates, 59% non-Hispanic white, 5% non-Hispanic black or African–American, 6% American Indian or Alaska Native, 4% Asian, 2% Native Hawaiian or Pacific Islander, 25% Hispanic.
Inclusion criteria: (a) English- or Spanish-speaking; (b) with children.
Focus: to reduce TV viewing by reducing mealtime TV and displacing with increased physical activity.
Study design: data collected from convenience sample during a 3-week period pre-intervention and from a different convenience sample in another 3-week period 6 months after the intervention commenced.
Data collections: baseline (week 0) and post-intervention (6 months).
Control condition: no control group.
Intervention: individual handouts and education approaches were based on client’s stage of change. WIC staff encouraged to use materials to accommodate local needs, to adopt healthy behaviours themselves and to share the materials with local partners. Two main modules both emphasized reduction in TV viewing (a) family meals; (b) family physical activity. Modules included background materials, staff-training materials, banners, posters, interactive handouts for clients, detailed plans for group sessions and other supportive materials (provided in English and Spanish).
Anthropometry: n/a.
Dietary intake: n/a.
Physical activity/sedentary behaviour: 2 or less hours of TV viewing per day was reported by 64% of WIC clients at baseline and 71% at 6 months (P < 0.001). The portion of non-high school graduates who watched 2 or less hours per day increased 9% (from 58% to 67% at post-test), whereas there was a 5% increase for high school graduates (P < 0.001). 82% of white staff reported watching 2 or less hours per day compared with 59% of non-white staff.
Other relevant outcomes: at baseline, 65% of respondents reported that they do not usually or never watch TV during meals; after 6 months, 69% reported not watching TV during meals (P < 0.001). These respondents are more likely to be white (P < 0.001), have larger families (P < 0.001), or have higher education levels (P < 0.001). 76% of white WIC staff reported that they seldom or never watch TV during meals in comparison with 59% of non-white WIC staff (P = 0.014).
General comments: this large population-based study provides some lower-level evidence that an intensive education programme delivered via clinic-based services may have some reducing television viewing times in socio-economically disadvantaged families.
Strengths: very large sample size. Followed over a 6-month period. Characterized those WIC participants most likely to be watching more than 2 h of TV and watching during mealtimes.
Weaknesses: no description can be given of individual level of intervention received. Between 7% and 82% of participants were exposed to messages in this intervention. No control group. Parental reports of child’s behaviour. Study compared TV viewing in February (winter; before the intervention) and August (summer; after the intervention); thus, differences may reflect existing seasonal patterns of viewing.
(a) Lagstrom et al. (24)
(b) Talvia et al. (25)
Finland STRIP – Special Turku Coronary Risk Factor Intervention Project
Age: (a) 8 months−4 years; (b) to 10 years of age.
No.: 522 control, 540 intervention (1062 children from 1054 families).
Retention: not reported.
Demographics: general population sample.
Inclusion criteria: child aged 7 months at recruitment, without severe illness.
Focus: reduction of fat intake and changes to fat profiles of child diets.
Study design: prospective randomized clinical trial.
Data collections: multiple.
Control condition: families were seen biannually and did not routinely receive any detailed counselling focused on the risk factors of atherosclerosis.
Intervention: parents seen by a nutritionist at 1–3-month intervals until their child was 2 years old, then twice yearly. Individualized dietary counselling focused on appropriate fat reduction and ratios of saturated to poly- and mono-unsaturated fats.
Anthropometry: no growth differences.
Dietary intake: intervention children’s intake of saturated fat and cholesterol decreased throughout the study and remained below that of the control group throughout follow-up (P < 0.001). Note that this study continues today and thus dietary data continue to be collected and continue to show positive dietary differences between intervention and control children. Physical activity/sedentary behaviours: n/a.
Other relevant outcomes: none.
General comments: this well-designed study shows that individualized infancy-onset dietary counselling directed at parents (and later children) favourably influenced children’s diet, with reductions in total fat intake and maintenance of recommended dietary intakes of other nutrients (however, desired ratio of fats was not achieved). Early changes (to 4 years of age) persisted between 4 and 10 years of age.
Strengths: high-quality prospective RCT study.
Weaknesses: regular interaction and dietary recording may change control group behaviours.
Table 4.  Pre-school/childcare-based interventions
StudySampleInterventionOutcomesEvaluation
  1. BMI, body mass index; n/a, not available; RCT, randomized control trial; SES, socio-economic status.

