Obesity has been associated with low diet quality and the suboptimal intake of food groups and nutrients. Two composite diet quality measurement tools are appropriate for Americans 2–18 years old: the Healthy Eating Index (HEI) 2005 and the Revised Children’s Diet Quality Index (RC-DQI). The five components included in both indexes are fruits, vegetables, total grains, whole grains, and milk/dairy. Component scores ranged from 0 to 5 or 0 to 10 points with lower scores indicating suboptimal intake. To allow direct comparisons, one component was rescaled by dividing it by 2; then, all components ranged from 0 to 5 points. The aim of this study was to directly compare the scoring results of these five components using dietary data from a nationally representative sample of children (NHANES 2003–2006,
The development of diet quality indexes began several decades ago when capturing the characteristics of complete diets, rather than consumption levels of specific nutrients, became a goal in nutrition monitoring. This progression away from the minimalist’s approach addressed the need for tools to measure differences between diets of individuals and the dietary intake recommendations or other dietary standards. Unlike in animal-based research, which can be conducted in settings that allow complete control, free living people consume foods at various locations and throughout the day—most often not following obvious patterns. Thus, total food intake is a complex construct, which cannot be described or evaluated based on nutrients measured in isolation of one another. Even within foods, nutrients have complex synergistic effects upon one another that are not always well understood [
Currently, a number of diet quality indexes exist. Several of them were adapted to reflect the nutritional needs of different population groups. Two of those are the Healthy Eating Index (HEI) [
The HEI 2005 has 12 components, representing both adequacy and moderation, with maximum scores ranging from 5 to 20 points, all adding up to a total of 100 possible points. All HEI components are assessed on a density basis, that is, percentage of calories per 1000 calories consumed, allowing for characterization of intake levels while controlling for total energy intake, which is highly correlated with the quantity of foods consumed. The use of energy-adjusted estimates was an important difference from the HEI 1995 version and addressed the premise that a person eating a lot of food would be much more likely to meet minimum food group or nutrient intake levels than a person consuming less food (and energy). Maximum points reflect meeting or exceeding the intake standards, zero points indicate that the individual did not consume the recommended level of the food group/nutrient, and scores between the two extremes are prorated in a linear fashion. The intake standards for adequate intake were based on MyPyramid recommendations. For moderation components, the standard was set at the 85th percentile of the population distribution in order to prevent a large proportion of the population from receiving a zero score [
Performance of the HEI-2005 was evaluated for content and construct validity and reliability by scoring the reported one-day intakes of 8,650 individuals and by scoring menus. Content validity was measured by comparing each HEI-2005 component to key diet quality recommendations in the 2005 Dietary Guidelines for Americans. Construct validity was measured by scoring exemplary menus, theoretically capturing a high quality diet, and reliability was assessed through tests of internal consistency. Results showed strong evidence that the HEI-2005 is a valid measure of diet quality and most components are independent of energy intake [
The original DQI was developed for Americans two years old and up in 1994 as the composite measure of diet quality with eight components, each scored from a zero to two with zero being the measure of meeting standards [
Children are physiologically different from adults, and having a high quality diet is challenging. Although the use of density variables, as applied in the HEI 2005, accounts for overall energy intake, children consume much less food than adults, and their nutrient or food group density must be higher to accommodate their physiological need for higher amounts of nutrients to support healthy growth and development. Therefore, using a “one size fits all” index (early childhood to late in life) to measure the diet quality may not be prudent because it does not account for the higher nutrient need along with a lower quantity of food intake in children. To address this issue, several indexes were developed or adapted for use in child and/or adolescent populations [
Neither of those indexes addressed the specific needs of young children. The first composite diet quality index for children under five years of age was developed by Kranz and colleagues in 2006, the Children’s Diet Quality Index (C-DQI) [
Both the HEI-2005 and RC-DQI was designed to measure total diet quality in Americans, including children 2–18 years old. A side-by-side comparison of the components and their scoring criteria for both indexes are presented in Table
(a) Scoring scheme of the Healthy Eating Index (HEI) 2005. (b) Components and scoring scheme for the Revised Children's Diet Quality Index (RC-DQI): girls. (c) Components and scoring scheme for the Revised Children's Diet Quality Index (RC-DQI): boys.
