Skip to main content

Table 6 Results of the reviews about determinants of dietary behavior among youth

From: Determinants of dietary behavior among youth: an umbrella review

Author, date Outcome measures Correlate measures Overall results of the reviews Overall limitations of the review Overall recommendations of the review
Williams et al., 2014 [31] Sugar-sweetened beverages, fast food, fruit and vegetables Food and retail outlets Little evidence for an association between retail food environment surrounding schools and food consumption. 1) Meta-analysis not possible due to different conceptualizations and measures of the food environment surrounding schools. 1) Longitudinal studies needed.
2) Integrate validated classification systems of the retail food environment, explore the capacity of alternative methods for validating exposure data.
2) Loss of detail; a review is dependent upon outcomes and analyses that individual papers reported.
3) Specify individual-level measures of exposure to the food environment.
4) Collecting complementary measures of both qualitative and quantitative measures of food access.
5) Collect outcome measures that are appropriate relative to the exposures.
6) Take age and ethnicity differences into account.
Gardner et al., 2011 [17] Sugar-sweetened soft drinks Habit strength The weighted habit–behavior correlation effect estimate for nutritional habits was moderate to strong in size (fixed: r + =0.43; random: r + =0.41), and effects were of equal magnitude across healthful (fixed: r + =0.43; random: r + = 0.42) and unhealthful (fixed: r + =0.42; random: r + =0.41) dietary habits. The medium-to-large grand weighted mean habit–behavior correlation (r + ≈0.45) suggests that habit alone can explain around 20% of variation in nutrition related behaviors (i.e. R2 ≈ 0.20). 1) While it was not possible to meta-analyze interaction effects, habit often moderated the relationship between intention and behavior, such that intentions had reduced impact on behavior where habit was strong. This finding must be interpreted cautiously as it may reflect a bias towards publication of studies which find significant interaction, and so an overestimation of the robustness of this effect. 1) Explorations of the role of counter-intentional habits on the intention–behavior relationship, such as the capacity for habitual snacking to obstruct intentions to eat a healthful diet, are needed.
2) Healthful behaviors can habituate. The formation of healthful (‘good’) habits, so as to aid maintenance of behavior change, thus represents a realistic goal for health promotion campaigns.
3) More methodologically rigorous research is required to provide more conceptually coherent and less biased observations of the influence of habit on action.
2) Many studies were cross-sectional, and so modeled habit as a predictor of past behavior. This fails to acknowledge the expected temporal sequence between habit and behavior, and is also conceptually problematic given that, at least in early stages of habit formation, repeated action strengthens habit. 4) A more comprehensive understanding of nutrition behaviors, and how they might be changed, will be achieved by integrating habitual responses to contextual cues into theoretical accounts of behavior.
3) Reliance on self-reports of behavior.
Pearson & Biddle, 2011 [19] Fruit, vegetables, fruit and vegetable intake combined, energy-dense snacks, fast foods, energy-dense drinks Sedentary behavior: screen time (TV viewing, video/DVD, computer use, sitting while talking/reading/homework) Sedentary behavior, usually assessed as screen time and predominantly TV viewing, is associated with unhealthy dietary behaviors in children and adolescents. There appears no clear pattern for age acting as a moderator. There appears to be more consistent associations between sedentary behavior and diets for women/girls than for men/boys. 1) Many studies were cross-sectional. 1) More studies using objective measures of sedentary behaviors and more valid and reliable measures of dietary intake are required.
2) Use of self-report measures of sedentary and dietary behaviors that lack strong validity.
2) Examine the longitudinal association between sedentary behavior and dietary intake, and the tracking of the clustering of specifıc sedentary behaviors and specifıc dietary behaviors. For example, it appears from the mainly cross-sectional evidence presented that TV viewing is associated with unhealthy dietary patterns. Much less is known about diet and either computer use or sedentary motorized transport. It is likely that the main associations will be with TV, but this needs testing.
3) Sedentary behavior is largely operationally defined as screen time, and this is mainly TV viewing, making it diffıcult to draw any conclusions regarding non-screen time and dietary intake.
4) Although “screen time” can include TV and computer use, this does not help in identifying whether it is TV, computer use, or both, that is associated with unhealthy diets.
3) A focus on sedentary behaviors and dietary behaviors that “share” determinants as well as determinants of the clustering of sedentary and dietary behaviors will aid the development of targeted interventions to reduce sedentary behaviors and promote healthy eating.
