PARENT FEEDING STRAT EGIES AND THEIR ASSOCIATION WITH PRESCHOOLERS™ WEIGHT STATUS AND DIET QUALITY IN LOW INCOME FAMILIES By Sulafa Hassan Elshowaya A THESIS Submitted to Michigan State University in partial fulfillment of t he requirements for the degree of Human Nutrition ŠMaster of Science 2015ABSTRACT PARENT FEEDING STRAT EGIES AND THEIR ASSOCIATION WITH PRESCHOOLERS™ WEIGHT STATUS AND DIET QUALITY IN LOW INCOME FAMILIES By Sulafa Hassan Elshowaya The purpose of the Eat Healthy study ( EH), A Parent ™s Guide to Raising a Healthy Eater , was to improve parent feeding strategies, as well as the preschooler™s diet quality and weight status. P ara professional educators delivered 1 -6 lessons to 152 Supplemental Nutrition Assist ance Program Education (SNAP -Ed) eligible parents in four Michigan counties using a combination of home visits and phone calls. They collected demographic, anthropometric , Parent Feeding Behavior Questionnaire (PFBQ) and Block Kids Food Screener (BKFS ) dat a at baseline and a t a 3-month follow -up. Data w ere analyzed to compare control and intervention groups for parental feeding strategies, children™s weight and diet quality at baseline and follow -up to examine associations . At the 3 -month follow up , the EH study succeeded in improving three of eight parent feeding strategies : high control, high contingency and , mealtime behaviors using a valid and reliable instrument . The parents became less controlling, used less contingency and rewarding, an d improved mealtime strategies such as no TV at meals and eating family meals . The EH was also successful in improving the diet quality of the preschooler s by increasing nutrient dense food s and decreasing energy dense food s. However the EH did not improv e the weight status of the preschooler. ` ACKNOWLEDGEMENTS I would like to thank my major professor, Dr. Sharon Hoerr, for mentoring me during the last two years. She taught me many things during th is time , such as ded ication and professionalism. She devoted her time on weekends and holidays coming to the lab in order to assist with finish ing the writing of my thesis . She empowered , supported , encouraged , and picked me up. As English is my second language, she went through my thesis uncountable times to ensure no grammatical or spelling mistakes were made . Dr.Hoerr was and will always be my role model. I would also like to thank my co -advisor , Dr. W eatherspoon , and my third committee member , Dr. Horodynski , for their guidance and constructive advice during the last year. I also would like to thank Tzufen Chang for her statistical help in the data analysis. I thank my family, my hu sband Amin, my three lovely kids Omar, Azza and Ahmad , my parents, and my sisters and brothers for their tremendous support and encouragement during my master s degree . iii TABLE OF CONTENTS LIST OF TABLES ........................................................................................................................vi LIST OF FIGURES ....................................................................................................................vii i KEY TO AB BREVIATIONS ......................................................................................................ix CHAPTER I . INTRODUCTIO N.................................................................................................1 Problem Statement and Rationale ...................................................................................1 Research Questions and Hypotheses ...............................................................................4 Operati onal Definitions ...................................................................................................5 CHAPTER II . LITERA TURE REVIEW ....................................................................................9 Diet Qual ity .....................................................................................................................9 Obesity ...........................................................................................................................11 Home Environment .......................................................................................................12 Parent Feeding Strategies Versus Parenting Styles or Feeding Styles ..........................13 Research on Parent Feeding Strategies Relating to Children's Diet Quality and Weight Status .....................................................................................................15 Interventi on Studies to Change Parent Feeding Strategies and Affect Child Weight and Food Intake ...........................................................................................16 CHAPTER III . METHODS ........................................................................................................23 Study Design................................................................................................................. 23 Sample and Recruitment ............................................ ...................................................24 Procedure s.....................................................................................................................24 Intervention ........................................................ ...........................................................25 Instruments, Measurements and Variables ....................................................................30 Demographic Data................................................................. ........................................30 Anthropometrics............................................................................................................30 Block Kid s Food Screener (BKFS) ................................................... ............................30 Parent Feeding Behavior Questionnaire (PFBQ) .............. ............................................33 Statistical Analysis ........................................................................................................34 CHAPTER IV . RESULTS ..........................................................................................................38 Demographics....................................................................................... .........................38 Research Question 1: Baseline Results.........................................................................48 Research Question 2: Post -Test Results........................................................................49 Research Question 3: Follow -Up Results.....................................................................51 CHAPTER V . DISCUSSION ......................................................................................................55 CHAPTER VI. STRENGTHS, LIMIT ATION S, AND IMPLICATIONS .............................60 Study Strengths ..............................................................................................................60 Limitations .....................................................................................................................60 iv Implications for Future Research .................................................................................. 62 Concl usion .....................................................................................................................62 APPENDICES ..............................................................................................................................64 Appendix A. Parent Feeding Strategy Questionnaire Listed by Construct.......................65 Appendix B. Parent Consent Form......................................................................... ...........67 Appendix C. IRB Consent Form........................................................................................71 Appendix D. Block Kids Food Screener (BKFS)..............................................................73 Appendix E. Glo ssary........................................................................................................75 REFERENCES .............................................................................................................................77 v LIST OF TABLES Table 1. Dietary Guidelines for Young Children for Energy and Food Groups, DGA (2010).....10 Table 2. Dietary Guidelines for Micronutrients by Young Children, DGA (2010)......................10 Table 3. Summary of Intervention Studies t o Change Parent Feeding Strategies and Affect Child Weight and Food Intake to Address Child Obesity in Young Children over Last Decade.....................................................................................20 Table 4 . A Quasi -Experimental Stud y Design and Group Comparisons......... .............................23 Table 5 . Types of Data Collected for Each Group Over Time of t he Lesson Delivery and Follow -Up...............................................................................................................25 Table 6. Curriculum Grid for Eat Healthy by Lesson, Concepts, Objectives and Activities ........27 Table 7. Categories and Indicators of Diet Quality from BKFS............................................. ......32 Table 8. Demographics and Weight Status for Parents and Children at Baseline........ ................38 Table 9. Descriptive Statistics and Internal Reliability of Parental Feeding Strategies by Construct at Baseline....................................................................................................39 Table 10. Descriptive of Child™s Food Intake from the KBFS Measured at Baseline Shown by Group at Baseline ..............................................................................................40 Tab le 11. Bivariate Correlations Between Feeding Strategies and Parent and Child BMI at Baseline .................................................................................................................42 Table 12. Regression Models for Association of Child Weight Status with Parental Feeding Strategies : Estimated Odds Ratios, Standard Errors, 95% CI of Parental Feeding Strategies at the Baseline....................................................................... ............................43 Table 13. Logistic Regression Models for Association of Child Weight Status with Parental Feeding Behavior: Estimated Odds Ratios, Standard Errors, 95% CI of Parental Feeding Strateg ies at Baseline..................... ....................................................................................45 Table 14. Pearson Correlations Between Parental Feeding Behavior and Diet Qua lity at Baseline.................................................................................................................. ........47 Table 1 5. Change in Feeding Strategies Between Pre- and Post -tests for the Treatment and Control Groups Using GEE for Repeated Measures and Controlling for Time and Group ............................................................................................................................. ....50 vi Table 1 6. Differences in Feeding Strategies Between Baseline and 3 Month Follow -Up for the Combined Sample of Both the Treatment and Control Groups after Receiving the EH Curriculum (N=91) ................................................................................................51 Table 1 7. Difference in Diet Quality Between Baseline and 3-Month Follow -Up for the Combined Sample of Both the Treatment and Control Groups (N =91) ...........................52 Table 1 8. Difference in Child BMI Per centile and Weight Status Between Baseline and Follow -Up Test for the Combined Sample of Both the Treatment and Control Groups (N =91) ................................................................................................................ ..53 Table 1 9. Healthy Eating Activities..............................................................................................68 vii LIST OF FIGURES Figure 1. IRB Consent Form Image..............................................................................................71 Figure 2. Block Kids Food Screener (BKFS) Questionnaire Image.............................................73 viii KEY TO ABBREVIATIONS BKFS: Block Kids Food Screener BMI: Body Mass Index DGA: Dietary Guidelines for Americans EH: Eat Healthy: Your Kids are Watching: A Parent Guide to Raising a Healthy Eater GEE: Generalized Estimating Equation PFBQ: Parent Feeding Behavior Questionnaire SNAP ŒEd: Supplemental Nutrition Assistance Program Education WIC: Special Supplemental Nutrition Program for Women, Infants and Children (WIC) ix CHAPTER I. INTRODUCTION Problem Statement and Rationale Preschool children in families with limited resources have poor quality diet s and a high rate of obesity that appear to relate , in part, to poor parental feeding strategies, but we do not know the degree to which such strategies can be improved. Childhood obesity has tripled in the last three decades in the United States (Ogden et al., 2010; Ogden and Carroll, 2012; Ogden et al., 2014) and women and children of low socio -economic status demonstrate poor diet quality and higher rates of obesity and chronic diseases compared to those w ith middle inco me (Darman & Drewnowski; 2008, & May et al., 2013). Low intakes of nutrient -dense foods such as fruits and vegetables, low fat dairy products, and whole grains, and high intake of energy dense, low -nutrient foods like sweets, snack chips, an d sweet beverages are the main dietary issues for low income families (Kranz, 2006; Power et al. 2006; Nicklas & Hayes, 2008). Healthy food environments play a major role in the diet quality of young children (Anderson & Whitaker, 2010) and pa rents play a central part in teaching their young children how to become healthy eaters (Fisher & Birch, 2007). Parents are role models during early childhood, model eating behaviors and food selection for their children , serve as the primary determinant of the food environment in providing the foods and the food structures in the home (Scaglioni & Salvioni, 2008). Parenting practices are, by definition, behaviors or strategies that parents use to get their children to do something specific; in the case of feeding strategies, this is to impact children ™s food choices and consumption . Parental strategies are embedded within the general parental style definition (Baumrind, 1989; Ventura & Birch, 2008), but are no t synonymous with styles. 