Dennison et al. (26) USAAge: 2.5–5.5 years. No.: 83 control, 93 intervention. Retention: 44%: 34 control and 43 intervention from eight intervention and eight control centres. Demographics: predominantly Caucasian families with highly educated and employed parents. Inclusion criteria: none.Focus: reduce television viewing. Study design: cluster-RCT. Data collections: baseline and approximately 6 months later. Control condition: 20-min monthly interactive education sessions for 8 months and complementary staff and parental materials on general health and safety. Intervention: 7 x 20-min interactive education sessions (six weekly sessions and a final session 1 month later) plus complementary materials and suggested classroom activities for staff and parents on reducing television viewing.Anthropometry: no differences between groups. Dietary intake: n/a. Physical activity/sedentary behaviour: intervention children watched significantly less TV at follow-up than control children (adjusted mean difference −4.7 h week−1, P = 0.02) and significantly fewer exceeded 2-h daily viewing (19% vs. 41%, P = 0.007). No differences for video/computer games. Other relevant outcomes: no differences between groups in frequency of snacking during viewing, eating dinner as a family or watching television during dinner.General comments: this programme appeared to be effective in reducing children’s television viewing time and increasing the number of children meeting recommended viewing limits. Predominantly high SES. Strengths: it was delivered within an existing infrastructure using a single external interventionist. Long-term follow-up. Weaknesses: part of a broader health promotion programme so unclear whether other aspects of the programme may have impacted the results. There was a small sample and poor retention. Timelines for intervention and control differed.
Fitzgibbon et al. (27) USA Hip Hop to Health JrAge: 3–5 years. No.: 212 control, 197 intervention. Retention: 71%, n = 289 at 1-year follow-up; 73%, n = 300 at 2-year follow-up. Demographics: predominantly low-income African–American families. Inclusion criteria: child enrolled in Head Start Program.Focus: overweight prevention (through healthy eating and physical activity). Study design: cluster-RCT. Data collections: baseline, post-intervention, 1- and 2-year follow-ups. Control condition: 14 x 20-min weekly education sessions covering general health topics. Weekly newsletter for parents. Intervention: 3 x 40-min sessions each week for 14 weeks (20-min education+20-min aerobic activity). Weekly newsletter for parents plus homework assignment requiring 15–35 min per week.Anthropometry: no difference post-intervention but intervention children had lower BMI increases than control group children at 1- and 2-year follow-ups. Dietary intake: intervention children had lower percentage of calories from saturated fat than control children at 1-year follow-up, but no differences post-intervention or at 2-year follow-up. No differences in total fat intake or dietary fibre intake. Physical activity/sedentary behaviour: no differences. Other relevant outcomes: none.General comments: intervention appeared to be effective over the longer term in reduction BMI increases, but did not impact significantly on dietary and physical activity outcomes. Focused on socio-economically disadvantaged families. Strengths: follow up to 2-year post-intervention. Delivered within an existing infrastructure. Weaknesses: low retention rate. Poor parental compliance. The measures of physical activity may not have been sensitive enough to detect changes.
Table 5.  Mixed setting interventions
StudySampleInterventionOutcomesEvaluation
  1. n/a, not available.

