- Email: [email protected]
The home environment: A mediator of nutrition knowledge and diet quality in adolescents
Appetite 105 (2016) 46e52
Contents lists available at ScienceDirect
Appetite journal homepage: www.elsevier.com/locate/appet
The home environment: A mediator of nutrition knowledge and diet quality in adolescents Tamara Tabbakh, Jean H. Freeland-Graves* Department of Nutritional Sciences, The University of Texas at Austin, Austin, TX 78701, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 December 2015 Received in revised form 31 March 2016 Accepted 5 May 2016 Available online 8 May 2016
The objective of this research was to assess adherence to the Healthy Eating Index-2010 of mothers and their adolescents (11e14 years old) and to examine the role of the home environment as a mediator of maternal nutrition knowledge and adolescent diet quality. It is hypothesized that mothers with greater knowledge impact the diet quality of their adolescents by creation of healthier home environments. A sample of 206 mother-adolescent dyads separately completed the Multidimensional Home Environment Scale, a Food Frequency Questionnaire, and a Nutrition Knowledge Scale. Body mass index-for-age percentiles were derived from weight and height measurements obtained by researcher; diet quality was estimated via the Healthy Eating Index (HEI)-2010. Percent of maximum score on nutrition knowledge for both mothers and adolescents were poor, with lowest scores on recommendations of healthy eating and physical activity (48% and 19%, respectively). A model of maternal nutrition knowledge (independent variable) and adolescent diet quality (dependent variable) indicated that greater knowledge was associated with higher scores on total fruit (p ¼ 0.02), whole grains (p ¼ 0.05), seafood and plant proteins (p ¼ 0.01), and overall diet quality (p < 0.01), as well as lower scores on empty calories (p ¼ 0.01). Inclusion of the home environment as a mediator yielded significant estimates of the indirect effect (b ¼ 0.61, 95% CI: 0.3e1.0). Within the home environment, psychological (b ¼ 0.46), social (b ¼ 0.23), and environmental (b ¼ 0.65) variables were all significant mediators of nutrition knowledge on diet quality. These results emphasize the importance of maternal nutrition knowledge and the mediating effect of the home environment on the diet quality of adolescents. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Nutrition knowledge Home environment Diet quality Adolescent Mediation
1. Introduction Adolescence is a time of physiological change which may affect both diet and weight status. Recommendations by experts for this stage of life suggest food choices such as fruits, vegetables, whole grains, low-fat dairy, and protein foods (U.S. Department of Agriculture, 2010). Yet, the majority of adolescents in the United States consume diets that are inadequate in fruits and vegetables (U.S. Department of Agriculture, 2015), but excessive in refined grains (U.S. Department of Agriculture, 2015) and sugar-sweetened beverages (Snacking patterns of U.S. Adolescents). Unhealthful diets during this critical time may be linked to dramatic changes in body shape, preoccupation with physical appearance, increased autonomy in decision-making, and a growing desire to conform to
* Corresponding author. Department of Nutritional Sciences, The University of Texas at Austin, 1 University Station, A2703, Austin, TX 78712, USA. E-mail address: [email protected] (J.H. Freeland-Graves). http://dx.doi.org/10.1016/j.appet.2016.05.002 0195-6663/© 2016 Elsevier Ltd. All rights reserved.
social pressures (Stang & Story, 2005). The barriers or opportunities that are presented in the surrounding environment at this time influence whether, and how, an individual engages in obesogenic behaviors. The home is one area that may impact the quality of adolescents’ diets. Also, the extent to which mothers are knowledgeable in nutrition may shape psychological, social, and environmental features of the home that affect diet quality in her adolescent (Pinard et al., 2014). The importance of maternal knowledge on diet quality has been reported in several studies (Campbell et al., 2013; Gibson, Wardle, & Watts, 1998). In preschoolers, nutrition knowledge of the mother has been established as a key modulator of diet quality of the child. For example, Vereecken et al. found that maternal knowledge was a significant predictor of dietary adequacy components (b ¼ 0.39, p < 0.001) of her child, but not excess components (b ¼ 0.40, p ¼ 0.42) of dietary guidelines (Vereecken & Maes, 2010). In older children, Gibson et al. assessed maternal knowledge using a three-category questionnaire containing
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
concepts of nutrient content, expert recommendations, and application of practical dietary choices. When mothers exhibited superior knowledge, their 9e11 year old children consumed more fruits (r ¼ 0.40, p < 0.001) but not fruit juices (r ¼ 0.01) or vegetables (r ¼ 0.03) (Gibson et al., 1998). A more recent study by Campbell et al. observed a similar pattern in 5e12 year old children. Mothers with superior nutrition knowledge had children whose diets were higher in fruits (b ¼ 0.09) and vegetables (b ¼ 0.17) and lower in salty snacks (b ¼ 0.09) and soft drinks (b ¼ 0.14) (Campbell et al., 2013). These relationships were mediated by the availability of foods in the home. Homes of mothers with greater knowledge had more healthy foods available, resulting in more desirable food choices in their children (Gibson et al., 1998). Burchi reported consistent findings; mothers with better nutrition knowledge reported selecting more varied diets for their children than did their lower knowledge counterparts (Burchi, 2010). Yet the extant literature on the role of nutrition knowledge on diet quality is mixed; some investigations showed weak or no effects (Blaylock, Variyam, & Lin; Dixon, Tershakovec, McKenzie, & Shannon, 2000; Covalito, Guthrie, Hertzler, & Webb, 1996; Nelson, Lytle, & Pasch, 2009). Colavito et al. demonstrated a significant association between maternal nutrition knowledge and fat intake (r ¼ 0.47) but not fiber intake, of her 2e5 year old child. However, this impact was only reported for foods consumed at home rather than in the total diet. Other research that involved knowledge of mothers and diet quality of their children (2e17 years) observed an age-specific effect (Blaylock et al.). Blaylock et al., found a significant positive relationship between diet-health awareness of mothers and the diet quality of their younger children (2e5 years) but not older ones (6e17 years) (Blaylock et al.). In a study focused on knowledge of energy balance of adolescents and parents, associations were not observed between greater knowledge and higher diet quality (lower fast food and sugar-sweetened beverage consumption) (Nelson et al., 2009). Also, a nutrition education program that incorporated parental involvement did not show any improvement in intakes of fruits and vegetables, complex carbohydrates, or calcium in children (4e10 years) at 3 months post-intervention (Dixon et al., 2000). It is plausible that the gap between translation of knowledge into consistent action may be related to indirect influences such as the home environment. The purpose of this research is to explore the role of the home as a mediator between maternal nutrition knowledge and diet quality of their adolescents. It is hypothesized that adolescents achieve more optimal diet quality when their mothers are knowledgeable in nutrition via more desirable psychological, social, and environmental features of the home environment.
2. Experimental methods 2.1. Study design and subjects Adolescents (11e14 years) and their mothers were recruited from local middle schools (n ¼ 206) and contacted by email. A link to a secure, research-based website, Qualtrics© was provided for administration of the questionnaires. Mothers were administered a demographics survey and both mothers and adolescents separately completed the Multidimensional Home Environment Scale, a food frequency questionnaire (FFQ) and a nutrition knowledge scale. Weight and height were measured as part of FITNESSGRAM, a standardized curriculum for assessment of physical fitness and activity of children in middle school. The risks and benefits were explained and this study was approved by The University of Texas at Austin Institutional Review Board.
