DIETARY QUALITY OF LOCAL AND NATIONAL LOW-INCOME PRESCHOOL CHILDREN INDICATED BY HEALTHY EATING INDEX-2005. By Wen Guo A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Human Nutrition 2012 ABSTRACT DIETARY QUALITY OF LOCAL AND NATIONAL LOW-INCOME PRESCHOOL CHILDREN INDICATED BY HEALTHY EATING INDEX-2005. By Wen Guo Dietary quality in early life affects growth and development throughout childhood and influences food preferences and risks for developing chronic diseases in later adulthood. There is a lack of recent data on the diet quality of low-income preschoolers. Updated data on diet quality is important for health professionals to monitor dietary trends and modify educational interventions. The objective of this study was to report current dietary quality as indicated by the Healthy Eating Index-2005 (HEI-2005) of both local and national preschoolers. This is a crosssectional study using two secondary datasets. One sample was the Head Start children (n=330) from Michigan Greater Lansing Area and the other was the 3-5 year old children (n=505) surveyed in the 2007-2008 National Health and Nutrition Examination Survey (NHANES). The NHANES sample was categorized into subgroups by the Head Start status and children’s family income. Both datasets showed that preschool children’s dietary intake of total fruits, whole fruits, total grains, and milk met or approached the recommendations. However, vegetables, whole grains, saturated fat, sodium, and calories from solid fats and added sugars still need improvements. Nationally, children (n=105) who enrolled in Head Start (HS) had the highest overall HEI-2005 score (p=0.03) comparing to non-HS high income and low-income children, after adjusted for age, gender, race, and weight status. The non-HS low-income children scored lowest on all HEI-2005 subgroups except for meat. Low-income preschool children’s diets still need improvements. Whether the Head Start program benefits low-income preschoolers nutritionally by itself needs further study. ACKNOWLEDGEMENTS I would like to thank my major professor, Dr. Sharon Hoerr, for mentoring me during the past four years. I learned knowledge of nutrition and research from her, gained skills on management though working with her, and most important I was deeply influenced by her dedication and her professionalism. Her unconditional help and support on forming this study and editing this thesis helped me in completing the most important piece of work of my Master’s degree. She is and will always be a great role model in my life. I would also like to thank my committee members Dr. Lorrain Weatherspoon and Dr. Katherine Alaimo for their constructive advice in perfecting my thesis. Dr. Stan Kaplowitz and Dr. Won Song also generously offered great advice on the statistical part of this thesis. I really want to thank some of my colleagues as well, Dr. Melissa Reznar, Dr. Megumi Murashima, Edda Lungu, Sumathi Venkatesh, Julie Plasencia, Simone Wilson, DaYeon Shin, Amy Saxe, Tara Fischer, Caroline Martin, and Kellie Mayfield for supporting me and being wonderful friends. The department of Food Science and Human Nutrition gave me a precious opportunity to study in the US and offered wonderful resources during my time here. I truly appreciate all the help that I received and will never forget my time at Michigan State University. I want to thank my parents too, Mrs. Lining Shen and Mr.Yaqian Guo for raising me, educating me, and motivating me to pursue higher education and to explore the world. Last but not least, I shall thank my beloved husband, Mr. Chen (Kevin) Zhou for loving me for who I really am, for bearing my temper, and for holding my hands while we walk though tough moments in life. iii TABLE OF CONTENT LIST OF TABLES ......................................................................................................................... vi LIST OF FIGURES ..................................................................................................................... viii LIST OF ABBREVIATIONS ........................................................................................................ ix CHAPTER 1 INTRODUCTION ................................................................................................... 1 Rationale...................................................................................................................................... 1 Specific Aims and Hypotheses .................................................................................................... 2 CHAPTER 2 REVIEW OF LITERATURE .................................................................................. 4 Dietary Recommendations for Preschool Children..................................................................... 4 Overall Dietary Pattern and Energy Intake ............................................................................. 4 Food Groups ............................................................................................................................ 5 Nutrients of Importance for Preschoolers................................................................................ 6 HEI-2005 is a Valid Indicator of Diet Quality ............................................................................ 8 Diet Quality of Preschool Children in the U.S. ......................................................................... 11 Diet Quality of Low-Income Preschoolers................................................................................ 13 Issues in Measuring the Dietary Intake of Preschoolers ........................................................... 16 CHAPTER 3 METHODS ............................................................................................................. 19 Study Design ............................................................................................................................. 19 Samples and Recruitment .......................................................................................................... 19 Overall Study Procedures .......................................................................................................... 22 Measurements and Variables .................................................................................................... 23 Michigan Head Start Children Sample .................................................................................. 23 NHANES Sample .................................................................................................................. 24 Healthy Eating Index-2005 (HEI-2005) ................................................................................ 29 Demographics ........................................................................................................................ 33 Analysis ..................................................................................................................................... 33 Data Preparation .................................................................................................................... 34 Calculating the Healthy Eating Index-2005 Scores............................................................... 35 Data Analyses by Each Aim .................................................................................................. 37 CHAPTER 4RESULTS ................................................................................................................ 39 Demographics ........................................................................................................................ 39 Diet Quality Indicated by HEI-2005 ..................................................................................... 40 For Aim 3 .................................................................................................................................. 44 Demographics ........................................................................................................................ 44 Ho3.1 ..................................................................................................................................... 47 Diet Quality Indicated by HEI-2005_Population Ratio(PR) Method ................................... 47 Ho3.2 ..................................................................................................................................... 54 Diet Quality Indicated by HEI-2005_Mean Score(MS) Method .......................................... 54 iv CHAPTER 5 DISCUSSIONS....................................................................................................... 59 Michigan Head Start Sample..................................................................................................... 59 2007-2008 NHANES Sample ................................................................................................... 61 Strengths and Limitations.......................................................................................................... 65 Strengths ................................................................................................................................ 65 Limitations............................................................................................................................. 66 CHAPTER 6 SUMMARY and CONCLUSIONS ........................................................................ 68 REFERENCES ............................................................................................................................. 71 v LIST OF TABLES Table 1 Estimated Average Requirement and Adequate Intake for key nutrients for young children……………………………………………………………………………………………7 Table 2 HEI-2005 components and standards for scoring….…….……………………………...30 Table 3 Example for selecting maximum standard for total fruit food group.…..……………....31 Table 4 Comparison of demographics and weight status for MIHS and HS children from 20072008 NHANES…………………………………………………………………………………..39 Table 5 Mean, SD for the total and subgroup HEI-2005 scores of MIHS children (n=330) and the possible maximum scores for the total and subgroup of HEI-2005.………………………..........40 Table 6 Mean, SD for daily total energy the key nutrients intake of MIHS children (n=330), percentage of MIHS children met the recommendations, and the EAR or AI for the key nutrients…………………………………………………………………………………………. 42 Table 7 Frequency and percentage of milk type chosen by MIHS children (n=330)..….…….....43 Table 8 Frequency and percentage of cold cereal type chosen by MIHS children (n=330)..........43 Table 9 Comparison of total and subgroup HEI-2005 scores between MIHS and NHANES HS children…………………………………………………………………………………………..44 Table 10 Demographics and weight status of NHANES subgroups at different income cut point……………………………………………………………………………………………...46 Table 11 Total and subgroup HEI-2005 scores using the population ratio method for all NHANES preschool children (n=505) and by gender.…………………..……………………....48 Table 12 Total and subgroup HEI-2005 scores using the population ratio method for NHANES preschool children by weight status………………………….…………………………………..49 Table 13 Total and subgroup HEI-2005 scores using the population ratio method for NHANESs children 3-5 year old, by Head Start attendance and income status at 1.0PL, unadjusted for gender, age or BMI percentile…………………………………………………………………...50 Table 14 Total and subgroup HEI-2005 scores using the population ratio method for NHANES children 3-5 year old, by Head Start attendance and income status at 1.3PL, unadjusted for gender, age or BMI percentile…………………………………………………….……………..51 Table 15 Total and subgroup HEI-2005scores for NHANESs children 3-5 year old, by Head Start attendance and income status at 1.85PL, unadjusted for gender, age or BMI percentile……………………………………….......…………………........................................52 vi Table 16 Total and subgroup HEI-2005 scores using the population ratio method for NHANESs children 3-5 years old, by Head Start attendance and income status at 1.3-1.85PL, unadjusted for gender, age or BMI percentile.......................................................................................................53 Table 17 HEI-2005 scores calculated using two methods: Mean Score Method and Population Ratio Method for all preschool children from 2007-2008 NHANES..........................................54 Table 18 HEI-2005 scores of NHANES children by Head Start status and 1.85 poverty level calculated using Mean Score Method and adjusted for age, gender, race/ethnicity, and BMI percentile………….…………………………………………………………………………….56 Table 19 Comparison of NHANES children’s nutrients intake (means + confidence interval) by Head Start and income status at 1.85 poverty level, controlled for age, weight, race/ethnicity, and BMI percentile…………………………………………………………………………………..58 vii LIST OF FIGURES Figure 1 Grouping of NHANES preschool children by Head Start and income status.................29 Figure 2 Formula for the Population Ratio Method used to calculate the Healthy Eating Index2005 scores.………………………………………………………………………………………35 Figure 3 Comparison of the subgroups HEI-2005 scores of MIHS children (n=330) to the possible maximum score for each subgroup of HEI-2005………………...……………………..41 viii LIST OF ABBREVIATIONS AHA: American Heart Association AI: Adequate Intake BKFS: Block Kids Food Screener CDC: Centers of Disease Control and Prevention CNPP: Center for Nutrition Policy and Promotion DGA: Dietary Guidelines for Americans DOVL: Dark green, orange vegetables and legumes DRI: Daily Reference Intakes EAR: Estimated Average Requirements Eq: Equivalent FDA: Food and Drug Administration FFQ: Food Frequency Questionnaire HEI-2005: Healthy Eating Index-2005 HS: Head Start Kcal: Kilocalorie mcg RAE: micrograms of retinol activity equivalents MI: Michigan MIHS: Michigan Greater Lansing Area Head Start MVUs: Masked Variance Units NHANES: National Health and Nutrition Examination Survey Non-HSHI: Non-Head Start high income group Non-HSLI: Non-Head Start low income group ix Non-HSmidI: Non-Head Start middle income group PL: Poverty Levels PSUs: Primary Sampling Units Sat. Fat: Saturated Fat Sod: Sodium SoFAAS: Calories from saturated fats, alcohol and added sugars SoFAs: Calories from solid fats and added sugars UI: Upper Limit USDA: United States Department of Agriculture x CHAPTER 1 INTRODUCTION Rationale Previous studies have shown that the diet quality of low-income preschool children is poor [Kranz, 2006; Bucholz, 2011]. This is a problem because dietary intake in early life not only affects growth and development throughout childhood, but also influences food preferences and risks for developing chronic diseases, such as obesity, in later adulthood [Van Duyn, 2000; Maynard, 2003; Ogden, 2008; De Koning, 2011; McCullough, 2011]. The preschool period is critical, because children 3-5 years of age usually have a different diet from that of infants, toddlers and school-aged children [Brown, 1998]. At this time, young children start to develop their own food preferences and eating behaviors [Harris, 2008; Liem, 2010]. Head Start (HS) is a federal program that provides early childhood education, health and nutrition services, and family support for preschoolers from low-income settings or for those with special needs. Nationwide, 983,809 children were enrolled in HS in 2009-10 [US Department of Health and Human Services, 2010b]. For several decades, nutrition researchers and educators have targeted HS families as a population of concern for both obesity and diet quality. To our knowledge, however, there is a lack of recent data on the dietary quality of HS preschoolers and few have used the Healthy Eating Index-2005 (HEI-2005) as a diet quality indicator [LaRowe, 2007; Bucholz, 2011]. Both policy makers and nutrition educators need current information on diet quality for designing interventions or in implementing new program emphases. New information on preschoolers’ diet quality can help educators and policy makers gain insights about which food groups and nutrients need improvement. Finally, it is also important to monitor preschoolers’ dietary intake to examine the effects of national policies and supplemental programs on their health and assist future policymaking. 1 The main goals of this study were to report the diet quality of 330 children in the Michigan Greater Lansing Area Head Start (MIHS) using the HEI-2005 and to report the diet quality of the same age group (3-5years of age, n=505) from the 2007-2008 National Health and Nutrition Examination Survey (2007-08 NHANES). Michigan has one of the highest unemployment and poverty rates in the nation [Michigan Department of Human Services, 2008]. Meanwhile, Michigan HS has a racially/ethnically diverse population with an unusually high prevalence of child overweight and obesity [Murashima, 2011]. Thus it is a critical area for researchers to study. The NHANES population was further categorized into different subgroups depending upon the children’s Head Start Status and their family’s income status. The family’s income status was indicated by poverty level (PL) in this study, and defined in methods section. The researcher made comparisons between those NHANES subgroups for differences in whole grains, dairy, fruits, vegetables, and solid fats/added sugars. The ultimate goal for the subgroup comparison was to examine the relation of participation in the Head Start program and of income status to the child’s dietary quality. The specific aims listed below were used to accomplish these goals. Specific Aims and Hypotheses Aim1. Describe preschoolers’ diet quality both as food groups using HEI-2005 and for selected key nutrients from food intake data of Michigan Head Start (MIHS) Children. Ho1.1: The subgroup and total HEI-2005 score of MIHS children’s diet will not meet the Mypyramid recommendations, especially for whole grain, low-fat dairy, total vegetables, dark green and orange vegetables, sodium, saturated fat, and solid fats and added sugars (SoFAs). 