Horodynski & Stommel (28)USA NEAT – Nutrition Education Aimed at ToddlersAge: mean 19.3 months. No.: 73 control, 62 intervention. Retention: 69% of intervention group and 71% of control group at 6-month follow-up. Demographics: families enrolled in Early Head Start programmes (designed to provide child and family development services to low-income pregnant women and families of infants or toddlers). Low-income families. 24% did not complete high school. 66% not employed. Predominantly Caucasian (84%) and rural families. 48% caregivers married.Focus: enhancing parent–toddler feeding practices and promoting healthy eating behaviours. Study design: quasi-experimental with repeated measures. Data collections: baseline (week 0), at 4 weeks (post-group lessons) and 6 months (post-individual reinforcement activities). Control condition: no treatment.Anthropometry: n/a. Dietary intake: n/a. Physical activity/sedentary behaviour: n/a. Other relevant outcomes: Child-Parent Mealtime Behaviour Questionnaire: no changes in toddler feeding self-regulation (over the three measurement points). Facts on Feeding Children Questionnaire: intervention group carers were more knowledgeable post-intervention than were controls (P < 0.01).General comments: this intervention highlights the fact that changes in knowledge do not necessarily result in changes in targeted behaviours, thus reinforcing the understanding that emphasis on skills development and descriptions of predictors of behaviours are paramount. While dietary data were collected, this was not reported. This study focused on socio-economically disadvantaged families.
Inclusion criteria: (a) low-income families (defined as family/household income at or below 100% of the Poverty Index); (b) toddlers aged 11–25 months at the time of intake to the study.Intervention: the NEAT intervention consisted of two components: (a) four group-based nutrition lessons: approximately 90 min long and offered weekly by trained nutrition instructors. Class sessions included such concepts as adult modelling of positive eating behaviours for toddlers, introducing new foods to toddlers, addressing parents’ concerns about what and how much their toddlers eat, and parenting skills; (b) 18 individually structured reinforcement activities: the reinforcement dealt with both cognitive and affective activities, presenting the caregiver with a scenario he or she would be likely to encounter in dealing with their toddler. During these activities, emphasis was given to the toddlers’ self-regulation, the division of responsibility in feeding and parent–child interactions.The Feeding Self-Efficacy Questionnaire: caregivers’ self-efficacy scores increased slightly from time 1 and time 3 in both the control and intervention groups (P < 0.05). Interviews and observations: the percentage of caregivers who left the TV on during mealtime reduced from 58% (time 1), to 54% (time 2), to 39% (time 3). Within the control group, these percentages remained stable 46% (time 1), to 48% (time 2), to 48% (time 3); however, the differences between the control group compared with the intervention group were not significant (P = 0.08).Strengths: moderately good retention of participants. Intervention delivered in existing infrastructure. Followed over a 6-month period. Assessments occurred at three time points. Weaknesses: two of the three measures used were not tested for their reliability – thus, we cannot be sure they were assessing what they intended to assess. No discussion of sample size and the capacity to detect change in the domains assessed.

Family/home setting

One randomized control trial (RCT) (19) directed at overweight Native American mothers [body mass index (BMI) > 25 kg m−2] of children aged 9 months−3 years compared generic parenting support (control group, n = 20) with parenting support focusing on improving children’s eating and physical activity behaviours (intervention group, n = 20). Mothers in both groups received weekly 1-h home visits for 16 weeks from an indigenous peer educator who delivered a parenting skills programme. For intervention group mothers, the focus was on using the parenting skills to develop healthy eating and exercise behaviours in their children. Positive changes were reported between baseline and 16-week follow-up for intervention compared with control group children with decreased weight-for-height z-scores (−0.27 vs. 0.31, P = 0.06), decreased total energy intake (−39.2 vs. 6.8 kcal kg−1 per day, P = 0.06) and reduced maternal use of restrictive child feeding practices (−0.22 vs. 0.08, P < 0.05). No differences were noted for children’s fat intake or physical activity, or for maternal BMI, diet or activity behaviours.

One pre-/post-intervention study (20) without control group assessed the nutrition education component of a broader intervention (Prevention-Oriented System for Child Health) in low-income, predominantly Hispanic (74%) families with children <72 months of age (mean 27 months at baseline) who had an identified risk factor. Of the 60 families in this study, approximately two-thirds were enrolled because of high child lead burden and one-third because of child psychosocial risk. A public health nurse developed an individualized parent-focused intervention strategy delivered through home visits (weekly, monthly or bimonthly dependent on needs). The nutrition education focused on increasing iron intake, supplementing diet with vitamins, increasing water and juice consumption, reducing high-fat snacks and increasing healthy snacks (e.g. yoghurt and fresh vegetables) plus topics focusing on reducing lead exposure. A 24-h dietary recall completed at baseline and approximately 8 months later indicated decreased total caloric intake (1698–1411 kcal, P < 0.05), but there was no change in macronutrient proportions, and vitamin supplement use increased (60–87%, P < 0.01).