47
2.2. Measurement of the home environment The Multidimensional Home Environment Scale (MHES) was used to measure the comprehensive home environment of adolescents (Tabbakh & Freeland-Graves, 2015). This novel scale incorporates psychological, social, and environmental parameters of the home environment using the Social Ecological Model as a framework (Foltz et al., 2012). Psychometric evaluation was conducted in a validation sample (n ¼ 114). The scale demonstrated construct validity (factor loadings: 0.3e0.9), internal consistency reliability (adolescent: a ¼ 0.82, mother: a ¼ 0.83) and test-retest reliability (adolescent: r ¼ 0.90, p < 0.01; mother: r ¼ 0.91, p < 0.01). Psychological factors were estimated using the sum of scores on healthy eating attitude, self-efficacy, and mindless eating. Social factors included frequency of family meals, social eating, social support, and injunctive norms for healthy eating. Environmental variables were the aggregate of availability of healthy foods, availability of unhealthy foods, accessibility to unhealthy foods, and physical characteristics of neighborhood. Variables that correspond to a negative feature of the home environment were reverse-coded such that a higher score reflect more desirable behaviors. A detailed description of development of the instrument is presented elsewhere (Tabbakh & Freeland-Graves, 2015). 2.3. Assessment of diet quality A slightly condensed version (116-items) of a 159-item food frequency questionnaire (FFQ) previously developed by the author was administered to all subjects for collection of dietary information (Chacko, Milani, Hans-Nuss, Kim, & Freeland-Graves, 2004). The FFQ exhibited good psychometrics (Cronbach’s a ¼ 0.72) and was validated by comparison to diet records (r ¼ 0.45) (Chacko et al., 2004). Mothers and adolescents were asked to report frequency, type, and amount of foods and beverages ingested over a one month period. A medium reference of a serving size was given to aid subjects in portion estimation. For example, a medium serving of orange juice was defined as 8 fluid ounces, and participants were instructed to select small, medium, large, or extra large if they had consumed half, one, one and a half, or two times the indicated medium serving. Mixed foods were deconstructed into their appropriate food groups and nutrients; analysis was conducted via MyDiet Analysis (Version 8.2.5, ESHA Research, Inc, Salem, OR). Diet quality was estimated using the Healthy Eating Index (HEI)-2010; with scores reflecting varying degrees of compliance to recommendations of the Dietary Guidelines for Americans (DGA) (Guenther et al., 2013). The HEI scores range from 0 to 100 points, utilizing a system based on energy density (per 1000 Kcal) that has been reported as valid and reliable (Guenther et al., 2013). Components measured were total fruit, whole fruit, total vegetables, greens and beans, whole grains, dairy, total protein foods, seafood and plant proteins, fatty acids, refined grains, sodium, and empty calories. 2.4. Evaluation of nutrition knowledge Nutrition knowledge was assessed using a 20-item tool modified from a scale previously developed by the author (Nuss, Freeland-Graves, Clarer, & Klohe-Lehman, 2007). This instrument was composed of multiple-choice questions classified into total and four categories of knowledge: macronutrient, micronutrient, MyPlate and USDA recommendations based on the 2010 dietary guidelines, and fast-food nutrition. Content validity of the scale was established using feedback from an expert panel in nutrition (n ¼ 10) and a focus group of the sample population tested (n ¼ 7). The scale demonstrated high internal consistency reliability
48
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
(adolescents: a ¼ 0.70, mothers: a ¼ 0.78) and test-retest reliability (adolescents: r ¼ 0.47, p ¼ 0.01, mothers: r ¼ 0.77, p < 0.01). 3. Measurement of anthropometrics Height and weight of adolescents were measured by the author using a stadiometer (Health-O-Meter® Professional Height Rod, McCook, IL) mounted on electronic scale (Health-O-Meter® Professional Mechanical Beam Scale, McCook, IL). Height was reported to the nearest 0.1 inch, with participants erect and barefoot, and weight to the nearest 0.01 pound. Body Mass Index (BMI) (kg/m2) was calculated as weight (kg)/height (m2). BMI-for-age percentile groups were classified as: healthy weight (5the85th), overweight (85the95th) or obese (>95th) according to standard growth charts (Center for Disease Prevention and Control). Mothers were asked to self-report weight and height and were categorized as healthy weight (18.5e24.9), overweight (25e29.9), or obese (30). 3.1. Statistical analyses All analyses were conducted using Statistical Product and Service Solutions (Version 20.0, SPSS, Armonk, NY, 2011). Mean scores for each nutrition knowledge category were computed and the percent of maximum score was reported for both mothers and adolescents. Comparisons across the two groups were determined via one-way analysis of variance. Statistical significance was set at p < 0.05. Univariate linear regression was used to assess associations between nutrition knowledge of mothers and adolescents and diet quality of adolescents using total and components scores of the Healthy Eating Index-2010. Regression analysis was utilized to predict the impact of maternal nutrition knowledge and home environment on diet quality and BMI of the adolescent. The home environment score was derived from the sum of psychological, social, and environmental factors and reflected positive characteristics in the home environment. This unidimensional scale defined as the composite home environment score, was used in the present analysis, as individual effects of factors in the home environment are the focus of another study. Subscales representing negative influences were reverse coded, such that higher scores represented more desirable outcomes. Mediation analysis involves testing the relationship between the independent variable and the dependent variable, accounting for the influence of a third variable (the mediator). The Sobel test, the traditional method, provides an estimate of the magnitude of mediation by computation of regression coefficients from the product of two pathways, one that tests the effect of the independent variable on the mediator variable and another that tests the effect of the mediator variable on the dependent variable. More recently, the Preacher and Hayes Multiple Mediation Model that uses bootstrapping has been recommended for assessment of mediation (MacKinnon, Lockwood, & Williams, 2004; Preacher & Hayes, 2008). Bootstrapping is a technique that involves sampling, with replacement, from the original data to obtain estimates of a sampling distribution curve. Because multivariate bootstrap extension of the Sobel test does not assume a normal distribution for the indirect effect, it provided a more robust estimate of confidence intervals than the Sobel test, particularly for the smaller sample size (MacKinnon et al., 2004; Preacher & Hayes, 2008). Variables tested by this model were nutrition knowledge, the home environment, and diet quality. 4. Results and discussion The mean age of adolescents and mothers in this sample was 12.5 years and 44.3 years, respectively. Participants were equally
distributed among boys (49.5%) and girls (50.5%), with approximately one-third of adolescents in the overweight or obese BMIfor-Age category. Mean BMI of mothers was 25.7 kg/m2, with 52% in the overweight or obese category. Roughly half of the mothers had a standard college education. Two-thirds of adolescents and mothers were White and 22.3% were Hispanic. The percent of maximum possible scores in total and the four areas of nutrition knowledge for mothers and adolescents are shown in Fig. 1. Overall, mothers in this sample displayed greater nutrition knowledge for micronutrients (p < 0.01), healthy eating and physical activity expert recommendations (p < 0.01), and fastfood nutrition (p < 0.01) than their adolescents. Total nutrition knowledge scores were 71.8% for mothers and 61.8% for adolescents. These findings are consistent with other research, with parents achieving higher scores (71.3%) on nutrition knowledge when compared with their adolescents (50%) (Nelson et al., 2009). The low degree of knowledge among adolescents noted in this study is of concern, as dietary practices established at this stage in life may persist in subsequent years (Pedersen, Holstein, Flachs, & Rasmussen, 2013). Mothers scored particularly high (96.8%) on the micronutrient portion of the survey. Two of the questions on the micronutrient segment involved the ability to identify that citrus fruits are a rich source of vitamin C and bananas as a rich source of potassium. These questions may have not been challenging enough for mothers, thus boosting the overall mean for the micronutrient portion. The lowest knowledge scores were observed for recommendations of healthy eating and physical activity, with 48.4% for mothers and 19.2% for adolescents. This finding could be responsible, in part, for the poor diet quality often reported in this population (U.S. Department of Agriculture, 2015) as it is exceedingly challenging to plan a nutritionally adequate diet without knowledge of the information present in recommendations of dietary guidelines. Nutrition knowledge scores of mothers and adolescents as compared to diet quality measured by components of the Healthy Eating Index-2010 are documented in Table 1. Significant relationships between mother and adolescent diet quality on most HEI2010 components were observed, with the exception of whole grains, dairy, and total protein. The higher dairy HEI scores observed for adolescents may be attributable to the greater consumption of milk and milk products by this population (Stang & Story, 2005). The strongest dietary similarities between mothers and adolescents were reported for total fruit, whole fruit, total vegetables, empty calories, and overall diet quality. This resemblance in diet is comparable to other reports on mother-child dyads, with respect to fruits and vegetables (322.6 vs. 302.4 g/d, respectively), as well as overall diet quality (50.3% vs. 48.9%, respectively) (Beydoun & Wang, 2009). The predicted differences in standard deviations in dietary outcomes associated with a one standard deviation increase in nutrition knowledge are seen as standardized b coefficients in Table 1. These estimates represent the effect size; these can be interpreted as: small ¼ 0.10e0.29, moderate ¼ 0.30e0.49, or large 0.50 (Cohen, 1988). Greater maternal nutrition knowledge was associated with higher scores of the adolescent on HEI-2010 components of total fruit (p ¼ 0.02), whole grains (p ¼ 0.05), seafood and plant proteins (p ¼ 0.01), and overall diet quality (p < 0.01), as well as lower scores on empty calories (p ¼ 0.01). Yet, the size of these effects was small (0.19e0.28). These findings are analogous to other reports, in which positive relationships were demonstrated between nutrition knowledge of the mother and her 9e11 year old child for fruit (r ¼ 0.40, p < 0.001) but not vegetable (r ¼ 0.03), consumption (Gibson et al., 1998). Campbell et al. also observed a link between maternal nutrition knowledge and fruits (b ¼ 0.09), vegetables (b ¼ 0.17), salty snacks (b ¼ 0.09), and soft drink
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
49
Fig. 1. Comparisons of percent maximum score in areas of nutrition knowledge between mothers and adolescents (n ¼ 103). aRecommendations were based on MyPlate and Dietary Guidelines for Americans.