Adriaanse et al., 2011 [15] Fruit and vegetable consumption Use of implementation intentions Considerable support was found for the notion that implementation intentions can be effective in increasing healthy eating behaviors, with twelve studies showing an overall medium effect size of implementation intentions on increasing fruit and vegetable intake. However, when aiming to diminish unhealthy eating patterns by means of implementation intentions, the evidence is less convincing, with fewer studies reporting positive effects, and an overall effect size that is small. NR 1) Although implementation intention instructions were not included as a moderator in the present meta-analysis due to the limited amount of studies, it seems prudent that future research takes into account the importance of using autonomy supportive instructions.
2) Stricter control conditions as well as better outcome measures are required.
3) Investigate efficacy of implementation intentions in diminishing unhealthy eating behaviors. In doing so, these studies should also compare the efficacy of different types of implementation intentions, as these may have differential effects on unhealthy food consumption.
McClain et al., 2009 [24] Fruit, juice and vegetable consumption, sugar, snacking, sweetened beverage consumption Psychosocial correlates like: attitude, availability, intention, knowledge, norms, self-efficacy, preferences, parental factors, and more. Perceived modeling and dietary intentions to make healthy or less healthy dietary changes (such as intentions to decrease consumption of sugary beverages or intentions to increase consumption of medium fat milk) have the most consistent and positive associations with dietary behavior. Other psychosocial correlates such as liking, norms, and preferences were also consistently and positively associated with dietary behavior in children and adolescents. Availability, knowledge, outcome expectations, self-efficacy and social support did not show consistent relationships across dietary outcomes. 1) Many studies were cross-sectional. 1) Future intervention research may benefit from the incorporation of findings from this review to create more effective adolescent and childhood dietary interventions by targeting the variables shown in this review that are most consistently associated with the various eating behaviors such as intentions, modeling, norms, liking, and preferences.
2) Not possible to conduct meta-analysis.
3) Authors combined conceptually similar psychosocial determinants into one category, which may have introduced bias.
4) Most studies relied on self-report of dietary intake. 2) Investigate variables that have been insufficiently examined to date, particularly the variables rooted in affective theories. It is quite plausible that affective factors, such as motivation, executive control, or meanings of behavior might drive the dietary behavior of children and adolescents.
5) Bias might have been introduced due to possible lack of validity or reliability of both dietary and psychosocial measures.
6) Certain studies reported only significant findings and did not address non-significant findings. 3) Investigate psychosocial correlates of several dietary behaviors that are known to influence weight and metabolic health such as fat and fiber that have been understudied.
7) Only studies included that were published in English in peer-reviewed journals (electronic databases).
8) This review did not separate children and adolescents into distinct categories, although research has suggested that children and adolescents exhibit different health behaviors.
Van der Horst et al., 2007 [29] Fruit intake, vegetable intake, juice intake, composite measure of fruit and vegetable intake, composite measure of fruit juice and vegetable intake, fast food consumption, snack food intake, pizza and snack, soft drink consumption. Correlates are categorized under home/household, educational institutions, neighborhood, city/municipality. Consistent evidence, for the relationship between parental intake and children’s fruit and vegetable intake, and for parent educational level with adolescent’s fruit and vegetable intake. A positive association was found for the relationship between availability and accessibility with children’s fruit and vegetable intake. Further positive associations were found for modeling (fruit/vegetable), parental intake (soft drink), parenting style (fruit/vegetable), family connectedness (fruit/vegetable) and encouragement to increase food intake (fruit/vegetable). 1) Many potential environmental determinants have been examined for a variety of dietary behaviors, but only few studies have been conducted on the same specific environmental factor—dietary behavior combination. 1) Replication of studies on the same specific environmental factors is necessary, to generate more compelling evidence for associations between environmental factors and dietary intake.
2) The finding that parental behavior is associated with child and adolescent intakes implies that interventions should take the behavior of parents into account, or desensitize adolescents for the (unfavorable) behavior of their parents. Parents should be more strongly encouraged to give the right example, especially where fat and energy intakes are concerned.
2) Many studies were cross-sectional.
3) Reliance on self-report measures.
4) Similar environmental determinants were collapsed conceptually into one category, although potential determinants in the same category were often dissimilar or measured in different ways.
3) Fruit and vegetable promotion should focus especially on adolescents from parents with lower levels of education.