One RCT (21) involved three conditions: control (n = 45), information (n = 48) and exposure (n = 50) to increase 34–82-month-old children’s liking for and consumption of a previously disliked vegetable. The sample contained predominantly Caucasian (74%), highly educated parents. All children undertook a preference taste-test in their home at baseline and post-intervention (2 weeks). The control group received no intervention. Information group parents were told it is recommended children consume five portions of fruits/vegetables daily and were provided with a leaflet containing advice and suggestions for increasing consumption. Exposure group parents were instructed to offer the target vegetable (vegetable ranked third most liked by the child in the pre-intervention taste-test of six vegetables) daily for 14 days. They were given suggestions for encouraging tasting and the importance of not offering a reward for tasting was stressed. They were also provided with a ‘vegetable diary’ for their child to record their liking for the vegetable after each tasting. There was poor compliance in the exposure group with 29% completing less than 10 tastings. The rated liking (‘like’, ‘OK’ or ‘dislike’) and preference (1–6) ranking of the target vegetable increased more for children in the exposure than information (P < 0.001 and P < 0.05 for liking and preference respectively) or control groups (P < 0.01 and not-significant for liking and preference respectively). At follow-up, 65% of the exposure group rated their target vegetable as liked and almost 30% as the most liked (preferred). Voluntary consumption of the target vegetable at the taste-test increased for the exposure group (47–77% of children, P < 0.01).

Group setting

One study (22) employed a pre-test–post-test, non-randomized control group design to assess the feasibility of implementing a parent-focused child obesity prevention programme in the established Special Supplemental Nutrition Program for Women, Infants and Children (WIC) clinics, and the capacity of a programme to promote targeted parental behaviours to prevent obesity in WIC clients’ children. Control group (n = 151) participants received the standard WIC programme over the course of a year (six attendances over the year of the study) and two individual sessions with a WIC nutritionist. Intervention group (n = 185) participants received the same number of contacts; however, education sessions were targeted to six specific behaviours to promote physical activity and to improve diet (i.e. increase physical activity, monitor mealtime behaviour, limit household television viewing, drink water instead of sweetened beverages, consume five fruits or vegetables daily and increase family activities to promote fitness). In addition, clinic staff were encouraged to participate in six ‘staff wellness challenges’ which ran concurrently with the group education classes. Emphasis was placed on the importance of clinic staff modelling the targeted behaviours. Health messages were reinforced in a variety of venues via provision of identical educational resources. Process evaluation confirmed the feasibility of implementing this programme in the established WIC clinics. It appeared that clinic staff sought to model targeted behaviours as requested, with intervention parents being significantly more (52% vs. 6%, P = 0.002) likely to report observing staff engaging in a variety of healthy behaviours. In addition, differences in the adoption of two of the targeted health behaviours, increased frequency of offering water over soft-drinks (P = 0.005) and increased likelihood of engaging in active play with child (P = 0.009) were found between the intervention and control groups.

Primary care setting

One large study (23) examined the feasibility of providing individual level education regarding television reduction to socio-economically disadvantaged families participating in the WIC programme. This intervention was diverse and there was not capacity within the study design to describe the dose of intervention received by individual participants. Participants included those who attended WIC services over a 3-week period (n = 10 204) prior to a global intervention which lasted for 6 months and separately, those who attended in a 3-week period after the global intervention had been delivered (n = 8977). It is possible therefore that the samples are unrelated to each other. The intervention focused on how to reduce television viewing times, and was delivered using key components of social marketing. Individual handouts and education approaches were based on client’s stage of change. Messages focused on reduction in television viewing by turning television off during family meals and by displacing television viewing with family physical activity. Resources used in this study can be found at http://Departmentswashington.edu/tvhealth/index.htm. Comparisons between the two independent samples assessed in this study (0 and 6 months) showed that the proportion of WIC clients reporting watching 2 or more hours of television per day increased significantly over the 6-month period (64.2–70.5%; P < 0.001), as did the proportion of respondents reporting that they never or rarely watch television during meals (65–69%; P < 0.001). A number of difficulties in study design, most particularly the potential impact of season on hours of television viewed (baseline data collected in winter, follow-up in summer), and the inability to compare individuals within groups across time support the view that more controlled studies need to be conducted to confirm the efficacy of this intervention.

