1. Introduction
Obesity among adolescence is a concern as 20.6% of US adolescents aged 12–19 have obesity [
1]. Obesity during adolescence is associated with many physical and mental health conditions and may have long-term impacts on health status in adulthood [
2]. Research has indicated that diet quality, consuming a healthy and balanced diet, may be an important strategy for obesity prevention due to its relationship with overweight or obesity in adolescents [
3,
4,
5]. However, only limited research has examined dietary quality in adolescents [
6,
7,
8], and, to our knowledge, only one study has explored racial/ethnic differences in dietary quality [
8]. This study assessed diet used the Healthy Eating Index-2010 (HEI-2010) for children and adolescents aged 2–18 and found that Mexican Americans had better dietary quality than other racial/ethnic groups while non-Hispanic blacks had the poorest dietary quality [
8]. Given racial/ethnic differences in obesity rates [
1], it is important to understand differences in dietary quality by race/ethnicity using current dietary guidelines. The 2015–2020 guidelines focus on the totality of healthy eating patterns and for the first time include a recommendation for limiting added sugars and saturated fats (<10% of total daily caloric intake) [
9]. The 2015–2020 guidelines also brought two additional changes to the HEI-2015, which is the latest version of dietary quality measure, regarding how energy from alcohol accounted for and how legumes allocated to different dietary components [
9]. No studies, to the best of our knowledge, have examined racial/ethnic disparities in adolescents’ dietary quality as assessed using the HEI-2015 which aligns with 2015–2020 dietary guidelines and MyPlate dietary recommendations.
Research has identified a negative relationship between body mass index (BMI) and eating a prudent dietary pattern, which is defined by high consumption of vegetables, legumes, nuts and seeds, fruits, and whole grains [
10]. Studies have identified a correlation between weight loss attempts and eating less fatty foods or fewer calories and protein among males only [
11,
12]. Studies have also identified an inverse relationship between weight dissatisfaction and the Western pattern diet consumption [
13]; and associations between higher levels of physical activity and lower intake of saturated fat and higher intake of fruits and vegetables, and dairy products [
14,
15]. Most of these studies, however, have assessed dietary quality using measures that are less comprehensive than the HEI-2015 [
10,
13], or assessed only selected aspects of diet [
11,
12,
14,
15], and/or did not control for sociodemographic characteristics [
10,
14,
15]. No identified studies have examined differences in dietary quality as assessed by the HEI-2015 by weight-related factors and physical activity controlling for sociodemographic characteristics.
Research is needed to understand dietary quality differences by race/ethnicity, and weight-related variables and physical activity among US adolescents using the HEI-2015. In addition, the influence of sociodemographic characteristics, and weight-related factors and physical activity on dietary quality in adolescents is poorly understood. This research may lead to more effective interventions for obesity prevention efforts designed to reduce racial/ethnic health disparities in adolescents. Accordingly, the purpose of the present study was to systematically explore racial/ethnic and weight-related and physical activity differences in dietary quality and its interactions and independent associations on dietary quality through a progressive set of hypotheses and exploratory analyses using a representative sample of US adolescents and the HEI-2015. Due to the previously identified sex effects [
4,
8], all analyses were stratified by sex.
3. Results
Of the 3373 respondents, approximately half (49.5%) were females, 37.1% were racial/ethnic minorities, 41.4% of the sample had parents who had a high school diploma or less, and 21.8% had a PIR below poverty level. In total, 17% of respondents were classified as being overweight, 20.9% had obesity, 51.3% were weight dissatisfied, 41.6% reported trying to lose weight in the last year, and 60% met the physical activity recommendation. Non-Hispanic blacks and Hispanics were more likely to live in a household with an income below the poverty level than non-Hispanic whites (
p < 0.001) and more likely to have parents with less than a high school education than non-Hispanic whites (
p < 0.001). See
Table 1 for sample characteristics stratified by race/ethnicity.
Controlling for age, sex, income, and parental education, a greater proportion of non-Hispanic whites had a lower BMI, were classified as normal weight, and met PA recommendations than non-Hispanic blacks and Hispanics. A smaller proportion of non-Hispanic whites were classified as being weight dissatisfied, and reported a weight loss attempt in the previous year than non-Hispanic blacks and Hispanics (see
Table 1). Additionally, there were differences in most dietary component scores and weight-related variables and physical activity by sex (see
Table S1).
3.1. Dietary Quality Differences by Race/Ethnicity
Dietary quality by race/ethnicity is presented in
Table 2. Overall, the weighted mean dietary quality scores, controlling for sociodemographic characteristics for all racial/ethnic groups, were low (total HEI-2015 score < 50 out of 100). Hispanic males (48.4 ± 0.5) and females (50.2 ± 0.4) had a higher total diet quality score than non-Hispanic whites [45.7 ± 0.6,
p = 0.005(males); 47.5 ± 0.5,
p < 0.001 (females)] and non-Hispanic blacks [45.5 ± 0.5,
p < 0.001 (males); 47.1 ± 0.5,
p < 0.001 (females)].