HEI 2005* | |||
---|---|---|---|
Component | Max points | Max score |
Min score |
Total fruit | 5 | ≥0.8 c equiv. | No fruit |
Whole fruit |
5 | ≥0.4 c equiv. | No whole fruit |
Total vegetables | 5 | ≥1.1 c equiv. | No vegetables |
Dark green and orange vegetables and legumes | 5 | ≥0.4 c equiv. | No dark green or orange vegetables or legumes |
Total grains | 5 | ≥3.0 oz equiv. | No grains |
Whole grains | 5 | ≥1.5 oz equiv. | No whole grains |
Milk | 10 | ≥1.3 c equiv. | No milk |
Meat and beans | 10 | ≥2.5 oz equiv. | No meat or beans |
Oils | 10 | ≥12 grams | No oil |
Saturated fat | 10 | ≤7% of energy | ≥15% of energy |
Sodium | 10 | ≤0.7 gram | ≥2.0 g per 1000 kcal |
SoFAAS** | 20 | ≤20% of energy | ≥50% of energy |
Component | Score | Age (years) | Scoring criteria | ||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | |||
Added sugara | 10 | ≤10% total energy | WHO | ||||||||||||||||
Fata | 2.5 | 30–40% | 25–35% total energy | AMDR | |||||||||||||||
Linoleic acid (18_2)a | 2.5 | ≤5–10% of total energy |
( | ||||||||||||||||
Linolenic acid (18_3)a | 2.5 | 0.6–1.2% of total energy | ( | ||||||||||||||||
DHA and EPAa | 2.5 |
≤10% of |
(More potent | ||||||||||||||||
Total grainsa,c | 5 | 3 | 3 | 4 | 4 | 4 | 4 | 4 | 6 | 6 | 6 | 6 | 6 | 7 | 7 | 7 | 7 | 7 | Food groups |
Whole grainsa,c | 5 | 1.5 | 1.5 | 2 | 2 | 2 | 2 | 2 | 3 | 3 | 3 | 3 | 3 | 3.5 | 3.5 | 3.5 | 3.5 | 3.5 | Food groups |
Fruita,c | 10 | 1.5 | 1.5 | 1.5 | 1.5 | 1.5 | 1.5 | 1.5 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | Food groups |
Vegetablea,c | 10 | 1 | 1 | 2 | 2 | 2 | 2 | 2 | 3 | 3 | 3 | 3 | 3 | 4 | 4 | 4 | 4 | 4 | Food groups |
Excess juicea | 10 | 6 | 6 | 6 | 6 | 6 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | AAP age-appropriate limit |
Dairya,c | 10 |
2 | 2 | 2 | 2 | 2 | 2 | 2 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | Food groups and AAP |
Ironb | ≤3.0 | ≤4.1 | ≤5.7 | ≤7.9 | ≤EAR = 0 points | ||||||||||||||
10 | 3.1–6.9 | 4.2–9.9 | 5.8–7.9 | 8.0–14.9 | EAR-RDA = 5 points | ||||||||||||||
≥7 | ≥10 | ≥8 | ≥15 | ≥RDA = 10 points | |||||||||||||||
TV/energy a,d | 10 | EER: 1072 | EER: 1080 | EER: 1133 | EER: 1189 | EER: 1247 | EER: 1298 | EER: 1360 | EER: 1415 | EER: 1470 | EER: 1538 | EER: 1617 | EER: 1684 | EER: 1718 | EER: 1731 | EER: 1729 | EER: 1710 | EER: 1690 | 2 hours TV |
| |||||||||||||||||||
Total Points |
|
TV score: ≤2 hours of TV = 10 points.
EER score = 0.9 * EER to 1.1 * EER = 10 points, otherwise for overconsumption 10 points − ((actual intake/highest EER point * 100))%, for underconsumption 10 − (actual intake/lowest EER point * 100)%.