A second pair of articles conducted in the primary care setting (24,25) provide high-quality evidence of the efficacy of individualized infancy-onset dietary counselling of parents (and later children) in promoting fat reduction in young children’s diets. The Special Turku Coronary Risk Factor Project for Babies study is a prospective randomized clinical trial which aims to reduce coronary heart disease risk factors in children over 1 year of age. It evaluates the impact of individualized and repeatedly given dietary counselling (parent) on fat intake and nutrient intake of children (i) aged 8 months to 4 years (24) and (ii) aged to 10 years (24,25). In this study, intervention group parents (n = 540) received repeated individualized dietary counselling by a dietitian, at 1–3-month intervals until the child was aged 2 years and then biannually after that. After the age of 7 years, separate sessions were organized for the child and the parents. Nutrition counselling was aimed at the reduction of the child’s saturated fat intake and the promotion of an unsaturated to saturated fatty acid ratio of 2:1, while maintaining adequacy of all other nutrients. The control group parents (n = 522) were seen biannually until the child was aged 7 years and thereafter, only once a year. Control families did not routinely receive any detailed counselling focused on the risk factors of atherosclerosis. Beyond the recommendation that children over 1 year of age should have milk with at least 1.5% fat, no detailed suggestions concerning the quality or quantity of fat were given, and dietary issues were discussed only briefly. At both 4 and 10 years of age, the intervention children consumed significantly less energy as fat than did control children (P < 0.001). These reductions in fat intakes were achieved with maintenance of recommended daily intakes of other nutrients.

Pre-school/childcare setting

One study reported a cluster-RCT (26) of an intervention aiming to reduce children’s television viewing within a broader health promotion intervention. Sixteen pre-school/childcare centres that enrolled 3–5-year-old children were involved. The intervention group (n = 43) received seven 20-min interactive education sessions (delivered over 6 weeks with the final session a month later) at the pre-school/childcare centre plus complementary materials for staff and parents. Topics included the importance of reading and television-free family mealtimes, alternative activities to television viewing and television time budgets. The control group (n = 34) received eight monthly education sessions and complementary materials on health and safety. Anthropometric and television viewing measures were taken pre-intervention and approximately 6 months later. Intervention group children spent less time watching television than control children post-intervention (adjusted mean difference −4.7 h week−1, P = 0.02) and fewer exceeded 2-h daily viewing (19% vs. 41%, P = 0.007). There were no differences between groups in time spent playing video/computer games, BMI, frequency of snacking while watching television, eating dinner together as a family, or watching television during dinner.

One cluster-RCT (Hip-Hop to Health Jr) (27) aimed to prevent obesity in minority 3–5-year-old children. Twelve Head Start pre-schools servicing predominantly African–American families were randomized to the weight control intervention or a general health programme (control group). The weight control intervention involved three 40-min sessions per week for 14 weeks, delivered by early childhood educators. Sessions comprised 20 min learning about healthy eating or physical activity and 20 min of aerobic physical activity. Parents received a weekly newsletter containing complementary information, weekly homework assignments and twice weekly aerobics classes. The control group received a general health programme involving 20-min education sessions on general health topics once per week for 14 weeks and parents received a weekly general health newsletter. Outcomes were assessed at baseline (n = 409), post-intervention (14 weeks), 1 (n = 289) and 2 years follow-up (n = 300). While BMI did not differ between intervention and control children post-intervention, intervention children demonstrated less increase in BMI at the 1- (0.02 kg m−2 vs. 0.64 kg m−2, P = 0.002) and 2-year follow-ups (0.48 kg m−2 vs. 1.14 kg m−2, P = 0.008). At 1-year follow-up, intervention children had lower saturated fat intake than control group children. No differences between groups were observed for total fat intake, dietary fibre intake, exercise frequency, exercise intensity or television viewing time.

Mixed settings (home and clinic)

One study (28) used a quasi-experimental design with repeated measures to assess the effectiveness of the Nutrition Education Aimed at Toddlers intervention, delivered to caregivers of toddlers aged 11–36 months. This intervention aimed to improve caregiver–toddler interactions by empowering adults to become responsive to children’s verbal and non-verbal behaviours. Control group participants received no programme. Intervention group participants engaged in four 90-min group-based nutrition lessons over 4 weeks. These sessions focused on feeding toddlers and dealt with adult modelling of positive eating behaviours, introducing new foods, parents’ concerns about what and how much to feed, and on parenting skills. In addition, they were provided with 18 individualized structured reinforcement sessions over a 6-month period. These were delivered during scheduled weekly home visits. Evaluation by repeated measures of constructs concerned with child-carer mealtime behaviour, with carer nutrition knowledge and with carers’ self-efficacy in feeding found significant differences between intervention and control only on scores of nutrition knowledge (P = not provided). Intervention group parents had significantly increased nutrition knowledge post-intervention when compared with control (P < 0.01). In addition, home observations of mealtime television watching revealed that while there was no reduction in television viewing during meals in the control group, this behaviour became progressively less likely, but not significantly so, in intervention homes (58% at baseline, 39% at follow-up; P = 0.08).

