Little evidence exists on whether effects of an early obesity intervention are sustainable.
To assess the sustainability of effects of a home-based early intervention on children’s body mass index (BMI) and BMI z score at 3 years after intervention.
Design, Setting, and Participants
A longitudinal follow-up study of the randomized clinical Healthy Beginnings Trial was conducted with 465 participating mothers consenting to be followed up at 3 years after intervention until their children were age 5 years. This study was conducted in socially and economically disadvantaged areas of Sydney, Australia, from March 2011 to June 2014.
No further intervention was carried out in this Healthy Beginnings Trial phase 2 follow-up study. The original intervention in phase 1 comprised 8 home visits from community nurses delivering a staged home-based intervention, with one visit in the antenatal period and 7 visits at 1, 3, 5, 9, 12, 18, and 24 months after birth.
Main Outcomes and Measures
Primary outcomes were children’s BMI and BMI z score. Secondary outcomes included dietary behaviors, quality of life, physical activity, and TV viewing time of children and their mothers.
In total, 369 mothers and their children completed the follow-up study, a phase 2 completion rate of 79.4% (80.9% for the intervention group and 77.7% for the control group). The differences between the intervention and control groups at age 2 years in children’s BMI and BMI z score disappeared over time. At age 2 years, the difference (intervention minus control) in BMI (calculated as weight in kilograms divided by height in meters squared) was −0.41 (95% CI, −0.71 to −0.10; P = .009), but by age 5 years it was 0.03 (95% CI, −0.30 to 0.37). No effects of the early intervention on dietary behaviors, quality of life, physical activity, and TV viewing time were detected at age 5 years.
Conclusions and Relevance
The significant effect of this early life home-visiting intervention on child BMI and BMI z score at age 2 years was not sustained at age 5 years without further intervention. Obesity prevention programs need to be continued or maintained during the early childhood years.
There is an upward trend in the prevalence of overweight or obese preschool-age children (age range, 0-5 years), from 4.2% in 1990 to 6.7% in 2010, that is occurring worldwide.1 Obesity in young children may progress to obesity in later childhood2,3 and potentially into adulthood.4-7 It is increasingly argued that childhood obesity prevention should begin in the early years.8 Tackling childhood obesity now represents an important opportunity to reduce the effect of heart disease, diabetes mellitus, and other serious preventable chronic diseases in the future while immediately improving the health of children.9 The importance of early obesity intervention has been emphasized by the increased number of intervention research studies10-15 in recent years.
Some promising results from recent randomized clinical trials suggest that prevention interventions directed to mothers in the first years of their child’s life can reduce bottle use and energy intake,13 support the development of healthy infant feeding practices,12 lead to reductions in sweet snack consumption and TV viewing,11 and reduce the mean body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared) by 0.38 (95% CI, 0.08-0.68) of children at age 2 years.10 The Healthy Beginnings Trial (HBT) phase 1 was the first randomized clinical trial to show the effectiveness of an early childhood obesity intervention delivered in the first 2 years of life.10,16 The provision of healthy infant feeding education, with increased active play and reduced TV viewing time, remains central to these early interventions. A 2014 systematic review of the effect of interventions to prevent obesity or improve obesity-related behaviors in preschool-age children (age range, 0-5 years) concluded that early intervention effects are modest but promising.15 The authors of the review called for high-quality studies with longer-term follow-up. Indeed, without a long-term follow-up, the sustainability of early intervention effects remains unknown.