2 Ho1.2: The intake of MIHS children will not meet the Estimated Average Requirements (EAR) for preschool children for selected key nutrients--fiber, folate, Vitamin A, Vitamin E, iron, zinc calcium, and potassium. Aim2. Compare the diet quality from food intake data obtained using the Block Kids Food Screener (BKFS) of MIHS children to the diet quality from one-day 24-hour dietary recalls of preschoolers in the 2007-08 NHANES. Aim3. Compare the diet quality of 2007-08 NHANES Head Start children to those of nonHead Start higher-income (non-HSHI) children and to those of non-Head Start low-income (non-HSLI) children from the 2007-08 NHANES using different poverty levels (PL). Ho3.1: When using the 1.85PL, the total and most subgroup HEI-2005 scores calculated using Population Ratio method will significantly differ between the HS children, nonHSLI, and non-HSHI children of the 2007-08 NHANES sample. Ho3.2: When using the 1.85PL, the total and most subgroup HEI-2005 scores calculated using Mean Score method will significantly differ between the HS children, non-HSLI, and non-HSHI children of the 2007-08 NHANES sample. 3 CHAPTER 2 REVIEW OF LITERATURE In this literature review section, the researcher summarized the current dietary recommendations for preschool children and described the dietary intake and dietary quality of this age group. The researcher also illustrated the importance of monitoring and improving preschoolers’ diet quality and the documented benefits from preschool programs for low-income populations, such as Head Start. Finally, the researcher established that there is a need to examine preschoolers’ most recent dietary quality. Dietary Recommendations for Preschool Children It is important to identify current dietary recommendations for preschoolers, in order to establish the criteria for the comparisons the present study will make. In this section, the researcher examines current recommendations for preschool children including, overall dietary patterns and energy intake, food groups, and key nutrients. Overall Dietary Pattern and Energy Intake The American Heart Association (AHA) has suggested that , “People who are two years and older should consume a diet mainly based on fruits, vegetables, legumes, whole grains, low-fat and non-fat dairy and lean meat or fish.” [Gidding, 2005]. This suggestion is consistent in principle with the Dietary Guidelines for Americans 2010 (DGA 2010), where more detailed recommendations were made for each age group for each food group and key nutrients [Dietary Guidelines for Americans, 2011]. The DGA 2010 and the MyPlate food guide [Dietary Guidelines for Americans, 2011; USDA, 2011a] made recommendations for preschool children to maintain their calorie intake at the level 4 of 1,000 kcal for ages 2-3 years and 1200-1400 kcal for ages 4-8. This recommendation was based on a sedentary lifestyle of less than 30 minutes daily of physical activity. The DGA also recommended an increase in energy intake when physical activity was increased [Dietary Guidelines for Americans, 2011]. When preschool children’s energy consumption was at the recommended level, it should ensure the children maintain an appropriate weight-for-age while permitting adequate nutrients for growth. Food Groups For specific food groups, the DGA recommended that children 2-3 years old consume 3 oz equivalents of grain products, 2 oz of lean meat, 1cup of fruits, and 1cup of vegetables per day. Children 4-8 years old should consume 4-5 oz equivalents of grain products, 3-4oz of lean meat, 1 to 1½ cups of fruit, and 1½ cups of vegetables daily per day [Dietary Guidelines for Americans, 2011]. For both age groups, children should have 2 cups of fat-free or low-fat milk each day. Fatfree and low-fat milk offer the same nutrients as does whole milk, but with less discretionary fat. The DGA and MyPlate also emphasized that at least half of the grain products should be whole grains and that children should increase their intake of whole fruits, and dark green and orange vegetables [USDA, 2011a]. For the meat group, even though the DGA, MyPlate, and AHA recommended increasing fish consumption as a meat source, The Food and Drug Administration (FDA) recommends avoiding certain types of fish, such as swordfish and king mackerel due to concern for mercury contamination [Gidding, 2005]. 5 Nutrients of Importance for Preschoolers The DGA 2010 also highlighted several nutrients of concern, such as sodium, saturated fat, sodium, iron, and calcium [Dietary Guidelines for Americans, 2011]. Although, sodium is a necessary nutrient, high intakes are strongly associated with high blood pressure conditions for adults as well as for children [National Research Council, 2005b]. Sodium exists widely in foods, but is especially high in commercially prepared foods. Yeast breads, mixed dishes with chicken, and pizza were the top three sources of sodium, in part, due to high frequency of consumption [Dietary Guidelines for Americans, 2011]. Therefore, the Adequate Intake (AI) for sodium is less than 1000mg/day sodium per day for children ages 1-3 years and less than 1200mg/day for those 4-8 years of age. The Upper Limit (UI) for sodium for children under 14 years old is 2300mg/day [National Research Council, 2005b]. Dietary fat is a complicated food component regarding its dietary benefits. The DGA 2010 recommend eliminating all trans fat in the diet, limiting saturated fat to 10% of total energy intake, and consuming more mono- and poly-unsaturated fats than is currently typical. Cholesterol intake should be less than 300mg per day [National Research Council, 2005a]. Important nutrients that most preschoolers in the US do not consume in adequate amounts include fiber, iron, zinc [Rose, 1995; Alaimo et al., 2001] calcium, potassium, Vitamins A and E [Bucholz, 2011], and folate [Kranz, 2006]. The table below provides the current Estimated Average Requirement (EAR) and Adequate Intake (AI) for key nutrients that will be analyzed in this study [National Research Council, 2005b; National Research Council, 2000; National Research Council, 2001; National Research Council, 2010]. 6 Table 1 Estimated Average Requirement and Adequate Intake for key nutrients for young children. Component Recommendations 1-3 years old 4-8 years old Fiber (g/d) 19 25 Cholesterol (mg/d) <300 < 300 Vitamin A, (mcg RAE /d) 210 275 Vitamin E (mg/d) 5 6 Total folate (mcg/d) 120 160 Calcium (mg/d) 500 800 Phosphorus (mg/d) 380 405 Magnesium (mg/d) 65 110 Iron (mg/d) 3.0 4.1 Zinc (mg/d) 2.4 4.0 Potassium (mg) 3000 3800 These nutrients were selected for examination in this study, first because they play an important role in health and development of chronic disease. For example, the balance of sodium and potassium in the body helps maintain normal blood pressure [National Research Council, 2005b]. Calcium and phosphorus contribute to bone health [National Research Council, 2011]. Another reason for inclusion is that some nutrients are frequently reported to be over consumed (like saturated fat, and sodium) [Krebs-Smith et al., 2010] or frequently deficient in diets of young children (like the vitamins and minerals just mentioned above) [Bucholz, 2011]. Practitioners, researchers and policymakers need current intake data on these nutrients to guide educational programs, policies and research. In order to understand why these nutrients might be low in the diets of preschoolers, it is important to examine their distribution in foods and beverages. Fiber is found mainly in whole grains, fruits and vegetables-foods that are typically found low in the diets of many preschoolers. Fortified grains, meat, orange juice, dark leafy greens, dried beans, and egg yolks are rich in folate. Due, in part, to inadequate intake, as well as high needs for growth, iron deficiency 7 anemia is prevalent among low-income preschoolers [Alaimo et al., 2001; Bucholz, 2011]. Calcium is found mainly in dairy products, fortified cereals and fortified fruit juices [World Food Programme, 2011]. National survey data from 1999-2004 showed that preschoolers did not have adequate intakes of some of these nutrient-dense foods, which likely contributed to the low intakes [Fungwe et al., 2009; Bucholz, 2011]. Such findings strongly support that children’s diet quality needs improvement. The criteria for nutrient adequacy will be the EAR from the Recommended Dietary Allowances as shown in Table1 [National Research Council, 2000; National Research Council, 2001; National Research Council, 2005b; National Research Council, 2010]. For evaluating diet quality, the Healthy Eating Index-2005 (HEI-2005) is the standard indicator and will be used. Intakes of key food groups from MyPyramid and of several nutrients (like sodium, added sugars, and solid fats) are components of HEI-2005. The next section describes the diet quality indicator in detail and why it is pertinent to the diets of young children. HEI-2005 is a Valid Indicator of Diet Quality Due to the need for a single number to describe the diet quality of individuals, instead of a list of percentages as one gets with single nutrients, nutrition researchers have sought to identify a dietary quality indicator for decades [Guenther, 2007; Guenther, 2008c]. Such an indicator would be valid, indicate diet quality and predict health risks. For this reason the U.S. Department of Agriculture’s (USDA) Center for Nutrition Policy and Promotion (CNPP) developed the Healthy Eating Index (HEI) [Guenther, 2008c] in 1995 to monitor changes in the diet of the US population. HEI 2005is an improved version of the original HEI. HEI-2005 includes 12 components compared to the original HEI with only 10 components. Other differences from the 8 original HEI include the elimination of the cholesterol, and a subcomponent that indicated the number of variety of foods that a person had, addition of whole fruits, dark green and orange vegetables (DOVL) and whole grains, and replacement of the total fat component with Calories from saturated fats, alcohol and added sugars (SoFAAS) [Guenther, 2007; Guenther, 2008c]. Because preschool children have no intake of alcohol, in this particular study, only calories from saturated fats and added sugars (SoFAs) were used for this component. The HEI-2005 score has undergone rigorous validation studies. The indicator was tested for content validity by checking the HEI-2005 components against the DGA 2005; the HEI-2005 reflected all key recommendations [Guenther, 2008c]. Construct validity was tested in three different ways. First, researchers calculated the scores for several expertly designed menus, including a MyPyramid sample menu [USDA, CNPP, 2006], the menu for Dietary Approaches to Stop Hypertension (DASH) [National Heart Lung and Blood Institute, 2006], the Harvard menu [Willett, 2005], and the AHA’s No-Fad Diet [American Heart Association, 2005]. The HEI-2005 provided a maximum, optimal score for the MyPyramid and DASH menus as expected, because the HEI-2005 and the DASH diet are based on MyPyramid. The AHA diet plan scored well for most HEI-2005 components too. However, the Harvard menu scored low for the Milk group, because that plan does not encourage milk consumption. Next, researchers examined how the HEI-2005 would distinguish between sample groups with known differences in their diet quality. A study was done using one-day dietary intakes of smokers vs. non-smokers. The results showed that nine out of 12 HEI-2005 components for the diets of smokers were significantly poorer than those of non-smokers. Finally, to ensure the index’s components did not depend on energy intake, Pearson correlations between the HEI-2005 component scores and energy intakes were performed. SoFAAs had the highest correlation with energy and other food 9 groups had correlations less than 0.11 [Guenther, 2007; Guenther, 2008c]. Thus, researchers demonstrated good validity for the HEI-2005. Several studies have demonstrated the HEI-2005 to be a good discriminator of diet quality [LaRowe , 2007; Fungwe et al., 2009; Beydoun, 2011; O’Neil, 2011a; O’Neil, 2011b; O’Neil, 2011c]. O’Neil et al. showed that 100% fruit juice consumption of 2-5 year old children was positively associated with a better HEI-2005 score [O’Neil, 2011b]. She also showed that mean HEI-2005 scores were lower in chocolate candy consumers [O’Neil, 2011c]. Diet quality assessed using the HEI has been related to weight status in different research studies of both children and adults. A Canadian study with adolescents revealed that the HEI-C (Canadian) score was low among students with weight concerns and who were dieting (p<0.001), and among students frequently skipping breakfast (p<0.001) [Woodruff, 2008]. Shah et al. reported that among postpartum low-income women, the HEI-2005 scores were inversely associated with BMI, Low Density Lipoproteins, and total cholesterol, but positively predicted serum High Density Lipoprotein levels [Shah, 2010]. The longitudinal Whitehall II study with London office workers who had metabolic syndrome revealed a high HEI score was associated with better odds (OR= 1.88) of reversing metabolic syndrome; that is, of not having it after 5years [Akbaraly, 2010]. In addition, the HEI-2005 significantly predicted depression, cardiovascular diseases and diabetes in other studies [Drewnowski, 2009; Kuczmarski, 2010; De Koning, 2011]. As a widely used and valid tool for indicating diet quality, HEI-2005 will be the indicator employed in this study to examine diet quality of preschool children. 10 Diet Quality of Preschool Children in the U.S. The diet quality of preschool children from all income levels in the U.S. needs improvement, especially for whole grains, vegetables, and discretionary energy [Krebs-Smith et al., 2010]. Low intake of fruits, vegetables, and whole grains, along with high intake of solid fats and added sugars, is strongly correlated with overall poor diet quality [Weissberg, 1997; Van Duyn, 2000]. National surveys showed that most children consistently consumed inadequate levels of whole grains, dark green and orange vegetables [Kantor, 2001, Fungwe et al., 2009, O’Neil, 2010, Krebs-Smith et al., 2010]. Intakes of whole grains have not improved in the decade from the mid 1990’s to 2005. In 1994-96, the average daily intake of whole grain was only one serving for children 2-19 years old [Harnack, 2003]. In 1999-2004, children 2-5 years old consumed an average of 0.45 servings of whole grain daily and only 8.4% of those 2-5years old consumed more than one and half servings of whole gains [O’Neil, 2010]. The NHANES data from 2001-04 further demonstrated that only 1% of 2-8 years olds met the recommendations for whole grains [Krebs-Smith et al., 2010]. Intake of vegetables also remained at a low level. During 2001-04, 20% of children 2-3 years old met the national dietary recommendation for vegetables, 3% met the recommendation for dark green vegetables, 11% ate enough orange vegetables, while half of the 2-3year olds ate more than the recommended amount of whole fruits [Krebs-Smith et al., 2010]. The trend of low intake of vegetables was worse among children 4-8 years old, with only 8%, 1.3%, and 5.5% meeting the recommendations for vegetables, dark green, and orange vegetables, respectively [Krebs-Smith et al., 2010]. A national study using Healthy Eating Index-2005 (HEI-2005) as an indicator for diet quality in 2003-04 NHANES further showed that fruit scores for all children 11 were close to the full score of 5 for achieving the recommendations [Fungwe et al., 2009]. However, children 2-5 years old surveyed scored only 2.2 out of 5, 0.6 out of 5, 0.8 out of 5 for total vegetables, dark green and orange vegetables, and whole grains intake, respectively. With the increasing attention on child obesity, the discretionary energy from solid fats and added sugars has become a major dietary concern due to the increasing prevalence of child obesity [KralTanja, 2010]. The DGA 2010 suggested that energy from solid fat, added sugars (SoFAs) should be no more than 5-15% of total daily energy intake for most people [USDHHS, 2010a]. Almost 99% of children’s had intakes of solid fats and added sugars (SoFAs) exceeding this maximum discretionary energy allowance in 2003-04 [Krebs-Smith et al., 2010]. The National Cancer Institute reported that the top food sources of solid fats in the diet of 2-3 year old American children were whole milk, sausage, hot dogs, bacon, ribs, and regular full fat cheese [NCI, 2010a]. For those children 4-8 years old, whole milk, pizza, and grain-based desserts were the top three contributors of solid fats [NCI, 2010a]. The top sources of added sugars in the diets of 2-8 American children were soda, energy drinks, and sports drinks, fruit drinks, and grain-based desserts [NCI, 2010b]. Children ages 2-5years who drank high-fat milk and fruit juice as their primary beverage had a higher energy intake than children who drank mixed beverages (children who drank all types of beverage--milk, soda, juice, to tea and coffee) or water [LaRowe, 2007]. Knol et al. reported that preschoolers from 1994-1998 Continuing Survey of Food Intake by Individuals data had original HEI score from 65.8 to 78.1 depending on their eating patterns. The original HEI scores of 51-80 indicated that the diets in the US needed improvement. The USDA reported the diet quality of children ages 2-17 years from the 2003-2004 NHANES. This brief report indicated that children 2-5 years of age had a total HEI-2005 score of 59.6 and those 12 children met recommendations for total fruit, total grain, and milk. However, their intakes of DGOV and whole grains were very low. To our knowledge, few researchers have used the HEI2005 as an indicator to report preschooler’s diet quality and none have used NHANES data after the 2003-2004 cycle. Diet Quality of Low-Income Preschoolers Preschool children from low-income families are at high risk of consuming a poor diet [Gibson et al., 1998; Hoerr et al., 2006a; Bucholz, 2011.]