1 Some consider parenting style to be an i nherent trait and not easily changed, whereas behavioral strategies can be altered (Hughes et al., 2005; Hughes et al., 2008). Many research studies in predominantly white middle income fam ilies have found that parental food restriction or pressure to eat , are feeding strategies which negatively relate to children ™s self -regulation and satiety (Fisher, 2002). Food restriction, a highly controlled feeding strategy, has been associated with children being overweight (Clark, 2007). On the other hand, parents with permissive feeding style such as indulgent or uninvolved had children with low intake s of nutritionally dense foods like fruits, vegetables and low fat dairy (Hoerr et al., 2009). Furthermore, permissive parenting styles have also been associated with increased obesity in children (Rhee et al., 2006). So far, several studies have attempted to change paren ts™ feeding strategies in order to address obesity in toddlers , preschoolers or older children (Dicken et al.,2014 ; Lent et al, 2011; Horodynsi et al., 2011; Ostbye et al., 2011; West et al.,2010; Harvey -Berino et al., 2003). Although difficult to compa re, it appears that interventions using both in -home and group lessons have been effective in improving some parent feeding strategies associated with improved child diet quality and lower child BMI Z -scores for those considered overweight or obese . Only one study, however, showed reduction in obese children™s BMI Z -score, and this was for middle income parents (West et al., 2005). The feeding strategies most frequently targeted appeared to be parents providing healthy foods in the househ old, role modeling, offering food choices within limits, and less restrictive feeding. In 2007, no evidence -based interventions were available for use with low -income families of preschoolers such as those in the Head Start program. In addition, almost a ll research with parent feeding strategies was done on white middle -income families (Birch 2001, Fisher et al., 1999; Fisher, 2002). Therefore, researchers in mid -Michigan conducted studies to determine 2 which parental feeding strategies were most assoc iated with good diet quality and healthy weight status in children (Mur ashima et al., 2011; Mur ashima et al., 2012). From these findings, intervention materials appropriate for low income and low literacy audiences were developed and pilot test ed (Reznar, 2012; Reznar et al., 2014). Following revisions to the materials, a random clinical trial was conducted 2013 -14 with 152 low income parents of preschoolers who received an intervention called, Eat Healthy (EH), Your Kids are W atching: A Parent Guide to Raising a Healthy Eater. Eat Healthy (EH), Your Kids are W atching: A Parent Guide to Raising a Healthy Eater was designed for low -income parents of preschool children to help them improve their feeding strategies. The materials target those feeding strategies found most associated with normal we ight status of the preschooler (Murashima, 2012). These strategies included availability of healthy foods, parent modeling of healthy foods, encouragement to try new foods, mealtime structures, and re gular timing of meals and snacks. The disadvantages of negative feeding strategies such as highly controlling feeding behaviors using contingency (reward and punishment), availability of unhealthy foods, and parent modeling of unhealthy foods were also inc luded in the lessons. EH consists of five lessons and 2 -3 minute videos that were delivered to Supplemental Nutrition Assistance Program Education (SNAP -Ed) eligible parents of preschool children age (2½ -5 years). Educators taught EH to parents in four M ichigan counties using a combination of home visits and phone calls. The lesson topics were: 1) Kids are what they eat ; 2) Be a good role model ; 3) Ways to praise at meals; 4) Making mealtime family time and fun ; 5) Learning to eat healthy. EH comprised t he intervent ion from which this secondary data analysis was conducted to determine the intervention™s efficacy to improve parent feeding strategies and child diet quality and weight status. 3 To achieve this goal, the following research questions a nd hypotheses were addressed: Research Questions and Hypotheses (See operational definitions that follow the hypotheses .) At baseline Research question 1.1 : What is the relationship between parent feeding strategies and child weight status and BMI percent iles at baseline ? Ho 1.1a: Preschoolers in low -income families whose parents have negative feeding strategies will be overweight /obese. Ho1.1b : Preschoolers in low -income families whose parents have positive feeding strategies will be less overweight /ob ese. Question 1.2 What is the relationship between parent feeding strategies and diet quality of their preschoolers? Ho 1.2a: Preschoolers in low -income families whose parents have negative feeding strategies will have poor diet quality. Ho 1.2b: Preschoo lers whose parents have positive feeding strategies will have improved diet quality. At the post - intervention Research question 2 : What impact does a six -week parent intervention, focusing on child feeding have on parent feeding strategies of preschool ers compared to parents who do not participate? 4 Ho 2.1: Feeding strategies of parents in the EH intervention group will improve compared to the control group, comparing the baseline and the post Œtest. At the 3 -month follow -up (after the end of the interve ntion) Research Question 3: What impact does a six -week parent intervention focusing on child feeding have on: (a) parent fee ding strategies of preschoolers; (b) the preschoolers™ diet quality; and (c) preschoolers™ weight status as compared to their basel ine measurements? Ho 3.1 : There will be an improvement in positive parent feeding strategies, and a decline in negative feeding strategies. Ho 3.2: There will be an improvement in dietary indicators reflecting good diet quality and a decline in those re flecting poor diet quality . Ho 3.3: There will be fewer a decline in average BMI percentiles for those who are overweight and or obese . Operational Definitions Preschooler™s BMI percentile Body Mass Index (BMI=wt in kg/ht in m2) in percentiles by age and gender according to CDC™s growth charts for children. (www: cdc .gov ) This variable can be continuous or categorized as follows: Underweight = 0 -4.99th percentile Normal weight = 5 -84.99th percentile Overweight =85 -94.99th per centile Obese =95 -100th percentile 5 Parental BMI This variable can be continuous or categorized as follows: (World Health O rganization - http://apps.who.int/bmi/index. ) Underweight = 0 -18.49 Normal Weight =18.5 -24.99 Overweight = 25 -29.99 Obese 1: 30 -34.99 Obese 2: 35 -39.99 Obese 3: 40 -49.99 Obese 4: 50 -100 Parental feeding strategies Derived from the 29 -item Likert Scaled Parental Feeding Behavior Questionnaire (PFBQ) (See Appendix A) scored as 1= never, 2=rarely, 3=sometimes, 4=most of t he time, 5=always, where high scores indicated a more positive behavior for each construct (Murashima et al., 2012; Reznar et al., 2014). The PFBQ was administered at baseline, post -test, and at a 3 -month follow -up. These items comprised six multi -items and two single -item constructs as follows, with the first four considered negative feeding strategies and the last four considered positive feeding strategies. 1) High Control feeding I beg my child to eat dinnerfl ; 6 2) High Contingency feeding child that I will take a food away if the child doesn™t eat , for example, fiIf you don™t finish your vegetables, you won™t get dessertfl; 3) Permissive feedi ng time= whenever he or she is hungryfl; 4) Indulgent mealtime behavior= play and watch TV during mealsfl; 5) Healthy Availability and Modeling= parents scored 4 for five items such as, fiI keep sweets, candy or salty snacks where my child can reach themfl (reverse scored); 6) Child -centered feeding = parents scored 4 for six items such as, fi I say something positive about the foodfl; 7) Milk modeling= parents scored 4 for one item, fi I drink milk in front of my childfl; 8) Fruit and vegetable modeling= parent scored 4 for one item, fiI eat fruits and vegetables in front of my child.fl Diet Quality The diet quality indices were developed from parental responses to the Block Kids Food Screener ( See Appendix D) for the preschooler™s food intake the previous week. Parents completed this screener twice Šboth at baseline and at the 3 month follow -up. Diet quality indices can be categorized as t hose reflecting a nutritionally balanced diet versus one that is energy -dense. Rather than a composite score for diet quality like the Healthy Eating Index -2005, a series of sentinel indicators were used. For example, a preschooler with poor die t quality might consume 2% or whole milk, sweetene d beverages , more than 6 fl oz of fruit juice a day, 7 sweetened cereals, and few fruits or vegetables. On the other hand, high diet quality for a preschooler might include: nonfat or 1% fat milk, frequent ve getable consumption, 1 -2 cups vegetables per day, and 1 -2 cups of fruits per day. Indicators of good diet quality (by frequency or cup equiv) included fruits, vegetables and low fat milk. Likewise, energy -dense diet indicators were for frequency of sugary beverages, sweets, snack chips, etc. 8 CHAPTER II. LITERATURE REVIEW This literature review covers several topics relevant to this research such as: children™s diet quality, child obes ity, the home food envi ronment and parent feeding strategies and styles. The review will conclude with the evaluation of intervention studies that attempted to improve parent feeding behaviors, the children™s diet quality and/or weight status. For purpose s of this study, young children are tho se less than five years of age, preschoolers are those ages 3 -5 years, and toddlers, 1 -3 years of age. The focus group of this study was families with limited resources who had a child from 2½ -5 years of age. Diet Qu ality The diet quality of children age 2-5 years is of great concern because rapid physical growth and development occur during this stage (Savage et al., 2007). The National Health and Nutrition Examination Survey (NHANES) revealed that the nutri ents in the diets of most children 2 -5 years of age were below what is recommended. Using the Healthy Eating Index -2010 to evaluate dietary intake for children ages 2 -17 years showed that all scores were below the standards, except for dairy and protein which were close to 100%. The diet of the children was low in whole grains, dark greens and green vegetables (Center for Nutrition Policy and Promotion, 2013). Other research on young children from low income families, found that they were consuming low amounts of dairy, fruits and vegeta bles (Hoerr et al., 2008, Patrick et al., 2005). The 2010 Dietary Guidelines for Americans (DGA), recommendations for preschool children are to maintain calorie intake at 1000 Kcal for those ages 2 -3 years and at 1200 -1400 Kcal for children age s 4-8 years. This recommendation was based on sedentary lifestyle of less 9 than 30 minutes of daily physical activity. The DGA also recommended that children 2 -3 years old consume 3 oz g rain, 2 oz lean meat, 1 cup fruits and 1 cup v egetables per day. Children ages 4-8 years should consume 4 -5 oz grain, 3 oz lean meat, 1 -1½ cups fruits, and 1½ cups vegetables (Dietary Guidelines for Americans, 2010). (See Tables 1 and 2). Table 1 . Dietary Guidelines for Young Children for Energy and Food Groups, DGA (2010) Category 2-3 years of age requirements 4-8 years of age requirements Total calories 1000 Kcal 1200 -1400 Kcal Grains 3 oz. 4-5 oz. Lean meat 2 oz. 3 oz. Fruits 1 cup 1-1 ½ cup Vegetables 1 cup 1 ½ cup Low fat milk 2 cups 2 cups Table 2. Dietary Guidelines for Micronutrients by Young Children, DGA (2010) Component Recommendation 1-3 years of age 4-8 years of age Fib er(g/d) 19 25 Cholesterol (mg/d) <300 <300 Vitamin A(mcg RAE/d) 210 275 Vitamin E (mg/d) 5 6 Total Folate 120 160 Calcium (mg/d) 500 800 Magnesium (mg/d) 65 110 Phosphorous (mg/d) 380 405 Iron (mg/d) 3.0 4.0 Zinc (mg/d) 2.4 4.0 Potassium (mg/d) 3000 3800 Krebs -Smith et al. (2010) using national dietary data found that the majority of the United States™ children did not meet the recommendations for food groups, except for grains, fat, meat, sugar, and beans. The same resear chers evaluated th e top dietary so urces of energy, solid fats, a nd added sugars among 2 -18 year old s in the United States. Top food sources for these 10 three were grain desserts, pizza and sugar sweetened beverages including soda and fruit drinks. The same s tudy reported that 40% of total en ergy consumption was from empty calories implying that poor diet quality was high in energy -dense foods and low in nutrient -dense foods. Only a few studies have examined children™s diet quality in combination with paren tal feeding strategi es among low income fam ilies. Murashima et al. (2012) and Hoerr et al. (2009) studied low income families and found that diet quality of Head Start children ages 3 -5 years with highly controlling parents was better than when the childre n had parents who used indulgent feeding pr actices. They also found that non directive feeding control (such as praising and encouraging to eat) was associated with heathier diet intake of the preschoolers. A cross -sectional study of low socio - economic status families in Israel with slightly older children age 5 -6 years old found that encouragement and parental modeling of healthy eating were associated with an increase of fruits and vegetables, reflecting a better diet quality (Entin et al.; 2008). Par ental food restriction was associated with consumption of low nutrient -density foods like sweets. Improving the diet quality of children from low income families is very important and relevant to this study on the effectiveness of the EH curriculum. Obes ity Childhood obesity rates in the United States have tripled in the last three decades. Approximately 32% of children and adolescents ages 2 -19 years are overweight or obese, with 17% of children obese (Ogde n et al., 2012 & Ogden et al., 2014). Today nearly one in three children is overweight or obese. In 2009 -2010, the prevalence of obesity was 12.1% among US children 2 -5 years of age with especially high rates among African Americans (18.9%) and Hispanics (16.2%) (Ogden et al., 2012). In Michi gan , the numbers decreased from 13.9% in 11 2009 to 13.3% in 2011 (May et al., 2013). Women and children in lo w income families are experienc ing higher rates of obesity and chronic diseases compared to higher socio -economic groups (Williams et al., 2008). In general, low socioeconomic status is related to high rates of obesity (Darman & Drewnowski, 2008). To help decrease the obesity among children, some investigators recommend addressing parent feeding strategies and the home food environment (Scaglioni & Sal vioni, 2008). Home Environment Parents provide environments for their children™s early experiences with food and eating (Birch & Davison, 2001). The home food environment includes two factors. First, is the physica l environment such as availability and accessibility of food. The second is the behavioral environment such as self -efficacy to change, self -regulation abilities, and feeding strategies parents use with their children (Martin -Biggers et al., 2014). Many studies have linked home environment to child obesity via provision of an unlimited, convenient supply of energy dense foods coupled with low levels of physical activity (Bryant & Stevens, 2006). This link between the home food environment and child obesity has been of such great concern tha t the Food and Nutrition Service, United States Department of Agriculture heavily promoted a campaign to improve the home food environment and provide tips on how to increase the intake of nutritious foods. The key messages included: 1) increase fat free milk and low fat milk and milk products, 2) increa se intake of whole grains; and 3) improve child feeding practices. The eating of family meals encourages parents to discuss with their children the benefits of each food item, for example milk builds bone s and muscles (www.fns.usda.gov/tipsformoms.htm). 