To address this evidence gap, children who participated in the HBT phase 1 were followed up for 3 years after intervention (ie, from age 2-5 years) to ascertain whether the benefits at age 2 years were sustained at age 5 years and to examine the quality of life of child participants. The HBT phase 1 used a home-based early intervention designed to improve family and behavioral risk factors for childhood obesity. The rationale and method for this follow-up study (HBT phase 2) were reported in a published protocol17 before study commencement and involved (1) a longer-term follow-up of the cohort of children at ages 3.5 and 5 years and (2) an economic evaluation of the intervention, which has been reported elsewhere.18
Box Section Ref ID
At a Glance
Prevention interventions directed to mothers in the first years of their child’s life can improve infant feeding practices and reduce mean BMI
Little evidence exists on whether effects of an early obesity intervention are sustainable
To address this evidence gap, a further follow-up of the Healthy Beginnings Trial was conducted 3 years after intervention
No effects of the early intervention on children’s mean BMI and other obesity-related behaviors were detected at age 5 years
In 2007, we commenced the HBT phase 1 with 667 first-time mothers (337 intervention and 330 control) to investigate the effectiveness of a home-based early intervention primarily on children’s BMI at age 2 years and, secondarily, on dietary behaviors, physical activity, and TV viewing time of children and their mothers. The intervention comprised 8 home visits from community nurses delivering a staged home-based intervention, with one visit in the antenatal period and 7 visits at 1, 3, 5, 9, 12, 18, and 24 months after birth. It was delivered from 2007 to 2010 in one of the most socially and economically disadvantaged areas of Sydney, Australia.10,16,19
Participants and Consent of the HBT Phase 2
In 2011, we started the follow-up HBT phase 2 study to investigate the long-term effects of the HBT intervention. We asked 497 participating families who had completed phase 1 (255 intervention and 242 control) whether they were willing to participate in a further 3-year study with no intervention from age 2 years onward. In total, 465 of 497 (93.6%) consented to follow-up. No further intervention was carried out in this study from March 2011 to June 2014. Written parental consent was obtained by mail or through home visits by 2 research assistants. Phase 2 baseline was taken as the home visit when the child was aged 2 years. Ethics approval was obtained for the study from Sydney Local Health District Research Ethics Review Committee X10-0312.
Two research assistants who were not involved in the HBT phase 1 collected the follow-up data and were blinded to intervention or control status. To ensure measurement consistency, they were trained by experienced early childhood nurses in anthropometry measures and semistructured interviews in the home setting. Data were collected when the children were ages 3.5 and 5 years. At each visit, the child’s anthropometric measurements were undertaken, and a face-to-face questionnaire interview was conducted with the mother.
The primary outcomes were children’s BMI and BMI z score. Height and weight were measured using standard techniques in the home using a portable stadiometer and electronic scales. Body mass index was calculated,2 and BMI z score was calculated using a software program (AnthroPlus; World Health Organization).20 We also categorized BMI as overweight, obese, or not overweight or obese based on internationally accepted criteria.21
For weight, measurements were taken using digital scales (TI1582136K; http://www.wedderburn.com.au) by a research assistant from children wearing lightweight clothes and no shoes. The measures were recorded to the nearest 0.1 kg.
For measuring height we used a portable stadiometer with a vertical backboard and movable headboard, the child stood erect against the backboard, and the back of the head, shoulder blades, buttocks, and heels made contact with the backboard of the stadiometer. Two measurements were taken by a research assistant and recorded to the nearest 0.1 cm. A third measure was taken if the first 2 measurements differed by 0.5 cm or more, and the mean of these 2 or 3 values was calculated.
The secondary outcomes included dietary behaviors, quality of life, physical activity, and TV viewing time of children and their mothers according to a published research protocol.17 We used a validated set of short questions22,23 to assess young children’s dietary patterns and physical activity. The TV viewing behavioral questions were based on those used in phase 1 of the HBT.10 Indicators of dietary patterns included the following: questions on daily servings of fruits and vegetables; quantity of specified beverages, including sugary drinks; frequency of snack foods; and eating behaviors, including eating when watching TV or in front of the TV. Indicators of physical activity and TV viewing time included mother-reported outdoor playtime and time spent on small screens on a typical weekday and weekend day.
Children’s health-related quality of life was assessed using parent proxy reports of a pediatric quality-of-life instrument (PedsQL; http://www.pedsql.org/) comprising 4 generic core subscales pertaining to physical, emotional, social, and school functioning. At 3.5 years, we used the parent report for toddlers (age range, 2-4 years). At 5 years, we used the parent report for young children (age range, 5-7 years).
Secondary outcomes also included the mothers’ eating habits and physical activity, as well as TV viewing time. Information on these outcomes was collected using existing survey instruments, including the New South Wales Adult Health Surveys,24 which were also used in phase 1 of the HBT.10
At phase 1 baseline, we collected sociodemographic data, including age, employment status, educational level, marital status, language spoken at home, and country of birth of mothers using the standard New South Wales Population Health Survey questions.24 At phase 2 baseline, we re-collected information on employment status, educational level, and marital status.
To maintain participants’ interest in the follow-up study, several retention strategies were used, including the following: (1) sending thank-you cards at the end of the HBT phase 1, (2) sending Christmas or New Year greeting cards, (3) obtaining contact numbers of participants’ relatives (updated at each contact visit), (4) providing brief feedback to participants on their child’s weight and height, and (5) text message reminders for follow-up visits. In addition, small gifts (less than US $8) of appreciation were sent to all participants after completing the 3.5-year data collection.