. In 2009, 20.7% of children less than 18 years old lived in poverty in the US [U.S. Census Bureau, 2009]. Low-income children are less likely to consume DOVL, whole grains, whole fruits, and low-fat dairy products and more likely to consume more sweetened soft drinks, added sugar, and sodium than those from higher income groups [Gibson et al., 1998; Hoerr et al 2006b; USDA, 2008; Guenther et al., 2008]. Data from a 2003-2004 NHANES study comparing diet quality between different poverty levels suggested that the low-income population over twenty years of age consumed significantly less total vegetables, dark green and orange vegetables and whole grains, but more SoFAAs compared to those of higher income groups [Guenther et al., 2008a]. However, this trend was not observed for children ages 2-18 years [Guenther et al., 2008a]. The reason might be that most low-income school age children and some preschoolers receive school meal programs, which contribute to good diet quality [Clark, 2009; Stevens, 2011]. For preschoolers, a cross-sectional study with 358 Head Start children using 1-day parental dietary recalls suggested that most Head Start children met the dietary recommendations for vitamins, minerals and protein, but their intake of total fat, saturated fat, and cholesterol exceeded recommendations [Bollella, 1999]. A cluster analysis of national data further supported 13 these findings [Knol, 2005]. Knol found that 48.6% of low-income 2-3 year old children tended to consume excess energy, mostly from refined grains and meat, and higher amounts of fat, sodium, and cholesterol [Knol, 2005] compared to children from higher income families using 1994-1996 and 1998 data from the Continuing Survey of Food Intakes by Individuals. Although 73.4% of 4-8 year old children were categorized into lower-energy intake patterns, discretionary fat and added sugars still contributed approximately 40% of the energy intake and those children rarely consume enough fruits, vegetables, and whole grains [Knol, 2005]. Therefore, it is important for researchers to focus on these specific dietary characteristics of low-income preschoolers. Diet quality related to health status. Monitoring low-income preschooler’s diet quality is important because diet quality is strongly associated with health status. Poor diet quality is associated with overweight and obesity, which affects 30% of low-income preschool children [USDA, 2008; Ogden, 2008]. Poor diet quality in early life stages has also been shown contribute to early-onset of diet-related diseases as well as later onset of such diseases. Such dietrelated chronic diseases include cardiovascular disease, diabetes, and cancer [Van Duyn, 2000; Maynard, 2003; De Koning, 2011; McCullough, 2011]. Preschooler’s dietary behaviors differ from other age groups. The preschool period, from age 3-5 years, is the time that children start to establish their own taste preferences and eating behaviors [Harris, 2008; Liem, 2010]. By the age of three years, most children have graduated from a high chair to join the family at the table [Brown, 1998]. Different from infants, who usually eat separately prepared foods, preschoolers typically share meals with adults [Brown, 1998]. The food environment, however, is often less than ideal in many low-income 14 families [Cooke, 2004], and the diet quality of many low-income caregivers’ is poor [Hoerr, 2008]. Preschoolers experience food neophobia and gradually develop acceptance and rejection of certain foods [Skinner, 2002; Kleinman, 2009]. Preschoolers learn to accept different foods and develop eating habits by observing and modeling behaviors of adults and peers [Cashdan, 1994]. Even though preschool children begin to exert control over their own eating, if there are only limited options for nutrient-dense foods available and the children model the poor eating behaviors of others [Hoerr, 2006a; Hoerr, 2009], then the dietary quality of those low-income preschoolers can gradually decline [Gable, 2000; Newby, 2007; Gorin, 2007; Brown, 2008]. Without a school meal program to provide at least one-third of the Dietary Reference Intake [USDA, 2012a], dietary intake of preschool children can be poorer that of school-age children. Exceptions might be the preschool children enrolled in preschool programs with federal oversight, such as Head Start [Bollella, 1999]. Considered together, these studies support that the preschool years are important in developing a healthy diet patterns and is an area that needs attention. A small body of evidence has supported the benefits of Head Start Program. Head Start is a federal preschool program serving 3-5 year old children from families with a gross income less than 100% of the poverty guidelines [US Department of Health and Human Services, 2010b]. Most researchers who have studied HS programs, examined the menus of HS programs but rarely the diet intake or diet quality [Oakley, 1995; Fox, 1997]. In two studies, however, the intakes of key nutrients were examined for HS children. Bucholz et al. suggested that Head Start children had lower intakes of protein, saturated fat, vitamin B2, phosphorous, and calcium 15 compared to children who were not in Head Start or who were previously in Head Start using 1999-2004 NHANES data [Bucholz, 2011]. Other research has compared the diet quality of Head Start children at home and at school [Bollela, 2005]. They found that the children who attended half day Head Start programs achieved one-fourth of the daily recommended amount of energy and nutrients and children who were in all day programs had at least one-third of the nutrient recommendations [Bollella, 2005]. Both of these studies examined HS children’s diet from the perspective of nutrient intake but not the overall dietary quality or food groups. It is important to monitor intake of foods, because foods provide nutrients, and nutritionists and dietitians recommend that the best way to meet nutrient intake is through improved food intake. This current study provides a different perspective the Head Start program’s benefits. The Head Start program is free to attendees, but it is a large financial cost for the government and faces budget cuts under the current economic crisis [US Department of Health and Human Services, 2011]. Evidence from this current study can add to the small body of literature on diet quality of HS preschoolers. Issues in Measuring the Dietary Intake of Preschoolers An accurate and efficient measurement of dietary intake of preschoolers is crucial to monitor trends and changes. Concerns about which instruments are best to assess dietary intake have occurred for decades and the choice is always a trade-off between validity, reliability, respondent burden, and the cost [Abdullah, 1993; Thompson, 1994; Serdula, 2001; Michels, 2009; Burrows, 2010]. At the group level, most dietary assessment methods provide a good estimation of the total energy intake [Burrow, 2010]. The accuracy of each method will vary among different study samples and partially depend on the study design. There is no one best 16 instrument that is suitable for all diet-related studies in every situation. In a review of 25 dietary assessment studies for preschoolers, 24-hour food recalls were the most used method in 12 studies, followed by Food Frequency Questionnaires (FFQ) (n=9) and food records (n=2) [Serdula, 2001]. Evidence supported that 24-hour dietary recalls can over or underestimate energy intake, especially when using parents as proxies for young children [Reilly, 2001; Montgomery, 2005]. Preschool children’s intake can vary widely from day to day, but remain relatively stable over a weekly period [Brown, 1998]. As Vucic et al. (2009) and Burrows et al. (2010) suggested, when using 24-hour dietary recalls, at least three days of interviews are needed to yield relatively accurate results for an individual’s intake. NHANES has been using two days of dietary recalls since 2004, because two day recalls is the best balance of intake accuracy versus survey cost [CDC, 2008b; CDC, 2008c]. Other than the workload of conducting several non-consecutive recalls, performing 24 hour dietary recalls requires highly trained personnel and can be challenging to perform in a population with low-literacy [Wolever, 1997; Tran, 2000]. Traditionally, a FFQ includes about 100 food items, but most shortened FFQ consist of a list of approximately 50 food items and usually take approximately 15 minutes to complete [Block, 1990; NutritionQuest, 2011]. FFQ have been sometimes reported to overestimated energy intake and worked best for rank-ordering food and nutrient intakes [Treiber, 1990; Serdula, 2001, Burrow, 2010]. Evidence also suggested that FFQ are better for assessing usual intake, because 24-hour recalls provide dietary intake only for the recalled time period [Zulkifle, 1992; Thompson, 1994]. Compared to the high respondent burden and high cost of multiple 24hr recalls [Willett, 1990], a food frequency questionnaire (FFQ) or screener completed by parents is easier to conduct than dietary recalls for young children [Blum, 1999]. As the outcomes of 17 interest have shifted over years from nutrients to foods, FFQ have gained increasing acceptance for assessing dietary intakes [IOM, 2002; Spurrier et al. 2008], especially when ranking of intake from low to high is desired [Willett, 1990; Thompson, 1994]. 18 CHAPTER 3 METHODS Study Design This was a cross-sectional study using secondary data from two different datasets that reported dietary intake of preschool children in the U.S. One was a local sample from Michigan and the other was a national dataset. The Michigan dataset (n=330) was from a study conducted by Murashima in the Greater Lansing Area with local Head Start programs [Murashima, 2010]. The national dataset (n=505) was from the 2007-2008 National Health and Nutrition Examination Survey (2007-08 NHANES). The first part of this study compared the diet quality of Michigan Head Start (MIHS) children to the current national recommendations. The second part reported the diet quality of preschoolers using NHANES data. Finally, for further dietary quality comparisons among groups, the preschool children from the NHANES dataset were categorized into subgroups depending on their family income and Head Start attendance. The diet quality of children belonging to different subgroups were compared. Research also compared the diet quality between male and female preschool children and compared the diet quality of children by weight status. Samples and Recruitment Michigan Head Start sample. The demographics of the Michigan Head Start sample used in this study was consistent with those of all the children 3-5years old enrolled in the Capital Area Community Services (CACS) Head Start Program of the greater Lansing area within Ingham, Eaton, Clinton, and Shiawassee Counties [Murashima, 2010; Murashima, 2011]. In the 2009-10 school years, 1,457 children were enrolled in Capital Area Head Start. The demographic distribution of all children in that program at that time was “58.5% white, 27.7% black, 11.8% 19 biracial, and 0.2% Asian, Native American, or Pacific Islander” as provided by Murashima [Murashima, 2010; Murashima, 2011]. Dyads of mother or female primary caregivers and their children (n=330) were recruited in a previously IRB approved cross-sectional study through Head Start from October 2009 to February 2010 [Murashima, 2010; Murashima, 2011]. Researchers recruited Head Start children and their parents by posting flyers and attending social events at Head Start sites like parents’ night[Murashima, 2010]. Researchers obtained consent forms from the mothers and female caregivers prior to assessments. Mothers or female primary caregivers served as proxies to report their children’s dietary intakes by completing the Block Kids Food Screener (BKFS). Details about the recruitment and power analysis were described in a previous study conducted by Murashima [Murashima, 2010]. National Health and Nutrition Examination Survey (NHANES). NHANES is a national survey designed to assess the health status and diet of infants, children, and adults of all ages in the U.S. The NHANES participants provide a representative sample for the non-institutional U.S. population and were selected based on age, gender, and racial/ethnic background through a series of statistical processes using the U.S. Census data [CDC, 2011a]. The National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC) conducted the survey, which was reviewed and approved by the NCHS Institutional Review Board [CDC, 2011a]. NHANES proxies consented for their children prior to any interview and assessment. Instead of a simple random sampling, NHANES used a stratified, multistage probability sampling method [CDC, 2011a]. They first selected primary sampling units that were then divided into smaller segments for a random drawing of specific household in those segments. The final step was to choose individuals from that household. Others could not volunteer to 20 substitute for the selected participants, if those who were identified were not eligible or chose not to participate. NHANES data were collected though two components, an in-home interview and an examination. Trained assistants conducted the interviews at participants’ homes for data on demographics and health behaviors. The dietary data, physical assessment data, and other examination data were collected at a Mobile Examination Center [CDC, 2011a]. The 2007-08NHANES is part of the Continuous NHANES, which started in 1999 and is released at a two-year cycle. For the Continuous NHANES researchers survey 30 counties in each cycle and 9,762 participants were individual for 2007-08. Participants in this sample were 40.7% non-Hispanic white, 21.9% non-Hispanic black, and 21.1% Hispanic-Mexican American, 11.7% Other Hispanic, and 4.5% Other [CDC, 2011a]. All Hispanic participants were over sampled in the 2007-08 NHANES and this must be accounted for with sample weights. The NHANES sample used in this study was a subset of children 3-5 years old (n=546) surveyed in the 2007-08 NHANES, which are the most recent data fully released on the CDC website [CDC, 2011a]. One participant was removed from the sample because the person failed to report if he/she was attending Head Start during the survey time period. Participants (n=36) who did not report their income status and therefore, had missing values for the income-poverty ratio were removed. Four children who did not have either height or weight information were also removed from this study. Thus, the final analytic sample size is 505 children aged 3 to 5. For this study, the researcher further divided the 505 children into several subgroups based on HS participation or not during the time of the survey and by the ratio of their family’s income to the poverty guideline as appropriate for their family’s size and age of children. 21 The researcher then conducted comparisons between diet quality of Head Start children and that of other Non-Head Start children in different poverty levels. Different income ratios were used to examine and confirm the trend of differences between subgroups. Income ratios and subgroups will be discussed more in depth in the section on variables. Overall Study Procedures The MI HS children’s dietary data had been de-identified as were the NHANES data. Because this study was a secondary data analysis using all de-identified data and there was no contact with human subjects, IRB approval from MSU was not required. This is according to the federal regulation 45 CFR §46.101(b). The first step was to obtain dietary, anthropometric, and demographic data from the two sources. The researcher requested the dietary, anthropometric and demographic data for the 330 MIHS children from its author [Murashima, 2010; Murashima, 2011]. Murashima collected her dietary data for children using a short food frequency called the Block Kids Food Screener (BKFS), described in the next section [NutritionQuest, 2009]. For the2007-08 NHANES sample, the researcher located the relevant variables in the specific datasets (dietary recalls for the first day, nutrients, demographics, and early childhood status) and downloaded these from the CDC website [CDC, 2011a]. Step two was to code the food items on the BKFS using the USDA food codes [USDA, 2010]. Step three was to download the SAS sample code from the USDA Center for Nutrition Policy and Promotion (CNPP) website for calculating the HEI-2005 individual scores for the 330 MIHS sample [USDA, 2011b] and the population score for the NHANES sample [USDA, 2011c]. Step four was to clean and prepare the datasets and calculate the HEI2005 scores, Body Mass Index (BMI) percentiles for both datasets, and the income ratios for the 22 NHANES dataset. Missing values were deleted. The final step was to conduct the comparisons outlined in specific aims. Measurements and Variables Michigan Head Start Children Sample Dietary intake. The mothers of HS children completed a Block Kids Food Screener (BKFS) for their child’s food intake over the past week. The Block Kids Food Screener 2007 for ages 2-17 years includes 39 food and beverage items grouped as fruit and fruit juices, vegetables, potatoes (fried and boiled), whole grains, meat/poultry/fish, dairy, legumes, saturated fat, added sugars [NutritionQuest, 2009]. For each food item or group, one to several food items were listed as examples. The BKFS asks how many days in the past week that the child ate a food group choosing from none, 1 day, 2 days, 3-4 days, 5-6 days, and every day. It also asks the amount of each food per day choosing from little, some, and a lot [NutritionQuest, 2009, Murashima, 2010]. To help proxies visualize the portion sizes, Murashima obtained the accurate amount for each item in each portion category from NutritionQuest and developed a 40-page visual aide booklet with pictures for each food item and the portions [Murashima, 2010; Murashima, 2011]. Parents usually take 10-12 minutes to complete the BKFS, a reasonable length of time for parents with limited resources like time and education. The 39-item version of Block Kids Food Screener is a shortened version of the original 80-item Block Kid FFQ [Block et al., 1990; Marshall et al., 2008]. The original version and several modified versions have been tested for validity and reliability [NutritionQuest, 2009]. To date, there have been no publications regarding the Block Kids Food Screener’s validity, although the Social and Health Research Center, San Antonio, Texas reports submitting an 23 evaluation study comparing the Block Kids 2004 FFQ and the Block Kids Food Screener for publication [personal communication]. Anthropometrics. Trained research aides measured the Head Start children’s heights and weights onsite using a portable stadiometer (SECA 214, Seca Corp., Hanover, MD) and digital weight scales (BWB800AS, Tanita, Tokyo, Japan) [Murashima, 2010; Murashima, 2011]. Student assessors followed standard protocols [Lohman et al., 1988], during the assessments. BMI percentile for age and gender was calculated by using the CDC method [CDC, 2012].Weight status was categorized based on the children’s BMI percentiles. The children’s weight status was defined using CDC criteria: Underweight: BMI for age <5th percentile; Normal weight: BMI for age 5th– 84.9th percentile; Overweight: BMI for age 85th– 95th percentile; and Obese: BMI for age >95th percentile. More details can be found in the previous study done by Murashima et al. [Murashima, 2010, Murashima, 2011]. NHANES Sample Dietary intake. The dietary intakes for participants in the 2007-08 NHANES were assessed using 24-hour dietary recalls for two non-consecutive days. Well-trained interviewers obtained the dietary intake data by one in-person interview for the first recall and one telephone or mail contact for the second dietary recall in either English or Spanish, 3-10 days after the initial assessment [CDC, 2008a]. Proxy respondents reported for children who were 5 years and younger [CDC, 2008a]. The NHANES data were obtained online at the CDC website and can be found at http://www.cdc.gov/nchs/nhanes.htm. Besides the information specific to each food and 24 beverage item, information regarding the recall day and participant's overall diet was also collected as part of the 24-hour recall procedure [CDC, 2008b]. Interviewers administered the 24-hour recalls using USDA’s Automated Multiple-Pass Method [Blanton, 2006], which can be found on the USDA website: http://www.ars.usda.gov/Services/docs.htm?docid=7710. This validated method for collecting dietary information has five standardized steps. 