12 Anderson and Whitaker (2010) found t hat preschool children in the US exposed to three routines had approximately 40% lower pre valence of obesity than those who were exposed to none of the three routines. These routines included regularly eating the evening meals as a family, adequate night time sleep, and having limited TV viewing time. This study supports the importance of the home food environment as a parent feeding strategy to develop a young healthy eater, which is an interest being addressed in this study. The EH intervention aimed to improve the home environment by increasing the availability of fruits and vegetables and by limiting children™s access to sweets, candy and sweetened beverage s. Parent Feeding Strategies Versus Parenting Styles or Feeding Styles Research on the effect of parent behaviors on their children™s food intake and weight status has been confounded somewhat by use of different terms. As described in the glossary, par enting styles are considered to be very stable characteristics that refer to overall parenting as measured by standardized instrument (Baumrind, 1989). Nearly a decade ago, Rhee et al. (2006) found that children of permissive parents had increased obesity . Since then child development researcher Sheryl Hughes validated a feeding specific instrument for parent feeding styles . Its use in several studies have resulted in consistently finding parents with permissive feeding styles (indulgent or non -involved) to have children with the poorest diet quality intake of nutrient dense foods such as fruits and vegetables, 100% juice and dairy (Hoerr, et al., 2009; Couch et al., 2014) and who were the most obese (Hughes, 2006; Couch et al., 2014; Frankel et al., 2014) . Frankel and colleagues further found that the relationship between permissive feeding style and high BMI -Z score to be mediated by the child™s lessened ability to self -regulate around food. These findings of permissive feeding styles and BMI Z -scores i n low income families are in contrast to 13 those of Birch et al. (1999) with middle income parents. Birch found the authoritarian parent feeding style was associated with a higher BMI for the children. If parenting styles or parent feeding styles are cons idered stable, however, then the focus of educators is best spent on parents™ feeding strategies that have potential for change. Studies suggest that parent feeding strategies have a great effect on children™s weight status and diet quality (Birch et al., 2001, Hughes et al. 2008). Child feeding strategies determine the availability of various foods and beverages, the portion sizes that children are offered, the frequency of eating occasions, and the social context in which eating occurs (Ogden et al., 200 6). Children 5 year s of age and younger start to learn when, where and how to eat ( Lent et al., 2012; Savage et al., 2007). Child feeding strategies are considered behaviors that parents use to get children to do so mething specific. Control in child f eeding is defined as fistrategies that the parent performs for the child to achieve healthy eating or consume the recommended amounts of nutrient dense foods and limited amounts of energy dense foodsfl (Hughes et al., 2006; Murashima et al., 2012). Control for child feeding strategies can be further divided into directive control and non -directive control . Directive feeding control is when parents put external pressure on the child to eat a healthy diet. Directive feeding control can also be divided into high control , where parents verbally, psychologically and physically pressure the child to eat and high contingency, where the parents threaten or reward the child to eat. Some examples are press ure to eat, monitoring, rewards/ threats and food restriction (O gden et al., 2006; Hughes et al. 2008; Megumi et al. 2012). Non - directive feeding control is when parents interact with the child to motivate him/ her to eat a healthy diet by internalizing the goal. Some examples are encouraging, complementing, modeling and reasoning (Hughes et al. 2008; Megumi et al. 2012). For non -14 directive feeding control , two sub -constructs are in this category: child centered feeding , such as rearranging foods to make them interesting, including healthy foods the child enjoys and complementing the child when she/he eats. The second is food environmental controls which are strategies where parents provide a healthy and organized home food environment with family rules around eating to help the child eat a healthy diet. This can be f urther divided into: food availability, where parents keep or do not keep certain types of foods in the house (Brown et al., 2008); mealtime behaviors, where parents set rules during meals such as sitting at a table, eating together and not viewing TV duri ng meals (Hoerr et al., 2005); and timing of the meals, where parents set regular meal and snack times for the child and family (Baughcum et al., 2001). Research on Parent Feeding Strategies Relating to Children™s Diet Quality and Weight Status Most, but not all, studies have shown that food restriction, a highly controlled feeding strategy, to be associated with children being overweight (Clark, 2007; Fisher, 1999; Thompson et al. 2009). Powers et al. (2006) studied 296 low income African America n women with preschool children. Among these low -income African Americans, there was a positive association between maternal food restriction and control in feeding and their preschoolers' BMI, but this was limited to obese mothers. On the other hand, pa rents with permissive feeding practices had children with low intake of nutritionally balanced foods like fruits, vegetables and low fat dairy (Hoerr et al., 2009). Entin et al. (2008) studied low socio - economic status families in Israel with children ag ed 5 -6 years and found food restriction to be associated with consumption of low density food such as sweets and junk food. Campbell et al. (2006) studied 5-6 year -old children in Australia and found increased TV viewin g time associated with children s™ i ncreased energy intake, increased sweet snack and high -energy drink consumption, 15 and deceased vegetable intake. Two studies found parent™s modeling of fruits and vegetables associated with their children™s intake of these foods. While this study did not investigate food restriction, it did examine high feeding control and high contingency feeding control as well as the other feeding strategies mentioned. Intervention Studies to Change Parent Feeding Strategies and Affect Child Weight and Food Intake So far, nine studies have been located that attempted to change parent feeding strategies in order to reduce risk for child obesity as shown in Table 3 . Each will be briefly summarized here. Harvey -Berino and Rourke (2003) studied overweight and obese Native American moms and their children ages 9 -36 months. The intervention included the delivery of 11 home lessons in 16 weeks. One of the lessons compared the restrictive vs permissive parental feeding strategies. The intervention group of parents reduce d their food restriction significantly compared to the control group, the weight -for -height Z -score also declined for the intervention group compared to those in the control group. This was a small study that had a control group, but no follow -up measures . Healthy Children, Healthy Families (HCHF): Parents Making a Difference! provided an eight week lesson series of 90 -minute sessions through EFNEP in New York State (Dickens et al., 2014; Lent et al., 2012). Participants were low income families (at 185% of poverty level or below) and their children age 3 -11 years old. Four categories of positive parenting practices to promote healthful eating and activity at home were chosen to assess parent behavior change: (1) demonstrating role modeling; (2) hel ping children feel good about themselves; (3) offering choices with limits; and (4) shaping home environment. Mean scores for parent behaviors 16 improved significantly and improvements were also seen in consumption of low fat dairy, fewer sweet beverages, l ess screen time, increased physical activity, increased family meals and increased fruit and vegetable intake. This study was notable for addressing low income families and having a large sample, but it had no control group or weight measures. Most recent ly, Skouteris et al., (2015), intervened with 201 middle income parent -child dyads of 2 -4 year old s with 10 weekly 90 minute workshops for small groups of 6 -10 child -parent dyads. They found significant positive group effects for increased vegetable and re duced snack food intakes, and child satiety immediately post intervention . At the 12 months follow -up, intervention children exhibited less neo -phobia than controls, but there was no change in we ight status and dietary changes were not maintained at follow -up. Fletcher et al., (2013) in New South Wales, Australia, studied 394 middle income parents of pre -schooners ages 3 -5 years. The intervention comprised four 30 -minute weekly phone calls for one month stressing parent modeling, the food environme nt and supportive food routines . Energy dense food like sweets, ice -cream and candy were assessed as was the home food environment and parental pressure to eat. The authors concluded that there was a significant decrease in the consumption of energy -dense foods in the in tervention group at the 2 month follow -up compared to the control group, but the difference lost significan ce at the 6 month follow -up. There were no weight measures taken. A pilot intervention study by Horodynski and Stommel (2005) was conducte d with 96 low income families of Early Head Start children age 11 -36 months. Four lessons, 90 -minutes each, were offered over six months stressing parent food modeling, introducing new foods and parenting skills. One of several instrumen ts used was the Chi ld Parent Mealtime Behavior 17 Questionnaire (C PMBQ). The self -regulat ory behavior of toddlers and their parents™ knowledge of child feeding improved in the intervention group vs the control group. Weight status was no measured, but one day diet recalls were conducted. There was no follow -up with this small pilot study. West et al. (2010) conducted an intervention study in Australia to improve parent feeding practices for prevention of obesity among 4 -11 year old children in white and mostly middle income fami lies. The 12 -week intervention consisted of 90 -minute group sessions and three 20 minute telephone sessions. Nutrit ion strategies, positive parenting strategies and physical activity strategies were taught to the parents in the intervention group. The 12 -week intervention was associated with significant improvement in children weight status and weight related problem beh avior. Diet was not assessed, however, and only 31 families participated in the one year follow -up. Ostbye et al. (2012) conducted an int ervention study targeting 400 obese women and their children 2 -5 years of age in North Carolina. The intervention group received eight monthly mailed interactive kits which were followed each month by 20 -30 minute phone calls. The kits provided and emphasi zed the following: (1) an authoritative parent style; (2) routines for sleep and meal times; (3) a supportive home environment; (4) role modeling of healthy eating and physical activity; and (5) improvement of feeding practices. This study resulted in a s ignificant improvement in feeding practices, maternal dietary intake, reduction in sweet beverage intake, increased consumption of fruits and vegetables, as well as fewer dinners and snacks eaten in front of the TV. The outcomes of the intervention study by Ostbye et al. are promising for decreasing obesity by improving the parent strategies. 18 Tabak et al. (2012) mailed four newsletters and had two goal setting phone calls to 43 parents of 2 -5 year olds in middle income families . Newsletters focused on he althy food availability, especially vegetables, offering food choice within limits, role modeling and encouraging vegetables. Compared to those in the control group, intervention parents reported increased availability of vegetables in the home, offering more fruits and vegetables as snacks and improved self -efficacy for managing their children™s food and activity behaviors. No weights were measured and there was no follow -up. To summarize , the intervention studies by Skouteris et al. (2015), Tabak et al. (2010) and Dickens et al. (2014) demonstrated improved diet quality and increased fruits and vegetable consumption of children . Likewsie, studies by Ostybe et al. (2012), Dickens et al. (2014) and Harvey -Beniro et al. (2003 ) showed some improved parent feed ing strategies. West et al. (2010) and Harvey -Beniro et al. (2003) demonstrated short term improvement in children weight status , but the study by West and colleagues was with middle income families in Australia and the Harvey -Ben iro study was with low income families and had no follow -up data . Fletcher et al . (2013), Dickens et al. (2014 ), Skouteris et al . (2015 ) demonstrated decreased use of sweets, snacks and sweet beverage. None of these studies used materials primarily targeting parent feeding strategies or instruments as comprehensive as the PFBQ. 19 Table 3 . Summary of Intervention Studies to Change Parent Feeding Strategies and Affect Chil d Weight and Food Intake to Address Child Obesity in Young Children over Last Decade Authors Participants Intervention Outcomes Comments Skouteris et al., 2015 201 mid -income parent child dyads, 2 -4 yo. New Zealand 10 weekly 90 min. workshop Groups of 6 -10 child -parent dyads Increased intake of vegetables and less snacks for intervention group compared to control group No wt change Assessments conducted at baseline, post intervention and 6 and 12 mo. Assessed food intak e and wt Dickens et al., 2014 500 EFNEP parents of children 3 -11 yo New York 2/3 Latino Healthy Children, Healthy Families, Parents Making a Difference 8 90-min lessons in small groups for 8 wk Role modeling Promoted Child self -esteem Food choices within limits Shaping th e food environment Improved scores for parents and children in: low fat dairy Less sweet beverage Less screen time Increase physical activity Increased family meals Increased FV intake Large sample Pre vs 3 mo FU data, but no control group No wt data Ava il in Spanish 20 Table 3 (cont'd) Fletcher et al., 2013 394 mid -income parents of children 3 -5 yo, Australia Four 30 -min weekly phone Focus on food modeling, meal structure and food environment Assessed energy dense foods Decrease in energy -dense foods by intervention group at the 2 month follow -up, but not significant at the 6 month FU. Wt not measured 2 mo. and 6 mo. follow up one month duration of intervention No wt measures Ostybe et al., 2012 400 White mid -income obese women and children 2 -5 yo, in North Carolina 8 lessons mailed monthly followed by 20 min phone calls Focu sed on authoritative parenting, Healthy food environment, Role modeling, Feeding practices No change in BMI Z score Mothers used less food rewards Reduced emotional feeding Children spent less time in front of TV More healthy foods in the home Large sam ple Had Intervention and control groups Tabak et al.,2012 White middle income in St. Louis, MO 