Using χ2 tests, we compared characteristics of the study participants between the intervention and control groups at baseline of the HBT phase 2 (at age 2 years) and ages 3.5 and 5 years. We also compared characteristics of those remaining in the study and those lost to follow-up at ages 3.5 and 5 years.
The primary outcome variables, BMI and BMI z score, were treated as continuous variables. Secondary outcomes, including eating habits and dietary behaviors, physical activity and outdoor play, and TV viewing time, were dichotomized as for phase 1 of the HBT.10 For example, cut points were based on the national guidelines for physical activity for children and adults, while the median intake of fruits and vegetables of the study participants was used.
To assess sustained short-term and long-term effects of the intervention on the primary outcomes (child BMI and BMI z score) and secondary outcomes (dietary behaviors, physical activity, and TV viewing time), outcomes were compared between the intervention and control groups at ages 3.5 and 5 years. For continuous variables, including BMI and BMI z score, the means were compared using t tests and then using multiple linear regressions to adjust for baseline differences (at 2 years). For categorical variables, χ2 tests were used, followed by multiple logistic regressions to adjust for baseline differences (at 2 years). The regression analyses were limited to those with 3.5-year or 5-year follow-up data. For the PedsQL quality-of-life subscales, Tobit models were used, specifying right censoring at the maximum subscale score of 100. All models were adjusted for mother’s marital status and employment status, which were found to be significantly different (P < .05) between the intervention and control groups at the beginning of the phase 2 study. Other variables (eg, child sex or when the mother was recruited) were assessed for their confounding effect and then dropped from the models. Adjusted regression coefficients (β levels) were calculated with 95% CIs and P values.
Data were analyzed using statistical software (Stata, version 12; StataCorp LP). Intent-to-treat principles were used in all primary analyses (ie, we analyzed the results based on participants’ initial group allocation at baseline of the HBT phase 1).
Of 497 participants who completed the HBT phase 1, 465 (236 intervention and 229 control) agreed to participate in this follow-up study, a consent rate of 93.6% (Figure 1 and Figure 2). Fifty mothers were lost to follow-up at 3.5 years and a further 46 mothers at 5 years. In total, 369 mothers (191 intervention and 178 control) remained at 5 years, giving an overall retention rate of 79.4% over the 3-year period. Of 96 lost to follow-up in phase 2, a total of 45 were from the intervention group, and 51 were from the control group. The main reasons for loss to follow-up were similarly distributed across both groups (Figure 1).
Characteristics of the Study Participants
Table 1 lists characteristics of the study participants at phase 2 baseline and follow-up of the children at ages 3.5 and 5 years. At baseline, most characteristics were similar in the 2 groups, but there were higher proportions of married or de facto partner and employed (93.9% and 64.2%, respectively) mothers in the control group than in the intervention group (88.1% and 57.4%, respectively) (P = .03 and P = .02, respectively). Comparisons of characteristics were also made between those retained in the study and those lost to follow-up. Regardless of group, mothers who were lost to follow-up were more likely to be younger, unmarried, and unemployed and have lower annual household income and lower educational levels.
At phase 2 baseline, the success of the phase 1 intervention meant that the mean BMI was lower in the intervention group (16.57) than in the control group (16.98). The difference (intervention minus control) was −0.41 (95% CI, −0.71 to −0.10) (P = .009) (Table 2). The mean BMI z score was also lower in the intervention group (0.55 vs 0.84, P = .006). However, there were no differences in these variables between the intervention and control groups when children were aged 3.5 years or 5 years. This was true when either unadjusted (eg, BMI difference at age 5 years was 0.03; 95% CI, −0.30 to 0.37) or after adjusting for baseline differences using multiple linear regression (eg, BMI difference at age 5 years was 0.27; 95% CI, −0.01 to 0.56). Figure 2 and Figure 3 show BMIs and BMI z scores for the groups at 2, 3.5, and 5 years based on observed values.
Body mass index or BMI z score at 2 years was positively associated with BMI (β level, 0.52; 95% CI, 0.42-0.61; P < .001) or BMI z score (β level, 0.52; 95% CI, 0.44-0.62; P < .001) at 3.5 years after controlling for maternal marital status and employment status at baseline. Similarly, BMI or BMI z score at 2 years was also positively associated with BMI (β level, 0.52; 95% CI, 0.44-0.60; P < .001) or BMI z score (β level, 0.51; 95% CI, 0.43-0.58; P < .001) at 5 years.