1) For the Quick list, participants report a uninterrupted list of foods and beverages that they ate 24hr before the interview. 2) For Forgotten foods, the interviewer asks a series of questions to remind the respondent of foods or beverages often missed. 3) For Time and Occasion, the interviewer asks what time the foods or beverages were consumed and the eating occasion. 4) For the Detailed Cycle, the interviewer probes for detailed information about each food’s preparation and the amount consumed. 5) For the Final Probe, the interviewer again asks about other foods and beverages consumed by tracking the respondent through daily activities. The follow-up dietary recall data collection was conducted using a similar procedure, but over the phone [CDC, 2008c]. Interviewers also inquired about any meal that the child had outside of home when the proxy was not present. Personnel who could recall this meal were contacted, when contact information was available from the proxy [CDC, 2008b; CDC, 2008c]. For this study, only the first day of dietary intake was used to calculate the HEI-2005 from the NHANES sample for any comparison between the Head Start sample and the NHANES sample, in part, because 20% of the children were missing a second dietary recall. Also, even though one day dietary recall is not accurate in evaluating an individual’s dietary intake, it can be used alone when reporting dietary intake and dietary quality of a population. 25 Anthropometrics. In NHANES, for children older than 3years, the standing height was taken for those who were able to stand unassisted. Standing height was measured with a fixed stadiometer with a vertical backboard, fixed floorboard, and movable headboard. All assessments were done by trained personnel and a standard protocol, which can be found at http://www.cdc.gov/nchs/data/nhanes/nhanes_07_08/manual_an.pdf. Participants were weighed using a digital weight scale wearing the standard examination gown. The procedures for obtaining the weight measurement were as follows: 1) Examiner introduced the procedures to participants. The participants were then guided to stand in the center of the scale in a fixed position and keep still. 2) When the readout on the device became stable, the recorder captured the results. Portable digital scales and stadiometers were used to measure the weight and height. The BMI percentiles for NHANES children were not provided by NHANES. In order to calculate children’s BMI percentile, their date of birth, date of assessments, and their height as well as weight are needed. The children’s date of birth was estimated based on children’s age in months and the examination date. These data were then exported into the CDC BMI percentile calculation sheet to determine children’s BMI percentile [CDC 2012]. Income. The total family income variable was released by 2007-08 NHANES as a income range value and was used to calculate the ratio of family income to poverty. NHANES asked the respondents to report total family income in dollar amount from different sources including “wages, retirement income, disability payments, interest income, and assistance programs” as described in the NHANES data file. If the participants refused to provide income information in dollar amount or did not know the answer, they were asked if their total family income was “less than $20,000” or “equal or greater than $20,000”. If the participants answered the previous 26 question, they were then asked to select an income range from a list of income ranges. NHANES used the midpoint of the income range as the total family income value to further calculate the ratio of family income to poverty. Poverty Level. In this study, the researcher defined poverty level as the ratio of family income to poverty status. This ratio was calculated by dividing family income by the poverty threshold, which varies among families. The U.S. Census Bureau developed and used the family appropriate poverty threshold, which counts money before taxes and excludes non-cash benefits, to define poverty status [U.S. Census Bureau, 2011]. The thresholds vary by family size and composition and also by the number and age of children [Vaughan, 1993]. For example, in 2008, the poverty threshold for a family of two adults and two children under age 18 was $21,834 [U.S. Census Bureau, 2011]. NHANES was not able to calculate the ratio when respondents were unable to define their income in a range with both minimum and maximum amounts. For example, if respondents only provided income< $20,000 or ≥ $20,000, the ratio was not computed under this circumstance. The ratio values were also not calculated if the family income data were missing. In this study, to compare the dietary quality, the researcher used different poverty levels to distinguish low-income and higher income. These poverty levels were the ratio of gross family income to the poverty threshold at 1.0, 1.3, and 1.85. The reason for using these three different poverty levels were that the cut points are the income eligibility criteria for federal funded lowincome supplemental program such as Head Start, Supplemental Nutrition Assistance program (SNAP), and Women, Infant, and children (WIC), respectively [USDHHS, 2010b; USDA, 2012b]. The ultimate goal was to see if the Head Start program benefited the low-income population. Thus by grouping low-income samples using different poverty levels, the researcher 27 could be sure: 1) the HS low-income population and non-HSLI population were comparable; and 2) the differences between groups were consistent when using different cut points. This can help the researcher dissect the influence of Head Start attendance from income status on children’s diet quality. Prior to grouping children by poverty level, the researcher first determined if the children were enrolled in Head Start or not. If the children were attending Head Start, then those children were categorized as Head Start. Then three different poverty levels (1.0, 1.30, and 1.85) were used in this study to group those children who were not attending Head Start by poverty level (non-HS). When below the income ratio, those children were categorized as non-Head Start lower income (non-HSLI) and when above the ratio, whose children were categorized as higher income population (non-HSHI). Those who were between 1.30PL and 1.85PL were categorized as the middle-income group (non-HSmidI). Because three different cut points were used in this study, the cut point will be add to each category to clarify comparisons. For example, when using the 1.85 cut point, the non-Head Start low-income population was simplified as “non-HSLI-1.85”. The comparison was conducted within groups between each category and was not across the groups. The researcher used the following three groups to conduct comparisons in order to achieve the goal that we discussed above. Figure 1 illustrates the grouping methods discussed above. i. HS vs. non-HSLI<1.0 vs. non-HSHI>1.0 ii. HS vs. non-HSLI<1.3 vs. non-HSmidI-1.3 vs. non-HSHI>1.85 iii. HS vs. non-HSLI<1.85 vs. non-HSHI>1.85 28 Figure1 Grouping of NHANES preschool children by Head Start and income status. Healthy Eating Index-2005 (HEI-2005) HEI-2005 Components. The HEI-2005 was employed in this study as an indicator of diet quality of the children’s dietary intake. The HEI-2005 assesses conformance to federal dietary guidance [Freedman, 2010]. Within each of the five food groups (12 components) as shown below in Table 2, some foods are more nutrient-dense than others and thus are weighted accordingly. The overall possible perfect index score is 100 [Guenther et al., 2008b]. Food groups including fruits, vegetables, grains, milk, meat, and oil were considered in the nutrient adequacy component. For such groups, individuals with an intake at the recommended level receive a perfect score of 5 or 10, and a score of 0 is given if no foods from a group are eaten. 29 Diet components considered in the nutrient excess portion were saturated fat, sodium, and discretionary calories from solid fat/ alcohol/added sugar (SoFAAs). When an individual consumed less from these groups, they earned a higher score. The SoFAAs group has a maximum score of 20 points [Guenther et al., 2008c]. Note that for this study, alcohol was not considered, so the discretionary calorie group was labeled as SoFAs. 1 Table 2 HEI-2005 components and standards for scoring . Component Maximum Standard for maximum Points score Total Fruit (includes 5 >0.8 cups/1000 Kcal 100% juice) Whole fruit (not juice) 5 >0.4 cups/1000 Kcal Total vegetables 5 >1.1 cups/1000 Kcal Dark green and orange 5 >0.4 cups/1000 Kcal vegetable and 2 legumes Total grains 5 >3.0 oz/1000 Kcal Whole grains 5 >1.5 oz/1000 Kcal 3 10 >1.3 cups/1000 Kcal Milk 4 10 >2.5 oz/1000 Kcal Meat and beans 5 Oils Saturated fat Sodium Calories from solid fat, alcohol, and added 6 sugar (SoFAAs) Standard for minimum score of zero No fruit No whole grain No vegetables No Dark green and orange vegetable and legumes No grains No whole grains No milk No meat or beans 10 >12 grams/1000 Kcal No oil 10 10 20 <7% of energy <0.7 grams/1000 Kcal <20% of energy >15% of energy >2.0 grams/1000 Kcal >50% of energy 1 Intakes between maximum and minimum levels are scored proportionally, except for the saturated fat and sodium (see note 5) 2 Legumes counted as vegetable only after meat and beans standard is met 3 4 Includes all milk products, such as fluid milk, soy milk, yogurt and cheese Includes non-hydrogenated vegetable oils and oils in fish, nuts and seeds 5 Includes all types of meats, as well as fish, egg, soybean product other than soy beverage and also nuts, seeds and legumes 6 Saturated fat and sodium get a score of 8 for the intake levels that reflect the 2005 Dietary Guidelines, < 10% of energy from saturated fat and 1.1grams of sodium/1000 Kcal, representatively. 30 The whole fruits group was a new component in the HEI-2005 because the dietary guidelines and pediatricians suggested that the intake of fruit juice should be limited due to concerns about dental caries. Whole grain is also a new component since the MyPyramid recommended half of grain intake should be whole grains. Each component included food that was directly eaten within the group as well as the ingredients that exist in mixed food items. For example, a fresh apple would be categorized as whole fruit. Milk and Meat groups only included the lowest fat portion of the product. The excessive fat portion of milk and meat products (not including fish, nuts, and seeds) was counted as solid fat. As indicated in Table 2, HEI-2005 counts the fat portion of fish, nuts, and seeds as an oil component. HEI-2005 Scoring. The HEI-2005 scoring is based on energy density. The scoring depends on the amount of the food intake per every 1000kcal, but not the absolute amount of food intake. For food groups, including fruits, vegetable, milk, meat and beans, and oils, the maximum score is the “least restrictive or easiest to achieve” amount per 1,000kcal in the 1,2002,400 kcal diet pattern recommended by MyPyramid. For example, MyPyramid recommended the fruit in cup equivalents per 1,000kcal at different levels for different energy intake levels as listed in the Table 3 below. The recommendations for fruit intake were similar across energy levels and the easiest to achieve amount as 0.8cup eq/1000kcal were used as the standard for a maximum score in fruit group. Table 3 Example for selecting maximum standard for total fruits food group. Food group Calorie Level 1200 1400 1600 1800 2000 2200 Fruit 0.8 1.1 0.9 0.8 1.0 0.9 (cup eq/1000kcal) 31 2400 0.8 Because of MyPyramid did not recommend how much whole fruit one should have each day, half of maximum standard for total fruits were used as the maximum standard for whole fruits. The dietary guidelines recommended half of the grains should be whole grains, thus half of the maximum standard of total grain was set as the standard for whole grains. MyPyramid recommended specific amounts on a weekly basis for different vegetables, thus the weekly recommendations were converted to a daily basis and the total of the vegetable subgroup was used as the maximum score. Several guidelines other than just MyPyramid Dietary Guidelines for Americans were considered to determine the standards for saturated fat and sodium. For saturated fat, 7% of total calories was chosen to receiving 10points as the maximum score, because 7% was recommended by the National Heart, Lung, and Blood Institute (NHLBI) and the American Heart Association (AHA) for managing hypertension and preventing heart disease, respectively. The dietary guidelines recommendation for saturated fat is less than 10% of total daily energy intake and this recommendation was also recognized in the scoring for saturated fat and assigned with 8 points. Similarly, the maximum score of 10 for sodium was assigned when one had less than 1,500mg (at 2,150kcal energy level) of sodium on daily basis, which is the Adequate Intake (AI) level, set by Food and Nutrition Board. A score of 8 was assigned for the intake of sodium at 2,300mg daily (at 2,150kcal energy level). The maximum scoring standard for SoFAAs is the “least restrictive or easiest to achieve of all the discretionary calorie allowances found in MyPyramid.” 32 Demographics Demographic data for both the MIHS and the NHANES children were used. These included age, gender, and race/ethnicity distribution in percentages. For the MIHS sample, the demographic categories were non-Hispanic white, non-Hispanic black, Hispanic and non-Hispanic mix and other races, which included people who identify themselves as more than one race or who were not white or black and were not Hispanic. NHANES demographic information was categorized in the same way. In NHANES, subjects self-identified if they were Mexican American, other Hispanic, non-Hispanic, non-Hispanic, and other non-Hispanic race including non-Hispanic multiracial. Mexican American and other Hispanic were combined into the Hispanic category. Analysis Two datasets were used in this study. The two datasets were collected using completely different sampling methods. Thus, the researcher used two different methods to calculate the HEI -2005. The two methods will be further discussed in the paragraph on calculating HEI-2005 score. NHANES used a complex, stratified multistage probability sampling design, so the SAS 9.3 software (SAS Institute Inc. Cary, NC: 2011) was employed to analyze most of the data [Gossett, 2006; CDC, 2011a]. T-tests and ANOVAs were performed. The significance level of alpha for all statistical analysis was set at 0.05. The Bonferroni method was used to adjust the critical value for the family of pairwise comparisons across the multiple levels (n>2) for subgroup comparison when ANOVA could not be performed [Ervin, 2011]. This method was mainly used to maintain the family wise error rate. Hypotheses were tested for each two group 33 comparison at a significance level of 1/n times. For example, the weight status was categorized into four subgroups, underweight, normal weight, overweight, and obese, depending on children’s BMI percentile. The significance level was set at p<=0.0125 due to the fact that this is a four group comparison and the overall significant level would be maintained at 0.05 level. Data Preparation First, the NHANES demographic, dietary, and anthropometry sub-datasets were merged. Next the scatter-plots of key variables were generated for the purpose of eliminating outliers. However, no obvious outliers were detected and therefore, no observations were deleted due to extreme values. One data record was deleted because of missing values. The sample weights for each variable collected in each setting (interview vs. mobile examination center) are available for each dataset [CDC, 2008d]. In this study, weights for the first 24hr recall were used. The variance of estimates (sampling errors) for all survey estimates were adjusted to aid in determining statistical reliability. The Taylor Series Linearization methods [Verma, 2011] were used, as suggested by NHANES analytic guidelines for variance estimation in NHANES data analysis [CDC, 2008d]. Masked Variance Units (MVUs) were created and provided on the demographic data files to mask the true primary sampling units (PSUs) for confidentiality purposes. The stratum variable and the MVUs variables were necessary for the analysis. SURVEYMEANS and SURVEYFREQ procedures were used to generate means, standard error, and frequencies as well as adjust for sample weights and study design to represent the national population [Bucholz, 2011]. SURVEYREG procedure was used to conduct ANOVA analysis and adjust for sample weights at the same time [Bucholz, 2011]. 