|Children and healthy eating: a systematic review of barriers and facilitators
|Thomas J, Sutcliffe K, Harden A, Oakley A, Oliver S, Rees R, Brunton G, Kavanagh J
This review evaluated the effectiveness of healthy eating interventions aimed at children aged between 4 and 10 years. The authors concluded that interventions can have a small, but significant, positive effect in increasing children's fruit intake by one-fifth and vegetable intake by nearly one-fifth of a portion per day. Their conclusions reflect the evidence presented and are reliably derived.
To evaluate the effectiveness of healthy eating interventions aimed at children. The following abstract focuses on this evaluation. In the same report, the authors explored intervention processes, and the views and experiences of children in the UK on this topic. The evaluations were subsequently combined in a cross-study synthesis.
Eligible studies written in English were sought from MEDLINE, EMBASE, Social Sciences Citation Index and PsycINFO (all from 1981 to 2001); CINAHL was searched from 1982 to 2001 and ERIC from 1985 to 2001; the search terms were reported. Specialist registers were also searched to 2001; these included Bibliomap, PrevRev, DARE, HealthPromis, the Cochrane Controlled Trials Register, the Cochrane Database of Systematic Reviews, and the Cochrane Heart Group trials register. Handsearching was carried out in Education and Health (1983 to 2002); Health Education Quarterly (1981 to 1996) and Health Education and Behavior (1997 to 2002). Reference lists were screened, and authors and relevant key organisations were contacted for further published and unpublished studies.
Study designs of evaluations included in the review
Studies employing a control or comparison group were eligible for inclusion in the review.
Specific interventions included in the review
Studies with the main focus on healthy eating were eligible for inclusion. The majority of the included interventions focused on promoting fruit and vegetable consumption; others included disease prevention or physical activity promotion. Most of the interventions were delivered in a school setting; others were offered at home, in a community, pre-school or health care setting, or by mass media. The intervention components included education, practical skills, media input, parental involvement and environmental adaptation, with some based on social learning theory. The interventions were largely delivered by parents, teachers and health promotion practitioners.
Participants included in the review
Studies of children aged between 4 and 10 years were eligible for inclusion. Children diagnosed with an illness or disability (e.g. diabetes, obesity or hypertension) were excluded. The age range of included children was 4 to 12 years, with a large proportion reported to be over 7 years. The reporting of ethnic and socioeconomic status was variable amongst the included studies. Although some were UK-based, the majority of the studies were conducted in the USA.
Outcomes assessed in the review
Studies measuring fruit and vegetable intake, or related attitude and behavioural outcomes, were eligible for inclusion. The most frequently reported outcomes were fruit and vegetable consumption (according to guidelines derived from the Health Survey for England 2001, see Other Publications of Related Interest), knowledge and preferences, self-efficacy, and willingness to try new fruit and vegetables. The majority of the included studies assessed outcomes immediately after the intervention, with no longer term follow-up.
How were decisions on the relevance of primary studies made?
Two independent reviewers selected the studies for inclusion in the review.
Assessment of study quality
Two independent reviewers carried out the validity assessment, using an established framework incorporating four criteria: the provision of pre-intervention data for all individuals (where appropriate); the provision of post-intervention data for each group; the reporting of results for each outcome measure as defined in the aims of the study; and the use of an equivalent control/comparison group. Where there was doubt regarding other sources of bias, some subcategorisation took place (details were provided in the report). Ultimately, reviewers judged the methodological quality of each study as 'high', 'medium' or 'not sound'. Any disagreements were resolved by consensus.
Two independent reviewers carried out the data extraction; any disagreements were resolved by consensus. Data were extracted in order to calculate standardised mean differences (SMDs) and 95% confidence intervals (CIs).
Methods of synthesis
How were the studies combined?
SMDs were calculated for 19 trials and pooled in a meta-analysis (a random-effects model was used in part). A narrative synthesis was provided on the 3 remaining studies.
How were differences between studies investigated?
Differences were explored in terms of study type and quality, population/setting, and intervention type. The Q statistic was reported. A sensitivity analysis was conducted to explore the impact of studies judged to be of 'medium' methodological quality. Further differences were explored in the text.
Results of the review
Twenty-two of 33 outcome evaluations were included in the synthesis. Nineteen of these were represented by 11 randomised controlled trials (RCTs) and 8 non-randomised controlled trials (16 employed a cluster design). Allocation units were schools, classes, families and children; the numbers in each were provided in the report. A further 3 studies were included (study designs not reported).