At 3.5 years, 27.4% of the intervention group and 29.6% of the control group were categorized as overweight or obese (P = .34). At 5 years, 22.6% of the intervention group and 21.7% of the control group were categorized as overweight or obese (P = .73).
As summarized in Table 3, χ2 tests and multiple logistic regression analyses found no differences between groups in dietary behaviors, physical activity, or TV viewing time of children and their mothers at 3.5 and 5 years. Children’s consumption of fruits and vegetables remained steady in both groups at 3.5 and 5 years, as did children’s physical activity and TV viewing time.
Quality of life, as measured by the median (interquartile range) total subscale score of the PedsQL, was the same for intervention and control participants at 3.5 years. These median (interquartile range) values at 3.5 years were 91.7 (85.7-95.2) for the intervention group and 91.7 (87.5-95.2) for the control group (P = .89). The median (interquartile range) values at 5 years were 93.5 (89.1-97.2) for the intervention group and 93.5 (88.9-97.2) for the control group (P = .92). Similarly, there were no differences between the 2 groups at 3.5 and 5 years in their physical, emotional, social, and school functioning subscale median scores.
Principal Findings of the Study
This follow-up study shows that the differences in children’s BMI and BMI z score between the intervention and control groups at age 2 years disappeared over time. By age 5 years, no difference was detected nor were any effects of the early intervention on dietary behaviors, quality of life, physical activity, and TV viewing time detected across time.
It is clear that the initial significant effect of this early life home-visiting intervention was not sustained without further intervention with the participants. The design of health promotion interventions and policies aimed at reducing the prevalence of childhood obesity needs to be considered within a life course framework, requiring frequent interventions addressing multiple factors at various stages of life.25,26 These factors can range from the social or built environment (macroenvironment) through behavior and physiology.26 Our results also reaffirmed that overweight or obesity is already established in the early years of life. Sustainable intervention programs are essential to address the social, cultural, economic, and environmental factors that are associated with childhood obesity over all life stages.
First, without the findings of the HBT phase 2, it is unlikely that appropriate recommendations for early childhood prevention programs could have been made purely based on the results of the HBT phase 1. The effectiveness and sustainability of this early intervention program in the Australian context would remain unknown. Our results are consistent with findings from a recent meta-analysis27 of early obesity intervention programs that involve parents, which concluded that interventions with parental involvement are effective at short-term follow-up but that effects were not retained in the long run.
Second, a Cochrane review of interventions for preventing obesity in children28 indicated that there were promising findings in those aged 0 to 5 years, particularly for interventions conducted in home or health care settings. At the same time, the research gaps evident from that review included the need to study the effectiveness of interventions for this age group. Our phase 2 study is the first to date to address this research gap of a lack of long-term follow-up studies. Other (possibly differently designed) intervention strategies are needed with longer-term follow-up.
Third, a systematic review of effective strategies for reducing screen time among young children29 has identified several research gaps, one of which was limited long-term follow-up data (>6 months). The review called for more work to be done to understand the potential for interventions in children younger than 6 years and in low-income and racial/ethnic or minority participants.
Unanswered Questions and Future Research
Early life interventions that target modifiable maternal and child risk factors hold significant promise, but potentially sustained benefits beyond the intervention period are possibly moderated by obesogenic environments. Long-term follow-up by similar studies is required to build a better understanding of whether weight-related behavioral changes attributed to intensive early child obesity interventions are sustainable. Potentially, the success of interventions such as the HBT diminishes across time, presumably because obesogenic factors within communities that put families and young children at risk of engaging in weight-related behaviors remain in place. Future research needs to explore early interventions beyond individual behaviors and family conditions that are related to overweight and obesity.
Strengths and Limitations
The HBT phase 1 study demonstrated the effectiveness of a home-based early childhood obesity intervention, with a mean reduction in BMI of 0.38 for children at age 2 years.10 The HBT phase 2 study, a further 3-year follow-up of the phase 1 participants, has generated new knowledge about the sustainability of the early intervention. Evidence produced by this study can enable policy makers and health care professionals to refocus on the life course approach to obesity prevention with a long-term commitment to tackle the obesity epidemic. A sustained and continuing support to mothers with young children is vital, particularly in socially and economically disadvantaged communities.