Examining the 13 dietary components, after controlling for sociodemographic variables and correcting for multiple comparisons, Hispanic males had the highest component scores for greens and beans (p < 0.001), seafood and plant proteins (p < 0.001), and added sugars (p < 0.001), as did Hispanic females. There was also difference in intake of refined grains by sex; non-Hispanic white and black males had higher scores than Hispanic males (p < 0.001) while non-Hispanic black females had higher scores than Hispanic females (p = 0.002).
3.2. Dietary Quality Differences by Weight-Related Variables and Physical Activity
There was no difference in total HEI-2015 scores by the examined weight-related variables by sex (see
Table 3). There also were no significant differences in total HEI-2015 scores in the effect of meeting physical activity guidelines for males. However, females who met the physical activity recommendation had a higher dietary quality score (49.3 ± 0.6) than those not meeting the recommendation (46.8 ± 0.5;
p = 0.004).
The multivariable analysis of the 13 components of dietary quality among males, after controlling for sociodemographic variables, found that male respondents with obesity scored higher in refined grains (5.4 ± 0.2) than those who were normal weight (4.7 ± 0.1, p = 0.003). Additionally, respondents who were weight dissatisfied had higher intake of total protein foods (p = 0.016) as well as seafood and plant proteins (p = 0.048) than those who were weight satisfied. Respondents who met physical activity recommendations had higher intake of total fruits (p = 0.004) than those not meeting the recommendation.
There were four dietary components scores that differed by the examined weight-related variables and physical activity in females. Females who were satisfied with their weight scored higher (5.2 ± 0.1) than those who were dissatisfied with their weight (4.6 ± 0.2, p < 0.001) for refined grains. However, respondents who were dissatisfied with their weight scored higher for added sugar than those who were satisfied with their weight (5.7 ± 0.1 vs. 5.1 ± 0.1, p = 0.002). Moreover, respondents who met physical activity recommendation scored higher than those who did not meet the recommendation for total fruits (2.5 ± 0.1 vs. 2.1 ± 0.1, p = 0.004) and whole fruits (2.4 ± 0.1 vs., 2.0 ± 0.1, p = 0.012).
3.3. Effect Modification of Weight-Related Variables and Physical Activity on Total HEI-2015 Score Differences by Race/Ethnicity
Meeting the physical activity recommendation in females modified dietary quality differences by race/ethnicity (
p = 0.011). There were no other significant interactions between the weight-related variables, physical activity, and race/ethnicity on total dietary quality score (see
Table 4).
Further analysis determined that non-Hispanic white females who did not meet the physical activity recommendation had lower total dietary quality by 6.43 points than Hispanic females, whereas non-Hispanic white females who met the recommendation had lower total dietary quality by 1.34 points (see
Table 5).
3.4. Percent Variability of Total HEI-2015 Scores Explained by Predictor Variables
Table 6 presents the associations between race/ethnicity and weight-related variables and physical activity on total HEI-2015 scores. Among males, race/ethnicity was the only significant independent predictor of dietary quality (ΔR
2 = 0.8%,
p = 0.0004) and the overall model only explained 1.7% of the variance. Among females, in addition to race/ethnicity (ΔR
2 = 1.3%,
p < 0.0001), other sociodemographic characteristics (ΔR
2 = 0.9%,
p = 0.0031), weight-related variables, and physical activity (ΔR
2 = 1.4%,
p = 0.0014) also were significant predictors and the overall model explained 3.6% of the variance in dietary quality.
To better understand the associations of sociodemographic, weight-related variables, and physical activity on total HEI-2015 scores, we conducted a stratified regression analyses (see
Table 7), which determined that only a small proportion of the variance of total HEI-2015 score was explained by the models. For males, there was only one significant predictor, weight dissatisfaction among non-Hispanic whites (ΔR
2 = 0.8%,
p = 0.0456). For females, sociodemographic variables (ΔR
2 = 3.6%,
p = 0.0008) and meeting physical activity recommendations (ΔR
2 = 2.6%,
p = 0.0004) were significant predictors of HEI-2015 score in non-Hispanic white females but no variables were significant for non-Hispanic black or Hispanic females.
4. Discussion
To the authors’ knowledge, this is the first study to systematically assess diet quality using the HEI-2015 in a representative sample of US adolescents and to examine the effect of race/ethnicity, weight-related factors and physical activity. The study determined that Hispanic adolescents, regardless of sex, had a higher total diet quality score than non-Hispanic whites or blacks, which reinforces findings from previous research [
8]. Meeting the physical activity recommendations was associated with improved dietary quality in females, and this was primarily driven by the associations among non-Hispanic white females. The overall variance in total HEI-2015 scores explained by study variables was in the low to moderate range [
29], but race/ethnicity was a significant independent predictor of dietary quality for males and females, whereas weight-related factors and physical activity were independent predictors of dietary quality for females only. This study extended previous research and findings can be used to inform interventions for adolescents. Specifically, the present study validated findings from previous research and extended these findings by using the HEI-2015, and focused on adolescents who have been found in other studies to have a lower dietary quality than children [
8,
30]. The present study also stratified analyses by sex due to consistent sex differences found in other studies [
4,
8]. Importantly, the present study controlled for sociodemographic characteristics and, through regression analyses determined the unique contribution of race/ethnicity and weight-related factors and physical activity to total HEI-2015 scores.
