Print version ISSN 0021-7557
J. Pediatr. (Rio J.) vol.84 no.1 Porto Alegre Jan./Feb. 2008
Use of the Revised Children's Diet Quality Index to assess preschooler's diet quality, its sociodemographic predictors, and its association with body weight status
Sibylle KranzI; Jill L. FindeisII; Sundar S. ShresthaIII
RD. Assistant professor, Department of Nutritional Sciences, Pennsylvania State
University, PA, USA. Research associate, The Population Research Institute,
Pennsylvania State University, PA, USA
IIPhD. Professor, Department of Agricultural Economics and Rural Sociology, Pennsylvania State University, PA, USA
IIIDepartment of Agricultural Economics and Rural Sociology, Pennsylvania State University, PA, USA
To determine the level of overall diet quality, sociodemographic predictors
of diet quality, and the association between diet quality and body weight status
in a nationally representative sample of preschoolers.
METHODS: Cross-sectional study using a sample of 2-5 years old with sociodemographic, dietary, and anthropometric data (n = 1,521) in the National Health and Examination Survey 1999-2002. Overall diet quality was determined using the Revised Children's Diet Quality Index. Sociodemographic predictors (age, sex, sociodemographic, ethnic group, household income, preschool attendance, federal food program participation) of diet quality were determined using multiple linear regression models in the total sample and stratified by household income for Food Stamp eligible (< 1.3 of the poverty income ratio) or Special Supplemental Program for Women, Infants, and Children eligible (poverty income ratio < 1.85). Association between diet quality and prevalence of childhood obesity was assessed with Pearson chi-square tests. Statistical significance was assumed at p £ 0.05. All analysis was conducted using complex survey design routines.
RESULTS: On average, preschooler consumed suboptimal levels of whole grains, fruits, vegetables, and dairy. Overall diet quality decreased with increasing age (beta-coefficient: -2.38, p < 0.001) but improved with increasing family income in the full sample (beta-coefficient: 1.22, p < 0.001) but not in the low-income subpopulations. Mexican American children had significantly better diet quality than non-Hispanic white children (beta-coefficient: 2.18, p < 0.033) especially in the low income group (beta-coefficient: 3.57, p < 0.006). Childhood obesity prevalence decreased significantly with increasing diet quality.
CONCLUSIONS: Preschooler's diet quality needs to be improved to support the prevention of childhood obesity early in life.
Keywords: Diet, diet surveys, obesity, food quality.
Childhood obesity rates have been increasing dramatically in the past decades in the USA while preschooler's intake of fruits and vegetables decreased.1 The number of children at risk for overweight (85th-94th percentile on the body mass index-for-age growth charts) or overweight (³95th percentile) has reached epidemic proportions.2 In addition to the health problems associated with high body weight, overweight children might also be at increased risk to suffer from the metabolic syndrome.3,4
To help prevent childhood obesity, it is of pivotal importance to understand the relationship between modifiable lifestyle factors, such as dietary intake patterns and the development of excessive body weight. Dietary intake levels of some food groups or nutrients have been found to be associated with overall diet quality or body weight status,5,6 however, there is lack of data indicating a direct association between overall diet quality, measured with a composite diet quality assessment tool specifically designed for children, and body weight status in American preschoolers. In an effort to close this gap, the aims of this study were to determine the level of overall diet quality in American children and to ascertain the sociodemographic predictors of overall diet quality, as well as to examine the association between diet quality and body weight status.
The Centers for Disease Control and Prevention (CDC) conducts the National Health and Nutrition Examination Survey (NHANES, http://www.cdc.gov/nchs/nhanes.htm), an ongoing survey using a multistage, stratified area design to obtain a sample of respondents that is representative of the civilian non-institutionalized American population. Certain population groups were over-sampled (e.g. young children, adolescents 12-19 years, African Americans, Mexican Americans, and low-income persons).
Although the data are released in 2-year increments, they were designed to be merged to multi-year data sets.7 The first 4-year data (NHANES 1999-2002) contains information on 21,004 individuals who provided interview data. Of these, 19,759 individuals also had medical examination data and the sample size for children 2-5 years old who provided sociodemographic, dietary, and body mass index (BMI) data, and who were not breastfeeding was 1,521. To allow the examination of the effect of federal food programs, such as the Food Stamp Program and the Special Supplemental Program for Women, Infants, and Children (WIC), multiple regression models were employed in two income-stratified subpopulations. Thus, linear regression models were developed for the total sample of preschoolers, the children who were income-eligible for the WIC program (poverty income ratio - PIR £ 1.85, n = 861) or the Food Stamp Program (PIR < 1.3, n = 676). All samples were nationally representative of the American preschool-age population.