43 parent child dyads, 2-5 yo 2 phone calls and 4 newsletters over 4 month Newsletters topics: Food availability, choice within limits, Making vegetables accessible, Snack preparation Increased availability of vegetable and vegetable offering for snacks in the intervention group No wt measures RD conducted the phone call No follow -up and small sample 21 Table 3 (cont'd) West et al., 2010 101 White mid - income parents of OW/obese 4 -11 yo in Australia 12 wk of 90 min group sessio ns 3 phone calls (20 min) Focused on modeling healthy eating and offering healthy Choices within limits Decreased BMI Z -scores Increased parents™ self -efficacy of managing their children weight related behavior Diet no assessed Control and intervention groups Only 31 in 1 year FU measures Hordynski et al., 2005 43 Early Head Start mothers of children 6 - 30mo in rural Michigan. Nutrition Education Aimed at Toddlers (NEAT) 4 lessons, 90 min each for 4 mo. Focus on parent modeling and introducing new foods Improved feeding self -regulation of toddlers in the intervention group more than the control group Improved parent knowledge about child feeding No change in wt or diet Intervention and control groups Small sample Low income, rural sample of diver se race -ethnicity Harvey -Beniro et al., 2003 43 obese, Native American mothers of children 9-36 mo 11 home lessons in 16 hr/wk Decreased Ht and Wt. Z -scores Decreased energy intake in children. Parental food restriction declined in Rx group. Control a nd intervention groups No follow -up 22 CHAPTER III. METHODS Study Design This study was a secondary data analysis. To collect the data, a quasi -experimental design was used wherein parents who volunteered to participate in a 6 -week interv ention plus a three month follow -up were randomly assigned to the intervent ion group or to a control group that received materials on general health. At the end of the first six -week intervention, an educator taught the EH program to the control group, making the study a crossover in design. Both groups re ceived pre and post -tests as well as three -month follow -up assessments. See Table 4 for study design. Table 4 . A Quasi -Experimental Study Design and Group Comparisons. Thin brackets show effect of EH compared to matched control gro up. Thick brack ets show EH program effect at 3 month follow -up for both groups and without a control group. Week Intervention group Control Group 0 Pretest Pretest 1-7 Eat Healthy Intervention Booklet on general health tips 8 Post -test Post -test 9-15 NA Eat Healthy Intervention 16 NA Post -test 20 Follow -up NA 32 Follow -up 23 Sample and Recruitment Beginning in April 2013, fami ly nutrition educators recruited SNAP -Ed eligible families living within four Michigan counties -Genesee, Kent, Ingham and Van Buren Šall in t he Lower Peninsula and within a 90 minute drive from Michigan State University. Eligibility criteria included being a parent or guardian over the age of 18 years , literate in English, with a 2½ to 5 year old child living at home who did not have a health problem that interfered with eating. Parents were recruited from Head Start in Ingham County and from hom e visiting programs for preschool ers via Intermediate School Districts in the oth er counties using home educators in the other counties . Procedures Upon Human Subjects app roval for this project from Michigan State University, MSU staff conducted two trainings for educators in the four counties: (See MSU Institutional Review (IRB) letter in Appendix C). All 22 educators completed the MSU Institutional Review training. When parents we re recruited they signed a consent form (Appendix B) and the educators measured their height and weight as well as that of their target child. Parents completed a 39 -item Block Kids Food Screener (BKFS) for their child™s food intake over the pa st week, a Likert -scaled questionnaire about their child feeding behaviors (29 items), and a demographic form. Then , the outside evaluator randomly assigned parents to the intervention group (n=81) or the control group (n=77) . At the post -tes t, educators collected feeding data from parents again using the PFBQ . At the three -month follow -up educators collected another BKFS, PBFQ, as well as heights 24 and weights from both the parent and child. This was do ne for both groups as shown in Table 5 . At the follow -up parents received $25 gift certificates to a local food store. Table 5 . Types of Data Collected for Each Group over Time of the Lesson Delivery and Follow -Up a Height and weight of pare nt and child. b Block Kids Food Screener, for frequency of FV, Sweet beverage, etc. c Parent Feeding Behavior Questionnaire, for high control, high contingency, availability of healthy foods, availability of sweets, meal ti me behaviors healthy modeling. Ht and Wt: Height and weight for BMI for adults and BMI percentiles for children. Group Baseline Wk 0 Wk 1 -7 Wk 8 Wk9 -14 Follow - up ( 3 months) Intervention (Rx) Ht and Wt a BKFS b PFBQ c Rx. starts PFBQ Post - test Ht and Wt BKFS PFBQ Wk 1 2 Control Ht and Wt BKFS PFBQ PFBQ Post - test RX starts Ht and Wt BKFS PFBQ Wk27 Intervention The EH intervention consisted of five lessons (one booklet each) and 2 Œ3 minute videos delivered to parents of preschool chil dren, 2½ -5 years. Paraprofessional educators (n=22) taught EH to pare nts in four Michigan counties using a combination of three home visits and three phone calls alternated every other week. The lessons topics were: 1) kids are what they eat; 2) be a g ood role model; 3) ways to praise at meals; 4) making mealtime family time and fun; and 5) learning to eat healthy. The feeding strategies targeted were those associated most with normal weight status of the child and optimal food behaviors. These strateg ies included availability of healthy foods, parent modeling of healthy foods, encouragement to try new foods, 25 mealtime structure, and timing of meals and snacks. The disadvantages of negative feeding strategies such as highly controlling feeding behaviors, using contingency (reward and punishment), availability of unhealthy foods, and parenting modeling of unhealthy foods were also taught. After a 6 -week intervention period with the first intervention group, 90 participants remained and were assessed for parent feeding behaviors at post -test and follow -up. After the post -test, educators also delivered the intervention to parents in the control group. Three experts in pediatric nutrition and two in developmental psychology reviewed all items for content validity according to the main constructs. Face validity was conducted wi th parents. The PFBQ was administered at baseline, post -test, and at a 3 -month follow -up. See Table 6 for a description of each lesson with teaching objectives and list of activities an d video clips for each one. 26 Tab le 6. Curriculum Grid for Eat Healthy by Lesson, Concepts, Objectives and Activities Topic 1a visit 1b phone 2 visit 3 phone 4 phone 5 visit Title Kids are what they eat Be a good role model Ways to praise at meals Making mealtime fun Learning to e at healthy Key Concepts Keep healthy foods in the home visible and available Portion sizes for preschoolers are smaller than for adults Be a good role model with food and drinks New foods can take time Labeled praise helps your child understand what he /she does right Make positive family mealtimes a priority Family meals benefit everyone Use mealtime rules to reduce struggles Teaching Objectives Learn healthy vs unhealthy food choices. Explain why water is a better beverage choice than juice and fruit favored drinks. Recall what make the food healthy. List ways to make healthy fruits and vegetables snacks visible and accessible for children. List the number of servings of fruits ,vegetables and low fat dairy foods that a 4 years old need each day List ways that you model healthy food eating for your child. Provide things you will say to influence your child eating behaviors. Identify how what you say influence your child™s eating behavior Recall the kinds of praise you give to you child for eating. Des cribe a child act where you can praise it. Give an example of child™s behavior that you can ignore Describe why preschooler likes rules. List 3 reasons why food should not be used as a reword List 3 benefits of letting your child help with food preparati on. List 3 ways to make mealtime enjoyable for your family List 3 strategies to encourage your child to learn to eat without pressure. List rules your family has for meal time. Explain how setting some rules at meal time reduce mealtime struggle 27 Topic 1a Topic 1a Topic 1a Topic Anchor Clip 1.1 Discuss the types of foods that parent keeps available Parent does food inventory pp4 -5 Clip 1.4 Discuss changes parents have seen in child™s appetite Clip 2.1 Discuss who modeled food habits for the parent Clip 3.1 Discuss how the parent encourages and praises their child eating Clip 4.1 Discuss typical food commands that parents use Clip 5.1 Discuss food struggles the parent has with their child Discuss when parent got their child to eat without press ure Add Healthy foods are fianytime foods,fl but fiSometimesfl are only for now and then, p 5,8 F/V don™t need to be fresh; frozen and canned are good. Be sure to rinse canned veggies to remove added sodium. Children need less food than adults, p19 - 20 Grow th spurts make appetite erratic Eating behavior milestones, p23 Everything a parent does is a lesson for their child It can take up to 15 tastes to learn to like a new food Praise the action, not the person. Be specific with praise Positive feedi ng pattern allows children choice within structure Indulgent/uninvolved feeding leads to poor diet and weight problems p3 -5 Reward children with attention and family activities, not food The 90 minute rule between meals & snacks can reduce struggles Karp S. Jamie, Process Evaluation of Implementing The Eat Healthy, Your Kids are Watching. A Parent‚s Guide to Raising a Heal thy Eater Program. East Lansing, MI: Department of Food Science and Human Nutrition, Michigan State University, 2014 Table 6 (cont'd) 28 1a visit 1b phone 2 visit 3 phone 4 phone 5 visit Title Kids are what they eat Be a good role model Ways to praise at meals Making mealtime fun Learning to eat healthy Away Parent switches a fisometimesfl food for an fianytimefl food, p7 Parent waits 90 minutes between meals & snacks Parent serves only small portions at first, p22 Parent let™s child pick a new fruit or vegetable at the store and tastes it, p10-11 Parent keeps track of labeled praise they use for next 2 -3 meals, p4 Parents choose two non -food rewards to use the next week, p11 Parent plays a sensory game with child to en courage tasting new food p6 Child Activity Child selects photos of healthy snack choices that can substitute for sometimes foods, p14 Child chooses a new fruit or vegetable to try. Child places a super taster sticker in activity book after trying it, p11 Parent enhances praise of child with touch, eye contact and smiles, p15 Parent and child choose and do a fun mealtime activity, p13 Parent helps child pick mealtime rules and track for several days, p13 Handouts EH binder; Topics 1 & 2 Topics 3,4,5 Nutrition Education Reinforcing Incentive Eat Healthy Magnets Supertast er stickers Healthy Snacks recipes or Fruit/Veg playing cards, MSU extension recipe book Karp S. Jamie, Process Evaluation of Implementing The Eat Healthy, Your Kids are Watching. A Parent‚s Guide to Raising a Heal thy Eater Program. East Lansing, MI: Department of Food Science and Human Nutrition, Michigan State University, 2014 Table 6 (cont'd) 29 Instruments, Measurements and Variables Demographi c Data. Demographic dat a included the children™s and parents™ gender, age and race -ethnicity. In addition, the parents reported their educational attainment, current relationship, living arrangement, employment, pregnancy, breastfeeding, transportation fo r and frequency of grocery shopping, and participation in Supplemental Nutrition Assistance Program (SNAP) and SNAP -Education (SNAP -Ed). Anthropometrics. A trained staff member measured the height and weight of each child and parent twice following standar d procedures as stated by Lohman et al., 1988. The height was measured to the closest 0.1 cm using a portable stadiometer (SECA 214, Seca corp., Hanover, MD). Weight was measured to the closest 0.2kg using a digital scale accurate to 200 kg (BWB00AS, Tanit a, Tokyo, Japan). The Body Mass Index (BMI) was calculated using the equation = weight (kg)/ height (m) 2. For childr en, percentile by age and gender specific BMI was obtained using CDC growth charts (www.cdc.org). For all measurements, the average was calculated before calculating the child™s BMI percentile or the parent™s BMI. Block Kids Food Screener (BKFS) The BKFS is a food frequency questionnaire for the child™s food intakes. Mothers reported the foods that children ate within the past week using the BKFS for children aged 2 Œ6 (Nutrition Quest Inc., Berkeley, CA). This food screener is a 39 -food item qu estionnaire developed from a validated 80 -item food frequency questionnaire to assess food and nutrient intakes in children 2 Œ17 years old (Block, 2008; Cullen, Watson, & Zakeri, 200 8). Of the 39 foods and beverage items, eight nutrient -dense diet variables and four energy -dense diet variables were selected for data analysis. Nutrient -dense foods were those that provided 30 substantial amounts of vitami ns and minerals and relatively few c alories, i.e., fruits without juice, non-fried vegetables without potatoes, whole grains, and low fat milk in frequency per week and as cup equivalents per day (American Academy of Pediatrics, 2001; Barlow, 2007 ). En ergy -dense foods were those that con tained greater than 25% of the food energy from added sugars, and/or greater than 35%of the food energy from fat per serving based on USDA™s food and nutrient database, i.e., sweets (ice cream, candy, cookies), sweet bev erages, chips/popcorn, and high fat milk (http://www.nal.usda.gov/US Department of Agriculture, national agricultural library). The units of measure for each indicator were: 1 cup equivalency= 8 fl oz and frequency consumed per week = 0 -7 times. See Tabl e 7 for a list of the sentinel indicators used separately to indicate diet quality. Note that some variables can be interpreted as indicators either nutrient density or energy densit y.31 Table 7. Categories and Indicators of Diet Quality fro m BKFS Category Variable Indicators Fruit 1.Total Fruit with juice freq. per wk 2. Fruit without juice (CE) per day 3. Fruit without juice frequency per wk Nutrient dense Nutrient dense Nutrient dense Vegetables 1. Vegetable freq. per week 2. Vegetabl e cup equivalency per day 3. No. different vegetables per wk Nutrient dense Nutrient dense Nutrient dense Milk 1. Milk freq. per week 2. Milk cup equivalency per day 3. Percentage drinking high fat milk 4. Percentage drinking low fat milk Energy dense Nutrient dense Sweets beverage 1. Sweet beverage frequency per wk 2. Sweet beverage (CE) per day Energy dense Energy dense Sweet and chips snacks 1. frequency per day 2. frequency per week Energy dense Energy dense Whole grain and fiber 1. Whole wheat bread frequ. per wk 2. Whole wheat bread amount per wk 3. Average daily grams of fiber 4. Percentage eating whole grain cereal 5. Percentage eating non -whole grain cereal Nutrient dense Nutrient den se Energy 1. Average daily Kcal 2. Diet energy density = grams food/kcal per day 3. Average daily grams fat 4. Average daily grams sugar CE=Cup equivalency . Freq.