Table 1 Nutrition knowledge scores of mothers and adolescents as compared to diet quality measures by components of the Healthy Eating Index-2010 (n ¼ 103). Healthy eating index-2010 componentsa
rb
Nutrition knowledge score Mother
Adequacy Total Fruit Whole Fruit Total Vegetables Greens & Beans Whole Grains Dairy Total Protein Seafood & Plant Proteins Fatty Acids Moderation Refined Grainsd Sodiumd Empty Caloriesd Total HEI-2010 a b c d
Adolescent
b (Std. Error)
p
b (Std. Error)
p
0.43** 0.52** 0.49** 0.33** 0.15 0.02 0.12 0.29** 0.20*
0.24 0.17 0.11 0.02 0.19 0.13 0.04 0.25 0.12
(0.05) (0.06) (0.05) (0.08) (0.06) (0.06) (0.04) (0.06) (0.04)
0.02 0.09 0.26 0.87 0.05 0.21 0.70 0.01 0.23
0.10 0.01 0.00 0.05 0.07 0.01 0.02 0.04 0.05
(0.05) (0.05) (0.05) (0.07) (0.05) (0.06) (0.03) (0.05) (0.04)
0.32 0.96 0.99 0.59 0.47 0.94 0.83 0.69 0.59
0.22* 0.21* 0.40** 0.35**
0.06 0.01 0.27 0.28
(0.10) (0.14) (0.12) (0.30)
0.54 0.95 0.01 0.00
0.18 0.11 0.18 0.17
(0.09) (0.12) (0.11) (0.29)
0.08 0.26 0.06 0.08
c
Linear regression with nutrition knowledge as independent variable and Healthy Eating Index-2010 components as dependent variable. Correlation between mother and adolescent for each Healthy Eating Index-2010 component. Standardized b coefficient representing effect size of the relationship between nutrition knowledge and dietary outcomes. Reverse coded such that higher score reflects better compliance with dietary guidelines.
(b ¼ 0.14) intake of her 5e12 year old child (Campbell et al., 2013). In the current research, adolescent nutrition knowledge was not significantly related to any measures of their own diet quality, although it approached significance for empty calories (b ¼ 0.18, p ¼ 0.06) and overall diet quality (b ¼ 0.17, p ¼ 0.08). Nelson et al. (2009) found similar results, with no effect of adolescent knowledge on consumption of fast foods (b ¼ 0.03, ns) or sugarsweetened beverages (b ¼ 0.15, ns) (Nelson et al., 2009). The link between maternal nutrition knowledge and adolescent diet quality (b ¼ 0.28, p < 0.01) was stronger than that of adolescent knowledge and dietary outcomes (b ¼ 0.17, p ¼ 0.08) in the present research. This finding is not surprising, as mothers function as the primary decision makers with regards to food and have greater control over meal purchasing and planning (Campbell et al., 2013). The impact of the comprehensive home environment on diet quality and BMI are shown in Fig. 2. The home environment explained 28% of the total variation in diet quality (b ¼ 0.53, p < 0.01) and 8% of the total variation in BMI (b ¼ 0.29, p < 0.01). These findings document the potential of the home as a setting to promote positive dietary behaviors in adolescents. Couch and colleagues (Couch, Glanz, Zhou, Sallis, & Saelens, 2014) found similar results, with associations between the presence of high-caloric foods in the home and lower diet quality. Couch et al. observed
that positive familial practices of eating in the home were related to fruit and vegetable intake (p < 0.05), higher DASH score (p < 0.01), and lower caloric beverage consumption (p < 0.01); whereas negative family practices were linked to increased odds of being overweight (p < 0.01) (Couch et al., 2014). The contribution of the home environment on adolescent weight-related behaviors has been reported by the author (Tabbakh & Freeland-Graves, 2015). Tabbakh and Freeland-Graves (2015) observed that availability of healthy foods and healthy eating attitude were positively associated with diet quality; availability of healthy foods and self-efficacy were inversely related to BMI (Tabbakh & Freeland-Graves, 2015). These results suggest the importance of the home environment in shaping diet quality of adolescents. The results of mediation analysis are presented in Fig. 3. Nutrition knowledge of the mother was the independent variable, the home environment was the mediator, and the effects are shown on the diet quality of the adolescents. This figure demonstrates pathway 1, which shows a significant influence of maternal nutrition knowledge on diet quality of the adolescent (b ¼ 0.89, p < 0.01), independent of the home environment. Pathways 2 and 3 illustrate the influence of nutrition knowledge of the mother on the home environment (b ¼ 0.70 p < 0.01) and the impact of the home environment on diet quality of the adolescent (b ¼ 0.82, p < 0.01),
50
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
Fig. 2. Impact of the comprehensive home environment on A) diet quality and B) body mass index of adolescents (n ¼ 103). a Home environment score was estimated using the sum of psychological, social, and environmental factors in the home.
Fig. 3. A mediation analysis for (A) the direct effect model representing association between the independent variable (maternal nutrition knowledge) and the dependent variable (diet quality of adolescent) and (B) the mediation model indicating the effect of independent variable (maternal nutrition knowledge) on the dependent variable (diet quality of adolescent), accounting for the mediator variable (total home score). aCI: confidence interval. *p < 0.05 level, **p < 0.01 level.
respectively. Estimates of the indirect effect derived from bootstrapping reflect the magnitude of mediation, i.e., how much of the effect of nutrition knowledge on diet quality was influenced by the home environment. In this analysis, proficient knowledge in nutrition appeared to be linked to a healthier home environment, and enhanced diet quality. Inclusion of the mediator (home environment) in the model yielded significant estimates of the indirect effect (b ¼ 0.61, 95% CI: 0.3e1.0), as confidence intervals do not cross zero, and a 65.2% reduction in the model, suggesting the importance of the home environment as a partial mediator. This outcome corroborates previous findings that competency in nutrition on its own may be insufficient for achievement of improvements in diet (Campbell et al., 2013). Also, the findings from the current research offer a likely explanation for the small effect size observed in Table 1, where nutrition knowledge and diet quality were examined in isolation. Thus, it is beneficial to explore the impact of nutrition knowledge on dietary patterns within the context of the home environment, as it seems to play a key role in modulation of those relationships. It remains unclear which factors in the home environment have the greatest impact on diet quality. Fig. 4 illustrates the influence of
nutrition knowledge on each of the specific domains measured (psychological, social, and environment) as they relate to diet quality. Estimates of the indirect effect of the amount of mediation revealed that psychological (b ¼ 0.46), social (b ¼ 0.23), and environmental features (b ¼ 0.65) of the home were all significant mediators of maternal nutrition knowledge on diet quality of the adolescent, as confidence intervals do not cross zero. Higher nutrition literacy in mothers was associated with improved psychosocial factors of the adolescent, as characterized by healthier attitudes toward eating and higher self-efficacy. In turn, these qualities were linked to higher compliance to dietary guidelines as measured by the Healthy Eating Index-2010. Therefore, mothers with higher nutrition presumably are fostering positive psychology with regards to eating for their adolescents; these behaviors may then translate to promotion of better diet quality. The combination of environmental factors, which include the sum of availability and accessibility of healthy and unhealthy foods in the home, showed a significant mediation effect, as seen in Fig. 4. Thus, mothers with greater knowledge in nutrition provided more healthy foods and made unhealthy foods less available in the home; this resulted in improved diet quality of the adolescent. Other
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
51
Fig. 4. A mediation analysis for the association between the independent variable (maternal nutrition knowledge), mediator variables (psychological, social, and environmental factors in the home), and dependent variable (diet quality of adolescent). Solid line represents total effect of independent variable on dependent variable without the mediator. Dashed line represents the influence of the independent variable on dependent variable with the mediator in the model. Magnitude of mediation represented in the b estimates of the indirect effect. aCI: confidence interval. *p < 0.05 level, **p < 0.01 level.