5) Only studies included that were published in English in peer-reviewed journals (electronic databases). 4) Studies are needed that target the environmental levels and factors that have found to be (nearly) empty in the ANGELO framework, such as physical, socio-cultural, economic and political factors in the school (e.g. school food policy and food prices), neighborhood (e.g. availability and accessibility of foods in shops) and city/ municipality environment (e.g. food policy, food prices, marketing). Factors such as availability and accessibility at home, school and neighborhood should be studied in relation to energy, fat, soft drink, snacks and fast food intake.
6) Studies were heterogeneous in the conceptualization, measurement of the environmental determinant and/or dietary intakes, samples and analyses used: not possible to assess the overall strength of associations.
7) Multiple environmental factors examined in one study were included in the review, so these associations are not independent.
5) Need for longitudinal studies with valid or objective measures.
Pearson et al., 2009 [25] Breakfast consumption, breakfast skipping Physical (availability and accessibility), food poverty, socio-cultural (e.g. two parent family, modeling, family communication, monitoring, food rules, parental presence), demographic (SES, parental education level and employment) This review reported support for three family variables: Parental breakfast eating and living in two parent families were positively associated with adolescent breakfast consumption; and socio-economic deprivation was inversely associated with breakfast consumption. 1) Several studies may not have been powered to detect significant associations between family correlates and breakfast behaviors. 1) Future studies should clearly define breakfast foods (e.g. breakfast cereal, breads, milk, snacks on the run) being measured as this will allow for an understanding of the healthfulness of this behavior and will provide scope for interventions to promote healthy breakfast consumption.
2) Diversity in the definition of breakfast across the literature.
2) Importance of family structure should be considered when designing programs to promote breakfast consumption.
3) Future qualitative studies are needed to further explicate the mechanisms of the complex relationship between SES and adolescent breakfast behaviors.
De Craemer et al., 2012 [21] Sweet beverages, fruit and vegetable intake combined, snacks, milk intake Demographic and biological variables, behavioral variables, physical environmental variables TV viewing was positively associated with the intake of sweet beverages, snacks and inversely associated with fruit and vegetable intake. Parental modeling was associated with fruit and vegetable intake. No association with fruit and vegetable intake was found for restriction of eating, and an indeterminate result was found for pressuring the child to eat. Food availability was not associated with fruit and vegetable intake and snacking but had an indeterminate result for sweet beverages. NR 1) Future research should investigate similar correlates of physical activity, sedentary behavior and eating behavior to develop more efficient interventions.
2) Future research should be on interventions to predict whether interventions targeting these correlates will have an impact.
3) Future research should focus on identifying the common correlates of physical activity, sedentary behavior and eating behavior in preschool-aged children so that better tailored interventions could be developed.
4) More longitudinal studies are needed.
Pearson et al., 2009 [26] Fruit and vegetable consumption separately. fruit, fruit juice and vegetable consumption combined. Physical (e.g. availability, accessibility), socio-cultural (parental modeling, parental intake, family rules), and demographic correlates (e.g. SES). Children: home availability, family rules (demand/allow) and parental encouragement were positively associated with children’s fruit and vegetable intake. Parental modeling and parental intake were positively associated with children’s consumption of fruit and fruit juice and vegetable intake. Adolescents: parental intake and parental occupational status were found to be positively associated with adolescents’ consumption of fruit. Parental intake was also positively associated with adolescents’ vegetable consumption. There is also evidence for a positive association between parental education and adolescents’ fruit juice and vegetable intake. 1) Diversity in character (e.g. measures used and correlates studied) 1) More longitudinal studies are needed.
2) More studies are needed to test understudied correlates to generate more convincing evidence for associations between correlates and dietary behaviors. 3) Studies should report the validity and reliability of measures used to assess predictor variables.
2) Difficult to assess overall consistency of associations.
3) Several studies may not have been powered to detect significant associations between family correlates and dietary behaviors.
4) Few studies have examined the same specific combination of family correlate and dietary behavior, thus limiting the possibilities of drawing strong or consistent conclusions.
5) Many studies were cross-sectional.
6) Reliance on self-report measures.
7) Little data on reliability and validity of measures of dietary outcomes and physical and socio-cultural family correlates.