= Frequency 32 Parent Feeding Behavior Questionnaire (PFBQ) The PFBQ lists the strategie s that the parent used to encourage children to eat. It was administered at ba seline, post -test , and at the 3 -month follow -up. Items were selected from other instruments designed to measure feeding strategies which had shown significant associations with children™s food intakes. Most of the original instruments from which the items were selected had been tested for vali dity and reliability (Baughcum et al., 2001; Brown et al., 2008; Hughes et al., 2006a, 2006b; Spurrier et al., 2008; van der Horst et al., 2007). Item selection was based on the association with the three main constructs Šdirective control, non -directive co ntrol , and food environmental control. Each construct included two or three sub -constructs (See PFBQ by construct -Appendix A). A five -point Likert scaled response category (never = 1 to always = 5) was used for each item. Items measuring undesirable behavi ors (e.g., keeping sweets and salty snacks in the home) were reverse - scored. High scores indicated a more positive behavior for each const ruct (Murashima et al., 2012; Reznar et al., 2014). The PFBQ , consist ing of 30 items (Q) , was divided into the following sub constructs: high control (3Q), high contingency (4Q), availabilit y of healthy foods (2Q) , Unhealthy modeling (2Q), meal time behaviors (6Q), timing of meals (2Q), child centered Œnon directive (6Q) and food modeling non directive (4Q), overweight (1Q). Questions 1, 2, 3, 4, 5, 6, 7, 9, 10, 12, 14, 15, 19, 28, and 29 were reverse -score d. These items comprised six multi -items and two single -item constructs as follows, with the first four considered negative feeding strategies and the last four considered positive feeding strategies. Examples of items: High Control feeding I beg my child to eat dinnerfl ; High Contingency feeding will take a food away if the child doesn™t eat , for example, fiIf yo u don™t finish your vegetables, 33 you won™t get dessertfl; Permissive feeding time= allow my child to eat whenever he or she is hungryfl; Indulgent mealtime behavior= parents w my child to play and watch TV during mealsfl; Healthy Availability and Modeling= parents scored 4 for five items such as, fiI keep sweets, candy or salty snacks where my child can reach themfl (reverse scored); Child -centered feeding = parents scored 4 fo r six items such as, fiI say something positive about the foodfl; Milk modeling= parents scored 4 for one item, fiI drink milk in front of my childfl; Fruit and vegetable modeling= parent scored 4 for one item, fiI eat fruits and vegetables in front of my chi ld.fl Statistical Analysis Data were cleaned before analysis. For this analysis, SPSS version 20, 2013, Armonk, NY: IBM Corp was used. Descriptive data analysis of demographic data and sample characteristics was conducted using mean and standard deviation. The data on diet quality, parent feeding strategies and preschoolers™ BMI percentiles were all checked for skewedness and kurtosis. If the skewedness is between -1.0 and 1.0 and the kurtosis between -1.0 and 2.0, the assumption is that the var iable is normally distributed (Rosner, 1995). All variables were normally distributed so none needed to be transformed. Paired T-Test analysis was used to compare control and intervention groups in parental feeding behavior and food dietary quality at the base line. Pearson correla tion matrix was constructed to test association s between BMI and parental feeding behavior, BMI and food dietary quality, and between dietary quality and parental feeding behaviors. Descriptive statistics, including means and freq uencies, were reported for demographics and weight of the parents and their children based on the baseline 34 dataset. Next, descriptive statistics and Cronbach™s Alpha (an index to measure reliability) were examined for parental feeding behavior. Research Question 1 was analyzed three ways. First, Pearson correlations were run between the parent feeding behavior strategy sub constructs and the preschoolers™ BMI percentile. Then the BMI percentiles were categorized into two groups within one variab le. Preschoolers whose BM I percentile was normal weight or underweight codes as zero and those who were overweight or obese were coded as 1. Finally Bi -serial correlations were run for these variables. Two types of regression models were conducted in exa mining effects of each measure of parental feeding behavior on child BMI separately: 1) to use general linear regression to regress child BMI percentile on each measure of parental feeding behavior; and 2) to use logistic regression to regress child weight status (underweight/normal weight versus overweight/obesity) on each measure of parental feeding behaviors. For both types of the models, parental age and BMI were controlled. The general linear regression and the logistic regression model are represented by the following equations for each measure of parental feeding behavior (FB), respectively: (1) (2) In Equation 2, p represents the probability of children with overweight/obesity. For Research Question 2 , a series of Gener alized Estimating Equation (GEE) models were conducted to evaluate whether changes from baseline to post -test in variables of parental feeding behavior were different between the control and intervention groups. Each parental 35 feeding behavior was predicted by three variables: group (0=control and 1=treatment), time (0=baseline and 1 = post -test), and interaction of group and time. The models are represented by the following equations for each measure of parental feeding behavior (FB): FBi = o+ 1Group i+ 2Time i+ 3Group i×Time i (3) Significant effects of interaction of group and time indicated that the control and intervention groups were different in the change from baseline to post -test in parental feeding behavior. The study was also inter ested in testing whether the change from baseline to post -test within each study group for each parental feeding behavior was significant. To that end, within each model, Z-tests were conducted to test differences in the marginal means (estimated from t he GEE model) between the baseline and the post -test for the control and treatment groups separately. For each model, the p values based on the Bonferroni adjustment were reported. The analysis plan for Research Question 3 was analyzed similar to the plan for Research Question 2. Specifically, a series of GEE models were conducted to evaluate whether there were significant changes from baseline to follow -up test in variables of parental feeding behavior and diet quality for the whole sample (including the c ontrol and intervention groups). Each parental feeding behavior or diet quality was predicted by time (0=baseline and 1 = follow -up test). The models are represented by the following equations for each measure of parental feeding behavior (PFB) and diet qu ality (DQ): FBi = o+ 1Time i (4) DQi = o+ 1Time i (5) 36 Significant effect of time indicated that there were the changes from baseline to follow -up test in parental feeding behavior and diet quality for the whole sample. 37 CHAPTER IV. RE SULTS Demographics Descriptive statistics for demographics and weight status for parents and children are shown in Table 8. A majority of parents were white (75.9%), and female (92.4 %), and the biological parents (91.1%). Close to half of the parents (48.7%) were obese with an average age of 32 years. Over half of the parents received SNAP and WIC. The child sample was composed of almost equal proportion s of males to females. The majority of the children were white (77.2%) and over half (55.1%) had nor mal weight. Table 8. Demographics and Weight Status for Parents and Children at Baseline Total N = 158 Parent Child Characteristics Mean±SD or % Mean±SD or % Age(yr) 32.02±8.29 3.50±.83 Gender , Female 92.40 49.40 Race/Ethnicity a White 75.90 77.20 Black 14.60 16.50 Hispanic 13.30 18.40 Mixed/others 5.00 3.20 Weight Status Underweight 1.30 1.90 Normal Weight 27.80 55.10 Overweight 19.60 19.00 Obese 48.70 20.30 Education Level Less than High School 8.2 High School 24.1 Any post high scho ol 67.1 Employment Part -time 20.9 Full -time 15.2 SNAP 55.1 SNAP -Ed 7.6 WIC 63.3 EFNEP 3.2 38 aSum of percentage was larger than 100% , because part icipant could be Hispanic and any race. Table 9 shows the descri ptive data from the PFBQ. An average parent had high scores for controlling feeding strategies, mealtime behaviors and modeling fruits and vegetables . Due to high scores indicating positive feeding strategies, these parents were not high ly controlling , their overwei ght concerns for their child were very low. Table 9. Descriptive Statistics and Internal Reliability of Parental Feeding Strategies by Construct at Baseline Construct N Mean SD Cronbach Alpha High Control 158 4.21 .74 .68 High Contingency 158 3.88 .87 .84 Mealtime Behavior 158 4.14 .64 .70 Healthy Food Availability 158 3.79 .72 .68 Timing of Meals 158 3.37 .77 .61 Child Centered 154 3.53 .68 .71 Concern for Child Overweight 154 1.30 .87 N/A Fruit/vege table Modeling 153 4.48 .72 N/A Milk Modeling 154 3.75 1.41 N/A Not e: Concern for child overweight, fruit/vegetable modeling, and milk modeling did not have reliability reported because they only had a single item each. A five -point Likert scaled re sponse category (never = 1 to always = 5) was used for each item. Higher scores indicated more positive behavior in each construct . The descriptive data for the chi ld™s food intake from the KBFS are in Table 10. The average amount of fruit without juice per day, the number of d ifferent vegetables per week, the percentage drinking low fat milk, and the pe rcentage eating whole grain cereal scored high, 39 suggested good diet quality for these indicators. The overall diet quality was low, however, as indicate d by high sweet be verage frequency per week and per day, and frequent intake of sweets and snack chips per week. The cups of vegetables per day was low compared to the DGA recommendation of 1 -1 ½ /day (DGA, 2010) . The average intake s of both milk and dieta ry fiber were low. Table 10. Descriptive of Child™s Food Intake from the KBFS Measured at Baseline Shown by Group at Baseline Food Item Total Intervention Control N=158 N=81 N=77 Mean± SD (%) Mean± SD (%) Mean± SD (%) Fruit with juice frequency/wk 3.4±1.4 3.5±1.4 3.2±1.3 Fruit with juice cup equiv/day 1.7±.8 1.7±.9 1.5±.8 Fruit without juice frequency/wk 3.2±1.5 3.4±1.6 2.9±1.4 Fruit without juice cup equiv/day 1.1±.6 1.2*±.7 .9*±.6 Vegetable freq/wk 1.4±.9 1.5±.9 1.4±.9 Veg etable cup/day .6±.5 .7±.6 .6±.5 No. different veg/wk 4.1±1.7 4.4±1.8 3.8±1.6 Milk freq/wk 5.6±2.1 5.7±2.0 5.5±2.2 Milk cup/day 1.6±.9 1.6±.9 1.5±.8 % Low fat milk 72.8 79.0 66.2 Sw Beverages freq/wk .8±1.3 .9±1.5 .7±1.2 Sw Bevarages per day .1±.2 .1±.21 .1±.2 Sweets (ice cream + candy + cookies) freq/wk 1.5±1.2 1.5±1.3 1.4±1.1 Snack & chips freq/wk 1.55±1.4 1.7±1.6 1.4±1.2 Whole Wheat bread freq/wk 3.07±2.4 3.3±2.3 2.8±2.4 Whole Wheat bread Amt/time eaten 1.8±.6 1.8±.5 1.8±.6 Whole wheat 1oz eq uiv/day .6±.5 .7±.5 .5±.4 Ave daily fiber in gm 11.0±6.2 11.9±6.8 10.1±5.5 % eating Whole grain cereal 61.4 63.0 59.7 Ave daily kcal 1209.5±737.6 1277.9±808.4 1137.5±652.5 Ave daily fat in gm 48.4±33.9 50.9±37.0 45.8±30.5 Ave daily sugar 79.4±39.7 83.6±45.4 75.0±32.4 Ave daily Retinol equiv, mcg 495.4±240.8 533.5±253.9 455.4±220.6 Ave daily folate, mcg 330.7±209.5 351.2±225.4 309.1±190.3 *Means significantly different at p<0.05 . 40 Results for Research Question 1 , Table 11 shows the bivariate results for correlations between parental feeding strategies and parents™ and children™s BMI™s. No measures of parents™ feeding behavior were associated with the child™s or parent™s BMI. The variance pre dicted by each model was negligible . One exception was that parents™ concern for child overweight was positively associated with child™s and the parent™s BMI. 41 Table 11. Bivariate Correlations Between Feeding Strategies and Parent and Child BMI at Baseline Variable 1 2 3 4 5 6 7 8 9 10 11 1. High Control 1 2. High Contingency .574** 1 3. Mealtime Behavior .319** .291** 1 4.Healthy Food Availability .157* .199* .288** 1 5.Timing of Meals .150 .132 .324** .111 1 6.Child Centered -.280** -.245** .074 .002 -.079 1 7.Co ncern for Child Overweight -.001 .015 -.094 -.051 .03 .095 1 8.Fruit/vegetable Modeling .044 .170* .362** .130 .011 .332** .137 1 9.Milk Modeling -.027 .020 .167* .012 -.009 .197* -.014 .228** 1 10.Child BMI -.047 .032 -.015 -.033 -.004 .093 .288** -.016 -.038 1 11. Parent BMI .013 .089 -.085 -.099 .039 .094 .264** -.021 .071 .268** 1 * Correlation was significant at the 0.05 level (2 -tailed). ** Correlation was significant at the 0.01 level (2 -taile d). 42 In Table 12, consistent with the bivariate results, the results of the general lin ear regression indicated that parents™ concern for child overweight predicted higher percentile of P = 0.004). This was when parental age and BMI were controlled. Compared to the normal weight children, parents of overweight children were twice as likely to be concerned about their children ™s weight status. Compared to the normal weight children, parents of overweight children were ½ as likely to structure mealtime behaviors like family meals or no TV. Table 12. Regression Models for Association of Child Weight Status with Parental Feeding Strategies : Estimated Odds Ratios, Standard Errors, 95% CI of Parental Feeding Strategies at the Baseline B SE R2 Model 1: High Control -.02 .03 -.06 .004 Model 2: High Contingency .01 .03 .02 .000 Model 3: Mea ltime Behavior .001 .035 .003 .000 Model 4: Healthy Food Availability -.01 .03 -.02 .000 Model 5: Timing of Meals -.02 .03 -.04 .002 Model 6: Child Centered .03 .03 .06 .004 Model 7: Concern for Child Overweight .08* .03 .24 .053 Model 8: Fruit/v egetable Modeling .004 .031 .010 .001 Model 9: Milk Modeling -.01 .02 -.04 .002 Notes: 1) In each model, parental age and BMI were controlled. To save sp ace, the regression coefficients were not reported. 2)* Coefficient was significant at 0.05 level (2 -tailed). 3) R 2 was the percentage of variance for child BMIpercentile explained by the parental feeding strategy in each model. 43 Results of the logistic r egression ( Table 13) indicated a similar relation between children™s weight status and parents™ concern for child overweight. Specifically, with each one unit increase in parents™ concern for child overweight, the odds of overweight or obese versus under or normal weight increased by 127% ( OR = 2.27, p = 0.006), when parental age and BMI were controlled. 