Component | Score | Age (years) | Scoring criteria | ||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | |||
Added sugara | 10 | ≤10% of total energy intake | WHO | ||||||||||||||||
Fata | 2.5 | 30–40% | 25–35% | AMDR | |||||||||||||||
Linoleic acid (18_2)a | 2.5 | ≤5–10% of total energy | ( | ||||||||||||||||
Linolenic acid (18_3)a | 2.5 | 0.6–1.2% of total energy | ( | ||||||||||||||||
DHA and EPAa | 2.5 |
≤10% of |
(More potent | ||||||||||||||||
Total grainsa,c | 5 | 3 | 3 | 5 | 5 | 5 | 5 | 5 | 7 | 7 | 7 | 7 | 7 | 9 | 9 | 9 | 9 | 9 | Food groups |
Whole grainsa,c | 5 | 1.5 | 1.5 | 2.5 | 2.5 | 2.5 | 2.5 | 2.5 | 3.5 | 3.5 | 3.5 | 3.5 | 3.5 | 4.5 | 4.5 | 4.5 | 4.5 | 4.5 | Food groups |
Fruita,c | 10 | 1.5 | 1.5 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 3 | 3 | 3 | 3 | 3 | Food groups |
Vegetablea,c | 10 | 1 | 1 | 2 | 2 | 2 | 2 | 2 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | Food groups |
Excess juicea | 10 | 6 | 6 | 6 | 6 | 6 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | AAP age-appropriate limit |
Dairya,c | 10 |
2 | 2 | 2 | 2 | 2 | 2 | 2 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | Food groups and AAP |
Ironb | ≤3.0 | ≤4.1 | ≤5.9 | ≤7.7 | ≤EAR = 0 points | ||||||||||||||
10 | 3.1–6.9 | 4.2–9.9 | 6.0–7.9 | 7.8–10.9 | EAR-RDA = 5 points | ||||||||||||||
≥7 | ≥10 | ≥8 | ≥11 | ≥RDA = 10 points | |||||||||||||||
TV/energya,d | 10 | EER: 1120 | EER: 1162 | EER: 1215 | EER: 1275 | EER: 1328 | EER: 1393 | EER: 1453 | EER: 1530 | EER: 1601 | EER: 1691 | EER: 1798 | EER: 1935 | EER: 2090 | EER: 2223 | EER: 2320 | EER: 2366 | EER: 2383 | 2 hours TV |
| |||||||||||||||||||
Total Points |
|
TV score: ≤2 hours of TV = 10 points.
EER score = 0.9 * EER to 1.1 * EER = 10 points, otherwise for overconsumption 10 points − ((actual intake/highest EER point * 100))%, for underconsumption 10 − (actual intake/lowest EER point * 100)%.
Imputation data, if TV/screen time was missing in the dataset:
2–4 year olds: 2 hours (28).
2–6 year olds: 2 hours/d.
7–9 year olds: 3.5 hours/d (includes video and videogames) (29).
8–10 year olds: 4.5 hours/d.
11–14 year olds: 4.5 hours/d.
15–18 year olds: 3.5 hours/d (30).
10–19 year olds: 3 hours/d (31).
6–19 year olds: 4–6 hours/d (32).
Use for RC-DQI coding:
2–6 year olds: 2 hours/d.
7–14 year olds: 4 hours/d.
15–18 year olds: 3.5 hours/d.
To date, a comparison of possible differences between the rankings of intake levels based on the two indexes to examine has not been conducted. The indexes both include the same five dietary intake recommendations for dairy, total grains, whole grains, fruits, and vegetables. Each index also includes additional components, but they are not comparable. Thus, an examination of total HEI 2005 scores and RC-DQI scores would not yield adequate information to allow a direct comparison or examination of differences. The purpose of this paper was to examine if the scoring schemes of the five components present in both indexes affect the ranking of the dietary intake level and in which manner this difference presents itself.