34 Calculating the Healthy Eating Index-2005 Scores There are two ways to calculate the HEI-2005 scores when only one day 24h recall data is available or used. The first method was referred as “Mean Score” method by Freedman et al. and it calculates the HEI-2005 component score first. Component scores for each individual were first calculated base on each individual’s 24hr recalls and then the means were taken over all the individuals’ component scores. The total HEI-2005 score was the sum of all component scores. The second method was used when the national data were used to report the entire population’s intake. This method was named “Population Ratio” method by Freedman et al. [Freedmen, 2008]. The Population Ratio method first calculates a population’s total intake of a HEI-2005 component (food group) or a nutrient, next calculated the population’s total energy intake, and then took the ratio of these two values. The formula for this method is shown in Figure 2. Figure 2 Formula for the Population Ratio Method used to calculate the Healthy Eating Index2005 scores. For MIHS children’s HEI-2005 score, the Mean Score method was used. To calculate the HEI-2005 score, for the MIHS children, the researcher identified the USDA Food Codes for all food items listed on the Block Kids Food Screener for each food group from the “What's In the Foods You Eat Search Tool 2.0” website. This tool can be found online at www.ars.usda.gov/Services/docs.htm?docid=17032. For each food group, if more than one food item was listed, all the food codes were obtained and the weight in grams for this particular food group was divided by the number of food items and equally assigned to each food item in the 35 group. For example, if a child consumed 90grams of fresh fruit, where the items listed as examples for fruits were apples, bananas, and oranges, 30grams were assigned to each fruit item. Then food items from the BKFS were assigned a USDA food code. The researcher used the USDA Food Codes to merge dietary data from the BKS with the Mypyramid Equivalents database 03-04 (MPED) for USDA food codes and CNPP 03-04 Mypyramid Equivalents database for whole fruits and fruit juice [Cook, 2004]. The HEI-2005 total and component scores were calculated for each child using the National Cancer Institute methods. The SAS code can be found at: http://www.cnpp.usda.gov/HealthyEatingIndex.htm. For the 2007-08 NHANES dataset, the Population Ratio method was first used. USDA has not released the 05-06 or newer version of MPED and CNPP. Thus 03-04 MPED and CNPP and an addendum for 05-06 & 07-08 MPED and CNPP databases were used. This addendum was combined with the 03-04 MPED and CNPP datasets to provide the selected USDA food codes for analyzing 07-08 NHANES data. This addendum can be found at: http://www.cnpp.usda.gov/OtherProjects.htm. The researcher used “procsurveymeans ratio” to calculate the ratio of the population’s 1 food intake to its energy intake as well as account for the complex sample design . The ratio was expressed on a density basis. The mean density ratio and standard error (SE) were generated from the code. Then the ratio and SE were exported into an Excel spreadsheet provided by CNPP to allow researchers to calculate each non-truncated HEI-2005 component score through the HEI-2005 standard scoring system. Because each HEI-2005 component score has a minimum and maximum score, the maximum score was assigned to each component, if the non-truncated 1 The SAS code for this method can be found at the National Cancer institute website at: http://riskfactor.cancer.gov/tools/hei/tools.html. The CNPP provided a SAS/SUDAAN code for calculation, but SUDAAN was not unavailable to this researcher, so another method was used. 36 score exceeded the maximum. The total HEI-2005 score was the summary of the truncated component score. The Standard Errors of the total HEI-2005 scores were calculated by the SAS code provided by CNPP. One problem with the Population Ratio method is that it will not generate the individuals’ score for comparison with control confounders. Thus, the Mean Score method was also employed to calculate the HEI-2005 score for the NHANES data sets. ANOVA was used to compare the scores between Head Start and non- Head Start groups controlling for gender, age, race/ethnicities, and BMI-percentiles. Data Analyses by Each Aim For Aim1_Describe the diet quality (both as food groups HEI-2005 and for selected key nutrients) from food intake data of MIHS Children. Descriptive data such as means and standard deviations for the total HEI-2005 score and for each subgroup of the HEI-2005 were reported. Selected nutrients, such as fiber, folate, Vitamin A, Vitamin E, iron, zinc calcium, and potassium were also reported as means, standard derivations, and the percentages of the Daily Reference Intakes. For Aim2_Compare the diet quality from food intake data from the BKFS of MIHS children to the diet quality from one-day 24-hour dietary recalls of the 2007-08 NHANES HS children. Weighted means, SE, and the 95% CI’s from NHANES were presented and compared descriptively to the means, SD, and 95% CI of the MIHS data. 37 For Aim3_Compare the diet quality of 2007-08 NHANES Head Start children to those of middle/high-income and to those of low-income non-HS children from the 2007-08 NHANES using HEI-2005. When HEI2005 scores were calculated using Population Ratio method, t-tests were conducted to compare the HEI-2005 total scores and subgroup scores among the Head Start and the other NHANES subgroups. Each two subgroups were tested independently as group 1 HS, group 2 non-HSHI, and group 3 non-HILI. When HEI2005 scores were calculated using Mean Score method, ANOVA was conducted for comparisons among HS, non-HSHI, and non-HSLI groups while controlling for age, gender, race/ethnicity, and weight status (BMI percentile were used). Instead of using one categorical variable to represent the group of HS, non-HSHI, and non-HSLI, dummy variables were created for the combination each group. For example, if children were in Head Start program then HS=1, else, HS=0. Meanwhile, race/ethnicity was also recoded to a dummy variable “race”, where if one’s race is white then race=1, else, race=0. Race/ethnicity was recoded, because we were not looking into how different race status affects the HEI-2005scores and the majority of the NHANES preschoolers (56.3%) were white after adjusting for sample weights and study design. Thus the formula for predicting the HEI2005 total score was: HEI-2005 = A1 + A2*non-HSHI + A3*non-HSLI + A4*age + A5*gender + A6*race + A7*BMI percentile In this formula, the HS group was the reference group and therefore was not introduced into the formula. Nutrient intakes of the NHANES subgroups were also examined by performing ANOVAs controlled for age, gender, race/ethnicity, and weight status (BMI percentile). 38 CHAPTER 4RESULTS Demographics In the MIHS sample, on average, the children were 4.2 years old and 49.1% female (Table 4). The majority of children were non-Hispanic white (40.6%). Proxies reported that over 20% of their children were Hispanic. Even though over half of the children were normal weight, nearly 40% were overweight or obese. In comparison, the NHANES HS children’s average age was 4.0 and 41.5% were female. The NHANES HS children were similar for race and ethnicity, except for having more African American and fewer children of mixed race/ethnicity. The majority of the NHANES children were normal weight and only 16.4% were overweight and obese. Table 4 Comparison of demographics and weight status for MIHS and HS children from 20072008 NHANES. Demographic characteristics MI Head Start Children NHANES Head Start Children and (n=330) (n=105) Weight status Age, yr 4.2+0.6 3.98+0.09 Gender, % female 49.1% 42.7 Race & Ethnicity, % Non-Hispanic white 40.6 42.2 Non-Hispanic black 21.8 35.7 Non-Hispanic mixed/other 17.0 2.5 Hispanic 20.6 19.6 1 Weight Status , % Underweight 1.2 2.9 Normal weight 58.8 80.8 Overweight 18.2 9.4 Obesity 21.8 6.9 1 The children’s weight status was defined using CDC criteria: th Underweight: BMI for age <5 percentile; th th Normal weight: BMI for age 5 – 84.9 percentile; th th Overweight: BMI for age 85 – 95 percentile; th Obese: BMI for age >95 percentile. 39 Diet Quality Indicated by HEI-2005 For Aims 1 and 2, the HEI2005 scores calculated using Population Ratio method were used for the comparison with the MIHS sample. The MIHS children had a total HEI-2005 score of 60.2 out of a possible maximum score of 100 (Table 5 and Figure 2). Among all components, total fruit (4.8/5), whole fruit (4.7/5), and milk (9.5/10) were close to the full score meaning MIHS children’s intake of these three groups were close to the recommended amount. Total grains (4.1/5), meat and beans (7.9/10), oil (3.5/10), and sodium (4.0/5) groups did not meet the recommendations, but the differences from recommendations were moderate. MIHS children’s average intakes of total vegetables (2.8/5) and DOVL (0.2/5) were far below the recommendations and intake of saturated fat (2.5/5) and SoFAs (13.2/20) exceeded the recommendations. These four subgroups need attention. Table 5 Mean, SD for the total and subgroup HEI-2005 scores of MIHS children (n=330) and the possible maximum scores for the total and subgroup of HEI-2005. Component Scores of MIHS Possible Maximum Mean + SD score Total HEI-2005 score 60.2+7.4 100 Total Fruit (cup/1000 kcal) 4.8+0.7 5 Whole Fruit (cup/1000 kcal) 4.7+1.0 5 Total Vegetables (cup/1000 kcal) 2.8+1.1 5 0.2+0.5 5 Dark Green and Orange Vegetables and Legumes (cup/1000 kcal) Total Grains (oz/1000 kcal) 4.1+0.9 5 Whole Grains (oz/1000 kcal) 3.0+1.6 5 Milk (cup/1000 kcal) 9.5+1.4 10 Meat and Beans (oz/1000 kcal) 7.9+2.3 10 Oils (gm/1000 kcal) 3.5+1.8 10 1 Saturated Fat (% of kcal) 2.5+2.6 10 Sodium (gm/1000 kcal)1 4.0+2.1 10 Calories from Solid Fat and Added Sugar 13.2+3.2 20 1 (% of kcal) 1 Those subgroups were scored reversely, which means the more children consumed for this subgroup, the lower they will score for this subgroup. 40 25 MI HEI-2005 score 20 Possible Max Score 15 10 5 0 Figure 3 Comparison of the subgroups HEI-2005 scores of MIHS children (n=330) to the possible maximum score for each subgroup of HEI-2005. *For interpretation of the references to color in this and all other figures, the reader is referred to the electronic version of this thesis. Table 6 reported the MIHS population’s intake of total energy and selected nutrients. The shaded items were those that did not meet recommendations. MIHS children had an average total energy intake of 1240Kcal per day. MIHS’s intake of nutrients not meeting the recommendations included fiber (11.6/19-25g), vitamin E (3.3/5-6g), and potassium (2022/3000-3600mg). MIHS children’s cholesterol intake was 179.4g which is less than the recommended maximum intake of 300g per day. The MIHS’s vitamin A and folate intake both surpassed the recommendations. Most minerals, for example, calcium (846.6/500-800mg), phosphorus (959.3/380-405mg), magnesium (180/65-110mg), iron (9.5/3-4.1g), and zinc (8.2/2.4-4g) of the MIHS children’s daily consumptions exceeded recommendations. 41 Table 6 Mean, SD for daily total energy the key nutrients intake of MIHS children (n=330), percentage of MIHS children met the recommendations, and the EAR or AI for the key nutrients. Component MI’s Mean + MIHS % met Recommendations SD recommendations 3 years old 4-5years old 1-3 years old 4-8 years old 1 Total Energy 1240.4+547.5 1200(female), 1000 1 intake (Kcal/d) 1400(male) 2 11.6+5.7 5% 2% 19 25 Fiber (g/d) Cholesterol (mg/d) Vitamin A, (mcg RAE /d) Vitamin E 2 (mg/d) Total folate (mcg/d) Calcium (mg/d) Phosphorus (mg/d) Magnesium (mg/d) Iron (mg/d) Zinc (mg/d) Potassium 2 (mg) 179.4+103.3 91.2% <300 548.0+225.2 90% 96% 210 275 3.3+1.7 5% 9% 5 6 275.2+124.6 89% 91% 120 160 846.6+316.2 78% 58% 500 800 959.3+385.9 95% 97% 380 405 180.0+73.4 99% 91% 65 110 9.5+4.7 8.2+3.8 2022.87 98% 99% 9% 99% 97% 2% 3.0 2.4 3000 4.1 4.0 3800 1 The recommendation for energy intake for those two age groups were based on sedentary life style. The recommendation should increase if the physical activity level increased. 2 Shaded items did not meet recommendations. DGA2010 recommended low-fat milk or skim milk for children because whole milk was a main contributor of solid fat in young children’s diet. The researcher further examined the type of milk and cereal that the MIHS children were fed. The majority of families chose reduced fat 2% milk (66.4%) over other types of milk (Table 7), for example whole milk (21.5%), and lowfat/skim milk (8.5%). 42 Table 7 Frequency and percentage of milk type chosen by MIHS children (n=330). Milk type Frequency (n) Percent (%) Whole milk 71 21.5 Reduced fat 2% 219 66.4 Low fat 1% 13 3.9 Non-fat 15 4.6 1 12 3.6 Other 1 Other milk type including chocolate milk (n=5), soy milk (n=5), lactaid milk (n=1), and don’t know (n=1). DGA 2010 also recommended half of the grains of American’s intake should be whole grain. Whole grain cereal could boost fiber intake [Dietary Guidelines for Americans, 2011]. Sweetened cereals, which are preferred by young children, increased added sugar in their diet. Table 8 revealed that the majority of families tended to choose whole grain sweetened cereal (45.4%). The second most commonly eaten was whole grain, not heavily sweetened (27.8%). Table 8 Frequency and percentage of cold cereal type chosen by MIHS children (n=330). Cold cereals type Frequency (n) Percent (%) Whole grain cereals, not heavily sweetened Whole grain cereals, sweetened Cereals, not whole grain, sweetened Cereals, not whole grain, not sweetened 91 27.6 150 45.4 60 18.2 29 8.8 Finally, the researcher listed the MIHS and the 2007-08 NHANES HS children’s HEI2005 scores side by side to observe the trends. As discussed before, due to the different study designs, the researcher could not conduct statistical comparisons. Table 9 indicates a similar 43 pattern of intake between these two groups of children. Both MIHS and 2007-08 NHANES children had high intakes of total fruits (4.8 vs. 5), whole fruits (4.7 vs. 5), total grains (4.1 vs. 5), and milk (9.5 vs. 10). The consumption of total vegetables (2.8 vs. 2.5), meat and beans (7.9 vs. 7.2), sodium (4.0 vs. 4.8), and SoFAs (13.2 vs. 11.8) was also fairly similar between the two groups. The MIHS children tend to have better intake of whole grains (3.0 vs. 1.4) than the national sample. The 2007-08 NHANES HS children tend to have better intake of dark green, orange vegetable and legumes (1.0 vs. 0.2), saturated fat (5.2 vs. 2.5), and oils (6.9 vs. 3.5) than MIHS children. Table 9 Comparison of total and subgroup HEI-2005 scores between MIHS and NHANES HS children. Component Mean (95% CI) MIHS NHANES HS n=330 n=105 Total HEI-2005 score 60.2 (59.4, 61.0) 65.9 (58.5, 73.7) Total Fruit (cup/1000 kcal) 4.8 (4.7, 4.8) 5.0 (4.3, 5.0) Whole Fruit (cup/1000 kcal) 4.7 (4.6, 4.8) 5.0 (3.6, 5.0) Total Vegetables (cup/1000 kcal) 2.8 (2.7, 2.9) 2.5 (2.2, 2.9) Dark Green and Orange Vegetables and 0.2 (0.2, 0.3) Legumes (cup/1000 kcal) 1.0 (0.5, 1.6) Total Grains (oz/1000 kcal) 4.1 (4.0, 4.2) 5.0 (5.0, 5.0) Whole Grains (oz/1000 kcal) 3.0 (2.9, 3.2) 1.4 (0.9, 2.0) Milk (cup/1000 kcal) 9.5 (9.3, 9.6) 10.0 (9.4, 10.0) Meat and Beans (oz/1000 kcal) 7.9 (7.7, 8.2) 7.2 (6.4, 8.1) Oils (gm/1000 kcal) 3.5 (3.3, 3.7) 6.9 (5.2, 8.7) 1 2.5 (2.2, 2.7) Saturated Fat (% of kcal) 5.2 (4.1, 6.3) 1 4.0 (3.7, 4.2) Sodium (gm/1000 kcal) 4.8 (4.1, 5.4) Calories from Solid Fat and Added 1 Sugar (% of kcal) 1 13.2 (12.9, 13.5) 11.8 (10.7, 13.1) These subgroups were scored reversely. For Aim 3 Demographics For all groups (Table 10), the average age was close to 4 years. The HS group had fewer 44 females (42.7%), more African American (35.7%), less Hispanic (19.6%) compared to the other non-HSLI groups. In all groups, except for non-HSLI-1.0PL group, the majority of children were non-Hispanic white. In the non-HSLI<1.0PL group, the majority of children (37.0%) were Hispanic. The non-HSHI>1.85PL group had the largest percentage of white (70.8%), fewer African Americans (7.8%) and Hispanic (11.4%) compared to the other groups. The nonHSHI>1.85PL group also had the highest percentage of normal weight children (83.5%) and lowest percentage of overweight and obese children (2.1% and7.8%, respectively). The three non-HSLI groups had higher percentages of overweight and obese children (25.1%, 25.6%, and 23.5%) compare to the HS (16.3%) and the non-HSHI>1.85PL (9.9%) groups. The research used the ratio of gross family income to the poverty threshold to indicate income in the results. The HS children’s average family income was 1.96PL, which is much higher than the other non-HSLI groups, no matter which cut points the researcher used. Among the other non-HSLI groups, as expected, the non-HSLI<1.85PL group had the highest average family income status (0.98 vs. 0.62 vs. 0.76). 45 Table 10 Demographics and weight status of NHANES subgroups at different income cut point. Demographic NHANES NHANES NHANES NHANES non-HSLI<1.0PL non-HSLI<1.3PL non-HSLI<1.85PL characteristics HS & n=105 n=140 n=197 n=247 Weight status Age, yr 3.98+0.09 4.02+0.12 3.99+0.09 3.98+0.09 Gender, % of female 42.7 54.4 52.8 45.2 Race & Ethnicity, % Non-Hispanic 42.2 35.8 41.0 45.8 white Non-Hispanic 35.7 19.5 15.0 14.1 black Hispanic 19.6 37.0 38.1 33.3 Non-Hispanic 2.5 7.6 5.9 6.8 mixed/other 1 Weight Status , % Underweight 2.9 7.3 5.8 4.3 Normal weight 80.8 67.6 68.6 72.3 Overweight 9.4 8.7 11.2 9.6 Obesity 6.9 16.4 14.4 13.9 Average Poverty 1.96 0.62 0.76 0.98 2 Level 1 NHANES non-HSHI>1.85PL n=153 4.02+0.14 46.9 70.8 7.8 11.4 8.0 6.5 83.5 2.1 7.8 3.8 The children’s weight status were defined follow the CDC criteria: th Underweight: BMI for age <5 percentile; th th Normal weight: BMI for age 5 – 84.9 percentile; th th Overweight: BMI for age 85 – 95 percentile; th Obese: BMI for age > 95 percentile. 2 Family income level was indicated by the average poverty level, which was defined as the ratio of gross family income to the poverty threshold 46 Ho3.1 Diet Quality Indicated by HEI-2005_Population Ratio(PR) Method For all 3-5 year old children from 2007-08 NHANES, the total HEI-2005 score was 62.0 out of 100 and they reached full scores for total fruit, whole fruit, total grains, and milk (Table 11). However, their intake for DOVL and whole grains scored only 1.0 out of 5.0. Their intake of saturated fat (4.8/10), sodium (4.5/10), and SOFAs (10.1/20) needed improvement. Those three subgroups of nutrients were reversely scored, which means the greater the intakes of those food groups, the lower the score. The researcher also examined differences in diet quality between genders (Table 11). The results showed that male preschoolers tended towards slightly higher HEI-2005 total scores compared to females (62.3 vs. 61.5), however, the difference was not significant. Male children consumed significantly more DOVL (1.1 vs. 0.7), meat and beans (7.8 vs. 7.3) compared to females. Female preschoolers’ consumption showed a trend towards more milk, oil, saturated fat, and more sodium compared to males. The results of comparing dietary quality by weight status, unadjusted by income, gender, age, and race/ethnicity, (Table 12) revealed that overweight, but not obese children, had the highest total HEI-2005 score (65.5/100). All four subgroups met or approximated the recommendations for total fruits, whole fruit, and milk. Normal weight and overweight children achieved full scores for total grains. Overweight children scored highest for several subgroups including total vegetables (3.2/5) and DOVL (1.8/5), however, the lowest for whole grains (0.7/5) and sodium (3.9/10), P<0.0125. Overweight children scored significantly higher on saturated fat than overweight and underweight but normal weight children. 