researchers have observed similar results, with home food availability mediating the impact of maternal nutrition knowledge on intakes of fruits (b ¼ 0.03, p < 0.001), vegetables (b ¼ 0.02, p < 0.05), salty snacks (b ¼ 0.02, p < 0.05), and soft drinks (b ¼ 0.03, p < 0.001) (Campbell et al., 2013). Although psychological variables were significant mediators in the current model, environmental features including availability of healthy foods, availability of unhealthy foods, accessibility of unhealthy foods, and physical features of the neighborhood showed stronger effects as mediators. The relatively small sample size is a limitation of this study. However, use of bootstrapping to test for mediation enables generation of robust estimates of the indirect effect, regardless of sample size. Additionally, the nutrition knowledge scale used in the study was not compared against other instruments. However, the scale used has been validated in a previous sample and has demonstrated high content validity as well as internal consistency and test-retest reliability, suggesting that it is psychometrically sound and suitable for use in this population. Assessment of dietary intake was conducted using FFQ, which may introduce a social desirability bias. Although there is no gold standard for measurement of diet quality (Perry et al., 2015), the HEI-2010 is regarded as a valid and reliable tool for use in various populations (Guenther et al., 2014). Exclusion of spouse/partner limits the generalizability of the findings to mother-adolescent interactions only. The equivocal findings related to the effect of nutrition knowledge may also be partly attributable to differences in the instruments used to measure nutrition knowledge as well as disparities in the demographics of the various study populations. In this research, the home environment scale was treated as an uni-dimensional scale. However, the independent effects of psychological, social, and environmental factors within the home, which are often conducted in the literature, (Campbell et al., 2013; Couch et al., 2014; Pinard
et al., 2014) is the aim of another study. Differences between mother and son vs. mother and daughter were not explored in this research, as dividing the sample into boys and girls would have diminished the power. Future studies should address possible sex differences between dyads. 5. Conclusions The present study explores the role of the home as a mediator between maternal nutrition knowledge and diet quality using the HEI-2010 of her adolescent. Rather than a direct association, this study proposes the home environment as a partial mediation of the influence of nutrition knowledge on diet quality. Maternal knowledge was associated with a healthier home environment, leading to improved diet quality of the adolescent. Thus, evaluation of education programs tailored for mothers should not assume a direct path between knowledge and dietary patterns; environmental variables should be explored as these may modulate diet quality as a result of greater knowledge. Conflict of interest The authors declare no conflict of interest. Acknowledgements Bess Heflin Centennial Professorship. References Beydoun, M. A., & Wang, Y. (2009). Parent-child dietary intake resemble in the United States: evidence from a large representative survey. Social Science & Medicine, 68(12), 2137e2144.
52
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
Blaylock J. R., Variyam J. N., Lin B. H.. Maternal nutrition knowledge and children’s diet quality and nutrient intakes. United States Department of Agriculture. Economic Research Service. http://www.ers.usda.gov/media/1787067/fanrr1. pdf. Accessed on: September 18. Burchi, F. (2010). Child nutrition in Mozambique in 2003: the role of mother’s schooling and nutrition knowledge. Economics & Human Biology, 8, 331e345. Campbell, K. J., Abbott, G., Spence, A. C., Crawford, D. A., McNaughton, S. A., & Ball, K. (2013). Home food availability mediates associations between mothers’ nutrition knowledge and child diet. Appetite, 71, 1e6. Center for Disease Prevention and Control. Data tables for BMI-for-age charts. http://www.cdc.gov/growthcharts/html_charts/bmiagerev.htm. Accessed on March 2, 2016. Chacko, G., Milani, T. J., Hans-Nuss, H., Kim, M., & Freeland-Graves, J. (2004). Development and validation of a semi-quantitative food frequency questionaire for young adult women in the southwestern United States. Nutrition Research, 24, 29e43. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum Asociates Publishers. Couch, S. C., Glanz, K., Zhou, C., Sallis, J. F., & Saelens, B. E. (2014). Home food environment in relation to chidren’s diet quality and weight status. Journal of the Academy of Nutrition and Dietetics, 114, 1569e1579. Covalito, E. A., Guthrie, J. E., Hertzler, A. A., & Webb, R. E. (1996). Relationship of diet-health attitudes and nutrition knowledge of household meal planners to the fat and fiber intakes of meal planners and preschoolers. Journal of Nutrition Education Behavior, 28, 321e328. Dixon, L. B., Tershakovec, A. M., McKenzie, J., & Shannon, B. (2000). Diet quality of young children who received nutrition education promoting lower dietary fat. Public Health Nutrition, 3, 411e416. Foltz, J. F., May, A. L., Belay, B., Nihiser, A. J., Dooyema, C. A., & Blanck, H. M. (2012). Population-level intervention strategies and examples for obesity prevention in children. Annual Review of Nutrition, 32, 391e415. Gibson, E. L., Wardle, J., & Watts, C. (1998). Fruit and vegetable consumption, nutritional knowledge, and beliefs in mothers and children. Appetite, 31, 205e228. Guenther, P. M., Casavale, K. O., Reedy, J., Kirkpatrick, S. L., Hiza, H. A. B, Kuczynski, K. J., et al. (2013). Update of the healthy eating index: HEI-2010. Journal of the Academy of Nutrition and Dietetics, 113, 569e580. Guenther, P. M., Kirkpatrick, S. I., Reedy, J., Krebs-Smith, S. A. M., Buckman, D. W., Dodd, K. W., et al. (2014). The Healthy Eating Index-2010 is a valid and reliable measure of diet quality according to the 2010 Dietary Guidelines for Americans. Journal of Nutrition, 144, 399e407. MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Confidence limits for the indirect effect: distribution of the product and resampling methods.
Multivariate Behavioral Research, 39, 99. Nelson, M. C., Lytle, L. A., & Pasch, K. E. (2009). Improving literacy about energyrelated issues: the need for a better understanding of the concepts behind energy intake and expenditure among adolescents and their parents. Journal of American Dietetic Association, 109, 281e287. Nuss, H., Freeland-Graves, J., Clarer, K., & Klohe-Lehman, M. T. J. (2007). Greater nutrition knowledge is associated with lower 1-year postpartum weight retention in low-income. Journal of American Dietetic Association, 107, 1801e1806. Pedersen, T. P., Holstein, B. E., Flachs, E. M., & Rasmussen, M. (2013). Male frequencies in early adolescents predict meal frequencies in late adolescents and early adulthood. BMC Public Health, 13, 445. http://dx.doi.org/10.1186/14712458-13-445. Perry, C. P., Keane, E., Layte, R., Fitzgerald, A. P., Perry, I. J., & Harrington, J. M. (2015). The use of a dietary quality score as a predictor of childhood overweight and obesity. BMC Public Health, 15, 581. Pinard, C. A., Yaroch, A. L., Hart, M. H., Serrano, E. L., McFerren, M. M., & Estabrooks, P. A. (2014). The validity and reliability of the comprehensive home environment survey (CHES). Health Promotion Practice, 1, 109e117. Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879e891. Snacking patterns of U.S. Adolescents. What we eat in America. National health and nutrition examination survey 2005-2006. http://ars.usda.gov/ba/bhnrc/fsrg. Accessed on: August 29. Stang, J., & Story, M. (2005). Adolescent growth and development. Guidelines for adolescent nutrition services. Minneapolis, MN: Center for leadership, education and training in maternal and child nutrition. Division of Epidemiology and Community Health, School of Public Health, University of Minnesota (Print). Tabbakh, T., & Freeland-Graves, J. (2015). Development and validation of the multidimensional home environment scale in adolescents and their mothers (Manuscript submitted for publication). U.S. Department of Agriculture, Agricultural Research Service. Percentage of selected nutrients contributed by food and beverages consumed away from home, by gender and age, what We Eat in America, National health and nutrition examination survey 2011-2012. http://www.ars.usda.gov/Services/ docs.htm?docid¼18349. Accessed August 28, 2015. U.S. Department of Agriculture. (2010). Report of dietary guidelines advisory committee on the dietary guidelines for Americans Accessed on: September 2 http:// dietaryguidelines. Vereecken, C., & Maes, L. (2010). Young children’s dietary habits and associations with mothers’ nutritional knowledge and attitudes. Appetite, 54, 44e51.