This study used sociodemographic and dietary intake data of a nationally representative sample of children 2–18 years old from the National Health and Nutrition Examination Survey (NHANES) 2003–2006 to explore if children’s dietary intakes were scored differentially on each of the five comparable components in the two diet quality indexes (the HEI-2005 and RC-DQI). The hypothesis tested is that although both indexes will discern between children with optimal versus suboptimal food group intake levels, the scoring of the RC-DQI, which was specifically developed to address children’s dietary needs, will provide more levels of differentiation between intake levels and therefore potentially provide a higher level of detail to the researchers employing the index.
Socioeconomic and dietary data from the combined survey years of 2003-2004 and 2005-2006 of the National Health and Nutrition Examination Survey (NHANES (available at
According to the interview responder’s categorization, race, and ethnicity were reported as American Indian or Alaskan Native, Asian, black or African American, Native Hawaiian or Pacific Islander, white, or other non-Hispanic, Mexican American, other Hispanic. These variables were recoded to reflect the cultural eating differences in Hispanic/other (Mexican Americans, other Hispanic, other/multiethnic), nonhispanic white, and nonhispanic black children. Household income was used to differentiate households by the income-eligibility cut points for USDA food assistance programs: high income are families with household incomes ≥3.5 of the poverty income ratio (PIR), medium income defined as 1.86–3.4 PIR, and low income defined as ≤1.85 PIR. The latter group are income eligible for participation in the USDA food assistance program [
Two 24-hour dietary recalls of food consumption data are available for both 2003-2004 and 2005-2006 survey years. It is noteworthy to point out that a correction for the distribution of intakes of food groups and nutrients, such as the NCI or the Iowa State method, is not warranted in this type of analysis for a number of reasons. First, while intake distributions may affect the association between levels of consumption and the risk for chronic diseases in a manner that may lead to misclassification of individuals, thus introducing error, the comparison of two diet quality indexes scores is not affected by that threat. The analysis is based on comparing the population’s intake levels of food groups twice, based on applying two distinct scoring systems to the diets of the same nationally representative sample. No individual data was generated. The aim of the study was not to rank individuals or compare individual’s intake levels to disease risk factors to discern those at higher risk from those at lower risk. One might have chosen to add a calculation to correct for interperson intake variation, but no additional information would have been gained in this particular project.
Detailed information on survey design and data collection of the NHANES data can be found elsewhere [
Usual dietary intake of food groups and nutrients was estimated by calculating two-day average intakes for each child. Mixed dishes and foods were disaggregated using the MyPlate Equivalents Database (MPED) and expressed in servings (ounces or cups). The components and total scores of the HEI 2005 and RC-DQI were calculated using Stata, version 10 [
The scoring of the HEI 2005 is described in detail in the technical report [
The comparison between the two indexes and their components is shown in Table
To allow comparison of HEI 2005 and RC-DQI component scores, the scoring of the RC-DQI was adjusted to reflect that of the HEI 2005, which had a maximum of 5 points for the fruits, vegetables, total grains, and whole grains components and a maximum of 10 points for the dairy component. Thus, The RC-DQI maximum scores for fruit and vegetables were reduced from 10 points to 5 points by dividing the maximum scores by two. The scores for total grains, whole grains, and dairy remained unchanged (total grains and whole grains had a maximum of 5 points and dairy a maximum of 10 points in both indexes). Although the dairy component was scored with a maximum of 10 points in both indexes, maintaining that value would have provided a larger weight of this component in the calculation of the total subscore of the five selected components; thus, in both indexes the dairy scores were divided by two, which results in a possible maximum score of 5 points. The total maximum subscore of the five HEI-2005 and RC-DQI components selected for this study was therefore 25 points; the possible minimum score was zero.