44 Table 13. Logistic Regression Models for Association of Child Weight Status with Parental Feeding Behavior: Estimated Odds Ratios, Standard Error s, 95% CI of Parental Feeding Strategies at Baseline Constructs Odds Ratio SE 95% CI of Odds Ratio Model 1: High Control .72 1.27 .455-1.150 Model 2: High Contin gency 1.14 1.22 .770-1.699 Model 3: Mealtime Behavior .58+ a 1.33 .336-1.011 Model 4: Healthy Food Availability 1.00 1.26 .642-1.570 Model 5: Timing of Meals .85 1.24 .552-1.302 Model 6: Child Centered 1.24 1.29 .752-2.042 Model 7: Concern for Ch ild Overweight 2.27**b 1.35 1.257-4.084 Model 8: Fruit/vegetable Modeling .92 1.28 .562-1.488 Model 9: Milk Modeling .91 1.13 .718-1.162 Note: The reference group for child weight status was under or normal weight. Parent age and BMI were controlled for the mode l. ** Odds ratio was significant at the 0.01 level (2 -tailed); + Odds ratio was marginally significant at 0.10 level (2 -tailed). a p=.055 b p=.006 45 Although the bivariate and the general linear regression results indicated that the mealtime behavior construct was not associated with child BMI, the logistic regress ion results indicated that mealtime behavior construct was associated with healthier child weight status with an approximate significant level. Particularly, with each one -unit increase in mealtime behavior, the odds of overweight or obese versus under or normal weight decreased by 42% ( OR = 0.58, p = 0.055) , when parental age and BMI were controlled. Except for mealtime strategies and parents™ concern for child overweight, the other parent feeding constructs were not significantly associated with child w eight status or child BMI, based on the general linear and the logistic regression results. In the results for correlations between parental feeding strategies and diet quality (see Table 14), high control (really low control), high contingen cy (really low contingency), mealtime behavior, and healthy food availability were associated with decreased children™s intake of sweet beverages, sweets, and snack and chips. Timing of meals was associated with children™s intake of fruit with juice in bot h less frequency per week and fewer cups per day, with fewer cups of milk per day, and with lower frequency of sweet beverages and snacks per week. Child centered feeding was associated with higher frequency of sweet beverages per week. Fruit/vegetable mod eling was associated with lower frequency of sweet beverages and snacks per week. 46 Table 14. Pearson Correlations Between Parental Feeding Behavior and Diet Quality at Baseline Fruit with Juice Frequency Fruit CE Milk cup Sweet Beverages cup Sweet Beverage Frequency Sweets frequency Snack and Chip Frequency High Control -.072 -.052 -.077 -.034 -.296** -.026 -.189* High Contingency -.015 -.055 -.028 -.172* -.394** -.163* -.326** Mealtime Behavior .034 -.051 .004 -.321** -.283** -.303** -.363** Healthy Food Availability .057 -.035 -.033 -.372** -.326** -.371** -.307** Timing of Meals -.193* -.263** -.161* -.022 -.243** -.023 -.186* Child Centered .111 .074 -.096 -.008 .166* -.004 -.010 Concern for C hild Overweight .024 .051 -.068 .041 -.001 .043 -0.005 Fruit/vegetable Modeling .144 .028 -.016 -.098 -.185* -.083 -.170* Milk Modeling .118 .113 .081 .052 -.013 .057 -.001 * Correlation was significant at p< 0.05 (2-tailed). ** Correlation was si gnificant at p< 0.01 (2-tailed). 47 Research Question 1: Baseline Results Research Question 1.1 : What is the relationship between parent feeding strategies and child BMI percentiles at the baseline ? (See Table 11) Ho 1.1a: Preschoolers in l ow-income families whose parents have negative feeding strategies will be overweight /obese. Ho 1.1a was not supported, because there was no relationship between parent practicing negative feeding strategies of high control and high contingency and their c hildren™s BMI™s. Ho1.1b : Preschoolers in low -incom e families whose parents have positive feeding strategies will be less overweight /obese. Ho1.lb was not supported by the correlational or general linear regression data, but was supported in part by the l ogistic regression when the parent™s BMI and age were both controlled and the child™s weight was dichotomized as underweight/normal (0) or overweight/obese (1). The parents who practiced the most positive mealtime behaviors like setting regular meal times, eating together , and avoiding TV distractions were slightly more likely to have children who were normal weight (p<0.055). Research Question 1.2: What is the relationship between parental feeding strategies and child diet quality of their preschooler? (See Table 12) Ho 1.2a: Preschoolers in low -income families whose parents have negative feeding strategies will have poor diet quality. Ho.1.2a was not supported Ho 1.2b: Preschoolers whose parents have positive feeding strategies will have improved diet quality. Ho.1.2b was supported , in part, because parents who practiced child centered feeding strategies had children who consumed sweet beverages and snacks less frequently. Also , parents who modeled eating fruits and vegetables in front of their childre n had children who 48 consumed sweet beverages and chips less frequently. However, Ho 1.2.b was not supported, in part, because parents who practiced child centered feeding strategies had children who more frequently consumed sweet beverages. Also , parents wh o practiced regular times for meals and snacks had children who consumed fewer fruit and drank less milk . Research Question 2: Post -Test Results Table 1 5 shows results for changes in parent feeding strategies from baseli ne to post -test by group. Both groups significantly improved their scores for feeding control, mealtime behaviors , and timing of meals and snacks. There was no effect of the EH intervention on the feeding strategies, however, as is shown in the last colu mn. 49 Table 1 5. Change in Feeding Strategies Between Pre - and Post -Test s for the Treatment and Control Groups Using GEE for Repeated Measures and Controlling for Time and Group Feeding Behavior 6 wk Post -test Difference test Rx N=55 Control N=35 Rx SE Control SE Difference between Rx Control High control .23* .07 .31* .10 Rx Control High Contingency -.09 .10 .07 .13 Rx Control Fruit/Veg Modeling .16 .09 .00 .14 Rx Control Milk Modeling .25 .15 -.03 .23 Rx Control Healthy Availability Modeling .16 .09 .10 .10 Rx Control Mealtime behaviors .16* .07 .19* .06 Rx Control Timing of meals .24* .09 .32* .13 Rx Control Child centered .19 .09 .14 .12 Rx Control Concern C hild Over wt -.07 .09 -.06 .04 Rx Control -tailed). Research Question 2 : How does a six -week parent feeding intervention (EH) change the parent feeding strategies of parents compared to the control group? (See Table 13). Ho 2.1: Feeding strategies of parents in the EH intervention group will improve compared to the control group, from baseline until the end of the post Œtest. Ho2.1 was not supported. 50 Research Question 3: Follow -Up R esults Table 16 shows the difference between baseline and follow -up data from the PFBQ. Parents became less controlling, high contingency decreased , and the mealtime behaviors were improved. Table 16. Differences in Feeding Strategies Between Baseline and 3 Month Follow -Up for the Combined Sample of Bo th the Treatment and Control Groups After Receiving the EH Curriculum ( N=91) Mean SE High control .20* .06 High Contingency .17* .07 Fruit/Veg Modeling .14 .08 Milk Modeling .05 .14 Healthy Availability Modeling .08 .08 Mealtime behaviors .13* .06 Feeding times .07 .07 Child centered .09 .07 Concern Child Wt. .00 .09 * p<0.0.05 level (2 -tailed) In Table 17, the vegetable frequency per week, vegetable cup equivalency per day, and number of different vegetables per week intake all increased significantly at the follow -up. The Folate was also increased significantly implying and indicating that the intake of vegetables was increased. Milk cup equivalency intake per day, and sweets such as ice cream, candy, and cookies frequency per week decreased significantly at the follow -up. Milk cup equivalency intake decrease was undesirable, but earlier milk modeling was correlated with mealtime behaviors. We saw fruit w ith juice cup equivalency, fruit with juice frequency, milk cup equivalency , and milk frequency intake decreased at the follow -up but they were not significant. The whole wheat bread frequency per week and amount increased after the 3 month follow -up. 51 The percentage drinking low fat milk increased from 72% at the baseline to 75% at the follow - up. Table 1 7. Difference in Diet Quality Between Baseline and 3-Month Follow -Up for the Combined Sample of Both the Treatment and Control Groups (N=91) Diet quality indicator Mean SE Fruit with juice freq./wk 0.01 .15 Fruit with juice cup equiv/day -0.07 .10 Fruit without juice frequency/wk -0.01 .16 Fruit without juice cup equiv/day -0.05 .08 Vegetable freq/wk 0.19* .08 Vegetable cup/day 0.08* .04 No. different veg/wk 0.47* .16 Milk freq/wk -0.17 .20 Milk cup/day -0.19* .09 % Low fat milk 75.0 Sw Beverages freq/wk -0.10 .16 Sw Bev erages per day -0.01 .02 Sweets (ice cream + candy + cookies) freq/wk -0.22* .10 Snack & chips freq/wk -0.07 .14 Whole Wheat bread freq/wk -0.01 .27 Whole wheat 1oz equiv/day 0.08 .05 Ave daily fiber ,gm 0.9 .51 % eating whole grain cereal 58.7 Ave daily kcal 43.0 42.7 Ave daily fat ,gm 2.2 1.8 Ave daily sugar, gm -3.8 3.4 Ave daily Retinol equiv, mcg 18.0 19.3 Ave daily folate equiv, mcg 33.7* 16.2 52 Table 1 8 shows the difference in child BMI percentile and weight status, where normal weight=0 and overweight/obese=1, between baseline and the 3 -month follow -up. There were no significant d ifferences. Table 1 8. Differenc e in Child BMI Percentile and Weight Status Bet ween Baseline and Follow -Up for the Combined Sample of Both the Treatment and Control Groups ( N=91) Mean SE Child BMI percentile -.22 2.57 Odds Ratio (OR) 95% CI of OR Weight status 1.03 .95, 1.12 Note: Parent age and BMI were controlled for this model Research Question 3: At the three month follow -up, what impact does a six -week parent intervention focusing on child feeding hav e on: (a) parent feeding strategies of preschoolers, (b) the preschoolers™ diet quality, and (c) pre schoolers™ weight status as compared to their baseline measurements? Ho 3.1 : There will be an increase in positive parent feeding strategies, and a decli ne in negative feeding strategies. Ho 3.1 Was supported because we did see a decrease in both hig h controlling and high contingency, that imply the EH intervention was successful in improving the feeding strategies. The parents use d less feeding control and less contingency with their preschoolers. Also , the mealtime behaviors improved at the follow -up. These included things like not watching TV, eating family meals, not eating before meals , and being seated while eating . 53 Ho 3.2: There will be an increase in dietary strategies reflecting good diet quality and a decline in those refl ecting poor diet quality . Ho3.2 was supported , in part, because there were improvements in amounts, frequency , and number of different types of vegetabl es that children ate, as well as an increase in folate. Also, the frequency of eating sweets declined. The only undesirable behavior seen was the decreased intake of milk. Ho 3.3: There will be fewer overweight and obese preschoolers and/or a decline in average BMI percentiles. Ho 3.3 was not supported. 54 CHAPTER V. DISCUSSION This study was designed to see if a six -week EH intervention with low income parents of preschoolers could improve their child feeding strategies, their children™s diet quality and their children™s weight status at the 3 -month follow -up. Eat Healthy was successful in improving three of eight parent feeding strategies that were targeted and measured Šhigh control, high contingency , and mealtime behaviors. By the end of the intervention, parents became less overtly controlling, use of rewards and punishments to get their child to eat had declined, and their mealtime behaviors , like eating as a family, no TV at meals, and not eating an hour before meals , had improved. For the children™s diets, EH was also successful in increasing the amounts of, the frequency of, and the number of di fferent vegetables, as well as decreasing the intake of sweets, snacks , and sweet beverages. EH did not , however, impact weight status of children at the three month follow -up measurements. This finding was not surprising perhaps because four to five months was too short to see change in weight status. Also, for preschoolers, the average change in weight an d height is 2 -3 inches and 5 -6 pounds per year (Brown, 2010). Although the bivariate and the general linear regression results from baseline data indicated that mealtime behavior was not associated with child BMI percentile, the logistic regression resu lts, wherein weight was categorized as not overweight versus overweight or obese, indicated that mealtime behavior was associated with healthier child weight status with an approximate significant level. With each one -unit increase in mealtime behavior, th e odds of overweight or obese versus under or normal weight decrease by 42% ( OR=0.58, p=0.055) when parental age and BMI were controlled. This finding is supported in part by other studies that found family mealtimes and/or limited TV to be associated with lower BMI™s (Jones et al, 2014; Wansink et al., 2014; Lehto et al, 2012; Chan et al, 2011). Except for mealtime behavior and 55 parents™ concern for child overweight, the rest of the measures for parents™ feeding strategies were not significantly associated with child weight status or child BMI percentiles. In contrast to parental feeding strategies and children™s™ BMI percentile, there were many significant associations at baseline for diet quality, an outcome more proximal to strategies than is weight s tatus. High control, high contingency, mealtime behaviors, and healthy food availability were associated with a decrease in children™s intake of sweet beverages, sweets, and snack and chips. Timing of meals was associated with less children™s intake of fr uit with juice in both frequency per week and cups per day; with fewer cups of milk per day; and with lower frequency of sweet beverage and snack and chips per week. Child centered feeding was associated with higher frequency of sweet beverages per week. F ruit/vegetable modeling was associated with lower frequency of sweet beverage and snack and chips per week. The baseline descriptive of child food intake from the food screener show high scores for the frequency of fruit and milk, percentage of chil dren drinking low fat milk, variety of different vegetables, and percentage of children eating whole grain cereal indicating good diet quality for these foods. These findings are in contrast to those found prior to new WIC food package implemented in 2010 (http://www.idfa.org/). Indicators of poor diet quality included few cups of vegetables per day (0.6 cups versus 1 -1½ cup recommended); low milk intake of 1.6 cups versus 2-3 cups recommended (Table 1 and Table 2 ). Dietary fiber averaged 11 grams compare d to 19 grams recommended for children age 1 -3, and 25 -26 grams for children age 4 -8 years ( Table 1 and Table 2 ). These finding are similar to Krebs -Smith et al. (2010) who analyzed national datasets. Furthermore, the overall diet quality was low as indi cated by high frequency of sweet beverages, sweets , and snack chips. 