Contents lists available at ScienceDirect
Appetite journal homepage: www.elsevier.com/locate/appet
The home environment: A mediator of nutrition knowledge and diet quality in adolescents Tamara Tabbakh, Jean H. Freeland-Graves* Department of Nutritional Sciences, The University of Texas at Austin, Austin, TX 78701, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 December 2015 Received in revised form 31 March 2016 Accepted 5 May 2016 Available online 8 May 2016
The objective of this research was to assess adherence to the Healthy Eating Index-2010 of mothers and their adolescents (11e14 years old) and to examine the role of the home environment as a mediator of maternal nutrition knowledge and adolescent diet quality. It is hypothesized that mothers with greater knowledge impact the diet quality of their adolescents by creation of healthier home environments. A sample of 206 mother-adolescent dyads separately completed the Multidimensional Home Environment Scale, a Food Frequency Questionnaire, and a Nutrition Knowledge Scale. Body mass index-for-age percentiles were derived from weight and height measurements obtained by researcher; diet quality was estimated via the Healthy Eating Index (HEI)-2010. Percent of maximum score on nutrition knowledge for both mothers and adolescents were poor, with lowest scores on recommendations of healthy eating and physical activity (48% and 19%, respectively). A model of maternal nutrition knowledge (independent variable) and adolescent diet quality (dependent variable) indicated that greater knowledge was associated with higher scores on total fruit (p ¼ 0.02), whole grains (p ¼ 0.05), seafood and plant proteins (p ¼ 0.01), and overall diet quality (p < 0.01), as well as lower scores on empty calories (p ¼ 0.01). Inclusion of the home environment as a mediator yielded significant estimates of the indirect effect (b ¼ 0.61, 95% CI: 0.3e1.0). Within the home environment, psychological (b ¼ 0.46), social (b ¼ 0.23), and environmental (b ¼ 0.65) variables were all significant mediators of nutrition knowledge on diet quality. These results emphasize the importance of maternal nutrition knowledge and the mediating effect of the home environment on the diet quality of adolescents. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Nutrition knowledge Home environment Diet quality Adolescent Mediation
1. Introduction Adolescence is a time of physiological change which may affect both diet and weight status. Recommendations by experts for this stage of life suggest food choices such as fruits, vegetables, whole grains, low-fat dairy, and protein foods (U.S. Department of Agriculture, 2010). Yet, the majority of adolescents in the United States consume diets that are inadequate in fruits and vegetables (U.S. Department of Agriculture, 2015), but excessive in refined grains (U.S. Department of Agriculture, 2015) and sugar-sweetened beverages (Snacking patterns of U.S. Adolescents). Unhealthful diets during this critical time may be linked to dramatic changes in body shape, preoccupation with physical appearance, increased autonomy in decision-making, and a growing desire to conform to
* Corresponding author. Department of Nutritional Sciences, The University of Texas at Austin, 1 University Station, A2703, Austin, TX 78712, USA. E-mail address: [email protected] (J.H. Freeland-Graves). http://dx.doi.org/10.1016/j.appet.2016.05.002 0195-6663/© 2016 Elsevier Ltd. All rights reserved.
social pressures (Stang & Story, 2005). The barriers or opportunities that are presented in the surrounding environment at this time influence whether, and how, an individual engages in obesogenic behaviors. The home is one area that may impact the quality of adolescents’ diets. Also, the extent to which mothers are knowledgeable in nutrition may shape psychological, social, and environmental features of the home that affect diet quality in her adolescent (Pinard et al., 2014). The importance of maternal knowledge on diet quality has been reported in several studies (Campbell et al., 2013; Gibson, Wardle, & Watts, 1998). In preschoolers, nutrition knowledge of the mother has been established as a key modulator of diet quality of the child. For example, Vereecken et al. found that maternal knowledge was a significant predictor of dietary adequacy components (b ¼ 0.39, p < 0.001) of her child, but not excess components (b ¼ 0.40, p ¼ 0.42) of dietary guidelines (Vereecken & Maes, 2010). In older children, Gibson et al. assessed maternal knowledge using a three-category questionnaire containing
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
concepts of nutrient content, expert recommendations, and application of practical dietary choices. When mothers exhibited superior knowledge, their 9e11 year old children consumed more fruits (r ¼ 0.40, p < 0.001) but not fruit juices (r ¼ 0.01) or vegetables (r ¼ 0.03) (Gibson et al., 1998). A more recent study by Campbell et al. observed a similar pattern in 5e12 year old children. Mothers with superior nutrition knowledge had children whose diets were higher in fruits (b ¼ 0.09) and vegetables (b ¼ 0.17) and lower in salty snacks (b ¼ 0.09) and soft drinks (b ¼ 0.14) (Campbell et al., 2013). These relationships were mediated by the availability of foods in the home. Homes of mothers with greater knowledge had more healthy foods available, resulting in more desirable food choices in their children (Gibson et al., 1998). Burchi reported consistent findings; mothers with better nutrition knowledge reported selecting more varied diets for their children than did their lower knowledge counterparts (Burchi, 2010). Yet the extant literature on the role of nutrition knowledge on diet quality is mixed; some investigations showed weak or no effects (Blaylock, Variyam, & Lin; Dixon, Tershakovec, McKenzie, & Shannon, 2000; Covalito, Guthrie, Hertzler, & Webb, 1996; Nelson, Lytle, & Pasch, 2009). Colavito et al. demonstrated a significant association between maternal nutrition knowledge and fat intake (r ¼ 0.47) but not fiber intake, of her 2e5 year old child. However, this impact was only reported for foods consumed at home rather than in the total diet. Other research that involved knowledge of mothers and diet quality of their children (2e17 years) observed an age-specific effect (Blaylock et al.). Blaylock et al., found a significant positive relationship between diet-health awareness of mothers and the diet quality of their younger children (2e5 years) but not older ones (6e17 years) (Blaylock et al.). In a study focused on knowledge of energy balance of adolescents and parents, associations were not observed between greater knowledge and higher diet quality (lower fast food and sugar-sweetened beverage consumption) (Nelson et al., 2009). Also, a nutrition education program that incorporated parental involvement did not show any improvement in intakes of fruits and vegetables, complex carbohydrates, or calcium in children (4e10 years) at 3 months post-intervention (Dixon et al., 2000). It is plausible that the gap between translation of knowledge into consistent action may be related to indirect influences such as the home environment. The purpose of this research is to explore the role of the home as a mediator between maternal nutrition knowledge and diet quality of their adolescents. It is hypothesized that adolescents achieve more optimal diet quality when their mothers are knowledgeable in nutrition via more desirable psychological, social, and environmental features of the home environment.
2. Experimental methods 2.1. Study design and subjects Adolescents (11e14 years) and their mothers were recruited from local middle schools (n ¼ 206) and contacted by email. A link to a secure, research-based website, Qualtrics© was provided for administration of the questionnaires. Mothers were administered a demographics survey and both mothers and adolescents separately completed the Multidimensional Home Environment Scale, a food frequency questionnaire (FFQ) and a nutrition knowledge scale. Weight and height were measured as part of FITNESSGRAM, a standardized curriculum for assessment of physical fitness and activity of children in middle school. The risks and benefits were explained and this study was approved by The University of Texas at Austin Institutional Review Board.