HEI 2005 and RC-DQI total subscore and component scores calculated for the population and the distribution of diet quality index points were examined in the following manner: first, the correlation of the components within each index was examined (Table
Correlation coefficients (
HEI 2005
Components | Fruits | Vegetables | Total grain | Whole grain |
---|---|---|---|---|
Total sample ( | ||||
Milk/dairy | 0.0891** |
|
0.0317 | 0.0545* |
Fruit |
|
0.0711** | 0.0726** | |
Vegetables | 0.1017** |
|
||
Total grains | 0.3007** | |||
Whole grains | ||||
| ||||
2–5 year olds ( | ||||
Milk/dairy | 0.0027 |
|
|
|
Fruit |
|
0.1079 | 0.1007** | |
Vegetables | 0.0603 |
|
||
Total grains | 0.3833** | |||
Whole grains | ||||
| ||||
6–11 year olds ( | ||||
Milk/dairy | 0.0198 | −0.0676* |
|
|
Fruit | 0.0183 | 0.1185* | 0.0722 | |
Vegetables | 0.1267* | 0.0207 | ||
Total grains | 0.3027** | |||
Whole grains | ||||
| ||||
12–18 year olds ( | ||||
Milk/dairy | 0.0531 | 0.0099 | 0.1157** | 0.1347** |
Fruit |
|
0.0311 | 0.0192 | |
Vegetables | 0.0981** | 0.0015 | ||
Total grains | 0.2661** | |||
Whole grains |
RC-DQI in the total sample and three mutually exclusive age groups
Components | Fruits | Vegetables | Total grain | Whole grain |
---|---|---|---|---|
Total sample ( | ||||
Milk/dairy | 0.1326** | 0.0938** | 0.2522** | 0.1585** |
Fruit | 0.2025** | 0.1513** | 0.2274** | |
Vegetables | 0.2056** | 0.1070** | ||
Total grains | 0.2560** | |||
Whole grains | ||||
| ||||
2–5 year olds ( | ||||
Milk/dairy | 0.0312 |
|
|
0.0060 |
Fruit | 0.2165** | 0.0779 | 0.1881** | |
Vegetables | 0.1496** | 0.0454 | ||
Total grains | 0.2203** | |||
Whole grains | ||||
| ||||
6–11 year olds ( | ||||
Milk/dairy | 0.0820 | 0.0488 | 0.2442** | 0.1980** |
Fruit | 0.0632 | 0.0760* | 0.1738** | |
Vegetables | 0.1305** | 0.0491 | ||
Total grains | 0.2510** | |||
Whole grains | ||||
| ||||
12–18 year olds ( | ||||
Milk/dairy | 0.0533 | 0.0362 | 0.2619** | 0.1092** |
Fruit | 0.0676 | 0.0926* | 0.1469** | |
Vegetables | 0.1950** | 0.0427 | ||
Total grains | 0.2067** | |||
Whole grains |
Correlation coefficients (
HEI 2005 | |||||
---|---|---|---|---|---|
Components | RC-DQI | ||||
Milk/Dairy | Fruits | Vegetables | Total grain | Whole grain | |
Total sample ( |
|||||
Milk/dairy | 0.3381** | 0.1217** | 0.0350 | 0.0337 | 0.1569** |
Fruit | 0.0297 | 0.5506** | 0.0838** | −0.0058 | 0.1418** |
Vegetables | −0.0724** | −0.0344 | 0.2769** | −0.0843** | −0.0275 |
Total grains | −0.0469* | 0.0024 | −0.1004** | 0.1659** | 0.0376 |
Whole grains | 0.0066 | 0.0536** | −0.0351 | 0.0453** | 0.4145** |
| |||||
2–5 year olds ( |
|||||
Milk/dairy | 0.2036** | 0.0045 | −0.0467 | −0.0772 | 0.0432 |
Fruit | −0.0252 | 0.5555** | 0.0481 | −0.0007 | 0.1871** |
Vegetables | −0.0709 | −0.0520 | 0.2601** | −0.0511 | −0.0417 |
Total grains | −0.0668 | 0.0186 | −0.0953* | 0.1867** | 0.1871** |
Whole grains | −0.0682 | 0.0462 | −0.1035* | 0.0686** | 0.4263** |
| |||||
6–11 year olds ( |
|||||
Milk/dairy | 0.2769** | 0.0535 | −0.0718 | −0.0263 | 0.0888 |
Fruit | −0.0247 | 0.4800** | −0.0334 | −0.0433 | 0.1057* |
Vegetables | −0.0492 | 0.0129 | 0.2948** | −0.0737 | −0.0239 |
Total grains | −0.0691 | 0.0636 | −0.0897** | 0.1573** | −0.0098 |
Whole grains | −0.0121 | 0.0918* | 0.0130 | 0.0613** | 0.3613** |
| |||||
12–18 year olds ( |
|||||
Milk/dairy | 0.3702** | 0.0251 | −0.0483 | 0.0003 | 0.1813** |
Fruit | −0.0062 | 0.5476** | 0.0039 | −0.0835* | 0.0029 |
Vegetables | −0.0659 | −0.0292 | 0.3633** | −0.0797* | 0.0053 |
Total grains | −0.0071 | −0.0246 | −0.1118** | 0.1942** | 0.0106 |
Whole grains | 0.0504 | 0.0115 | −0.0525 | 0.0246 | 0.4781** |
The distribution of children with zero (minimum), between one and 24 points, and 25 points (maximum) was 12%, 88%, and 0% for the HEI-2005 and 0%, 100%, and 0% for the RC-DQI.
