There were not enough studies that compared dietary interventions for children with and without parental components to adequately answer whether parent involvement enhanced program effectiveness (Research Question 1). Despite variability in the quality of reporting of the RCTs reviewed to address Research Question 2 (What type of parent involvement was most effective in achieving dietary outcomes?), interesting patterns emerged. Studies that used direct methods to engage parents were more likely to report positive or mixed results compared with those studies that used more indirect methods. Further, those studies that used indirect methods to involve parents but required children engage their parent in an activity were also more likely to report positive or mixed results, suggesting an intensity level that is adequate to result in significant change in children's dietary intake.
When dietary outcomes were cross-tabulated with methodological characteristics, there was no apparent pattern. It was unclear whether ‘direct’ interventions attracted those most interested in change (i.e. more motivated participants), or the intensity of the contact was sufficient to break through barriers preventing behavior changes (an intervention design factor). Similar trends were recently reported among interventions with family components intended to promote physical activity in children (O'Connor, et al., 2009
), suggesting that direct involvement of parents in interventions targeting child dietary behavior need to be further evaluated via well-designed, adequately powered, randomized controlled trials.
Of concern was the lack of comprehensive and transparent reporting among the published interventions - only four of the reported studies met at least 70% of the CONSORT criteria for non-pharmacologic randomized controlled trials (Boutron, et al., 2008
). Empirical evidence suggests that omitting information captured by the CONSORT checklist is associated with biased estimates of treatment effect, making it difficult to determine the reliability or relevance of findings (Altman, et al., 2001
). We were unable to ascertain whether this was an issue in our review because of the limited number of publications meeting a majority of the reporting criteria.
There is a great need for development of more valid and reliable approaches for assessing dietary intake among children of all ages. Lack of uniformity in measurement of child dietary intake was also a troubling pattern that emerged from this review. While four commonly used methods of assessing dietary intake were used by the studies reviewed here (24-hour recalls, food frequency questionnaires, diet records, and staff observation), each varied slightly in methods, resulting in almost as many different measures of dietary outcome as studies reported. Different measurement methods yield different results (Stevens et al., 2007
). The selection of a method for obtaining food intake data should be based on the research question, study design, and additional criteria regarding potential sources of error and problems that may occur due to socio-cultural characteristics of the participant population. For example, the food frequency questionnaire was designed to measure typical patterns of food intake and not necessarily intended to provide accurate quantitative measures of energy and nutrient intakes on an individual basis (Thompson and Byers, 1994
). Further, while parents can report somewhat accurately on their child's behalf (Linneman et al., 2004
), they report less accurately when children eat in settings outside the home (e.g. at school or in childcare settings) (Baranowski et al., 1991
Interventions designed to impact child diet have largely taken place in school settings, which allows for large numbers of children to be reached, but with limited effects (Thomas, 2006
). Strategies are needed that reach and impact a majority of children at a substantial and meaningful level. This review suggests that such strategies should aim to directly engage parents in ways to help support their child have more healthy dietary consumption patterns. Designing an effective nutrition intervention requires an understanding of psychosocial or environmental determinants of diet (Baranowski et al., 1997
). Parents remain attractive targets for nutrition intervention programs because they act as nutrition “gatekeepers,” providing their children with ability and opportunity to make healthy food choices.
A potential barrier to implementing an effective parent-focused dietary intervention is a lack of theory-driven research that systematically evaluates the effects of specific parenting strategies (and in what context they are used) on child dietary behavior and weight. Research has begun to explore “effective food parenting” (O'Connor et al., 2010
)with an emphasis on feeding styles (Hughes et al., 2005
) and parenting practices (Hendy et al., 2009
, Musher-Eizenman and Holub, 2007
) and linking these strategies to child intake (O'Connor et al., 2009
). Improving our understanding of this could inform policy and guide public health efforts.
This review has several limitations. Only published articles were reviewed which may bias the selection to more favorable outcomes, since interventions with null findings are less likely to be published than those with a positive effect (Doak et al., 2006
). Only studies published in the English language were included, limiting the number of studies included outside of English-speaking countries. The CONSORT criteria allow for review of the quality of reporting, and are not a direct assessment of study design and analysis. Using this checklist to assess quality of reporting is also somewhat subjective given the possibility that users may interpret the criteria differently (despite definitions provided by CONSORT statement authors). Since the method of parental involvement that was reported in the publications only specified intensity of contact, intervention targets and intervention content must also be considered. It was impossible to ascertain either of these factors from the majority of studies in this review because of a lack of detailed reporting, often the result of word limits set by many medical and public health journals.