The study plan was transparent with a published research protocol.17 The participation rate was high, with 93.6% of families who remained at the end of the HBT phase 1 being recruited for phase 2. The retention rate of 79.4% over the 3-year follow-up period after intervention was reasonable. Characteristics of the study participants were similar in the 2 groups except for marital status at the end of this study. The main outcome measures were assessed using validated, well-developed, and widely used population survey tools.21-23 Blinding to treatment allocation was applied for data collection, data entry, and analysis.
However, the study has several limitations. First, conducting an overall 5-year longitudinal study of the HBT phases 1 and 2 provides challenges for maintaining the study cohort.30 As a result, loss to follow-up may have led to incomplete study results and may have biased the results, although the main reasons for loss to follow-up were similar across both groups. Second, the study was limited by nonblinded participating mothers and self-reported behavioral measures. Third, we did not assess many other potential factors (eg, environmental, school, and community) that may also have influenced mothers’ and children’s physical activity and dietary behaviors. Fourth, the results may have been substantially biased because only those participants who completed all of phase 1 and phase 2 were included. However, it is likely that if sustained effects were to be seen, they would be in this most highly engaged group. Nevertheless, with the analysis limited to this group, no detectable effect at age 3.5 or 5 years was noted.
In conclusion, the effect of this early life home-visiting intervention on child BMI was not sustained without further intervention. With a substantial proportion of young children being overweight or obese, there is a need for ongoing research to identify intervention programs.
Accepted for Publication: January 28, 2015.
Corresponding Author: Li Ming Wen, MD, MMed, PhD, Health Promotion Service, Sydney Local Health District, New South Wales, Level 9 N, King George V Bldg, Camperdown, New South Wales, Australia 2050 (email@example.com).
Published Online: April 20, 2015. doi:10.1001/jamapediatrics.2015.0258.
Author Contributions: Dr Wen had full access to all the data in this study and takes responsibility for the integrity of the data and the accuracy of the data analyses.
Study concept and design: Wen, Baur, Rissel.
Acquisition, analysis, or interpretation of data: Wen, Simpson, Xu, Hayes.
Drafting of the manuscript: All authors.
Obtained funding: Wen, Baur, Simpson, Hayes, Hardy, Rissel.
Administrative, technical, or material support: Wen, Baur, Simpson, Xu, Hayes, Hardy, Rissel.
Conflict of Interest Disclosures: None reported.
Funding/Support: This project is funded by grants 393112 and 1003780 from the Australian National Health and Medical Research Council. Dr Hayes was supported by grant 571372 from the Australian National Health and Medical Research Council.
Role of Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank Vicki Flood and the members of the steering committee for their advice and support. We also thank the project team, including Lauren Viney, Maxine Goodwin, Maria Domenico, Emma Wood, Karen Wardle, and Angela Balafas. We thank all the families for their participation in this study.
E. Global prevalence and trends of overweight and obesity among preschool children. Am J Clin Nutr
. 2010;92(5):1257-1264.PubMedGoogle ScholarCrossref
TJ. Contribution of early weight gain to childhood overweight and metabolic health: a longitudinal study (EarlyBird 36). Pediatrics
. 2009;123(1):e67-e73. doi:10.1542/peds.2008-1292.PubMedGoogle ScholarCrossref
RA, van Buuren
JC. Body size and growth in 0- to 4-year-old children and the relation to body size in primary school age. Obes Rev
. 2011;12(8):637-652.PubMedGoogle ScholarCrossref
CG. Rapid growth in infancy and childhood and obesity in later life: a systematic review. Obes Rev
. 2005;6(2):143-154.PubMedGoogle ScholarCrossref
SL. Racial/ethnic differences in early-life risk factors for childhood obesity. Pediatrics
. 2010;125(4):686-695.PubMedGoogle ScholarCrossref