47
2.2. Measurement of the home environment The Multidimensional Home Environment Scale (MHES) was used to measure the comprehensive home environment of adolescents (Tabbakh & Freeland-Graves, 2015). This novel scale incorporates psychological, social, and environmental parameters of the home environment using the Social Ecological Model as a framework (Foltz et al., 2012). Psychometric evaluation was conducted in a validation sample (n ¼ 114). The scale demonstrated construct validity (factor loadings: 0.3e0.9), internal consistency reliability (adolescent: a ¼ 0.82, mother: a ¼ 0.83) and test-retest reliability (adolescent: r ¼ 0.90, p < 0.01; mother: r ¼ 0.91, p < 0.01). Psychological factors were estimated using the sum of scores on healthy eating attitude, self-efficacy, and mindless eating. Social factors included frequency of family meals, social eating, social support, and injunctive norms for healthy eating. Environmental variables were the aggregate of availability of healthy foods, availability of unhealthy foods, accessibility to unhealthy foods, and physical characteristics of neighborhood. Variables that correspond to a negative feature of the home environment were reverse-coded such that a higher score reflect more desirable behaviors. A detailed description of development of the instrument is presented elsewhere (Tabbakh & Freeland-Graves, 2015). 2.3. Assessment of diet quality A slightly condensed version (116-items) of a 159-item food frequency questionnaire (FFQ) previously developed by the author was administered to all subjects for collection of dietary information (Chacko, Milani, Hans-Nuss, Kim, & Freeland-Graves, 2004). The FFQ exhibited good psychometrics (Cronbach’s a ¼ 0.72) and was validated by comparison to diet records (r ¼ 0.45) (Chacko et al., 2004). Mothers and adolescents were asked to report frequency, type, and amount of foods and beverages ingested over a one month period. A medium reference of a serving size was given to aid subjects in portion estimation. For example, a medium serving of orange juice was defined as 8 fluid ounces, and participants were instructed to select small, medium, large, or extra large if they had consumed half, one, one and a half, or two times the indicated medium serving. Mixed foods were deconstructed into their appropriate food groups and nutrients; analysis was conducted via MyDiet Analysis (Version 8.2.5, ESHA Research, Inc, Salem, OR). Diet quality was estimated using the Healthy Eating Index (HEI)-2010; with scores reflecting varying degrees of compliance to recommendations of the Dietary Guidelines for Americans (DGA) (Guenther et al., 2013). The HEI scores range from 0 to 100 points, utilizing a system based on energy density (per 1000 Kcal) that has been reported as valid and reliable (Guenther et al., 2013). Components measured were total fruit, whole fruit, total vegetables, greens and beans, whole grains, dairy, total protein foods, seafood and plant proteins, fatty acids, refined grains, sodium, and empty calories. 2.4. Evaluation of nutrition knowledge Nutrition knowledge was assessed using a 20-item tool modified from a scale previously developed by the author (Nuss, Freeland-Graves, Clarer, & Klohe-Lehman, 2007). This instrument was composed of multiple-choice questions classified into total and four categories of knowledge: macronutrient, micronutrient, MyPlate and USDA recommendations based on the 2010 dietary guidelines, and fast-food nutrition. Content validity of the scale was established using feedback from an expert panel in nutrition (n ¼ 10) and a focus group of the sample population tested (n ¼ 7). The scale demonstrated high internal consistency reliability
48
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
(adolescents: a ¼ 0.70, mothers: a ¼ 0.78) and test-retest reliability (adolescents: r ¼ 0.47, p ¼ 0.01, mothers: r ¼ 0.77, p < 0.01). 3. Measurement of anthropometrics Height and weight of adolescents were measured by the author using a stadiometer (Health-O-Meter® Professional Height Rod, McCook, IL) mounted on electronic scale (Health-O-Meter® Professional Mechanical Beam Scale, McCook, IL). Height was reported to the nearest 0.1 inch, with participants erect and barefoot, and weight to the nearest 0.01 pound. Body Mass Index (BMI) (kg/m2) was calculated as weight (kg)/height (m2). BMI-for-age percentile groups were classified as: healthy weight (5the85th), overweight (85the95th) or obese (>95th) according to standard growth charts (Center for Disease Prevention and Control). Mothers were asked to self-report weight and height and were categorized as healthy weight (18.5e24.9), overweight (25e29.9), or obese (30). 3.1. Statistical analyses All analyses were conducted using Statistical Product and Service Solutions (Version 20.0, SPSS, Armonk, NY, 2011). Mean scores for each nutrition knowledge category were computed and the percent of maximum score was reported for both mothers and adolescents. Comparisons across the two groups were determined via one-way analysis of variance. Statistical significance was set at p < 0.05. Univariate linear regression was used to assess associations between nutrition knowledge of mothers and adolescents and diet quality of adolescents using total and components scores of the Healthy Eating Index-2010. Regression analysis was utilized to predict the impact of maternal nutrition knowledge and home environment on diet quality and BMI of the adolescent. The home environment score was derived from the sum of psychological, social, and environmental factors and reflected positive characteristics in the home environment. This unidimensional scale defined as the composite home environment score, was used in the present analysis, as individual effects of factors in the home environment are the focus of another study. Subscales representing negative influences were reverse coded, such that higher scores represented more desirable outcomes. Mediation analysis involves testing the relationship between the independent variable and the dependent variable, accounting for the influence of a third variable (the mediator). The Sobel test, the traditional method, provides an estimate of the magnitude of mediation by computation of regression coefficients from the product of two pathways, one that tests the effect of the independent variable on the mediator variable and another that tests the effect of the mediator variable on the dependent variable. More recently, the Preacher and Hayes Multiple Mediation Model that uses bootstrapping has been recommended for assessment of mediation (MacKinnon, Lockwood, & Williams, 2004; Preacher & Hayes, 2008). Bootstrapping is a technique that involves sampling, with replacement, from the original data to obtain estimates of a sampling distribution curve. Because multivariate bootstrap extension of the Sobel test does not assume a normal distribution for the indirect effect, it provided a more robust estimate of confidence intervals than the Sobel test, particularly for the smaller sample size (MacKinnon et al., 2004; Preacher & Hayes, 2008). Variables tested by this model were nutrition knowledge, the home environment, and diet quality. 4. Results and discussion The mean age of adolescents and mothers in this sample was 12.5 years and 44.3 years, respectively. Participants were equally
distributed among boys (49.5%) and girls (50.5%), with approximately one-third of adolescents in the overweight or obese BMIfor-Age category. Mean BMI of mothers was 25.7 kg/m2, with 52% in the overweight or obese category. Roughly half of the mothers had a standard college education. Two-thirds of adolescents and mothers were White and 22.3% were Hispanic. The percent of maximum possible scores in total and the four areas of nutrition knowledge for mothers and adolescents are shown in Fig. 1. Overall, mothers in this sample displayed greater nutrition knowledge for micronutrients (p < 0.01), healthy eating and physical activity expert recommendations (p < 0.01), and fastfood nutrition (p < 0.01) than their adolescents. Total nutrition knowledge scores were 71.8% for mothers and 61.8% for adolescents. These findings are consistent with other research, with parents achieving higher scores (71.3%) on nutrition knowledge when compared with their adolescents (50%) (Nelson et al., 2009). The low degree of knowledge among adolescents noted in this study is of concern, as dietary practices established at this stage in life may persist in subsequent years (Pedersen, Holstein, Flachs, & Rasmussen, 2013). Mothers scored particularly high (96.8%) on the micronutrient portion of the survey. Two of the questions on the micronutrient segment involved the ability to identify that citrus fruits are a rich source of vitamin C and bananas as a rich source of potassium. These questions may have not been challenging enough for mothers, thus boosting the overall mean for the micronutrient portion. The lowest knowledge scores were observed for recommendations of healthy eating and physical activity, with 48.4% for mothers and 19.2% for adolescents. This finding could be responsible, in part, for the poor diet quality often reported in this population (U.S. Department of Agriculture, 2015) as it is exceedingly challenging to plan a nutritionally adequate diet without knowledge of the information present in recommendations of dietary guidelines. Nutrition knowledge scores of mothers and adolescents as compared to diet quality measured by components of the Healthy Eating Index-2010 are documented in Table 1. Significant relationships between mother and adolescent diet quality on most HEI2010 components were observed, with the exception of whole grains, dairy, and total protein. The higher dairy HEI scores observed for adolescents may be attributable to the greater consumption of milk and milk products by this population (Stang & Story, 2005). The strongest dietary similarities between mothers and adolescents were reported for total fruit, whole fruit, total vegetables, empty calories, and overall diet quality. This resemblance in diet is comparable to other reports on mother-child dyads, with respect to fruits and vegetables (322.6 vs. 302.4 g/d, respectively), as well as overall diet quality (50.3% vs. 48.9%, respectively) (Beydoun & Wang, 2009). The predicted differences in standard deviations in dietary outcomes associated with a one standard deviation increase in nutrition knowledge are seen as standardized b coefficients in Table 1. These estimates represent the effect size; these can be interpreted as: small ¼ 0.10e0.29, moderate ¼ 0.30e0.49, or large 0.50 (Cohen, 1988). Greater maternal nutrition knowledge was associated with higher scores of the adolescent on HEI-2010 components of total fruit (p ¼ 0.02), whole grains (p ¼ 0.05), seafood and plant proteins (p ¼ 0.01), and overall diet quality (p < 0.01), as well as lower scores on empty calories (p ¼ 0.01). Yet, the size of these effects was small (0.19e0.28). These findings are analogous to other reports, in which positive relationships were demonstrated between nutrition knowledge of the mother and her 9e11 year old child for fruit (r ¼ 0.40, p < 0.001) but not vegetable (r ¼ 0.03), consumption (Gibson et al., 1998). Campbell et al. also observed a link between maternal nutrition knowledge and fruits (b ¼ 0.09), vegetables (b ¼ 0.17), salty snacks (b ¼ 0.09), and soft drink
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
49
Fig. 1. Comparisons of percent maximum score in areas of nutrition knowledge between mothers and adolescents (n ¼ 103). aRecommendations were based on MyPlate and Dietary Guidelines for Americans.