Of the 22 studies included in the synthesis, 12 were judged to be of 'high' methodological quality and 10 were accepted following the sensitivity analysis. In general, interventions targeting families at high risk of cardiovascular disease were the most effective. Those which reduced the emphasis on fruit and vegetables in a multi-component intervention were less effective. Full results for studies promoting fruit and vegetables (in isolation, or in combination), along with comparisons of different intervention intensities and follow-up periods, were given in the report.
Fruit and vegetable consumption.
Pooled results from 10 studies measuring fruit consumption showed a small, but statistically significant positive effect (SMD 0.10, 95% CI: 0.03, 0.17). The range of effect sizes was equivalent to an increase of one-fifth to two-thirds of a portion of fruit per day. There was no significant heterogeneity between the studies. Pooled results from 12 studies measuring vegetable consumption also showed a small, but statistically significant positive effect (SMD 0.23, 95% CI: 0.11, 0.34). The range of effect sizes was equivalent to an increase of nearly one-fifth to one-half of a portion of vegetables per day. There was statistically significant heterogeneity between the studies (p=0.005). Pooled results from 13 studies measuring fruit and vegetable consumption combined showed a statistically significant positive effect (SMD 0.23, 95% CI: 0.11, 0.35). The range of effect sizes was equivalent to an increase of one-half to 2 portions of fruit and vegetables per day. There was statistically significant heterogeneity between the studies (p=0.002).
Knowledge, preferences and self-efficacy.
Pooled results from 7 studies revealed a statistically significant increase in children's knowledge (SMD 0.67, 95% CI: 0.54, 0.79, p<0.001), reported to be equivalent to an improvement of one GCSE grade in English compulsory subjects. Similar improvements were noted for the impact on preferences for fruit and vegetables (SMD 0.65, 95% CI: 0.38, 0.91) from 3 studies. Small, but statistically significant improvements were also reported for self-efficacy outcomes (SMD 0.09, 95% CI: 0.00, 0.17) from 7 studies. There was no statistically significant heterogeneity between the pooled studies on each of these outcomes.
Experimentation with new fruit and vegetables.
Two of the 3 studies presented in a narrative synthesis revealed that interventions containing modelling (children observing an adult eating targeted foods) and reward initiatives were effective (results were not provided).
Sensitivity analysis assured the robustness of results for knowledge, preferences and self-efficacy. Studies of fruit consumption showed higher effects from non-randomised trials, but the overall pooled result was unaffected when a random-effects model was used. The analysis of consumption of vegetables (and fruit and vegetables combined) suggested that heterogeneity was unrelated to methodological differences.
Interventions in general can have a small, but significant positive effect in increasing children's fruit intake by one-fifth of a portion per day and vegetable intake by nearly one-fifth of a portion per day. The benefits will require translation into estimates of health gain and clinical significance, along with potential savings for health care services.
The review question was clear and supported by broad inclusion criteria, in respect of which a higher age range of participants was noted amongst the included studies. The comprehensive search strategy incorporated attempts to identify unpublished material. However, the authors acknowledged that the latter was not extensive and publication bias was a possibility. The restriction to English language papers also meant that language bias could not be ruled out. A validity assessment was carried out and the higher quality studies were identified for statistical synthesis. All aspects of the review process were carried out with adequate steps to minimise bias and error.
An extensive account of the included studies was provided. The exploration of heterogeneity and the provision of sensitivity analysis results justified the authors' cautious approach to statistical synthesis. The authors highlighted some important issues in this topic area: in particular, problems associated with the reliability of outcome measures and the measurement of the sustainability of intervention effects. Their conclusions reflect the evidence presented and are reliably derived.
Implications of the review for practice and research
Practice: The authors stated that focused interventions involving the contribution of parents, teachers and children, and using strategies which are relevant to addressing specific groups are likely to be most effective.
Research: The authors stated that rigorously conducted RCTs (including cluster designs) incorporating process evaluation and consideration of health inequalities are required.
Thomas J, Sutcliffe K, Harden A, Oakley A, Oliver S, Rees R, Brunton G, Kavanagh J. Children and healthy eating: a systematic review of barriers and facilitators. London: University of London, Institute of Education, Social Science Research Unit, EPPI-Centre. EPPI Report. 2003
Other publications of related interest
Doyle M, Hosfield N. Health Survey for England 2001: Fruit and vegetable consumption. London: HMSO; 2003.
Subject indexing assigned by CRD
Child; Food Habits; Health Behavior; Health Education; Health Promotion
Date bibliographic record published
Date abstract record published
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.