47 Table 11 Total and subgroup HEI-2005 scores using the population ratio method for all NHANES preschool children (n=505) and by gender. All NHANES children n=505 Male n=286 Female n=219 Mean (SE) Total HEI-2005 62.0 (1.77) 62.3(2.33) 61.5(1.86) Total Fruit 5.0 (0.06) 5.0 (0.08) 5.0 (0.08) Whole Fruit 5.0 (0.06) 5.0 (0.08) 5.0 (0.08) Total Vegetables 2.2 (0.02) 2.3 (0.04) 2.2 (0.03) Dark Green & Orange Veg. 0.9 (0.01) 1.1 (0.02) * 0.7 (0.01) Total Grains 5.0 (0.06) 5.0 (0.08) 5.0 (0.09) Whole Grains 1.0 (0.02) 1.0 (0.03) 1.0 (0.04) Milk 10.0 (0.04) 9.9 (0.04) 10.0 (0.05) Meat & Beans 7.5 (0.07) 7.8 (0.09) * 7.3 (0.11) Oils 5.9 (0.33) 5.6 (0.43) 6.1 (0.44) Sat. Fat 4.8 (0.19) 4.9 (0.22) 4.6 (0.24) Sodium 4.6 (0.02) 4.5 (0.03) 4.6 (0.03) SoFAs 10.1 (0.42) 10.1 (0.59) 10.1 (0.78) *Significant differences between Male and Female (p<0.05) Underweight children scored significantly higher for whole grains compared to overweight and obese weight groups (Table 12). Underweight children scored significantly lower for saturated fat compare to normal weight and overweight but obese children. This means that underweight children had the higher intake of saturated fat than that two groups. It is worth to notice that underweight group scored lowest on total grain and oil, which were major energy contribute groups. 48 Table 12 Total and subgroup HEI-2005 scores using the population ratio method for NHANES preschool children by weight status. Underweight n=20 Normal weight n=369 Overweight n=49 Obese n=67 Mean (SE) 4 Total HEI-2005 56.7 (1.69) 62.3 (2.13) 65.5 (1.69) 59.2 (2.76) Total Fruit 4.7 (0.06) 5.0 (0.07) 4.8 (0.06) 5.0 (0.11) Whole Fruit 5.0 (0.04) 5.0 (0.07) 5.0 (0.09) 5.0 (0.06) 1,4 Total Vegetables 2.4 (0.03) Dark Green & Orange Veg. 0.3 (0.003) 4 2 0.9 (0.02) 3.2 (0.08) 2 2.2 (0.02) 1.8 (0.03) 0.5 (0.01) 3,5,6 Total Grains 4.4 (0.08) 5.0 (0.08) Whole Grains 1.3 (0.09) 1.0 (0.02) Milk 10.0 (0.03) 10.0 (0.04) 10.0 (0.13) 10.0 (0.15) Meat & Beans 7.1 (0.15) 7.6 (0.09) 7.8 (0.18) 7.0 (0.16) Oils 5.4 (0.81) 5.9 (0.39) 6.0 (0.87) 5.5 (0.90) 5.0 (0.22) 5.3 (0.44) 3.7 (0.58) 2 3.9 (0.05) 1,4 Sat. Fat 2.6 (0.65) Sodium 5.3 (0.05) 1,4 4.5 (0.03) 5.0 (0.17) 3,5 2.2 (0.05) 2,3,4 0.7 (0.04) 8.5 (1.19) 10.1 (0.53) 12.0 (0.79) SoFAs 1 significant differences between underweight and normal weight (P < 0.0125) 2 significant differences between normal weight and overweight (P < 0.0125) 3 significant differences between overweight and obese (P < 0.0125) 4 significant differences between underweight and overweight (P < 0.0125) 5 significant differences between underweight and obese (P < 0.0125) 6 significant differences between normal weight and obese (P < 0.0125) 4.9 (0.17) 5 0.9 (0.06) 3,6 3 5.1 (0.04) 3,5,6 9.3 (1.16) In the comparison of Head Start children to other subgroups without adjustments (Table 13-16), The Head Start children had the highest HEI-2005 score (65.9/100). All groups achieved or nearly achieved the maximum scores for total fruits, whole fruits, and total grains. The HS and the non-HSHI children also achieved full score for milk. All the non-HSLI children groups did not have adequate milk intake, but still had good scores. The HS children had the highest score 49 on total vegetables (2.5/5) and the score was significantly more than the non-HSHI and the nonHSLI children. The non-HSHI group had significantly more DOVL (1.1/5) than the other groups. The only exception was that the-non-HS children, who had a family income between 1.3-1.85PL, had higher DOVL intake than the non-HSHI group. The HS group had significantly higher DOVL intake compare to the non-HSLI group. On average the DOVL intakes of all groups were far below the recommendations. Table 13 Total and subgroup HEI-2005 scores using the population ratio method for NHANESs children 3-5 year old, by Head Start attendance and income status at 1.0PL, unadjusted for gender, age or BMI percentile. NHANES HS children n=105 Non-HSHI >1.0PL n=260 Non-HSLI<1.0PL n=140 Mean (SE) Total HEI-2005 65.9 (3.62) 61.5 (2.30) 59.1 (1.94) Total Fruit 5.0 (0.15) 5.0 (0.07) 5.0 (0.09) Whole Fruit 5.0 (0.15) 5.0 (0.07) 4.9 (0.06) 2.2 (0.03) 2.2 (0.03) 1.1 (0.02) c 0.4 (0.01) 5.0 (0.09) 5.0 (0.10) c 0.7 (0.04) c a, b Total Vegetables 2.5 (0.04) Dark Green & Orange Veg. 1.0 (0.02) Total Grains a, b 5.0 (0.14) a, b Whole Grains 1.5 (0.08) Milk 10.0 (0.09) 10.0 (0.05) 8.8 (0.07) Meat & Beans 7.2 (0.10) b 7.6 (0.14) 7.8 (0.09) Oils 6.9 (1.05) 5.6 (0.40) 5.6 (0.46) Sat. Fat 5.2 (0.34) 4.6 (0.29) 5.0 (0.26) Sodium 4.8 (0.04) 4.3 (0.03) c 4.9 (0.05) SoFAs 11.8 (0.87) 10.1 (0.49) 8.9 (0.79) 1.0 (0.02) b a a significant differences between HS and non-HSHI (P < 0.0167) b significant differences between HS and non-HSLI (P < 0.0167) c significant differences between non-HSHI and non-HSLI (P < 0.0167) 50 The HS children also appeared to have the best intake of whole grains, which were significant higher than the other non-HS groups. The non-HSLI group had significantly higher intake of meat and beans than the other groups. The HS children scored high in the sodium component, which means that HS children had the lowest sodium intake. The same trends were also observed for the saturated fat and the SoFAs components, NS where HS children tended to be lower. Table 14 Total and subgroup HEI-2005 scores using the population ratio method for NHANES children 3-5 year old, by Head Start attendance and income status at 1.3PL, unadjusted for gender, age or BMI percentile. HS children n=105 Non-HSHI >1.3PL n=203 NonHSLI<1.3PLn=197 Mean (SE) Total HEI-2005 65.9 (3.62) 61.5 (2.72) 60.9 (1.29) Total Fruit 5.0 (0.15) 5.0 (0.08) 5.0 (0.06) Whole Fruit 5.0 (0.15) 5.0 (0.08) 5.0 (0.06) 2.2 (0.03) 2.3 (0.03) 1.1 (0.02) c 0.6 (0.01) 5.0 (0.10) 5.0 (0.09) c 0.8 (0.03) 10.0 (0.04) c 9.4 (0.07) a, b Total Vegetables 2.5 (0.04) Dark Green & Orange Veg. 1.0 (0.02) Total Grains a, b 5.0 (0.14) a, b Whole Grains 1.5 (0.08) Milk 10.0 (0.09) Meat & Beans 7.2 (0.10) 7.4 (0.16) c 8.0 (0.08) Oils 6.9 (1.05) 5.6 (0.47) 5.5 (0.58) Sat. Fat 5.2 (0.34) 4.8 (0.32) 4.5 (0.26) Sodium 4.8 (0.04) 4.4 (0.04) c 4.7 (0.05) SoFAs 11.8 (0.87) 10.0 (0.59) 9.4 (0.68) 1.0 (0.03) b b a a significant differences between HS and non-HSHI (P < 0.0167) b significant differences between HS and non-HSLI (P < 0.0167) c significant differences between non-HSHI and non-HSLI (P < 0.0167) 51 Table 15 Total and subgroup HEI-2005scores for NHANESs children 3-5 year old, by Head Start attendance and income status at 1.85PL, unadjusted for gender, age or BMI percentile. HS children n=105 Non-HSHI >1.85PL n=153 Non-HSLI<1.85PL n=247 Mean (SE) Total HEI-2005 65.9 (3.62) 61.3 (2.59) 60.6 (1.74) Total Fruit 5.0 (0.15) 5.0 (0.09) 5.0 (0.07) Whole Fruit 5.0 (0.15) 5.0 (0.08) 5.0 (0.06) 2.2 (0.04) 2.2 (0.02) 1.1 (0.02) c 0.8 (0.02) 5.0 (0.10) 5.0 (0.10) c 0.8 (0.03) 10.0 (0.04) c 9.4 (0.06) a, b Total Vegetables 2.5 (0.04) Dark Green & Orange Veg. 1.0 (0.02) Total Grains a, b 5.0 (0.14) a, b Whole Grains 1.5 (0.08) Milk 10.0 (0.09) Meat & Beans 7.2 (0.10) 7.2 (0.18) c 8.2 (0.07) Oils 6.9 (1.05) 5.7 (0.57) 5.4 (0.56) Sat. Fat 5.2 (0.34) 4.5 (0.33) 4.9 (0.28) 4.4 (0.04) c 4.6 (0.03) 10.1 (0.67) 9.4 (0.64) Sodium SoFAs 1.1 (0.03) b b a, b 4.8 (0.04) 11.8 (0.87) a significant differences between HS and non-HSHI (P < 0.0167) b significant differences between HS and non-HSLI (P < 0.0167) c significant differences between non-HSHI and non-HSLI (P < 0.0167) 52 Table 16 Total and subgroup HEI-2005 scores using the population ratio method for NHANESs children 3-5 years old, by Head Start attendance and income status at 1.3-1.85PL, unadjusted for gender, age or BMI percentile. NonHSHI>1.85PL n=153 HS n=105 Non-HSMidI 1.3-1.85PL n=50 NonHSLI<1.3PL n=197 Mean (SE) Total HEI2005 65.9 (3.62) 61.3 (2.59) 61.9 (6.48) 60.9 (1.29) Total Fruit 5.0 (0.15) 5.0 (0.09) 5.0 (0.21) 5.0 (0.06) Whole Fruit 5.0 (0.15) 5.0 (0.08) 5.0 (0.18) 5.0 (0.06) 1,2,3 2.2 (0.04) 2.0 (0.07) 2.3 (0.03) 1,2,3 1.1 (0.02) 1.4 (0.06) 5.0 (0.10) 5.0 (0.25) 1.1 (0.03) 0.8 (0.06) 0.8 (0.03) 10.0 (0.04) 9.5 (0.09) 9.4 (0.07) 7.2 (0.18) 8.6 (0.21) Total Vegetables 2.5 (0.04) Dark Green & Orange Veg. 1.0 (0.02) Total Grains Whole Grains 5.0 (0.14) 1,2,3 1.5 (0.08) Milk 10.0 (0.09) Meat & Beans 7.2 (0.10) 1,2 6 6 6 4 4,5 0.6 (0.01) 5.0 (0.09) 5 4,5 8.0 (0.08) Oils 6.9 (1.05) 5.7 (0.57) 5.0 (0.93) 5.5 (0.58) Sat. Fat 5.2 (0.34) 4.5 (0.33) 6.1 (0.63) 4.5 (0.26) 4.4 (0.04) 4.3 (0.07) Sodium 2,3 4.8 (0.04) 4,5 4.7 (0.05) 11.8 (0.87) 10.1 (0.67) 9.2 (1.77) 9.4 (0.68) SoFAs 1 significant differences between HS and non-HSLI-1.3 (P < 0.0125) 2 significant differences between HS and non-HSMidI-1.3-1.85 (P < 0.0125) 3 significant differences between HS and non-HSLI-1.85 (P < 0.0125) 4 significant differences between non-HSLI-1.3 and non-HSMidI-1.3-1.85 (P < 0.0125) 5 significant differences between non-HSLI-1.3 and non-HSLI-1.85 (P < 0.0125) 6 significant differences between normal non-HSMidI-1.3-1.85 and non-HSLI-1.85 (P < 0.0125) 53 Ho3.2 Diet Quality Indicated by HEI-2005_Mean Score(MS) Method Even though, Freedman from USDA Center for Nutrition Policy and Promotion recommends using the Population Ratio method to report the average dietary quality of populations by the HEI-2005. The method, however, does not generate HEI-2005 scores for each individual. For the purpose of comparison among multiple groups and correlation analysis controlling for confounders, the Population Ratio method cannot be used. To examine the dietary quality of preschool children nationwide related to their HS and income status, the Mean Score method was used as discussed earlier in Methods. Table 17 HEI-2005 scores calculated using two methods: Mean Score Method and Population Ratio Method for all preschool children from 2007-2008 NHANES. Mean Score method Population Ratio method Mean(SE) n=505 Total HEI-2005 54.6 (0.85) 62.0 (1.77) Total Fruit 3.5 (0.12) 5.0 (0.06) Whole Fruit 2.9 (0.24) 5.0 (0.06) Total Vegetables 2.2 (0.08) 2.2 (0.02) Dark Green & Orange Veg. 0.8 (0.08) 0.9 (0.01) Total Grains 4.4 (0.06) 5.0 (0.06) Whole Grains 1.1 (0.06) 1.0 (0.02) Milk 7.9 (0.19) 10.0 (0.04) Meat & Beans 6.5 (0.20) 7.5 (0.07) Oils 5.2 (0.18) 5.9 (0.33) Sat. Fat 5.1 (0.25) 4.8 (0.19) Sodium 4.7 (0.23) 4.6 (0.02) SoFAs 10.4 (0.22) 10.1 (0.42) 54 Table 17 shows the HEI-2005 scores of all NHANES preschool children using the Mean Score Method and using the Population Ratio Method. These two sets of data were not compared statistically due to the different methods used in calculations. None of the means of HEI-2005 subcomponents from Mean Score Method reached the full score. Using the Mean Score Method, the overall HEI-2005 score and most subcomponent scores were lower than the scores from the Population Ratio Method. The most noticeable differences were these for fruits, total grains, and milk. Table 18 demonstrates the differences of HEI-2005 total and subcomponent scores among three NHANES subgroups after adjusting for age, gender, race, and BMI percentile. HS children had significant higher total HEI-2005 scores compared to other groups (p=0.03 for the model). The results also revealed that low income non HS children had the lowest milk intake compared to other two groups (p=0.0001 for the model). The HS children also scored the highest for SoFAs, which means HS children had the lowest intake of SoFAs. 55 Table 18 HEI-2005 scores of NHANES children by Head Start status and 1.85 poverty level calculated using Mean Score Method and adjusted for age, gender, race/ethnicity, and BMI percentile. HS children n=105 NonHSHI >1.85PL n=153 NonHSLI<1.85PL n=247 Mean (SE) P value for the model Total HEI2005 57.3 (1.15) 55.0 (1.17) 53.0 (1.00) 0.03 Total Fruit 3.3 (0.25) 3.7 (0.19) 3.4 (0.10) 0.11 Whole Fruit 2.5 (0.34) 3.2 (0.27) 2.7 (0.28) 0.49 Total Vegetables 2.3 (0.18) 2.1 (0.17) 2.2 (0.10) 0.01 Dark Green & Orange Veg. 0.7 (0.18) 0.9 (0.12) 0.6 (0.12) 0.38 Total Grains 4.5 (0.08) 4.4 (0.13) 4.4 (0.06) 0.44 Whole Grains 1.4 (0.22) 1.1 (0.11) 0.8 (0.10) 0.39 Milk 8.1 (0.27) 8.2 (0.26) 7.5 (0.36) 0.0001 Meat & Beans 6.7 (0.39) 6.0 (0.46) 7.0 (0.15) 0.09 Oils 5.8 (0.60) 5.4 (0.38) 4.8 (0.38) 0.01 Sat. Fat 5.2 (0.40) 5.2 (0.35) 5.1 (0.35) 0.93 Sodium 4.9 (0.29) 4.5 (0.33) 4.7 (0.27) 0.08 SoFAs 11.7 (0.60) 10.4 (0.29) 9.8 (0.43) 0.0003 56 Nutrients intake among subgroups Because of the previous analysis revealed that the non-HSLI children had the most similar demographic distribution to the HS children. The researcher only conducted comparisons of nutrients for the three 1.85PL groups (Table 19). The NHANES HS children had significantly higher energy intake (1629.3 Kcal) compared to the other two groups (1618.4, 1477.8 Kcal) and higher sodium intake (p=0.04). The Non-HSLI-1.85PL group had the highest intake of cholesterol (p=0.0014), total folate (p=0.006), and zinc (p=0.02) compared to the other two groups. All groups met dietary recommendations for cholesterol, Vitamin A, calcium, magnesium, iron, and zinc. All groups had intakes of sodium higher than the recommended level, except the non-HSHI-1.85 group who didn’t exceed the UL. None of the groups met the recommendations for potassium. Table 12 listed the intakes of nutrients from the three groups using the population ratio method. 57 Table 19 Comparison of NHANES children’s nutrients intake (means + confidence interval) by Head Start and income status at 1.85 poverty level, controlled for age, weight, race/ethnicity, and BMI percentile. Component HS children NonNonPn=105 HSHI >1.85PL HSLI<1.85PL value n=153 n=247 Mean (95% CI) Energy (kcal) 1629.3 (1511.1, 1477.8 (1405.8, 1618.4 (1545.1, 0.0056 1747.0) 1549.8) 1691.6) Total 21.3 (19.3, 23.2) 20.0 (18.5, 21.5) 21.2 (19.4, 22.9) 0.14 saturated fatty acids (g) Sodium (mg) 2385.2 (2153.7, 2229.4 (2100.9, 2374.1 (2198.5, 0.04 2616.6) 2358.0) 2549.7) Dietary fiber 12.6 (10.5, 14.6) 10.8 (9.8, 11.9) 11.3 (10.5, 12.2) 0.14 (g) Cholesterol 186.9 (152.6, 152.2 (127.0, 177.3) 213.7 (186.1, 0.0014 (mg) 221.3) 241.2) Vitamin E as 4.9 (4.4, 5.5) 4.4 (4.1, 4.7) 4.9 (4.3, 5.4) 0.17 alphatocopherol (mg) Vitamin A, 665.0 (470.2, 594.2 (509.7, 678.8) 527.9 (487.6, 0.41 (mcg RAE /d) 859.8) 568.1) Total Folate 312.0 (275.9, 279.8 (247.9, 311.7) 332.0 (301.1, 0.006 (mcg) 348.2) 362.9) Vitamin D 7.0 (5.8, 8.0) 5.6 (5.0, 6.2) 5.6 (5.0, 6.3) 0.13 (D2 + D3) (mcg) Calcium (mg) 1008.8 (838.9, 955.2 (864.6, 916.3 (843.6, 0.37 1178.8) 1045.8) 988.9) Magnesium 214.2 (184.9, 183.5 (171.4, 195.7) 193.2 (183.6, 0.09 (mg) 243.5) 202.8) Iron (mg) 12.1 (10.5, 13.7) 10.7 (9.4, 12.0) 11.9 (10.8, 13.0) 0.15 Zinc (mg) 8.6 (7.7, 9.6) 7.7 (7.1, 8.3) 9.0 (8.1, 9.8) 0.02 Potassium 2187.1 (1888.7, 1857.7 (1723.0, 1977.0 (1886.3, 0.11 (mg) 2485.5) 1992.5) 2067.6) 58 CHAPTER 5 DISCUSSIONS To the researcher’s knowledge, this is the first study using the 2007-2008 NHANES to calculate the HEI-2005 scores to interpret preschool children’s diet quality. The most unique findings in this study were that in the NHANES survey: 1) Most of the income subgroups met the recommendations for fruits, total grain, and milk using Population Ratio method; 2) The Head Start children had better dietary quality not only compared to that of lowincome groups but also to higher income groups; 3) The Head Start children had a highest family income to poverty guideline ratio compare to all low-income groups at different income cut-points; and 4) Overweight, but not obese, preschool children had a better diet quality than preschool children in the other weight categories. Michigan Head Start Sample The MIHS and the NHANES children had or approached recommendations for the total and whole fruits, total grains, and milk. These findings are similar to what has been shown in other studies of preschool children [Fungwe, 2009]. However, this trend of high intakes of fruits and milk in preschoolers has been shown to decrease later in childhood [Lorson, 2009; Fungwe, 2009]. Similar to what has been documented elsewhere, MIHS children had low a intake of vegetables, especially dark green and orange vegetables [Fungwe, 2009]. Vegetables of all kinds are good sources of vitamins, some minerals and phytochemicals and are considered nutrient-dense foods, because the calories ware often low. For example, dark 59 green and orange colored vegetables are more nutrient dense compared to vegetables such as lettuce or French fries. In general, high intakes of vegetables are associated with better diet quality and appear to have health benefits [De Bock, 2012]. The DGA2010 emphasized the importance of reducing dietary saturated fat, and solid fats and consuming low-fat dairy foods to reduce the risk of developing cardiovascular disease. Related to this advice, in this study most MIHS families (67%) provided reduced fat (2% fat) to their children. Although this is still not the low fat milk recommended in the 2010 DGA, it is a big change from the whole milk seen in past studies [O'Connor, 2006; Fox, 2010]. To the researcher’s knowledge, this is the highest percentage of children consuming reduced fat milk that has been reported in preschool population [O'Connor, 2006; Fox, 2010]. The choice of 2% milk was an improvement from choosing whole milk among preschool children. It helps to reduce the discretionary fat that young children consume from whole milk and might reflect the new pediatric guidelines on choosing reduced or low fat milk after two years of age [Kleinman, 2009]. WIC and Head Start have changed their guidelines to provide only 1% fat or fat-free milk for all children after two years of age [USDA, 2012b]. But, in the present study, only 9% children drank low fat or skim milk. In the past, skim milk and low-fat milk were not preferred by families potentially due to taste preferences [Johnson, 1991The DGA2010 also emphasized the importance of reducing added sugar and increasing whole grains. Most MIHS children (45.4%) consumed sweetened whole grain cereal. It might be hard for children to consume un-sweetened cereal due to the fact that humans naturally prefer sweets [Brown, 1998; Ventura, 2011]. Therefore, it is important to teach parents to choose naturally sweet foods such as fruits, like bananas, raisins, and other dry fruits to 60 add natural sweetness to cereals. For most nutrients, MIHS children’s average intakes met or exceed the nutrient recommendations. It is likely that fortified food items, such as milk, yogurt, juices, bread, and cereal partially contributed to the improved intake of these key nutrients. However, MIHS children’s intake of fiber and vitamin E were still low. This observation is likely due to the fact that MIHS children’s vegetable intake was low as likely was their intake of nuts and seeds. Nuts and seeds are two of the 10 main food sources of allergies [Brown, 1998] for young children and can also be a choking hazard [Brown, 1998].Unfortunately, the intake of nuts and seeds was not captured independently by the HEI-2005. When comparing the diet quality of MIHS to that of the national HS children, similar trends were observed for most HEI-2005 subgroups. DOVL, oil, and saturated fat group had bigger variations than other food groups, which might be due to different dietary assessment methods that were used for two samples. That is FFQ for MIHS versus one day dietary recall for NHANES. 