56 At the post test, the finding was that both the control and intervention groups were less controlling, had improved mealtime behaviors , and better timi ng of meals and snacks. Because parents in both groups volunteered for this study on child feeding strategies, they were motivated to address some issues and make a significant time commitment. This might have affected the post -test results. Furthermore, the educators in the out -state co unties (other than Ingham) had regular contact with families in both groups. Because the educators were excited about the EH curriculum, they might have unintentionally contaminated the study by starting to give parents in the control group some advice on feeding strategies before the control group was intended to receive the intervention . A different and cleaner design to avoid this issue could be done by first evaluating the knowledge of the parents at the baseline , and second ly chang ing the educat ors for the control group when the intervention start s. At the three month follow -up, there was no change in either parent or child weight status. Of the intervention studies reviewed, only the one by Harvey -Bernino et al . (2003) reported a reduction in w eight status at the post test, but not at the follow -up when working with low -income families. The other studies that result ed in weight reduction at post -test but not follow -up were with middle income families ( West et al. , 2010). The majority of parent s were overweight or obese; a finding similar to other studies ( Ost ybe et al. , 2012). At the three month follow -up, scores had improved for high control , similar to Harvey -Beniro et al.,2003 findings , reduction in high contingency was similar to that of Ostybe et al., 2012. Two other studies demonstrated improvements in mealtime strategies such as less TV with meals (Ostybe et al., 2012; & Dickens et al. , 2014) although they used different measurements . These were the same parent feeding strategies that correlated with children™s food intake at baseline. 57 The sub -constructs of high control, high contingency , and child centered feeding were developed and validated by (Hughes et al. 20 05, Hughes et al. , 2006) and can be considered standardized. The construc t of mealtime behavior was developed and validated by Murashima et al. (20 12). All four of these sub constructs had good internal reliability and should be considered robust indicators for parent feeding strategies. To our knowledge, this study is the fir st to use these validated sub -constructs in an intervention study and find improvements. The findings at follow -up of increased vegetable frequency, amount , and variety were similar to those of three other intervention studies (Skouteris et al., 2015; Dick ens et al., 2014 & Tabak et al., 2012). Our findings differed, however, in that these researchers also found an increase in fruit intake where EH did not. Because folate also increased significantly this suggests that the vegetable increase was for green l eafy vegetables, a very positive finding. Sweets such as ice cream, candy, and cookies frequency per week and milk cup equivalency intake per day decreased at follow -up. The decline in such energy -dense foods following interven tion were similar to those f rom several studies (Skouteris et al., 2015 ;Dickens et al., 2014; Fletcher et al., 2013; Horodynski et al., 2012 and Harvey et al., 2003). The decline in milk intake was undesirable, although earlier milk modeling by parents correlated with their mealtime behaviors like avoiding distractions during meals. The decrease in milk consumption, because it correlated to this mealtime strategy, might explain this decline. For example, if children remain seated during meals without distractions like TV, they should be more likely to eat well, feel full , and drink fewer fluids , especially between meals. That there was a trend towards less fruit juice might support this possibility. On the positive side, these preschoolers consumed a higher percentage of low fat milk than has been reported in the past with young children from low income families (Hoerr et al., 2006; Hoerr et al., 2009). Changes in the WIC 58 program package in 2010 might have contributed to this difference ( http://www.idfa.org/ ). The majority of children in this study (63%) were on WIC. The EH study differed from the other parent intervention studies reviewed in that the EH focused specifically on several parents™ feeding strategies (Skouteris et al., 2015; Dickens et al., 2014; Fletcher et al.,2013; T aback et al., 2012) both in regards t o the intervention, as well as to the measurement. The EH intervention was successful in improving several negative feeding behaviors such as decreasing the high control and high contingency strategies and improving the mealtime behaviors using m ulti -item validated constructs. 59 CHAPTER VI. STRENGTHS, LIMITATIONS, AND IMPLICATIONS Study Strengths The EH intervention study was unique in combining all of the following aspects: 1) it targeted low income families with young children; 2) was home based using video clips and full color booklets with engaging activities for both parents and children; 3) captured the diet quality by using a well -known dietary screener for children; 4) used paraprofessional educators in conv eying educational concepts to parents within the context of their own homes; 5) measured the heights and weights with research quality instruments and t rained educators; and 6) used valid and reliable multi -item constructs for parent feeding strategies. The participating educators attended two 3 hour training sessions and were in weekly follow -up contact with the researchers to maint ain program fidelity. All the parents were low income families, a group who has the highest rate of obesity in the U.S., yet o nly a few intervention studies have targeted low income families (Harvey -Beniro et al. , 2003; Horodynski et al., 2005 ; Dickens et al., 2104). Most other intervention studies have targeted middle income families (Skouteris et al. , 2015 , Tabak et al., 2012; Ostobe et al., 2012; West et al., 2010 and Fletcher et al., 2013). Limitations The biggest limitation was that the only time for comparing the two groups was at the post test due to funding constraints and the fact that our funder hired an outside rev iewer to conduct independent assessments. Therefore, the 3 -month follow -up measurements became a gross impact analysis of the EH intervention. Another limitation was that while we had multi -item validated constructs for six feeding strategies, two sub -constructs Šmilk modeling and fruit and vegetable modeling , remained single 60 items due to poor Cronbach alphas for the original multi -item variables. Also, the PFBQ was self -reported by the parents. The accuracy of the information depended on parent memory a nd lack of social bias. The BFKS used for dietary intake of the children is less accurate than the multiple 24 hours dietary recall (Nelms et al., 2011). The BKFS may underestimate their dietary intake (Murshima et al.,2012). The BKFS might not ha ve capt ured all ethnic food consumed by families from ethnicities other than Black and White Americans. The study did not collect the parents or mother food intake which is a very good indicator of the child s diet (Hoerr et al ., 2006). Because we did not fully ca ptured in what previous nutrition education programs parents had participated at base line, it is unknown to what degree differences in nutrition knowledge might have affected the outcomes. Although overall this study was successful with t wo of three objec tives at the three month follow -up, it is not known if such improvements will continue. Some intervention studies did not see the improvement at the 6 mont h follow -up, e.g. Fletcher et al., (2013) a 6 or 12 month follow -up would strengthen the study, but would require significant financial incentives to maintain sample size. The study duration was not long enough for a 6 or 12 month follow -up as is recommended. Still a 3 month follow -up was better than none as in the Harvey -Berniro et al. (2003 ) study. Alt hough there was a 30% drop -out rate, it was about the same as two other studies that worked with low income families (Harvey -Berniro et al., 2003; Horodynski et al., 2005). Finally, the EH did not target nor evaluate physical activity or sleep of pre schoolers. Both are known to affect weight status (Firouzi et al.,2014 and Golley et al.,2013 ). There fore, it is recommended for future research to add the physical activity and sleep patterns as measurements. 61 Implications for Future Research The stud y demonstrated that a home -based intervention targeted to parent feeding strategies could successfully change three of them and that these were associated with positive changes in their children™s diet quality . Future studies should use the EH curriculum w ith different groups to determ ine generalization of findings. There is a need to translate the materials into Spanish for use with Hispanic populations. Because obesity is also associated with physical activity ( Firouzi et al.,2014 ) and poor sleep quality (Golley et al.,2013 ), these two aspects mi ght be added to the materials. Finally, it is highly recommend ed that researchers use the sub -constructs for parent feeding strategies used in this study. Conclusion At the 3 month follow up the EH study was suc cessful in improving three of the eight parent feeding strategies, high control, high contingency , and mealtime behaviors. The parents became less controlling, were using less contingency and rewarding, and had improved mealtime strategies. The EH was also successful in improving the diet quality of the preschooler s by increasing nutrient dense food and decreasing energy dense food. The preschoolers also had a decreased intake of sweets which included ice -cream, candy, and chips. However the EH was not succ essful in improving the weight status of the preschooler s. EH was unique compared to other studies. First, the study focused specifically on feeding strategies related to child weight status and diet quality. Second, it targeted low income fa milies. Third, the curriculum was designed to 3 rd to 5 th grade reading levels and did not require high literacy. Fourth, interactives materials were in full color and the short video clips were from real parents who exhibited ethnic and racial diversity . The message s delivered from the videos were 62 from real experiences and real daily life interactions. Fifth , the education material was delivered in person , not mailed to the parents , as compared to sev eral other studies. 63 APPENDICES 64 Appendix A. Parent Feeding Strategy Questio nnaire Listed by Co nstru ct HC1 1I beg my child to eat dinner. R12345HC2 2I spoon-feed my child to get him or her to eat dinner. R12345HC3 3I physically struggle with my child to get him or her to eat (for example, putting my child in the chair so he or she will eat). R12345Hi Cont1 4 I warn my child that you will take away something other than food if he or she doesn™t eat (for example, fiIf you don™t finish your meal, there will be no TV tonight after dinnerfl). R12345Hi Cont2 5I promise my child to something other than food if he or she eats (for example, fiIf you eat your beans, we can play ball after dinnerfl). R12345Hi Cont3 6I encourage my child to eat something by using food as a reward (for example, fiIf you finish your vegetables, I™ll get you some ice creamfl). R12345Hi Cont4 7I warn my child that I will take a food away if the child doesn™t eat (for example, fiIf you don™t finish your vegetables, you won™t get dessertfl). R12345HA1 8I keep fruits and vegetables available that my child can eat. 12345UHM1 9I keep sweets, candy or salty snacks where my child can reach them. R12345UHM2 10I keep sugar-sweetened beverages* where my child can reach them. *Drinks like Coke, 7-Up, Sunny Delight, Hawaiian Punch, or aguas frescas (DO NOT include 100% fruit juice and diet soda) R12345HA2 11I limit my child™s access to sweets, candy, salty snacks or sweetened beverages by not having them readily available. 12345MB1 12I allow my child to play and watch TV during meals. R12345MB2 13We eat dinner together as a family. 12345TM1 14I allow my child to eat whenever he/she is hungry during a day. R12345Mt f Scored: 1=never; 2=rarely; 3=sometimes; 4= most of the time; 5= always Construct Item no Items Never Rarely Sometimes Most of the time Always 65 Never Rarely Sometimes Most of the time Always TM2 15I allow my child to decide when to eat meals and snacks. R12345MB3 16I allow my child to eat an hour before meals. 12345MB4 17I set regular meal times for my child. 12345MB5 18I have my child sit down at home while eating. 12345MB6 19I allow my child to eat while standing or walking. R12345NonDir1 20I say something positive about the food my child is eating during dinner. 12345NonDir2 21I reason with my child to get him or her to eat (for example, fiMilk is good for your health because it will make you strongfl). 12345NonDir3 22I help my child to eat dinner (for example, cutting the food into smaller pieces). 12345NonDir4 23I compliment the child for eating food (for example, fiWhat a good boy! You™re eating your beansfl). 12345NonDir5 24I encourage my child to eat by arranging the food to make it more interesting (for example, making smiley faces on the pancakes). 12345NonDir6 25I ask my child questions about the food during dinner. 12345ModND1 26I eat fruits and vegetables in front of my child. 12345ModND2 27I drink milk in front of my child. 12345ModND3 28I eat sweets, candy or salty snacks in front of my child. R12345ModND4 29I drink sweetened beverages in front of my child. R12345OVWT 30I worry that my child is overweight right now 12345HC= High Control Hi Cont= High Contingency HA= Availability of Healthy Foods UHM= Unhealthy Modeling ModND= Modeling, Non-Directive NonDir= Child Centered, Non-Directive R=Reverse score (1=Always, 2=Most of the time, 3=Sometimes, 4=Rarely, 5=Never MB= Mealtime Behaviors TM= Timing of Meals 66 Appendix B . Parent Consent Form 67 Table 1 9. Healthy Eating Activities 68 69 70 Appendix C. IRB Consent Form Figure 1 . IRB Consent Form Image 71 Figure 1 (cont'd) 72 Appendix D. Block Kids Food Screener (BKFS) Figure 2. Block Kids Food Screener (BKFS) Questionnai re Image 73 Figure 2 (cont'd) 74 Appendix E . Glossary Child centered feeding is rearranging foods to make them interesting, and complimenting the child when sh e/he eats. (Hughes et al., 2006) . Food availability is where parents do or do not provide access to certain types of foods in the house (Brown et al., 2008). Food environmental controls are strategies where parents provide a healthy and organized home fo od environment and family rules around eating to help the child eat a healthy diet. High contingency is where parents threaten or reward the child to eat. (Hughes et al., 2006) . High control is where parents verbally, physiologically, and physically pres sure the child to eat (Hughes et al., 2006). Mealtime behaviors are where parents set rules during meals such as sitting at a table, eating together, and not viewing TV during meals (Hoerr et al., 2005). Modeling is where parents demonstrate the preferre d eating strategies in front of the child, e.g., eating fruits and vegetables and not eating high fat /sugar foods (Vander Horst et al., 2007). Parenting strategies are behaviors that parents use to get the child to do something specific, in this case to influence children™s eating. Examples of controlling feeding strategies are: restriction, monitoring, pressure to eat, rewarding, threats...etc. Parenting styles are stable characteristics of parenting reflecting both the degree of demands on/control of the child as well as parental responsiveness to child needs (Baumrind, 1989). Timing of the meals is where parents set a regular mealtime for the child and family (Baughcum et al., 2001). 