The present study found that Hispanics had a higher HEI-2015 score than non-Hispanic whites and blacks. These results are consistent with Gu and Tucker’s analyses of 1999–2012 NHANES data using the HEI-2010 [
8] that determined that Mexican Americans children had higher HEI-2010 scores than non-Hispanic white and black children [
8]. The present study also found that Hispanic males and/or females had dietary patterns that are indicative of a greater consumption of fruits, greens and beans, seafood and plant proteins and less added sugars, but less healthful grain choices than non-Hispanic whites and blacks. It should be noted that the HEI-2015, unlike previous versions, included legumes in total vegetables, greens and beans, total protein foods, and seafood and plant proteins as well as a new category for added sugars [
9,
20]. Nevertheless, study results are broadly consistent with previous research [
31,
32,
33]. It is possible that the pattern of higher dietary quality in Hispanics is related to foods traditionally consumed (e.g., legumes, fruits, and vegetables) by Hispanics [
33,
34]. It also is important to note that most respondents had total HEI-2015 scores and the majority of dietary components scores less than 50% of the recommendations, indicating a pattern of poor adherence to the 2015–2020 Dietary Guidelines for Americans [
35]. The scores for greens and beans and whole grains were particularly low with around 20% of maximum scores. These findings indicate a need for interventions in this age group to develop healthier eating habits that may persist into adulthood.
Furthermore, the present study found HEI-2015 scores differed by physical activity in females, whereas intake of some dietary components (e.g., total vegetables, total protein foods, and refined grains) differed by weight-related variables and physical activity in both males and females. Previous research has found dietary quality among adolescents is associated with weight-related factors (e.g., weight dissatisfaction and weight loss attempts) and physical activity. However, due to the use of different dietary quality measures, it is challenging to make cross-study comparisons [
10,
11,
12,
13,
14,
15]. This study extended existing research by exploring whether weight-related variables and physical activity modified the dietary quality differences between racial/ethnic groups and its independent associations on dietary quality. Analyses determined that meeting physical activity recommendation modified dietary quality difference between non-Hispanic white females and Hispanic females. Analyses also determined that meeting physical activity recommendation was associated with dietary quality among non-Hispanic white females. This finding suggests that non-Hispanic white females who were physically active might be more aware of the need for physical activity and healthy diet than their underactive counterparts. No other weight-related variables were found to moderate dietary quality differences between racial/ethnic groups, although weight dissatisfaction had a direct association with non-Hispanic white males’ dietary quality. A possible explanation for this finding is that non-Hispanic white males who are weight dissatisfied pay greater attention to their dietary quality, but such differences did not moderate dietary quality differences between non-Hispanic white males and other racial/ethnic groups. Further research is needed to investigate the potential influence of examined weight-related factors on racial/ethnic disparities in dietary quality in adolescents. Nevertheless, study findings indicated race/ethnicity is associated with dietary quality for both males and females, whereas meeting physical activity recommendation had an interaction and independent association with dietary quality only among females.
A practical implication of the present study is that it provides information on factors that might be related to dietary choices, which can be used to inform intervention to promote a healthful diet and help prevent obesity in adolescence thus reducing health disparities related to race/ethnicity [
3,
4,
5]. Health practitioners can work with adolescents, parents, and policy makers to advocate for more healthful food choices and to increase awareness of healthful food options. Results of the present study also indicate that intervention efforts need to address racial/ethnic differences on dietary choices, suggesting a need for specific message tailoring. For example, Hispanic adolescents could be reinforced for their more healthful dietary patterns but encourage to shift to towards greater consumption of whole grains. The present study also raises a need for research examining the influence of weight-related factors and physical activity on dietary quality between racial/ethnic groups. A longitudinal design could examine whether positive changes in weight-related factors and physical activity in adolescents could improve dietary quality among racial/ethnic groups over time.
Study findings should be considered within the context of its strengths and limitations. A strength of this study is the use of nationally representative data on adolescents. The study adds to the existing literature on racial/ethnic disparities in dietary quality by demonstrating moderation of weight-related variables and physical activity. Another strength is utilizing the HEI-2015 to measure diet quality, which was aligned with the 2015–2020 Dietary Guidelines for Americans [
9]. This study also has several important limitations. As a cross-sectional study, it is not possible to determine causal relationships [
36]. Dietary intake was assessed through recall methodology, which may underestimate intake and there were limitations in assessment of weight-related variables and physical activity. Weight dissatisfaction and weight loss attempt in the previous year were assessed by validated single items, which are less reliable than scales [
37]. Physical activity was collected via self-report; however, it was assessed by a widely used and validated instrument [
21]. Finally, to clarify distinctions between racial/ethnic groups, the current study excluded the “Other” race category.