Sociodemographic information, such as age, sex, race, ethnicity, preschool participation, and total household income was reported by the adult completing the household interview during the NHANES survey. Age was used as continuous variable in this study.
Children's race is determined based on the interview responder's categorization as American Indian or Alaskan Native, Asian, black or African American, Native Hawaiian or Pacific Islander, white, or other. In addition, self-reported ethnic background is determined by whether the child is Mexican American, other Hispanic or Latino, both (Mexican and other Hispanic), or not Hispanic. In an effort to capture cultural differences of children living in the USA, the variables were employed to define four mutual exclusive ethnic groups: non-Hispanic white, non-Hispanic black, Mexican Americans, and other.
To estimate the relative income of the households with preschool-age children in the sample, the method suggested by the USA Census Bureau to calculate the PIR was employed.8 The PIR is an indicator of the total household income in relation to the number of individuals living in the household. Based on annually updated threshold incomes, families' incomes are compared to the threshold. In 2006, for instance, a weighted average threshold for a family of four was US$ 20,614, thus families with a combined income above this value were considered as not living in poverty. 9 Total household income was used in this study to represent individuals eligible for the federal Food Stamp Program (PIR < 1.3), as well as for the WIC (PIR < 1.85), or children living in families with medium income (1.85-3.4 PIR) and children in high income families (3.5-5.0 PIR). The PIR variable was capped at 5.0, so that an investigation of very high-income families was not possible using this data set. Two dichotomous variables were created to categorize children as Food Stamp or WIC participants (compared to income-eligible non-participants). An interaction term (Food Stamp vs. WIC) was created to examine the relationship between participating in either or both of the programs.
NHANES dietary data was collected with one interviewer administered 24-hour recall. Respondents were asked to report dietary intake during the past 24 hours using a multiple-pass approach.10 Caretakers reported the diets for children less than 6 years old. Intake information was disaggregated to provide dietary intake data for total energy (kcal per day), macro or micronutrients (g or mg per day), and MyPyramid11 food groups (in cups or ounces per day).
Overall diet quality was assessed using the Revised Children's Diet Quality Index (RC-DQI),12 an index based on national dietary intake recommendations, such as the Dietary Reference Intakes (DRI) for macronutrients and iron,13,14 MyPyramid,11 and position papers from the American Dietetic Association (ADA),15 and the American Academy of Pediatrics (AAP).16,17 The RC-DQI has a maximum of 90 points and consists of 13 components: added sugar, total fat, linoleic and linolenic fatty acids, docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), total grains, whole grains, vegetables, fruits, 100% fruit juice, dairy, and iron intake, as well as a component representing a proxy for energy balance (time spent watching television interacted with total daily energy intake). A full description of the RC-DQI can be found elsewhere.12 In short, usual dietary intake information is collected from the child's caretaker and scores assigned for each of the 13 components. The calculation of the score is based on under- as well as overconsumption of intake. The formula to calculate the points scored for each component is based on underconsumption - point score = (actual intake/ideal intake) * maximum points - or overconsumption - point score = maximum component points - (actual intake/ideal intake) * 100%. For instance, for a 2-year-old boy the recommended intake of fruit is 1.5 cups per day. If the child ate only 0.5 cups of fruits, he scores 3 points out of the possible 10 component scores: 10 - (0.5/1.5) * 100. Conversions to the metric system from the units of intake in MyPyramid (ounces and cups) can be based on the assumption that 1 ounce is equivalent to 28.4 g (e.g. one slice of bread) and one cup is equivalent to 240 mL.
The proportion of children who met the dietary income recommendations of the components and scored maximum points for each component was calculated and results described for the total sample. Total RC-DQI score as continuous value from zero to 90 points was the dependent variable in all linear regression models, whereas total RC-DQI score was divided into quartiles to examine the association between diet quality and the prevalence of childhood obesity.