75 REFERENCES 76 REFERENCES Adamson M, Morawska A, Sanders MR. Childhood feeding difficulties: A randomized controlled trial of a groups -based parenting intervention. J Dev Behav Pediatr. 2013;34:293 -302. Anderson SE, Whitaker RC. Household routines and obesity in US preschool Œ aged children. Pediatric s. 2010;125:420 -428. Baumrind D. Rearing competent children in child development today and tomorrow. San Francisco, CA: Jossey -Bass; 1989. Berg F, Buechner J, Parham E. Guidelines for childhood obesity prevention programs: Promoting healthy weight in child ren. J Nutr Educ Behav . 2003 ; 35:1 -4. Biggers MB, Spaccarotella K, Berhaupt -Glickstein A, Hongu N, Worobe J. Come and get it! A discussion of family mealtime literature and factors affecting obesity risk. Adv Nutr . 2014; 5:235 -247. Birch L, Davison K. Fam ily environmental factors influencing the developing behavioral controls of food intake and childhood overweight. Pediatr Clin North Am. 2001; 48(4):893 Œ907. Birch L, Fisher JO, Grimm -Thomas K, Markey CN, Sawyer R, Johnson SL. Confirmatory factor analysis of the child feeding questionnaire: A measure of parental attitudes, beliefs and practices about child feeding and obesity proneness. Appetite. 2001;36:201 -210. BMI classification. World Health Organization. http://apps.who.int/bmi/index.jsp?introPage=intro_3.html . Accessed May 20, 2015. Bryant M, Stevens J. Measurement of food availability in the home. Nutr Rev . 2006;64(2):67 -76. Campbell KJ, Crawford DA, Ball K. Family food environme nt and dietary behaviors likely to promote fatness in 5 -6 year -old children. Int J Obes . 2006;30:1272 -1280. Child BMI in low -income African -American preschoolers. Obesity. 2006;14:2026 -2033. Clark HR, Goyder E, Bissel P, Blank L, Peters J. How do parents™ child -feeding behaviors influence child weight? Implications for child obesity policy. Am J Public Health . 2007;29:132 -141. Chan JC., Sobal J., Family meals and body weight. Analysis of multiple family members in family units. J.Appet. 2011.57:517 -524.doi:10.1016/j.appet.2011.07.001. Clark HR, Goyder E, Bissel P, Blank L, Walters SJ, Peters J. A pilot survey of socio -economic differences in child -feeding behaviors among parents of primary -school children. Public Health Nutr . 2007;11:1130 -1136. 77 Cohen JFW, K raak VI, Choumenkovitch SF, Hyatt RR, Economos CD. The CHANGE study: A healthy -lifestyles intervention to improve rural children™s diet quality. J Acad Nutr Diet . 2014; 114: 48 -53. Couch S, Glanz K, Zhou C, Sallis JF, Saelens BE. Home food environment in r elation to children™s diet quality and weight status. J Acad Nutr Diet . 2014;114(10):1569 -1579. http://www.ncbi.nlm.nih.gov/pubmed/25066057. Accessed May 11, 2015. Darmon N, Drewnowski. Does social class predict diet quality? Am J Clin Nutr . 2008;87:1107 -1117. Daniels LA, Mallan KM, Nicholson JM, Battistutta D, Magarey A. Outcomes of an early feeding practices intervention to prevent childhood obesity. Pediatrics. 2013;132:109 -118. Diana B. Current pattern of parental authority. Developmental Psychology. 1 971;4(1):1 -103 Dicken KL, Hill TF, Dollahite JS. Practice based evidence of effectiveness in an integrated nutrition and parenting education intervention for low -income parents. J Acad Nutr Diet . 2014;114(6):945 -950. Dietary guideline for Americans, 201 0. US Department of Agriculture. http://www.health.gov/dietaryguidelines/dga2010/DietaryGuidelines2010.pdf. Accessed August 1, 2014. Draxten M, Fulkerson JA, Friend A, Flattum CF, Schow R. Parental role modeling of fruits and vegetables at meals and snacks is associated with children's adequate consumption. Appetite. 2014;78C:1 -7. Duncanson K, Burrows T, Collins C. Study protocol of a parent -focused child feeding and dietary intake intervention: The feeding healthy food to kids randomized controlled trial. BMC Public Health. 2012;12:564. http://biomedcenteral.com/147 -2458/12/564. Accessed April 20, 2015. Entin A, Kaufman -Shriqui V, Naggan L, Vardi H, Shahar DR. Parental feeding practices in relation to low diet quality and obesity among LSES Children. J Am Coll Nutr . 2014;33(4):306 -314. Firouzi S., Poh BK.,Ismail MN., Sadeghilar A.,J Obes Res &Clinc pract 2014.8:e70 -e78. Doi http://dx.doi.org/101016/j.orcp.2012.12.001. Fisher JO, Birch L. Restriction access to a palatable food affects children™s behavior re sponse, food selection and intake. Am J Clin Nutr . 1999;69:1264 -1272. Fisher JO, Birch L. Restriction access to foods and children™s eating. Appetite. 1999;32:405 -419. Fletcher A, Wolfenden L, Wyse R, Bowman J, McElduff P. A randomized controlled trial and mediation analysis of the healthy habits, telephone -based dietary intervention for preschool children. IJBNPA. 2013;10:43. http://www.ijbnpa.org/content/10/1/43. Accessed February 22, 2015. 78 Ford CN, Slining MM, Popkin BM. Trends in dietary intake among US 2-6 year old children, 1989-2008. J Acad Nutr Diet . 2013;113:35 -42. Frankel LA, O™Connor TM, Chen T, Nickolas T, Power TG, Hughes SO. Parents™ perceptions of preschool children™s ability to regulate eating feeding style differences. Appetite. 2014;76:166 -174. Frankel LA, Hughes SO, O™Connor TM, Power TG, Fisher JO, Hazen NL. Parental influences on children™s self -regulation of energy intake: Insights from developmental literature on emotion regulation. J Obes . 2012;2012:1 -12. Galley RK., Maher CA., Maticci ani L., and Olds TS. Sleep duration or bedtime? Exploring the association between timing behaviors, diet and BMI in children and adolescents. Int J Obes. 2013. 37:546 -551. Han E, Powell LM. Consumption patterns of sugar -sweetened beverages in the United Sta tes. J Acad Nutr Diet . 2013;113:43 -53. Harvey -Berino J, Rourke J. Obesity Prevention in Preschool Native -American Children: A Pilot Study Using Home Visiting. Obes Res. 2003;11:606 -611. Haszard JJ, Skidmore PML, Williams SM, Taylor RW. Associations betwee n parental feeding practices, problem food behaviors and dietary intake in New Zealand overweight children aged 4 -8 years. Public Health Nutr. 2014;8(6):1036 -1043. Doi:10.1017/S1368980014001256. Accessed March 1, 2015 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 . 2006;106:1766 -1773. Hoerr SL, Hughes SO, Fisher JO, Nicklas TA, Liu Y, Shewchuk RM. Associations among parental feeding styles and ch ildren™s food intake in families with limited incomes. IJBNPA . 2009;6:55, http://www.ijbnpa.org/content/6/1/55. Accessed April 3, 2015. Hoerr SL, Tsuei E, Liu Y, Franklin FA, Nicklas TA. Diet quality varies by race/ethnicity of head start mothers. J Am D iet Assoc . 2008;108:651 -659. Horodynski MA, Baker S, Coleman G, Auld G, Lindau J. The healthy toddlers trial protocol: An intervention to reduce risk factors for childhood obesity in economically and educationally disadvantaged populations. BMC. 2011;11:58 1. http://www.biomedcentral.com/1471 -2458/11/581. Accessed May 3, 2015 Horodynski MA, Stommel M. Nutrition education aimed at toddlers: An intervention study. J Spec Pediatr Nurs . 2005;35:364 -372. Hughes SO, Power TG, Fisher JO, Mueller S, Nicklas TA. Rev isiting a neglected construct: parenting styles in a child -feeding context. Appetite. 2005;44:83 -92. Hughes SO, Anderson CB, Power TG, Micheli N, Jaramillo S, Nicholas TA. Measuring feeding in low Œ income African American and Hispanic parents. Appetite. 2 006;46:215 -223. 79 Hughes SO, Shewchuk RM, Baskin M, Nichlas TA, Qu H. Indulgent feeding style and children weight status in preschool. J Dev Behave Pediatc. 2008;29:403 -410. International dairy food s association .http://www.idfa.org/news -views/headline -news/a rticle/2014/03/21/final -wic -rule -excludes -reduced -fat -milk. Accessed 6/14/15 . Jones BL., Fiese BH., Parent routines. Child routines, and family demog raphics associated with obesity in parents and preschool Œaged children. Frontiers in Psychology 2014. 5:374. Doi:10.3389/fpsyg.2014.00374. Karp JS. Process evaluation of implementing the eat healthy, your kids are watching: A parent‚s guide to raising a h ealthy eater program [ dissertation ]. East Lansing, MI: Michigan State University; 2014. Khandpur N, Blaine RE, Fisher JO, Davison KK. Fathers™ child feeding practices: A review of the evidence. Appetite. 2014;78:110 -121. Krebs -Smith SM, Guenther PM, Subar AF, Kirkpatrick SI, Dodd KW. Americans do not meet federal dietary recommendations. J Nutr. 2010;140:1832 -1838. Laura K. The Science of Psychology: An Appreciative View . New York(NY): McGraw -Hills; 2008. Let‚s Move: American Move to raise a healthier ge neration. http://www.letsmove.gov/learn -facts/epidemic -childhood -obesity. Accessed August 1, 2014. May AL, Pan L, Sherry B, et al. Vital signs: obesity among low -income, preschool -aged children. MMWR. 2013;62(31):629 -634. Moore JB, Gaffney KF, Pawloski LR, Jaimovich SP, Campos MC. Maternal -child nutrition practices and pediatric overweight/obesity in the United States and Chile: A comparative study. J Pediatr Nurs . 2012;27:e45 -e49. Murashima M, Hoerr SL, Hughes SO, Kaplowitz S. Confirmatory factor analysis of questionnaire measuring control in parental feeding practices in mothers of Head Start children. Appetite. 2011;56:594 -601. Murashima M, Hoerr SL, Hughes SO, Kaplowitz SA. Feeding behaviors of low -income mothers: directive control relates to a lower BMI in children, and a nondirective control relates to a healthier diet in preschoolers. Am J Clin Nutr . 2012;95:1031 -1037. Murashima M, Hoerr SL, Hughes SO, Kattelmann KK, Phillips BW. Maternal parenting behaviors during childhood relate to weight status and fruit and vegetable intake of college students. J Nutr Educ Behav . 2012;44:556 -563. National Agricultural Library. United States Department of Agriculture, http://www.nal.usda.gov. Accessed May 25, 2015. Nelms M, Sucher KP, Lacey K, Roth SL. Nutrition Therapy & Pathophysiology. 2ed ed. Columbus (OH): Wadworth -Cengage Learning; 2011. 80 Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of childhood adult obesity in the United States, 2011 -2012. JAMA. 2014;311:806 -814. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of Obesity and trends in Body Mass Index Among US Children and Adolescents, 1999 -2010. JAMA. 2012;307:483 -490. Ogden CL, Carroll MD, Curtin LR, Lamb MM, Flegal KM. Prevalence of high body mass index in U.S. children and adolescen ts, 2007 -2008. JAMA. 2010;303(3):242 -249. Ogden J, Reynolds R, Smith A. Expanding the concept of parental control: A role for overt and covert control in children™s snacking behavior? Appetite. 2006;47:100 -106. Ostbye T, Krause K, Stroo M et al. Parent foc used change to prevent obesity in preschools: Results from the KAN -DO study. Am J Prev Med . 2012;55:188 -195. Patrick H, Nicklas TA. The benefits of authoritative feeding style: Caregiver feeding styles and children™s food consumption patterns. Appetite. 2005;44:243 -249. Patrick H, Nicklas T. A review of family and social determinants of children™s eating patterns and diet quality. J Am Coll Nutr . 2005;24:83 -92. Powers SW, Chamberlin LA, Van Schaick KB, Sherman SN, Whitaker RC. Maternal feeding strategies, child eating behaviors, and BMI in low -income African -American preschoolers. Obesity. 2006;14:2026 -2033. Reedy J, Krebs -Smith SM. Dietary sources of energy, solid fats, and added sugars among children and adolescents in the United States. J Am Diet Assoc . 2010;110:1477 -1484. Reznar MM, Carlson JS, Hughes SO, Pavandkar AS, Scott MK, Hoer SL. An interactive parents™ guide for preschool -aged children: pilot studies for improvement. J Acad Nutr Diet. 2014;114(5):788 -795. Reznar M. Application of behavior chang e and persuasion theories to a multimedia intervention designed to improve the home food environment and diet quality of resource - limited parents with young children [Dissertation]. East Lansing, MI: Michigan State University; 2012. Rhee, K. Childhood Ove rweight and the Relationship between Parent Behaviors, Parenting Style, and Family Functioning. Ann Am Acad Pol Soc Sci . 2008;615:12 -37. Roe LS, Meengs JS, Birch LL, Rolls BJ. Serving a variety of vegetables and fruit as a snack increased intake in presch ool children. Am J Clin Nutr . 2013;98:693 -699. Rosner B. Fundamentals of Biostatics . 4th ed. Daxbury Press; 1995. Scaglioni S, Salvioni M, Galmiberti C. Influence of parental attitudes in the development of children eating behavior. Br J of Nutri .2008;99( 1):S22 -S25. Skinner AC, Skelton JA. Prevalence and trends in obesity and severe obesity among children in the United States, 1999 -2012. JAMA Pediatr . 2014; 168: 561 -566. 81 Skouteris H, McCabe M, Swinburn B, Hill B. Healthy eating and obesity prevention for preschoolers: a randomized controlled trial. BMC public health. 2010;10:220. Http://www.biomedcentral.com/1471 -2458/10/220. Accessed June 1, 2015. SNAP -Ed interventions: A toolkit for states. NCCOR. March 2013:1 -15. Spera C. A review of the relationship amo ng parenting practices, parenting styles, and adolescent school achievement. Int J Sch Educ Psychol . 2005;17:125 -145. Spruijt -Metz D, Lindquist CH, Birch LL, Fisher JO, Goran MI. Relation between mothers™ child -feeding practices and children™s adiposity. Am J Clin Nutr . 2002;75:581 -586. Tabak RG, Tate FD, Stevens J, Siega -Riz AM, Ward DS. Family ties to healthy program: A randomized intervention to improve vegetable intake in children. J Nutr Educ Behav . 2012;44(2):166 -171. Tamayo T, Herder C, Rathmann W. Impact of early psychosocial factors (childhood socioeconomic factors and adversities) on future risk of type 2 diabetes, metabolic disturbances and obesity: a systematic review. BMC Public Health. 2010;10:525. http://www.biomedcentral.com/1471 -2458/10/525 . Accessed June 3, 2015. Taylor RW, Brown D, Dawson AM, et al. Motivational interviewing for screening and feedback and encouraging lifestyle changes to reduce relative weight in 4 -8 year old children: design of the MInT study. BMC Public Health. 2010;10:2 71. http://biomedcentral.com/1471 -2458/10/271. Accessed May 5, 2015. Thompson AL, Adair LS, Bentley ME. Pressuring and restrictive feeding styles influence infant feeding and size among a low -income African -American sample. Obesity. 2013;21:562 -571. Access ed April 10, 2015. Timperio A, Fulkerson JA. Food -related parenting issues and challenges. Public Health Nutr . 2014;17:957 -959. Wansink B., Kleef EV., Dinner ritual that correlate with child and adult BMI.J Ped Obes 2014.22:e91 -e95. Doi:10.1002/oby.20629. West F, Sanders MR, Cleghorn GJ, Davies PSW. Randomized clinical trial of a family -based lifestyle intervention for childhood obesity involving parents as the exclusive agents of change. Behav Res Ther. 2010;48(12):1170 -1179. Wolfenden L, Wyse RJ, Britton BI et al. Interventions for increasing fruit and vegetable consumption in children aged 5 years and under. Cochrane Database Syst Rev . November 2012:1 -48. Van der Horst K, Kremers S, Ferreira I, Singh A, Oenema A, Brug J. Perceived parting style and prac tices and the consumption of sugar -sweetened beverages by adolescents. Heath Educ Res. 2007;22(2):295 -304. Ventura AK, Birch LL. Does parenting affect children™s eating and weight status? IJBNPA. 2008;5:15. 82