Table 1 Nutrition knowledge scores of mothers and adolescents as compared to diet quality measures by components of the Healthy Eating Index-2010 (n ¼ 103). Healthy eating index-2010 componentsa
rb
Nutrition knowledge score Mother
Adequacy Total Fruit Whole Fruit Total Vegetables Greens & Beans Whole Grains Dairy Total Protein Seafood & Plant Proteins Fatty Acids Moderation Refined Grainsd Sodiumd Empty Caloriesd Total HEI-2010 a b c d
Adolescent
b (Std. Error)
p
b (Std. Error)
p
0.43** 0.52** 0.49** 0.33** 0.15 0.02 0.12 0.29** 0.20*
0.24 0.17 0.11 0.02 0.19 0.13 0.04 0.25 0.12
(0.05) (0.06) (0.05) (0.08) (0.06) (0.06) (0.04) (0.06) (0.04)
0.02 0.09 0.26 0.87 0.05 0.21 0.70 0.01 0.23
0.10 0.01 0.00 0.05 0.07 0.01 0.02 0.04 0.05
(0.05) (0.05) (0.05) (0.07) (0.05) (0.06) (0.03) (0.05) (0.04)
0.32 0.96 0.99 0.59 0.47 0.94 0.83 0.69 0.59
0.22* 0.21* 0.40** 0.35**
0.06 0.01 0.27 0.28
(0.10) (0.14) (0.12) (0.30)
0.54 0.95 0.01 0.00
0.18 0.11 0.18 0.17
(0.09) (0.12) (0.11) (0.29)
0.08 0.26 0.06 0.08
c
Linear regression with nutrition knowledge as independent variable and Healthy Eating Index-2010 components as dependent variable. Correlation between mother and adolescent for each Healthy Eating Index-2010 component. Standardized b coefficient representing effect size of the relationship between nutrition knowledge and dietary outcomes. Reverse coded such that higher score reflects better compliance with dietary guidelines.
(b ¼ 0.14) intake of her 5e12 year old child (Campbell et al., 2013). In the current research, adolescent nutrition knowledge was not significantly related to any measures of their own diet quality, although it approached significance for empty calories (b ¼ 0.18, p ¼ 0.06) and overall diet quality (b ¼ 0.17, p ¼ 0.08). Nelson et al. (2009) found similar results, with no effect of adolescent knowledge on consumption of fast foods (b ¼ 0.03, ns) or sugarsweetened beverages (b ¼ 0.15, ns) (Nelson et al., 2009). The link between maternal nutrition knowledge and adolescent diet quality (b ¼ 0.28, p < 0.01) was stronger than that of adolescent knowledge and dietary outcomes (b ¼ 0.17, p ¼ 0.08) in the present research. This finding is not surprising, as mothers function as the primary decision makers with regards to food and have greater control over meal purchasing and planning (Campbell et al., 2013). The impact of the comprehensive home environment on diet quality and BMI are shown in Fig. 2. The home environment explained 28% of the total variation in diet quality (b ¼ 0.53, p < 0.01) and 8% of the total variation in BMI (b ¼ 0.29, p < 0.01). These findings document the potential of the home as a setting to promote positive dietary behaviors in adolescents. Couch and colleagues (Couch, Glanz, Zhou, Sallis, & Saelens, 2014) found similar results, with associations between the presence of high-caloric foods in the home and lower diet quality. Couch et al. observed
that positive familial practices of eating in the home were related to fruit and vegetable intake (p < 0.05), higher DASH score (p < 0.01), and lower caloric beverage consumption (p < 0.01); whereas negative family practices were linked to increased odds of being overweight (p < 0.01) (Couch et al., 2014). The contribution of the home environment on adolescent weight-related behaviors has been reported by the author (Tabbakh & Freeland-Graves, 2015). Tabbakh and Freeland-Graves (2015) observed that availability of healthy foods and healthy eating attitude were positively associated with diet quality; availability of healthy foods and self-efficacy were inversely related to BMI (Tabbakh & Freeland-Graves, 2015). These results suggest the importance of the home environment in shaping diet quality of adolescents. The results of mediation analysis are presented in Fig. 3. Nutrition knowledge of the mother was the independent variable, the home environment was the mediator, and the effects are shown on the diet quality of the adolescents. This figure demonstrates pathway 1, which shows a significant influence of maternal nutrition knowledge on diet quality of the adolescent (b ¼ 0.89, p < 0.01), independent of the home environment. Pathways 2 and 3 illustrate the influence of nutrition knowledge of the mother on the home environment (b ¼ 0.70 p < 0.01) and the impact of the home environment on diet quality of the adolescent (b ¼ 0.82, p < 0.01),
50
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
Fig. 2. Impact of the comprehensive home environment on A) diet quality and B) body mass index of adolescents (n ¼ 103). a Home environment score was estimated using the sum of psychological, social, and environmental factors in the home.
Fig. 3. A mediation analysis for (A) the direct effect model representing association between the independent variable (maternal nutrition knowledge) and the dependent variable (diet quality of adolescent) and (B) the mediation model indicating the effect of independent variable (maternal nutrition knowledge) on the dependent variable (diet quality of adolescent), accounting for the mediator variable (total home score). aCI: confidence interval. *p < 0.05 level, **p < 0.01 level.
respectively. Estimates of the indirect effect derived from bootstrapping reflect the magnitude of mediation, i.e., how much of the effect of nutrition knowledge on diet quality was influenced by the home environment. In this analysis, proficient knowledge in nutrition appeared to be linked to a healthier home environment, and enhanced diet quality. Inclusion of the mediator (home environment) in the model yielded significant estimates of the indirect effect (b ¼ 0.61, 95% CI: 0.3e1.0), as confidence intervals do not cross zero, and a 65.2% reduction in the model, suggesting the importance of the home environment as a partial mediator. This outcome corroborates previous findings that competency in nutrition on its own may be insufficient for achievement of improvements in diet (Campbell et al., 2013). Also, the findings from the current research offer a likely explanation for the small effect size observed in Table 1, where nutrition knowledge and diet quality were examined in isolation. Thus, it is beneficial to explore the impact of nutrition knowledge on dietary patterns within the context of the home environment, as it seems to play a key role in modulation of those relationships. It remains unclear which factors in the home environment have the greatest impact on diet quality. Fig. 4 illustrates the influence of
nutrition knowledge on each of the specific domains measured (psychological, social, and environment) as they relate to diet quality. Estimates of the indirect effect of the amount of mediation revealed that psychological (b ¼ 0.46), social (b ¼ 0.23), and environmental features (b ¼ 0.65) of the home were all significant mediators of maternal nutrition knowledge on diet quality of the adolescent, as confidence intervals do not cross zero. Higher nutrition literacy in mothers was associated with improved psychosocial factors of the adolescent, as characterized by healthier attitudes toward eating and higher self-efficacy. In turn, these qualities were linked to higher compliance to dietary guidelines as measured by the Healthy Eating Index-2010. Therefore, mothers with higher nutrition presumably are fostering positive psychology with regards to eating for their adolescents; these behaviors may then translate to promotion of better diet quality. The combination of environmental factors, which include the sum of availability and accessibility of healthy and unhealthy foods in the home, showed a significant mediation effect, as seen in Fig. 4. Thus, mothers with greater knowledge in nutrition provided more healthy foods and made unhealthy foods less available in the home; this resulted in improved diet quality of the adolescent. Other
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
51
Fig. 4. A mediation analysis for the association between the independent variable (maternal nutrition knowledge), mediator variables (psychological, social, and environmental factors in the home), and dependent variable (diet quality of adolescent). Solid line represents total effect of independent variable on dependent variable without the mediator. Dashed line represents the influence of the independent variable on dependent variable with the mediator in the model. Magnitude of mediation represented in the b estimates of the indirect effect. aCI: confidence interval. *p < 0.05 level, **p < 0.01 level.