8) Only studies included that were published in English
Verloigne et al., 2012 [30] Breakfast consumption, soft drink consumption. Family and school environment: physical, socio-cultural, economic, political correlates. Parental descriptive/ injunctive norms and control/supervision were positively related to breakfast. Parental catering on demands, avoidance of negative modeling behavior, permissiveness, and area deprivation were inversely related. School SES was negatively related and teacher injunctive norms was positively related to breakfast. Availability at home, parental soft drink, and permissive parenting style were positively related to soft drink. Having family dinners, household income, parental employment status, and limits were inversely related. Availability of soft drinks at school and intake at school were positively related with soft drink. Participation in healthy school lunches was inversely related. 1) Only studies included that were published in English 1) More longitudinal studies are needed.
2) Interventions could help parents to create a supportive environment for their children to promote healthy behavior.
2) Did not take possible moderators and covariates into account.
3) More research is needed to focus on important school-environmental factors when developing an intervention program.
3) Not all existing studies on this topic were covered.
4) Focused on the consistency of the association and not on the strength of the association.
5) Conceptually similar variables were combined into a single category, even if variables were measured in a different way.
6) Many studies were cross-sectional.
Ford et al., 2012 [22] Vegetables, fruit and vegetable intake, (non)-fruit juice, high-energy/sugar-sweetened drinks, whole or 2% milk, fast foods, breakfast TV, video, and computer time in minutes. Eleven of the 12 included studies reported significant associations between TV and adverse dietary behaviors in young children. Six studies reported significant inverse relationships between TV viewing and fruit and vegetable intake. 1) Reliance on parent-reported methods to assess child TV viewing. 1) Guidelines for TV viewing use in young children should be further delimited.
2) Many studies were cross-sectional. 2) More longitudinal studies are needed.
3) Direct measurement of TV use.
Caspi et al., 2012 [16] Fruit and/or vegetables intake, 100% fruit juice consumption. Food environment: 5 dimensions of food access (availability, accessibility, affordability, accommodation, acceptability). Moderate evidence in support of the causal hypothesis that neighborhood food environments influence dietary health. Perceived measures of availability were consistently related to multiple healthy dietary outcomes. NR 1) More standardized/validated measures for food environment assessment needed.
2) Develop/refine understudied measures.
3) Abandon purely distance-based measures of accessibility, and combine multiple environmental assessment techniques.
4) Researchers should continue to expound upon the conceptual definitions of food access as they develop and refine new combinations of measure for the food environment.
Moore & Cunningham, 2012 [18] Daily fruit and vegetable consumption, snacking, breakfast consumption, soda consumption, meat intake. Social status, stress. Higher stress is related to less healthy dietary behaviors. The majority of studies reported that higher social position is related to healthier diet. 1) Only studies included that were published in English. 1) More quantitative dietary assessment tools such as FFQ, repeated 24-hr recalls, and food diaries are needed.
2) Many studies were cross-sectional. 2) More longitudinal studies are needed.
3) Because obesity results from a prolonged period of positive energy imbalance, assessment of dietary behaviors at a single point in time makes inferences related to diet and obesity difficult. 3) Important to acknowledge additional factors that influence energy intake, such as SES and stress levels.
4) Implementing appropriate monitoring and evaluation is essential to identifying successful, holistic strategies that can be used to improve quality of care.
4) Heterogeneity of measures.
Lawman & Wilson, 2012 [23] Fruit, vegetables, fast food, soft drink, dairy, milk, breakfast. Parenting (parental support for health behaviors, parenting style and parental monitoring surrounding health behaviors) and/or environmental factors (home availability/access, neighborhood availability/access and the built environment, neighborhood safety, neighborhood social factors). The current review found support for some parenting and physical environmental factors for health behaviors, particularly parental monitoring and neighborhood social factors. 1) Many studies were cross-sectional. 1) More longitudinal studies are needed on at-risk youth.
2) Reliance on self-report measures. 2) More objective measures of health behaviors or multiple reporters, who may hold different perspectives are needed, when objective measures are not feasible.
3) Future research should be conscious of reporting results in a way that facilitates systematic review of the literature
4) Examine additional levels of the bio-ecological model such as interpersonal and other macro- or society and policy level factors.
5) Future research should explore the relation between home/environment and health behaviors, particularly neighborhood social contextual factors such as social cohesion, and how factors at multiple bio-ecological levels may be influencing them (e.g. moderators).
6) More research monitoring is needed.
7) Development of more valid measures of parenting, family, and home environment variables is warranted.