2007-2008 NHANES Sample Overall, American preschool children’s diet quality still does not meet the recommendations for either nutrients or food groups. These results were confirmed by using both the Population Ratio and Mean Score methods. Nevertheless, their HEI-2005 scores were higher than in past years. All 2007-08NHANES preschool children (35years old) scored an average 62.0 for the HEI-2005 total score using Population Ratio method, higher than that of children aged 2-18 years from both low-income (56.4/100) and high-income families (55.4/100) in 2003-04 NHANES [Guenther, 2008a]. The 2007- 61 08 NHANES preschool children’s score is also higher than the average HEI-2005 score of all Americans 2 years and older (58.2/100) from the 2001-2002 NHANES [Guenther, 2008d]. In the original HEI scoring system, scores of 51 to 80 were considered as “need improvement” [Guenther, 2007]. In HEI-2005, no such scoring categories were used. Considering that the possible maximum score was 100 and by interpreting scores for each food component, it is clear that American preschool children’s diet still needs improvement for most HEI-2005 food components. However, the results have demonstrated that on average, American preschool children’s diet have met the recommendations for a few food components, including total fruits, whole fruits, and total grains using the Population Ratio method. Different income cut-points were used to analyze the diet quality differences between groups. Due to the fact that the HS group in NHANES averaged higher than normally eligible poverty levels, using the 1.85 cut-points was considered as the best for comparing differences between subgroups. The cut-point of 1.85 seemed to best distinguish a low-income non-HS population that is similar to the HS population. The results revealed that changing income cut-points, does not appeared to impact the HEI-2005 score trend that we observed among groups. The trend was that lowincome children enrolled in the HS program had a better overall diet quality compared to other groups no matter which cut-points were used. The HS children’s diet quality also surpassed that of the non-HS low income population on several food groups as well as that of the non-HS higher income children. That higher income preschool children do not necessarily have better diet quality than low-income children has been reported elsewhere 62 [Guenther, 2008a]. However, the higher income population did maintain a higher intake of DOVL than HS and non-HSLI children. The higher intake of DOVL might be attributed to both higher income and higher education level of this population. Contrary to previous findings using NHANES 1999-2004 where researchers reported that “The diet of HS preschoolers do not appear to be on par with those of other preschool-aged children” [Bucholz, 2011], the present study revealed that low-income preschool children in the Head Start program demonstrated better overall dietary quality indicated by HEI-2005.These findings were confirmed using two calculation methods and under different income cut-points. In the Mean Score method, after controlling for age, gender, race/ethnicity, we still see significance difference among the means of total HEI-2005 scores, total fruits, milk, and SoFAs. Bucholz et al. reported that on average HS children met the recommendation for all key nutrients except for vitamin E, which is consistent with the findings in the present study. Bucholz further suggested that higher percentages of HS children were not meeting the recommendations for protein, riboflavin, niacin, and calcium. In addition, HS children had worse serum profiles compare to other groups for folate, vitamin B2, and selenium [Bucholz, 2011]. There are some important issues to consider when evaluating the differences in results from Bucholz’s study versus the present study. 1) Bucholz categorized children differently from this study and used an older data set of 1999-2004 NHANES. Bucholz focused on comparing HS children to children who were not in HS but in other preschool programs, who used to be in preschool programs, or who had no preschool program at all. The preschool status was the main concern for the Bucholz study. In the present study, the researcher focused on the both Head Start status and 63 family income status. 2) Bucholz et al. used several studies on preschool program menus to support their findings. Those studies focused on the quality of preschool menus and indicated that HS program menus failed to provide enough and consistent nutrients for HS children, but none of those studies were published after 1999. After 12 years, it is possible that both the menus and practice of HS has improved. In the present study, the results on children’s nutrients intake were also similar to what has been reported by “What We Eat in America, NHANES 2009-2010” [USDA, 2012]. What We Eat in America reported similar total energy intake of low income (185% poverty income ratio) children 2-5 year old from 2009-2010 NHANES to that of NHANES population in this study. The present study also reported similar results for all nutrients as reported in the “What We Eat in America, NHANES 2009-2010”. In this study, the Head Start children’s average family poverty income ratio was higher (196% of poverty guideline) than the enrollment criteria of 100% of poverty guideline for 90% of enrollees. The researcher contacted personnel at the CDC’s National Center for Health Statistics about this issue and received the following explanations. 1) Due to the fact that the family income was self-reported, participants might potentially over-report their income for NHANES compared to what they report for Head Start. 2) Head Start participants might have under-reported their income to be eligible for the Head Start program. There are also several other reasons that might contribute to this discrepancy. 3) For NHANES, the family income was calculated by taking the mid-point of a reported income range and this could influence of accuracy of family income. 4) Children from state funded preschool programs often share the same classroom and 64 facility with Head Start children. Families might mistake in which program their children participated [http://www.michigan.gov/mde/0,1607,7-140-6530_6809_50451---,00.html]. And the state funded preschool programs usually have much higher income eligibility criteria than Head Start[http://www.michigan.gov/documents/mde/GSRP_Income_Eligibility_Guidelines_3 80766_7.pdf]. 5) Head Start programs can admit 10% of children with special needs who can from any income level. Therefore, it is likely that all these factors contributed to the high family income to poverty guideline ratio of Head Start families in NHANES. Another interesting point was that, in the present study, even when the researcher used income cut-points at1.0PL, there were still 141 low-income children who were not enrolled in Head Start and most of them were of Hispanic origin. Head Start not only prepares low-income children to do well in elementary school, but also provides good quality meals to children and nutrition education to the caregivers. It is important for the Head Start program to continue to reach the low-income communities and to increase the acceptance of Head Start among minority groups, especially the Hispanic population and other ethnic and culture groups. Strengths and Limitations Strengths This study used the 2007-08 NHANES data. CDC has released the 2009-08 NHANES, but the dietary component has not yet been released to the public. The Mypyramid Equivalents Database has not released the current version for researchers to 65 analyze any version of NHANES that is newer than the 2003-2004. The researcher used the Mypyramid Equivalent Database (MPED) Addendum along with the 2.0 version of the MPED, which allowed calculation of the HEI-2005 using the 2007-08 NHANES. In a previous study analyzing the 2005-06 NHANES, trained registered dietitians manually coded the food items that had the newer USDA food codes unavailable in the MPED, using food codes for similar items. Manually coding the food items by highly trained professionals is a possible way to address the problem. However, it is still less accurate than our methods. The researcher used the Population Ratio Method to calculate and report the HEI2005 score of the NHANES population. This is a method recommended and used by most researchers reporting the national population HEI-2005 scores [Freedman, 2008; Guenther, 2008a; Guenther, 2008d]. Limitations. The dietary intake information of the MIHS children was collected using the Block Kids Food Screener (BKFS). This provides less accurate intakes than if collecting dietary information by multiple 24hr dietary recalls [Vucic, 2009; Burrows, 2010]. Also, parents reported their children’s dietary intakes and did not ask for the children’s intake at school or specify with parents on sources and locations of the foods [Murashima, 2011]. This might lead to misreporting or underreporting. The average daily food in grams for each food item on BKFS was assigned equally to the items listed for each food group on the screener. This method was based on the assumption that NutritionQuest weighted all food items equally (as how frequently 66 they were consumed by young children) when calculating the average daily grams for each food groups. The researcher contacted NutritionQuest for their opinion on this method. NutritionQuest agreed that this method was an acceptable compromise, but less accurate than having them calculate the grams based the weight they gave to each items in a food group. NutritionQuest would provide more detailed grams for each food items listed on screener, but at a high cost. Another limitation was that during the statistical analysis, the interaction terms among variables was not examined. 67 CHAPTER 6 SUMMARY and CONCLUSIONS American’s preschool children still need to improve their diet quality, especially for key food components and nutrients that were emphasized by DGA 2010 and were shown to be low in this study. The Low-income non-HS preschool children are particularly at great risk of a poor diet and possible weight issues, and health problems as a consequence of such diet. A large percentage of low-income Hispanic children have not been reached by Head Start programs as revealed in this study. It is outreach efforts to this population are crucial to increase the diversity of Head Start programs and reduce health disparities. Although researchers have reported on the diet quality of low-income preschool children as being at risk, higher-income children did not have a better dietary quality. Not enough attention has been paid to the middle and higher income preschool children to determine why their diet quality is not better than it is. The two methods of calculating HEI-2005 scores should be carefully considered when interpreting this study. As recommended by other literature, the Population Ratio method was employed in this study to report diet quality of preschool children nationwide. The Mean Score method was employed for conducting comparisons while controlling for confounders. As the result of such comparisons, low-income preschool children in the Head Start program had the highest total HEI-2005 scores compared to non-Head Start children. However, this study is not enough to explain whether the Head Start program benefits children’s diet quality. To determine, whether Head Start program really benefit the low-income population, longitudinal studies are needed. It would be 68 interesting to follow Head Start children from enrollment to graduation to explore possible changes in the children’s diet quality and taste preferences as well as their parents’ diet quality and knowledge of nutrition. It important to notice that when using the 1.85 cut point, children who were potentially eligible for SNAP and WIC were included in the non-HSLI subpopulation. WIC implemented their new food package later in 2009, which emphasized intake of low-fat dairy and whole grains. This NHANES dataset was collected in2007-2008 before the new WIC food package implementation. It will be interestingly to see what impact the new WIC food package might have on children’s diet quality. 69 REFERENCES 70 REFERENCES Abdullah M, Ahmed L. Validating a simplified approach to the dietary assessment of vitamin A intake in preschool children. Eur J Clin Nutr. 1993;47:115–122. Akbaraly TN, Singh-Manoux A, Tabak AG, Jokela M, Virtanen M, Ferrie JE, Marmot MG, Shipley MJ, Kivimaki M. Overall Diet History and Reversibility of the Metabolic Syndrome Over 5 Years. Diab Care. 2010;33(11):2339-2341. Alaimo K, Olson CM, Frongillo EA, Briefel RR. Food insufficiency, family income, and health in US preschool and school-aged children. Am J Public Health. 2001;91:781-786. American Heart Association, 2005. No-Fad Diet Sample Menu Plan: 2,000 calories. Accessed July, 2011. Available at: http://lswhs.leesummit.k12.mo.us/FACSWebquest/AmericanHeartExample.pdf. Beydoun MA, Powell LM, Chen X, Wang Y. Food prices are associated with dietary quality, fast food consumption, and body mass index among U.S. children and adolescents. J Nutr. 2011;141(2):304-311. Blanton CA, Moshfegh AJ, Baer DJ, Kretsch MJ. The USDA automated multiple-pass method accurately estimates group total energy and nutrient intake. J Nutr. 2006; (10):2594-2599. Block G, Hartman AM, Naughton D. A reduced dietary questionnaire: development and validation. Epidemiology. 1990;1(1):58-64. Blum RE, Wei EK, Rockett HR, Langeliers JD, Leppert J, et al. Validation of a food frequency questionnaire in native American and Caucasian children 1 to years of age. Mat Child Health J.1999;3:167-172. Bollella MC, Spark A, Boccia LA, Nicklas TA, Pittman BP, Williams CL. Nutrition intake of Head Start Children: Home VS. School. J Am Coll Nutr. 1999;18(2):108-114. Brown JE. 1998. Nutrition through the life cycle. 3rd ed. ISBN-10: 0495116378. Publisher: Wadsworth Publishing. Bucholz EM, Desai MM, Rosenthal MS. Dietary Intake in Head Start vs. Non-Head Start Preschool-Aged Children: Results from the 1999-2004 National Health and Nutrition Examination Survey. J Am Diet Assoc. 2011;111(7):1021-1030. Burrows TL, Martin RJ, Collins CE. A systematic review of the validity of dietary assessment methods in children when compared with the method of doubly labeled water. J Am Diet Assoc. 2010;110(10):1501-1510. 71 Cashdan E. A sensitive period for learning about food. Human Nature. 1994;5(3):279291. Centers for Disease Control and Prevention (CDC) National Center for Health Statistics (NCHS).Interview Procedure Manual. 2008a. Accessed July 2011. Available at: http://www.cdc.gov/nchs/data/nhanes/nhanes_07_08/HouseholdInterviewer_07.pdf. Centers for Disease Control and Prevention (CDC) National Center for Health Statistics (NCHS).MEC In-Person Dietary Interviewers Procedures Manual 2007-2008. 2008b. Accessed July 2011. Available at: http://www.cdc.gov/nchs/data/nhanes/nhanes_07_08/manual_dietarymec.pdf. Centers for Disease Control and Prevention (CDC) National Center for Health Statistics (NCHS). Phone Follow-up Dietary Interviewers Procedures Manual 2007-2008. 2008c. Accessed July 2011. Available at: http://www.cdc.gov/nchs/data/nhanes/nhanes_07_08/manual_dietarypfu.pdf. Centers for Disease Control and Prevention (CDC) National Center for Health Statistics (NCHS).Analytic Guidelines 2007-2008. 2008d. Accessed July 2011. Available at: http://www.cdc.gov/nchs/nhanes/nhanes2003-2004/analytical_guidelines.htm. Centers for Disease Control and Prevention (CDC) National Center for Health Statistics (NCHS).National Health and Nutrition Examination Survey. Accessed July 7, 2011a. Available at: http://www.cdc.gov/nchs/nhanes.htm. Centers for Disease Control and Prevention (CDC) National Center for Health Statistics (NCHS).National Health and Nutrition Examination Survey Anthropometry Procedure Manual. Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention. Accessed July 7, 2011b. Available at: http://www.cdc.gov/nchs/data/nhanes/nhanes_07_08/manual_an.pdf. Centers for Disease Control and Prevention (CDC) National Center for Health Statistics (NCHS).Children's BMI Tool for Schools. Accessed Feb, 2012. Available at: http://www.cdc.gov/healthyweight/assessing/bmi/childrens_bmi/tool_for_schools.html Cook AJ, Friday JE. Pyramid Servings database for USDA Survey Codes Version 2.0. 2004. Accessed July, 2011. Available at: http://www.ars.usda.gov/Services/docs.htm?docid=8634. Cooke LJ, Wardle J, Gibson EL, Sapochnik M, Sheiham A, & Lawson M. Demographic, familial and trait predictors of fruit and vegetable consumption by pre-school children. Pub Health Nutr. 2004;7(2):295-302. Clark MA, Fox MK. Nutritional quality of the diets of US public school children and the role of the school meal programs. J Am Diet Assoc. 2009;109(2):S44-S56. 