Measured height and weight as well as calculated BMI are available in the NHANES data set. Weight (kg) was obtained as the individual stood on a digital scale. Standing height (m) was measured with an electronic stadiometer in individuals who were at least 2 years old. CDC's BMI-for-age and sex-specific growth charts were used to create four distinct groups: underweight (less than 5th percentile), healthy weight (5th to 84th percentile), at risk for overweight (between the 85th and 94th percentile), and overweight (³ 95th percentile). As expected, anthropometric data was not normally distributed, thus, the method by Cole et al. to construct a smoothed curve using the calculated power (L), mean (M), and coefficient of variation (S) to provided standards in terms of centiles was employed to determine children's normalized growth centile standards.18 Binary dummy variables were created that equaled "1" when children were classified as at risk for overweight or overweight and "0" otherwise.
All analysis was conducted using complex sample survey routines (version 9.2; StataCorp LP, College Station, TX, USA19) to maintain the nationally representative character of the data. Descriptive statistics, such as means and standard errors (SE), proportions, were calculated. The interrelationships between participation in food assistance programs and income required the assessment of endogeneity, thus, endogeneity between the income and income-related variables (such as child's age, sex, age squared, ethnicity, sociodemographic, body weight status, and federal food program participation) were tested using the Hausman-Wu test.20
The status of overall diet quality was described by calculating the population's mean RC-DQI score and the percentage of children receiving full point score for each of the 13 RC-DQI components. Multiple linear regression models were designed to examine the sociodemographic predictors of overall diet quality in the total sample as well as in the two subsamples of children who were income-eligible for the WIC or the Food Stamp Program. Forward and backward deletion process was employed with a p < 0.2 and < 0.25 respectively, indicating significance for the models. Likelihood ratio tests were conducted to examine the importance of each added/removed term to the model. Results were reported for the final models as beta-coefficient, 95% confidence interval and level of statistical significance (p). Quartiles of total RC-DQI scores were created to assess the relationship between level of diet quality and the prevalence of children being at risk for overweight or overweight. Pearson's chi-square test was employed to determine the significant difference in proportion of children at risk for overweight or overweight between the quartiles of RC-DQI scores. Statistical significance was assumed at a p < 0.05. The Institutional Review Board (IRB) in the Office of Research Protection at Pennsylvania State University granted approval for the study based on the use of secondary data with no person identifiers.
The descriptions of the samples can be found in Table 1. The proportion of non-Hispanic white children was highest in the total sample while more children in the low income samples were Mexican American or non-Hispanic black. Within the income-eligible subpopulations, approximately half of the children were enrolled in either program. One third of the preschoolers were either at risk for overweight or overweight.
The overall diet quality of the population was low, in that the population average total RC-DQI score was 59 points, ranging from 21-86 points of the total maximum of 90 points (Table 2). With the exceptions of the DHA and EPA, the total grain, or the iron component, less than half of the children met the intake recommendations of the RC-DQI components and achieved full points. Only 8% of the sample met the intake recommendation for whole grains.
Hausman-Wu test results showed that there was no endogeneity between the available sociodemographic variables and the children's household income variable. Thus, additional stratification by income-related covariates was not necessary. The regression models indicated that increasing age predicted lower overall diet quality (Table 3) while Mexican-American preschoolers had significantly better diet quality than non-Hispanic white children. No other ethnic group was significantly different from the referent group. Boys in the WIC income-eligible subsample had significantly better diet quality than girls ( Tables 4 and 5) and Mexican-American children scored almost 4 points more on the total RC-DQI than their non-Hispanic white counterparts. Food Stamp Program participation was not a significant predictor of overall diet quality in either subsample whereas a beneficial, although not statistical significant effect of WIC participation was indicated. The interaction between WIC and Food Stamp participation was not statistically significant in either model.
The proportion of children at risk for overweight or overweight decreased significantly between the lowest and the highest RC-DQI total score quartile (Figure 1). Although this trend was not consistent in the second and third quartile, the decrease of the numbers of children at risk for overweight between the first, second, and forth quartile of RC-DQI score was statistically significant whereas the difference in the proportion of children who were overweight was only statistically lower between the first and the second, the first and the fourth, as well as the third and the fourth RC-DQI score quartile.
Results of this study showed that preschooler's diet leaves much room for improvement. Particularly the consumption of whole grains, fruits, vegetables, and dairy were suboptimal.21 A 1-year increase in children's age was associated with a loss of approximately two points in the diet quality index. This phenomenon is likely due to the increasing independence in food choice and dietary intake with increasing age. While most children are still in the parent-guided transition period to table foods at 2 years of age,22 older children will have become independent eaters and choose their own foods for meals and snacks.