8) Examine how parenting style is related to health behavior outcomes.
Sleddens et al., 2011 [28] Vegetables, fruit, sugar-sweetened beverages, soft drinks, breakfast, snacks/sweets General parenting In many studies significant associations with general parenting were found. Generally, children raised in authoritative homes were found to eat healthier. 1) Reliance on questionnaires and parental self-report measures. 1) Additional research is needed to further study the influence of mediating and moderating factors influencing the general parenting - child weight relationship, preferably employing a longitudinal design with more extended follow-up periods.
2) Differences in conceptualization of parenting constructs across studies.
3) Different categorizations to classify parents into styles across studies. 2) More longitudinal studies are needed using diverse ethnic samples and age groups.
4) Heterogeneity of measurements across studies and lacking information about distribution of independent and outcome variables. 3) Larger samples of fathers should be included to allow for comparisons between mothers and fathers.
5) Few studies examined the role of general parenting as a contextual factor that can influence the effectiveness of food-related parenting practices in predicting children’s dietary intake behaviors (moderation analyses). 4) Intervention developers should increase their attention to the family context as it is an important factor influencing outcomes of overweight interventions for children.
Berge, 2009 [20] Fruit and vegetable consumption, sugar sweetened beverages, dairy products, breakfast consumption, etc. Parental domain (e.g. parenting style, parenting practices), family functioning domain (e.g. family meals, family emotional closeness/ connection, family weight teasing). Parental domain: authoritative parenting style is positively associated with dietary intake. Family functioning domain: from cross-sectional and longitudinal research there is convincing evidence that family meals have an enduring protective factor for children and adolescents, girls and boys, and across diverse ethnic groups related to healthy dietary intake. 1) Many studies were cross-sectional. 1) More longitudinal, experimental and direct observational research in all family domains is needed.
2) Many studies used single-group designs.
3) Reliance on self-report measures. 2) Beneficial to incorporate mixed qualitative and quantitative designs.
4) Many studies used single informant measures to measure family-level data
3) There is a need for more within-in family measurements that utilize multi-level and multi-measurement approaches.
5) Many studies used single item measures.
6) Many studies adjusted for gender, SES and ethnicity as covariates, but left out other influential covariates such as maternal BMI and parental perception of child/adolescent weight. 4) There is a need to use systemic outcome variables. More family system variables should be studies.
5) Examine possible mediator or moderator effects of the family domains.
6) Important to include covariates when studying familial correlates of child/adolescent obesity.
Rasmussen et al., 2006 [27] Fruit and/or vegetable intake Socio-demographic factors, personal factors, family-related, friends-related factors, school-related factors, meal patterns, TV watching, eating fast food. The determinants supported by the greatest amount of evidence are social-economic position, preferences, parental intake, and home availability/accessibility. For nutritional knowledge, self-efficacy and shared family meals the evidence for positive associations is rather convincing. 1) Publications may have been missed due to the search strategy. 1) More studies on the influence of the family setting for influencing fruit and vegetable intake among children and adolescents are needed to enable health promoters to make evidence based decisions.
2) Within this review only significant associations are considered.
2) Observational studies analyzing fruit and vegetable intake in a school setting are still lacking.
3) Many papers include analyses based on small study samples and samples that are non-representative or only representative of a restricted geographical area.
3) Future international comparative surveys should enable investigations of national level factors of importance e.g. price levels, policy, guidelines, supply, and exposure to mass media and commercials.
4) Often the validity of the applied instruments are only considered very superficially or not mentioned at all.
4) Future research should study the influence of e.g. local access to fruit and vegetables through grocery stores, local food policies, exposure to mass media and commercials, and fruit and vegetable availability in leisure time facilities for children and adolescents, like for instance local sport clubs.
5) There is insufficient confounder control.
6) Large variety of approaches for conceptualizing, operationalizing, measuring and coding the outcome variable(s) exist.
5) Future research would benefit from improvements in design and methodology.
6) More longitudinal studies of children and adolescents’ fruit and vegetable intake are needed.
7) Lack of knowledge about predictors of FVI among children and adolescents from non-western parts of the world.
8) Future studies should keep a very broad and comprehensive theoretical scope, in order not to exclude important etiological components of importance for child and adolescent FVI.
  1. Note: The overall results/limitations/recommendations of the reviews that are reported are those reported by the review authors themselves.