et al. Relationship of childhood obesity to adult obesity: a 20-year longitudinal study from birth in Ishikawa Prefecture, Japan [in Japanese]. Nihon Koshu Eisei Zasshi
. 2003;50(12):1125-1134.PubMedGoogle Scholar
et al; National Institute of Child Health and Human Development Early Child Care Research Network. Identifying risk for obesity in early childhood [published correction appears in Pediatrics
. 2006;118(5):2270]. Pediatrics
. 2006;118(3):e594-e601. doi:10.1542/peds.2005-2801.PubMedGoogle ScholarCrossref
et al. Early-life prevention of non-communicable diseases. Lancet
. 2013;381(9860):3-4.PubMedGoogle ScholarCrossref
VM. Effectiveness of home based early intervention on children’s BMI at age 2: randomised controlled trial [published correction appears in BMJ
. 2013;346:f1650]. BMJ
. 2012;344:e3732. doi:10.1136/bmj.e3732.PubMedGoogle ScholarCrossref
et al. A parent-focused intervention to reduce infant obesity risk behaviors: a randomized trial. Pediatrics
. 2013;131(4):652-660.PubMedGoogle ScholarCrossref
A. Outcomes of an early feeding practices intervention to prevent childhood obesity. Pediatrics
. 2013;132(1):e109-e118. doi:10.1542/peds.2012-2882.PubMedGoogle ScholarCrossref
C. Bottle-weaning intervention and toddler overweight. J Pediatr
. 2014;164(2):306-12.e1, 2. doi:10.1016/j.jpeds.2013.09.029.PubMedGoogle ScholarCrossref
RL; Greenlight Study Team. “Greenlight study”: a controlled trial of low-literacy, early childhood obesity prevention. Pediatrics
. 2014;133(6):e1724-e1737. doi:10.1542/peds.2013-3867.PubMedGoogle ScholarCrossref
KJ, van der Pligt
et al. The impact of interventions to prevent obesity or improve obesity related behaviours in children (0-5 years) from socioeconomically disadvantaged and/or indigenous families: a systematic review. BMC Public Health
. 2014;14:779.PubMedGoogle ScholarCrossref
VM. Effectiveness of an early intervention on infant feeding practices and “tummy time”: a randomized controlled trial. Arch Pediatr Adolesc Med
. 2011;165(8):701-707.PubMedGoogle ScholarCrossref
et al. Healthy Beginnings Trial Phase 2 study: follow-up and cost-effectiveness analysis. Contemp Clin Trials
. 2012;33(2):396-401.PubMedGoogle ScholarCrossref
K. Economic evaluation of “healthy beginnings”: an early childhood intervention to prevent obesity. Obesity (Silver Spring)
. 2014;22(7):1709-1715.PubMedGoogle ScholarCrossref
JM. Early intervention of multiple home visits to prevent childhood obesity in a disadvantaged population: a home-based randomised controlled trial (Healthy Beginnings Trial). BMC Public Health
. 2007;7:76.PubMedGoogle ScholarCrossref
WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ
. 2000;320(7244):1240-1243.PubMedGoogle ScholarCrossref
et al. Reliability and validity of a short food frequency questionnaire to assess diet among 2-5 year olds. Obes Res Clin Pract
. 2010;4(suppl 1):S61.Google ScholarCrossref
LM, van der Ploeg
C. A validation study of assessing physical activity and sedentary behavior in children aged 3 to 5 years. Pediatr Exerc Sci
. 2010;22(3):408-420.PubMedGoogle Scholar
Centre for Epidemiology and Research. 2006 Report on Adult Health From the New South Wales Population Health Survey. Sydney: NSW Dept of Health, Australia; 2007.
et al. Evidence Update on Obesity Prevention: Across the Life-Course: Prepared for NSW Ministry of Health. Sydney, Australia: Physical Activity Nutrition Obesity Research Group; 2012.
EM. Lifecourse approach to racial/ethnic disparities in childhood obesity. Adv Nutr
. 2012;3(1):73-82.PubMedGoogle ScholarCrossref
HM, van Ijzendoorn
J, van der Veek
S. Interventions aimed at reducing obesity in early childhood: a meta-analysis of programs that involve parents [published online October 8, 2014]. J Child Psychol Psychiatry
. doi:10.1111/jcpp.12330.PubMedGoogle Scholar
E, de Silva-Sanigorski
et al. Interventions for preventing obesity in children. Cochrane Database Syst Rev
. 2011;(12):CD001871.PubMedGoogle Scholar
et al. Systematic review of effective strategies for reducing screen time among young children. Obesity (Silver Spring)
. 2012;20(7):1338-1354.PubMedGoogle ScholarCrossref
M. Tackling the challenges in conducting a 5-year longitudinal study of the Healthy Beginnings Trial. Health Promot J Austr
. 2014;25(1):65-66.PubMedGoogle ScholarCrossref