Increasing family income was predictive of better diet quality scores in the total population. The significance of family income on preschooler's diet quality was removed in the two low-income subpopulations. This finding indicates a threshold level in the relationship between family income and dietary intake. Family economic power has been found to predict the purchases of foods with high nutritional quality, such as whole grain starches (e.g. bread and pasta), fresh fish, fruits and vegetables.23 However, it has been found that WIC or Food Stamp Program participation increase the diet quality of children.24,25 Hence, federal food program participation appears to reflect change in behavior in that more resources are spent on the purchase of high-quality foods.
The large proportion of fruit and high-fiber vegetables, such as beans, that are traditionally consumed in the Mexican diet were likely the reason for the observed increased diet quality in Mexican-American children. This effect of ethnic background was particularly high in the WIC income-eligible subgroup. Hence, it appears that especially low-income children may benefit from adopting traditional Mexican diets. The importance of cultural background and family income on the foods provided in households with children has been established.26,27 However, the emergence of Mexican American ethnicity as the strongest positive predictor of good diet quality indicates the urgent need to further explore the potential underlying factors for this beneficial effect.
Although this study was limited by the use of one single 24-hour recall to estimate usual intake, it is based on a large, nationally representative data set that was designed for nutrition and health surveillance in the population. Clinicians may choose to ask the caretakers of young children to report children's usual diet, rather than intakes of the previous day alone.
The results presented here contribute new evidence to the diet-childhood obesity relationship. Although this study focused on the difference between children with Hispanic and non-Hispanic background, results are applicable to other ethnic groups as well in various countries. While large proportions of children in South America might not be considered Mexican but non-Hispanic white, dietary intake patterns are likely more similar to the Mexican American population than to non-Hispanic white American children. Furthermore, non-Hispanic white American children have similar rates of childhood obesity as Latin American and Caribbean children.28 However, while parts of the preschool population in South America are at high risk for childhood obesity, others are likely to develop medical manifestations of malnutrition, such as stunting.
While several foods and nutrients have been found to be associated with chronic disease risk and dietary intake recommendations are based on these diet-disease relationships,29,30 the direct association between lower diet quality and increased risk childhood obesity was demonstrated in this study. Results emphasized the need to improve overall diet quality, for instance by increasing whole grain, fruit, and vegetable intake. Certain ethnic groups are likely to have much better diet quality than others, especially in the low-income population. Hence, low-income children should be encouraged to consume foods common in the traditional Mexican diet, such as high proportions of fruits and vegetables to increase overall diet quality and decrease the risk for childhood obesity in preschoolers.
1. Kranz S, Siega-Riz AM, Herring AH. Changes in diet quality of American preschoolers between 1977 and 1998. Am J Public Health. 2004;94:1525-30. [ Links ]
2. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of overweight and obesity in the United States, 1999-2004. JAMA. 2006;295:1549-55. [ Links ]
3. Goodman E, Dolan LM, Morrison JA, Daniels SR. Factor analysis of clustered cardiovascular risks in adolescence: obesity is the predominant correlate of risk among youth. Circulation. 2005;111:1970-7. [ Links ]
4. Huang TT, Ball GD, Franks PW. Metabolic syndrome in youth: Current issues and challenges. Appl Physiol Nutr Metab. 2007;32:13-22. [ Links ]
5. Kranz S, Smiciklas-Wright H, Francis LA. Diet quality, added sugar, and dietary fiber intakes in American preschoolers. Pediatr Dent. 2006;28:164-71. [ Links ]
6. Epstein LH, Gordy CC, Raynor HA, Beddome M, Kilanowski CK, Paluch R. Increasing fruit and vegetable intake and decreasing fat and sugar intake in families at risk for childhood obesity. Obes Res. 2001;9:171-8. [ Links ]
7. United States Department of Health and Human Services. The National Health and Nutrition Examination Survey. >Hyattsville, MD: NCHS; 2004. [ Links ]