Discussion

This review evaluated the published literature regarding interventions which may positively influence young children’s body weight via changes to dietary intake, to physical activity and/or to sedentary behaviours. The focus was on interventions involving children under the age of 5 years, an important group given the current emphasis by government and others on these early years.

It is perhaps surprising that at a time when the prevalence of childhood fatness is rapidly increasing, and in an environment where early childhood is considered to be a key time for the development of health behaviours, that relatively few studies were found that addressed these issues. Traditionally, the literature in this area has focused upon interventions targeting school-aged children. However, growing acknowledgement of the importance of the early years as a time for the development of both overweight and its precursor behaviours is evidenced by the fact that most of the studies identified in this review have been published since 2003. The lack of identification of interventions in this review does not necessarily mean work with children under 5 years is not being conducted. It is possible that programmes seeking to effect dietary, physical activity and/or sedentary behaviour change in this age group have been, or are currently being, conducted but have not yet been published in the peer-reviewed literature.

Nine studies inform this review. Of these, the majority were studies from the USA and many targeted socio-economically at-risk families through well-established health service infrastructures such as the WIC. While these studies varied widely in their objectives, designs and subsequent quality, overall most studies were able to show some level of effectiveness on some obesity-promoting behaviours in children. This finding is encouraging; however, it is important to recognize that publication bias is likely to favour interventions that show positive impacts and thus that less successful studies simply may not be reported in the peer-reviewed literature.

While the designs of studies varied substantially, some common themes emerged which may be useful in considerations of the most effective strategies by which to support healthy weight outcomes in young children. Many studies showed some success in changing some behaviours in children [e.g. reductions in fat intake (24), reductions in total energy intake (19), reductions in television viewing (26)]. Thus, we are reminded that families can be supported to make effective changes which seem likely to influence the propensity for fatness in their children.

Although the mode of delivery and the focus of messages differed across studies, most interventions can be classified as high-intensity interventions; that is, parents were seen many times and in a range of settings by their health service provider or the interventionist. These settings allowed for the repetition of targeted messages, often through different modes (tailored individual feedback, group education settings, use of pamphlets and posters in community facilities). These high-intensity interventions resulted in small but potentially meaningful behaviour changes; however, we found no evidence to support the premise that low-level interventions would result in similar changes. It is important to acknowledge that most studies were underpinned by social behavioural theory and were thus interested to impact not only on knowledge, but on parental skills and competencies.

A number of methodological issues limit our capacity to accurately interpret the impact of these studies. First, few studies mention sample size calculations; hence, we are unable to ascertain whether the sample size was large enough to show effect. Inclusion of such information in future published studies would assist in the development of new intervention programmes. Second, it is acknowledged in the literature that the measurement of diet, physical activity and sedentary behaviours is difficult; however, the limitations of these measures are rarely discussed in individual studies. Finally, it is interesting to note that none of the studies report data on the cost-effectiveness of the intervention programmes evaluated. Given the rigorous/high-intensity nature of many of the interventions reported, it is likely that they were also high-cost, a factor that is likely to limit more global adoption. However, given the potential impact of obesity prevention on healthcare costs, it would be useful to provide some projections of the anticipated cost savings vs. cost outlays.

Differences in ethnicities and socio-economic position are likely to limit the generalizability of the studies reported in this review. For example, studies that have had positive outcomes in one ethnic group may not be effective in another. Having said this, there are likely to be points of overlap in groups that are disadvantaged in any community, for example, single mothers, and those who have complex social histories, such as indigenous communities. A further limitation on potential generalizability relates to the differing health service infrastructure within and between countries. The capacity to replicate these studies in other settings may be limited.

These studies support, at a range of levels, the premise that parents are receptive to and capable of some behavioural changes that may promote healthy weight in their young children. The small quantity of research identified provides promising insights regarding interventions targeting young children and their parents and heralds the need, particularly given the potential for early intervention to have long-lasting impacts on individual and population health, to build in a substantial way upon this evidence base.

Conflict of Interest Statement

No conflict of interest was declared.

Acknowledgements

This review was commissioned and funded by the Centre for Health Promotion (SA). Karen Campbell is funded by the National Heart Foundation of Australia. Kylie Hesketh is funded jointly by National Heart Foundation of Australia and the National Health and Medical Research Council. We would like to thank Anna Sztendur for her considerable contribution to data collection.

Ancillary