researchers have observed similar results, with home food availability mediating the impact of maternal nutrition knowledge on intakes of fruits (b ¼ 0.03, p < 0.001), vegetables (b ¼ 0.02, p < 0.05), salty snacks (b ¼ 0.02, p < 0.05), and soft drinks (b ¼ 0.03, p < 0.001) (Campbell et al., 2013). Although psychological variables were significant mediators in the current model, environmental features including availability of healthy foods, availability of unhealthy foods, accessibility of unhealthy foods, and physical features of the neighborhood showed stronger effects as mediators. The relatively small sample size is a limitation of this study. However, use of bootstrapping to test for mediation enables generation of robust estimates of the indirect effect, regardless of sample size. Additionally, the nutrition knowledge scale used in the study was not compared against other instruments. However, the scale used has been validated in a previous sample and has demonstrated high content validity as well as internal consistency and test-retest reliability, suggesting that it is psychometrically sound and suitable for use in this population. Assessment of dietary intake was conducted using FFQ, which may introduce a social desirability bias. Although there is no gold standard for measurement of diet quality (Perry et al., 2015), the HEI-2010 is regarded as a valid and reliable tool for use in various populations (Guenther et al., 2014). Exclusion of spouse/partner limits the generalizability of the findings to mother-adolescent interactions only. The equivocal findings related to the effect of nutrition knowledge may also be partly attributable to differences in the instruments used to measure nutrition knowledge as well as disparities in the demographics of the various study populations. In this research, the home environment scale was treated as an uni-dimensional scale. However, the independent effects of psychological, social, and environmental factors within the home, which are often conducted in the literature, (Campbell et al., 2013; Couch et al., 2014; Pinard
et al., 2014) is the aim of another study. Differences between mother and son vs. mother and daughter were not explored in this research, as dividing the sample into boys and girls would have diminished the power. Future studies should address possible sex differences between dyads. 5. Conclusions The present study explores the role of the home as a mediator between maternal nutrition knowledge and diet quality using the HEI-2010 of her adolescent. Rather than a direct association, this study proposes the home environment as a partial mediation of the influence of nutrition knowledge on diet quality. Maternal knowledge was associated with a healthier home environment, leading to improved diet quality of the adolescent. Thus, evaluation of education programs tailored for mothers should not assume a direct path between knowledge and dietary patterns; environmental variables should be explored as these may modulate diet quality as a result of greater knowledge. Conflict of interest The authors declare no conflict of interest. Acknowledgements Bess Heflin Centennial Professorship. References Beydoun, M. A., & Wang, Y. (2009). Parent-child dietary intake resemble in the United States: evidence from a large representative survey. Social Science & Medicine, 68(12), 2137e2144.
52
T. Tabbakh, J.H. Freeland-Graves / Appetite 105 (2016) 46e52
Blaylock J. R., Variyam J. N., Lin B. H.. Maternal nutrition knowledge and children’s diet quality and nutrient intakes. United States Department of Agriculture. Economic Research Service. http://www.ers.usda.gov/media/1787067/fanrr1. pdf. Accessed on: September 18. Burchi, F. (2010). Child nutrition in Mozambique in 2003: the role of mother’s schooling and nutrition knowledge. Economics & Human Biology, 8, 331e345. Campbell, K. J., Abbott, G., Spence, A. C., Crawford, D. A., McNaughton, S. A., & Ball, K. (2013). Home food availability mediates associations between mothers’ nutrition knowledge and child diet. Appetite, 71, 1e6. Center for Disease Prevention and Control. Data tables for BMI-for-age charts. http://www.cdc.gov/growthcharts/html_charts/bmiagerev.htm. Accessed on March 2, 2016. Chacko, G., Milani, T. J., Hans-Nuss, H., Kim, M., & Freeland-Graves, J. (2004). Development and validation of a semi-quantitative food frequency questionaire for young adult women in the southwestern United States. Nutrition Research, 24, 29e43. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum Asociates Publishers. Couch, S. C., Glanz, K., Zhou, C., Sallis, J. F., & Saelens, B. E. (2014). Home food environment in relation to chidren’s diet quality and weight status. Journal of the Academy of Nutrition and Dietetics, 114, 1569e1579. Covalito, E. A., Guthrie, J. E., Hertzler, A. A., & Webb, R. E. (1996). Relationship of diet-health attitudes and nutrition knowledge of household meal planners to the fat and fiber intakes of meal planners and preschoolers. Journal of Nutrition Education Behavior, 28, 321e328. Dixon, L. B., Tershakovec, A. M., McKenzie, J., & Shannon, B. (2000). Diet quality of young children who received nutrition education promoting lower dietary fat. Public Health Nutrition, 3, 411e416. Foltz, J. F., May, A. L., Belay, B., Nihiser, A. J., Dooyema, C. A., & Blanck, H. M. (2012). Population-level intervention strategies and examples for obesity prevention in children. Annual Review of Nutrition, 32, 391e415. Gibson, E. L., Wardle, J., & Watts, C. (1998). Fruit and vegetable consumption, nutritional knowledge, and beliefs in mothers and children. Appetite, 31, 205e228. Guenther, P. M., Casavale, K. O., Reedy, J., Kirkpatrick, S. L., Hiza, H. A. B, Kuczynski, K. J., et al. (2013). Update of the healthy eating index: HEI-2010. Journal of the Academy of Nutrition and Dietetics, 113, 569e580. Guenther, P. M., Kirkpatrick, S. I., Reedy, J., Krebs-Smith, S. A. M., Buckman, D. W., Dodd, K. W., et al. (2014). The Healthy Eating Index-2010 is a valid and reliable measure of diet quality according to the 2010 Dietary Guidelines for Americans. Journal of Nutrition, 144, 399e407. MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Confidence limits for the indirect effect: distribution of the product and resampling methods.
Multivariate Behavioral Research, 39, 99. Nelson, M. C., Lytle, L. A., & Pasch, K. E. (2009). Improving literacy about energyrelated issues: the need for a better understanding of the concepts behind energy intake and expenditure among adolescents and their parents. Journal of American Dietetic Association, 109, 281e287. Nuss, H., Freeland-Graves, J., Clarer, K., & Klohe-Lehman, M. T. J. (2007). Greater nutrition knowledge is associated with lower 1-year postpartum weight retention in low-income. Journal of American Dietetic Association, 107, 1801e1806. Pedersen, T. P., Holstein, B. E., Flachs, E. M., & Rasmussen, M. (2013). Male frequencies in early adolescents predict meal frequencies in late adolescents and early adulthood. BMC Public Health, 13, 445. http://dx.doi.org/10.1186/14712458-13-445. Perry, C. P., Keane, E., Layte, R., Fitzgerald, A. P., Perry, I. J., & Harrington, J. M. (2015). The use of a dietary quality score as a predictor of childhood overweight and obesity. BMC Public Health, 15, 581. Pinard, C. A., Yaroch, A. L., Hart, M. H., Serrano, E. L., McFerren, M. M., & Estabrooks, P. A. (2014). The validity and reliability of the comprehensive home environment survey (CHES). Health Promotion Practice, 1, 109e117. Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879e891. Snacking patterns of U.S. Adolescents. What we eat in America. National health and nutrition examination survey 2005-2006. http://ars.usda.gov/ba/bhnrc/fsrg. Accessed on: August 29. Stang, J., & Story, M. (2005). Adolescent growth and development. Guidelines for adolescent nutrition services. Minneapolis, MN: Center for leadership, education and training in maternal and child nutrition. Division of Epidemiology and Community Health, School of Public Health, University of Minnesota (Print). Tabbakh, T., & Freeland-Graves, J. (2015). Development and validation of the multidimensional home environment scale in adolescents and their mothers (Manuscript submitted for publication). U.S. Department of Agriculture, Agricultural Research Service. Percentage of selected nutrients contributed by food and beverages consumed away from home, by gender and age, what We Eat in America, National health and nutrition examination survey 2011-2012. http://www.ars.usda.gov/Services/ docs.htm?docid¼18349. Accessed August 28, 2015. U.S. Department of Agriculture. (2010). Report of dietary guidelines advisory committee on the dietary guidelines for Americans Accessed on: September 2 http:// dietaryguidelines. Vereecken, C., & Maes, L. (2010). Young children’s dietary habits and associations with mothers’ nutritional knowledge and attitudes. Appetite, 54, 44e51.