8. Census Bureau. How the census bureau measures poverty (official measure). http://www.census.gov/hhes/www/poverty/povdef.html; 2004. Acesso: October 16, 2007. [ Links ]
9. United States Census Bureau. How the census bureau measures poverty. In: Census Bureau, Housing and Household Economic Statistics Division. http://www.census.gov/hhes/www/poverty/poverty.html. Acesso: October 16, 2007. [ Links ]
10. Guenther PM, De Maio UJ, Ingwersen LA, Berlin MC. >The multiple-pass approach for the 24-hour recall in the Continuing Survey of Food Intakes by Individuals. 994-6. FASEB J 1996;10:A198. [ Links ]
12. Kranz S, Hartman T, Siega-Riz AM, Herring AH. A diet quality index for American preschoolers based on current dietary intake recommendations and an indicator of energy balance. J Am Diet Assoc. 2006;106:1594-604. [ Links ]
13. Institute of Medicine Food and Nutrition Board. Dietary reference intakes for vitamin a, vitamin k, arsenic, boron, chromium, copper, iodine, iron, molybdenum, nickel, silicon, vanadium, and zinc. Washington DC: National Academy Press; 2001. [ Links ]
14. Institute of Medicine of the National Academy of Sciences. Dietary reference intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and amino acids (macronutrients). Report. Washington, DC: National Academy Press; 2002. [ Links ]
15. Nicklas T, Johnson R, American Dietetic Association. Position of the American Dietetic Association: dietary guidance for healthy children ages 2 to 11. J Am Diet Assoc. 2004;104:660-77. [ Links ]
16. American Academy of Pediatrics Committee on Public Education. Children, adolescents, and television. Pediatrics 2001;107:423-426. [ Links ]
17. American Academy of Pediatrics. Committee on Nutrition. American Academy of Pediatrics: The use and misuse of fruit juice in pediatrics (re0047). Pediatrics. 2001;107:1210-3. [ Links ]
18. Cole TJ. The LMS method for constructing normalized growth standards. Eur J Clin Nutr. 1990;44:45-60. [ Links ]
19. STATA corporation. Stata statistical software: Release 9.0. In. 9.0 ed. College Station, TX; 2005. [ Links ]
20. Wooldridge JM. Econometric analysis of cross section and panel data. Cambridge: MIT Press; 2002. [ Links ]
21. Field AE, Gillman MW, Rosner B, Rockett HR, Colditz GA. Association between fruit and vegetable intake and change in body mass index among a large sample of children and adolescents in the United States. Int J Obes Relat Metab Disord. 2003;27:821-6. [ Links ]
22. Briefel RR, Reidy K, Karwe V, Jankowski L, Hendricks K. Toddlers' transition to table foods: Impact on nutrient intakes and food patterns. J Am Diet Assoc. 2004;104:s38-44. [ Links ]
23. Drewnowski A, Darmon N. Food choices and diet costs: an economic analysis. J Nutr. 2005;135:900-4. [ Links ]
24. Siega-Riz AM, Kranz S, Blanchette D, Haines PS, Guilkey DK, Popkin BM. The effect of participation in theWICc program on preschooler's diets. J Pediatr. 2004;144:229-34. [ Links ]
25. Rose D, Habicht JP, Devaney BL. Household participation in the Food Stamp and WIC programs increases the nutrient intakes of preschool children. J Nutr. 1998;128:548-55. [ Links ]
26. Kranz S, Siega-Riz AM. Sociodemographic determinants of added sugar intake in preschoolers 2 to 5 years old. J Pediatr. 2002;140:667-72. [ Links ]
27. Patrick H, Nicklas TA. A review of family and social determinants of children's eating patterns and diet quality. J Am Coll Nutr. 2005;24:83-92. [ Links ]
28. Duran P, Caballero B, de Onis M. The association between stunting and overweight in Latin American and Caribbean preschool children. Food Nutr Bull. 2006;27:300-305. [ Links ]
29. Gidding SS, Dennison BA, Birch LL, Daniels SR, Gillman MW, Lichtenstein AH, et al; American Heart Association; American Academy of Pediatrics. Dietary recommendations for children and adolescents: A guide for practitioners: consensus statement from the American Heart Association. Circulation. 2005;112:2061-75. [ Links ]
30. Pereira MA, Jacobs DR, Jr., Van Horn L, Slattery ML, Kartashov AI, Ludwig DS. Dairy consumption, obesity, and the insulin resistance syndrome in young adults: the CARDIA Study. JAMA. 2002;287:2081-9 [ Links ]
Department of Nutritional Sciences, Population Research Institute
Pennsylvania State University
5 G Henderson Building
16802, University Park, PA - USA
Tel.: +1 (814) 865.2138
Fax: +1 (814) 865.5870
Manuscript received Jul 26 2007, accepted for publication Oct 31 2007.
Sources of support
for this study came from the USDA, Economic Research Service Small Grant no.
#K-981834-09, and a Pennsylvania State University Seed Grant.
No conflicts of interest declared concerning the publication of this article.