Examination of the correlation between the five component scores under investigation within each index is shown for the total population and the three age groups in Table
Analysis of the correlation between component scores in the RC-DQI showed that all comparisons were positive and statistically significant (
The correlation between component scores in both indexes (Table
Analyses of the distribution of component scores by age group and ethnic group are depicted in Figures
(a) Distribution of mean HEI component scores by age group. (b) Distribution of mean RC-DQI component scores by age group.
(a) Distribution of mean HEI component scores by ethnic group. (b) Distribution of mean RC-DQI component scores by ethnic group.
Proportion of children with the maximum, the minimum, or between the maximum and minimum diet quality index scores after applying the HEI 2005 and RC-DQI.
Measurement of DQ is complex and requires a thorough understanding of the nutritional issues affecting the population of interest [
Intake guidance is population sensitive and should not be applied to individuals living in other food cultures [
The HEI 2005 and the RC-DQI were scored to express increased diet quality with higher component or total scores. Based on this analysis, most of the children were classified as having diets between the ideal (25 points) and very low scores (zero points) using either of the indexes; however, the HEI-2005 subscore indicated that 12% of the population had zero scores, meaning that they scored no points in either of the five components (dairy, total or whole grain, fruits, and vegetables) whereas using the RC-DQI, none of the children were scored with zero. This difference is due to the scoring scheme of the RC-DQI components, which are based on the concept of “deviation from the recommended intake.” Thus, only children who did not consume any servings of the food groups received a zero score—partial consumption resulted in proportional deduction of points. It is important to point out that children who had missing dietary data were excluded from the study analysis (they would have received a zero score).
The examination of the associations between component scores within the HEI 2005 showed that several of the components were not related to each other and, although not significant, an increase in one component could be associated with a decrease of another; that is, dairy, whole grains, or fruit scores were associated with lower vegetable scores. The association between the RC-DQI components, on the other hand, showed that increased component scores were consistent in all five components and had at least moderate and statistically significant correlation coefficients for all comparisons.
Interestingly, the correlation coefficients between the two indexes showed a wide range of associations, including positive, significant, negative, and nonsignificant associations. For instance, while the association between the component scores of whole grains and fruit was at least moderate, the association between total grains or vegetable scores were much lower. Thus, although the data analysis was based on the same sample of a nationally representative group of children 2–18 years old and their estimated two-day average dietary intake of food groups, the scoring scheme of the two indexes resulted in differential evaluation of the diet quality.
Results from this study demonstrate the importance of understanding the relationship between dietary intake of food groups and nutrients. Based on the scoring method employed, the calculated dietary quality index scores varied greatly in the evaluation of the population’s diet quality. The RC-DQI scores led to higher proportions of the population with scores between the two extremes of maximum and minimum scores for the components or the subscore. Thus, depending on the research question, researchers interested in estimating overall quality of children ages 2–18 must consider the desired level of differentiation of diet and select the appropriate index. Furthermore, depending on the aims of the study, it might be more appropriate to employ only individual component scores rather than the total score of a diet quality index.
In recent years, the diet quality of the American population has become a public health issue of great concern [