72 De Bock F, Breitenstein L, Fischer JE. Positive impact of a pre-school-based nutritional intervention on children's fruit and vegetable intake: results of a cluster-randomized trial. Pub Health Nutr. 2012; 15(3):466-475. De Koning, L., Chiuve S. E., et al. Diet-Quality Scores and the Risk of Type 2 Diabetes in men. Diab Care. 2011;34(5):1150-1156. Dietary Guidelines for Americans 2010. US Department of Agriculture web site. Accessed December, 2011. Available at: http://www.cnpp.usda.gov/dgas2010policydocument.htm. Drewnowski A, Fiddler EC, Dauchet L, Galan P, Hercberg S. Diet quality measures and cardiovascular risk factors in france: applying the Healthy Eating Index to the SU.VI.MAX study. J Am CollNutr. 2009;28(1):22-29. Ervin RB. Healthy Eating Index-2005 total and component scores for adults ages 20 and over: National Health and Nutrition Examination Survey, 2003-2004. National Health Statistics Report; no 44. Hyattsville, MD: National Center for Health Statistics.2011. Fox MK, Glantz FB, Geitz L, Burstein N. Early childhood and child care study: nutritional assessment of the CACFP. Cambridge, MA: Abt Associates Inc; 1997. Fox MK, Condon E, Briefel RR, Reidy KC, Deming DM. Food Consumption Patterns of Young Preschoolers: Are They Starting Off on the Right Path? J Am Diet Assoc. 2010;110:S52-S59. Freedman LS, Guenther PM, Krebs-Smith SM, Kott PS. A population's mean Healthy Eating Index-2005 scores are best estimated by the score of the population ratio when only one 24-hour recall is available. J Nutr. 2008;138(9):1725-1729. Freedman LS, Guenther PM, Krebs-Smith SM, Dodd KW, Midthune D. A population's distribution of Healthy Eating Index-2005 component scores can be estimated when more than one 24-hour recall is available. J Nutr. 2010;140(8):1529-1534. Fungwe T, Guenther PM, Juan WY, Hiza H, &Lino M. The Quality of Children’s Diets in 2003-04 as Measured by the Healthy Eating Index-2005.Nutrition Insight 43. Alexandria, VA: US Department of Agriculture; 2009. Gable S, & Lutz S. Household, Parent, and Child Contributions to Childhood Obesity. Family Relations. 2000;49(3):293-300. Gossett JM, Jo CH, Simpson PM. U.S. Health and Nutrition: SAS® Survey Procedures and NHANES. 2006. Accessed July, 2011. Available at: http://www2.sas.com/proceedings/sugi31/140-31.pdf. 73 Gorin AA, Raynor HA, Niemeier HM, & Wing RR. Home grocery delivery improves the household food environments of behavioral weight loss participants: results of an 8-week pilot study. Int J Behav Nutr Phys Act. 2007;4:58. Gidding SS, Dennison BA, Birch LL, Daniels SR, Gilman MW, Lichtenstein AH, Rattay KT, Steinberger J, Stettler N, Horn LV. Dietary recommendations for children and adolescents. A guide for practitioners: Consensus statement from the American Heart Association. Circulation. 2005;112:2061-2075. Guenther PM, Reedy J, Krebs-Smith SM, Reeve BB. Development and Evaluation of the Healthy Eating Index-2005: Technical Report. 2007. Center for Nutrition Policy and Promotion, U.S. Department of Agriculture. Available at http://www.cnpp.usda.gov/HealthyEatingIndex.htm. Guenther PM, Juan WY, Hiza H, Lino M et al. Diet quality of low-income and higher income Americans in 2003-04 as measured by the Healthy Eating Index-2005. Nutrition Insight 42. Alexandria, VA: US Department of Agriculture; 2008a. Guenther PM, & Krebs-Smith SM. Development of the Healthy Eating Index-2005. J Am Diet Assoc. 2008b;108(11):1896-1901. Guenther PM, Reedy J, Krebs-Smith SM, Reeve BB. Evaluation of the Healthy Eating Index-2005. J Am Diet Assoc. 2008c;108(11):1854-1864. Guenther PM, Juan WY, Reedy J, Britten P, Lino M, Carlson A, Hiza H, Krebs-Smith SM. Diet Quality of Americans in 1994-96 and 2001-02as Measured by the Healthy Eating Index-2005. Nutrition Insight 37. Alexandria, VA: US Department of Agriculture; 2008d. Hamidi M, Tarasuk V, Corey P, Cheung AM. Association between the Healthy Eating Index and bone turnover markers in US postmenopausal women aged >= 45 y. Am J Clin Nutr. 2011;94(1):199-208. Harnack L, Walters SA, Jacobs DR Jr. Dietary intake and food sources of whole grains among US children and adolescents: data from the 1994-1996 Continuing Survey of Food Intakes by Individuals. J Am Diet Assoc. 2003;103(8):1015-1019. Harris G. Development of taste and food preferences in children. Curr Opin Clin Nutr Metab Care. 2008;11(3):315-319. Hoerr SL, Horodynski MA, Lee SY, & Henry M. Predictors of nutritional adequacy in mother-toddler dyads from rural families with limited incomes. J Am Diet Assoc. 2006a;106(11):1766-1773. 74 Hoerr SL, Lee SY, Schiffman RF, Horodynski MO, & McKelvey L. Beverage consumption of mother-toddler dyads in families with limited incomes. J Pediatr Nurs. 2006b;21(6):403-411. Hoerr SL, Tsuei E, Liu Y, Franklin FA, Nicklas TA. Diet quality varies by race/ethnicity of Head Start mothers. J Am Diet Assoc. 2008;108(4):651-659. Hoerr SL, Hughes SO, Fisher JO, Nicklas TA, Liu Y, Shewchuk RM. Associations among parental feeding styles and children's food intake in families with limited incomes. Int J Behav Nutr Phys Act. 2009;13(6):55. Institute of Medicine, Food and Nutrition Board. Dietary risk assessment in the WIC program. Washington, DC, National Academy Press. Accessed Feb, 2012. Available at: http://www.fns.usda.gov/ora/menu/published/wic/FILES/WICDietaryRisk.pdf. Johnson SL, McPhee L, Birch LL. Conditioned preferences: young children prefer flavors associated with high dietary fat. Physiol Behav. 1991 Dec;50(6):1245-1251. Kantor LS, Variyam JN, Allshouse JE et al. Choose a variety of grains daily, especially Whole grains: a challenge for consumers. J Nutr. 2001;131, Suppl., S473–S486. Kleinman RE. Pediatric Nutrition Handbook (6th ed.). American Academy of Pediatrics, Elk Grove Village, IL. 2009;1294-1295. Knol LL, Haughton B, & Fitzhugh EC. Dietary patterns of young, low-income US children. J Am Diet Assoc. 2005;105(11):1765-1773. KralTanja VE, Moore RH, Stunkard AJ. Adolescent eating in the absence of hunger and relation to discretionary calorie allowance. J Am Diet Assoc. 2010;110(12):1896-1900. Kranz S. Meeting the dietary reference intakes for fiber: Socio demographic characteristics of preschoolers, with high fiber intakes. Am J Pub Heal. 2006; 96(9):1538-1541. Krebs-Smith SM, Guenther PM, Subar AF, Kirkpatrick SI, & Dodd KW. Americans do not meet federal dietary recommendations. J Nutr. 2010;140(10):1832-1838. Kuczmarski MF, Sees AC, Hotchkiss L, Cotugna N, Evans MK, Zonderman AB. Higher Healthy Eating Index-2005 scores associated with reduced symptoms of depression in an Urban population: findings from the healthy aging in neighborhoods of diversity across the life span (HANDLS) study. J Am Diet Assoc. 2010;10(3):383-389. LaRowe TL, Moeller SM, Adams AK. Beverage patterns, diet quality, and body mass index of US preschool and school-aged children. J Am Diet Assoc. 2007;107(7):11241133. 75 Liem DG, Zandstra L, Thomas A. Prediction of children's flavor preferences. Effect of age and stability in reported preferences. Appetite. 2010;55(1):69-75. Lohman TG, Toche AF, Martorell M. Anthropometric standardization reference manual, champayne, IL., Human Kinerics. 1988. Lorson BA, Melgar-Quinonez HR, Taylor CA. Correlates of fruit and vegetable intakes in US children. J Am Diet Assoc. J Am Diet Assoc. 2009;109:474-478. Marshall TA, Eichenberger Gilmore JM, Broffitt B, Stumbo PJ, & Levy SM. Relative validity of the Iowa Fluoride Study targeted nutrient semi-quantitative questionnaire and the block kids' food questionnaire for estimating beverage, calcium, and vitamin D intakes by children. J Am Diet Assoc. 2008;108(3):465-472. Maynard M, Gunnell D, Emmett P, Frankel S, Davey Smith G. Fruit, vegetables, and antioxidants in childhood and risk of adult cancer: The Boyd Orr cohort. J Epi Comm Health. 2003;57:18-25. McCullough ML, Patel AV, Kushi LH, Patel R, Willett WC, Doyle C, Thun MJ, Gapstur SM: Following Cancer Prevention Guidelines Reduces Risk of Cancer, Cardiovascular Disease, and All-Cause Mortality. Can Epid Bioma Prev. 2011;20(6):1089-1097. Michels KB, Willett WC. Self-administered semi-quantitative food frequency questionnaires: patterns, predictors, and interpretation of omitted items. Epidemiology. 2009;20(2):295-301. Michigan Department of Human Services, 2008.Accessed July, 2012. Available at: http://www.michigan.gov/dhs/0,4562,7-124-7691_7752-193936--,00.html. Montgomery C, Reilly JJ, Jackson DM, Kelly LA, Slater C, Paton JY, Grant S. Validation of energy intake by 24-hour multiple pass recall: comparison with total energy expenditure in children aged 5-7 years. Br J Nutr. 2005;93(5):671-676. MyPlate for Kids homepage. Accessed July 7, 2011. Available at http://www.choosemyplate.gov/ Murashima M. The relationship of parental feeding control practices to food intake of 35yr children in families with limited incomes [PhD Dissertation]: Michigan State University. 2010. Murashima M, Hoerr SL, Hughes SO, et al. Confirmatory factor analysis of a Questionnaire Measuring control in parental feeding practices in mothers of Head Start children. Appetite. 2011; 56(3):594-601. National Cancer Institute. NCS Dietary Assessment Literature Review. Accessed July, 2011.Available at: http://riskfactor.cancer.gov/tools/children/review/ 76 National Cancer Institute. Sources of solid fat in diets of the U.S. population ages 2years and older, NHANES 2005-2006. Rick factor monitoring and methods, Cancer Control and Population Sciences. Updated December 21, 2010a. Accessed July, 2011. Available at: http://riskfactor.cancer.gov/diet/foodsources/solid_fats/table1a.html. National Cancer Institute. Sources of added sugar in diets of the U.S. population ages 2years and older, NHANES 2005-2006.Rick factor monitoring and methods, Cancer Control and Population Sciences. Updated December 21, 2010b. Accessed July, 2011. Available at: http://riskfactor.cancer.gov/diet/foodsources/added_sugars/table1a.html. National Heart Lung and Blood Institute. Your guide to lowering your blood pressure with DASH. 2006. National Research Council. Dietary Reference Intakes for Vitamin C, Vitamin E, Selenium, and Carotenoids . Washington, DC: The National Academies Press, 2000. National Research Council. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc . Washington, DC: The National Academies Press, 2001. National Research Council. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients) . Washington, DC: The National Academies Press, 2005a. National Research Council. Dietary Reference Intakes for Water, Potassium, Sodium, Chloride, and Sulfate. Washington, DC: The National Academies Press, 2005b. National Research Council. Dietary Reference Intakes for Calcium and Vitamin D . Washington, DC: The National Academies Press, 2011. Newby PK. Are dietary intakes and eating behaviors related to childhood obesity? A Comprehensive review of the evidence. J Law Med Ethics. 2007;35(1):35-60. NutritionQuest, 2011. Accessed July, 2011.Available at: http://www.nutritionquest.com/. Oakley CB, Bomba AK, Knight KB, Byrd SH. Evaluation of menus planned in Mississippi children care centers participating in the Child and Adult Care Food Program. J Am Diet Assoc. 1995;95:765-768. O'Connor TM, Yang SJ, Nicklas TA. Beverage intake among preschool children and its effect on weight status. Pediatrics. 2006;118(4):e1010-1018. Ogden CL, Carroll MD, Flegal KM. High body mass index for age among US children and adolescents, 2003– 2006. JAMA. 2008;299(20),2401–2405. 77 O’Neil CE, Niciklas TA, Zanovec M, Cho SS, Kleinman R. Consumption of whole grains is associated with improved diet quality and nutrient intake in children and adolescents: the National Health and Nutrition Examination Survey 1999–2004. Public Health Nutrition.2011a;14(2),347- 355. O'Neil CE, Nicklas TA, Zanovec M, Fulgoni VL 3rd. Diet quality is positively associated with 100% fruit juice consumption in children and adults in the United States: NHANES 2003-2006. Nutr J. 2011b;13:10-17. O'Neil CE, Fulgoni VL 3rd, Nicklas TA. Association of candy consumption with body weight measures, other health risk factors for cardiovascular disease, and diet quality in US children and adolescents: NHANES 1999-2004. Food Nutr Res. 2011c;55. Reilly JJ, Montgomery C, Jackson D, MacRitchie J, Armstrong J. Energy intake by multiple pass 24 h recall and total energy expenditure: a comparison in a representative sample of 3-4-year-olds. Br J Nutr. 2001;86(5):601-605. Rose D, Smallwood D, Blaylock J. Socioeconomic factors associated with the iron intake of preschoolers in the United States. Nutr Res. 1995;15:1298-1309. Serdula MK, Alexander MP, Scanlon KS, Bowman BA. What are preschool children eating? A review of dietary assessment. Annu Rev Nutr. 2001;21:475-489. Shah BS, Freeland-Graves JH, Cahill JM, Lu HX, Graves GR. Diet Quality as Measured by the Healthy Eating Index and the Association with Lipid Profile in Low-Income Women in Early Postpartum. J Am Diet Assoc. 2010;110(2):274-279. Skinner JD, Carruth BR, Bunds W, Ziegler PJ. Children's food preferences: A longitudinal analysis. J Am Diet Assoc. 2002;102, pp:1638–1647. Spurrier NJ, Magarey AA, Golley R, Curnow F, Sawyer MG. Relationships between the home environment and physical activity and dietary patterns of preschool children: a cross-sectional study. Inter J Beh Nutr Phy. 2008;5:31-42. Stevens L, Nelson M. The contribution of school meals and packed lunch to food consumption and nutrient intakes in UK primary school children from a low income population. J Hum Nutr Diet. 2011;24:223–232. Tardivo AP, Nahas-Neto J, Nahas EA, Maesta N, Rodrigues MA, Orsatti FL. Associations between healthy eating patterns and indicators of metabolic risk in postmenopausal women. Nutr J. 2010;9:64. Thompson FE, Byers T. Dietary assessment resource manual. J Nutr.1994;124(S11): 2245S-2317S. 78 Tran KM, Johnson RK, Soultanakis RP, Matthews DE. In-person vs. telephone administered multiple-pass 24-hour recalls in women: Validation with doubly labeled water. J Am Diet Assoc. 2000;100(7):777-783. Treiber FA, Leonard SB, Frank G, Musante L, Davis H. Dietary assessment instruments for preschool children: reliability of parental responses to the 24h recall and a food frequency questionnaire. J Am Diet Assoc.1990; 90:814-820. U.S. Census Bureau. 2009. Income, Poverty, and Health Insurance Coverage in the United States: 2008. Accessed July, 2011. Available at: http://www.census.gov/prod/2009pubs/p60-236.pdf. U.S. Census Bureau. Poverty. Last Revised: January 14, 2011. Accessed July, 2011.Available from: http://www.census.gov/hhes/www/poverty/data/threshld/index.html U.S. Department of Agriculture, Agricultural Research Service. What We Eat in America, NHANES 2009-2010 Tables: Nutrient Intakes from food: Mean Amounts Consumed per Individual, by Income (% poverty threshold). (2012, July).Accessed August 2012.Available at: http://www.ars.usda.gov/Services/docs.htm?docid=18349. U.S. Department of Agriculture and Nutrition Database for Dietary Studies, 4.1. Beltsville, MD: U.S. Department of Agriculture, Agriculture Research Services, Food Services Research Group. 2010. U.S. Department of Agriculture, USDA. Center of Nutrition Policy and Promotion (CNPP). 2006. Sample Menus for a 2000 calorie food pattern. Accessed July 2011. Available at http://www.cnpp.usda.gov/Publications/MyPyramid/print%20materials/MyPyramidSam leMenu.pdf U.S. Department of Agriculture, USDA. The Food Assistance Landscape, 2008 Annual Report. U.S. Department of Agriculture, USDA.Myplate.gov. 2011a.Accessed July, 18, 2011. Available from: http://www.myplate.gov/ U.S. Department of Agriculture, USDA. Center for Nutrition Policy and Promotion. Healthy eating Index. 2011b. Available at: http://www.cnpp.usda.gov/HealthyEatingIndexSupportFiles0102.htm U.S. Department of Agriculture, USDA. Center for Nutrition Policy and Promotion. Healthy eating Index -2005. 2011c. Available at: http://www.cnpp.usda.gov/HealthyEatingIndex-2005report.htm 79 U.S. Department of Agriculture. National School Lunch Program: participation and lunches served. Available at: http://www.commodityfood.usda.gov/pd/slsummar.htm.Accessed April, 2012a. U.S. Department of Agriculture. Food and Nutrition Service.WIC Food Packages. Available at: http://www.fns.usda.gov/wic/benefitsandservices/foodpkg.HTM. Accessed April, 2012b. U.S. Department of Health and Human Services, US Department of Agriculture, & US Dietary Guidelines Advisory Committee. Dietary guidelines for Americans, 2010. Washington, D.C.: G.P.O.; 2010a. U.S. Department of Health and Human Services. Office of Head Start. 2010b [updated 2010; cited January 10, 2011]. Available at: http://www.acf.hhs.gov/programs/ohs/index.html. U.S. Department of Health and Human Services. 2010c. the poverty guidelines updated periodically in the Federal Register by the U.S. Department of Health and Human Services under the authority of 42 U.S.C. 9902(2). Available at: http://aspe.hhs.gov/poverty/08poverty.shtml. Accessed July, 2011. Van Duyn MS, Pivonka E. Overview of the health benefits of fruit and vegetable consumption for the dietetics professional: selected literature. J Am Diet Assoc. 2000;100: 1511–1521. Vaughan DR. Exploring the use of the public’s views to set income poverty thresholds and adjust them over time. Soc Secur Bull. 1993;56(2):22-46. Ventura AK, Mennella JA. Innate and learned preferences for sweet taste during childhood. Curr Opin Clin Nutr Metab Care. 2011 Jul;14(4):379-384. Verma V, Betti G. Taylor linearization sampling errors and design effects for poverty measures and other complex statistics. J Appl Stat 2011;38(8):1549-1576. Vucic V, Glibetic M, Novakovic R, Ngo J, Ristic-Medic D, Tepsic J, Ranic M, Serra Majem L, Gurinovic M. Dietary assessment methods used for low-income populations in food consumption surveys: a literature review. Br J Nutr. 2009;101(S2):S95-101. Weissberg RP, Gullotta TP, Adams GR, Hampton RL, Ryan BA, eds. Enhancing Children’s & Wellness: Healthy Children 2010. Issues in Children’s & Families’ Lives. Volume 8. Thousand Oaks, CA: Sage;1997; 214-249. Willett W. Nutrition Epidemiology, 1sted. New York, NY: Oxford University Press; 1990. 80 Willett WC. Eat, Drink, and Be Healthy: The Harvard medical School Guide to Healthy Eating. New York: Free Press. 2005. Wolever TMS, Hamad S, Gittelsohn J, Hanley AJG, Logan A, Harris SB, Zinman B. Nutrient intake and food use in an Ojibwa-Cree community in Northern Ontario assessed by 24H dietary recall. Nutr Res. 1997;17(4):603-618. Woodruff SJ, Hanning RM, Lambraki I, Storey KE, McCargar L. Healthy Eating Index C is among adolescents with body weight concerns, weight loss dieting, and meal skipping. Body Image. 2008;5(4):404-408. World Food Programme. Accessed July, 2011. Available at http://foodquality.wfp.org/Home/tabid/36/Default.aspx. 81