Anthropometric, Parental, and Psychosocial ... - Wiley Online Library

Anthropometric, Parental, and Psychosocial Correlates of Dietary Intake of African-American Girls Karen W. Cullen,* Tom Baranowski,* Lisa M. Klesges,† Kathy Watson,* Nancy E. Sherwood,‡ Mary Story,§ Issa Zakeri,* Deborah Leachman-Slawson,¶ and Charlotte Pratt//

Abstract CULLEN, KAREN W., TOM BARANOWSKI, LISA M. KLESGES, KATHY WATSON, NANCY E. SHERWOOD, MARY STORY, ISSA ZAKERI, DEBORAH LEACHMANSLAWSON, AND CHARLOTTE PRATT. Anthropometric, parental, and psychosocial correlates of dietary intake of African-American girls. Obes Res. 2004;12:20S–31S. Objective: This paper identifies the anthropometric, parental, and psychosocial characteristics and meal practices (e.g., breakfast skipping and number of meals and snacks consumed) associated with consumption of total energy, percent energy from fat, fruit, 100% fruit juice, vegetables, sweetened beverages, and water among 8- to 10-year-old African-American girls. Research Methods and Procedures: This study included 114 8- to 10-year-old African-American girls and a parent or primary caregiver. Girls and a parent or primary caregiver completed several dietary questionnaires. Two 24hour dietary recalls were conducted with each girl. Height and weight were measured. Separate hierarchical regression analyses were conducted for each dependent dietary variable; potential field center differences were examined. Results: The number of meals and snacks consumed was correlated with energy intake. Lower BMI was related to higher vegetable consumption, and the number of snacks consumed was positively related to sweetened beverage

*Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, Texas; †Department of Preventive Medicine, University of Tennessee Health Science Center, Memphis, Tennessee; ‡HealthPartners Research Foundation, Minneapolis, Minnesota; §Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, Minnesota; ¶Center for Community Health, University of Memphis, Memphis, Tennessee; and 储Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute, Bethesda, Maryland. Address correspondence to Karen W. Cullen, Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, 1100 Bates Street, Houston, TX 77030-2600. E-mail: [email protected] Copyright © 2004 NAASO

20S

consumption. Greater low-fat food preparation practices reported by parents were related to lower consumption of fat as a percentage of total energy. Discussion: Dietary behavior differed across geographic areas. Low-fat food preparation practices in the home seemed to be an important influence on the percentage of energy consumed from fat. Greater vegetable consumption was associated with lower BMI. Interventions to prevent excessive weight gain in African-American girls should encourage low-fat food preparation in the home and greater consumption of vegetables. Key words: diet, BMI, African-American girls, parent, psychosocial

Introduction Obesity is a major problem for youth in the United States (1). Dietary behaviors, such as increased consumption of fruit, juice, and vegetables, lower dietary fat intake (2,3), and sweetened beverage consumption, may be important factors in balancing caloric intake and maintaining a healthy weight (4). Specific meal-related practices such as skipping breakfast and the number of meals and snacks eaten per day may also be related to weight (5– 8). Interventions to change dietary behaviors must target modifiable factors known to influence them (9,10). However, little is known about the correlates of diet among 8- to 10-year-old African-American girls to assist in designing targeted interventions. Environmental and parental influences, as well as children’s attitudes toward food, were associated with children’s food consumption (11–13). Home food availability and accessibility, and cooking or other food preparation practices, influenced children’s consumption (11–13). Parent-reported barriers to their families eating fruit, juice, and vegetables were negatively related to fruit consumption reported by 9- to 12-year-old children (12). Children also

OBESITY RESEARCH Vol. 12 Supplement September 2004

Anthropometric and Psychosocial Correlates of Diet, Cullen et al.

tended to consume only what they enjoyed or preferred (14). The Girls Health Enrichment Multisite Studies program, sponsored by the National Heart, Lung, and Blood Institute, targeted weight gain prevention among 8- to 10-year-old African-American girls. Four field centers in the United States developed separate obesity prevention interventions for pilot testing. Common questionnaires were used to assess diet-related psychosocial factors such as home food preparation techniques, barriers, preferences, and availability of target foods in a pilot study to test the feasibility of unique interventions. Each participating girl and her parent or caregiver at three of the field centers completed these questionnaires at baseline, and the girls also reported 2 days of dietary intake. This paper presents exploratory analyses of the anthropometric, parental, and psychosocial measures and specific meal practices that were hypothesized to be associated with dietary intake.

Research Methods and Procedures Study Design Pilot feasibility studies were conducted at four field centers, located at the Baylor College of Medicine, Houston, TX; University of Memphis, Memphis, TN; University of Minnesota, Minneapolis, MN; and Stanford University, Palo Alto, CA. George Washington University Biostatistics Center was the coordinating center and provided support and coordination for key study activities. The National Heart, Lung, and Blood Institute also participated in the study. A total of 210 African-American girls, 8 to 10 years of age, were randomly assigned to an active intervention (N ⫽ 115) or comparison group (N ⫽ 95). Each field center designed and developed its own 12-week active and control intervention. All four field centers measured height, weight, and dietary intake using a common protocol. However, only three of the field centers completed some of the diet-related psychosocial questionnaires. The results of the baseline assessment for these three centers were used in this paper. All parents or guardians gave written informed consent, and girls gave their assent to participate in the study. Each field center’s study was approved by its institutional review board. Other details of the studies have been described elsewhere (15). Data from the baseline measures were used in this paper. Anthropometric Measures Body weight was measured using a calibrated scale (Seca 770 model scale; Vogel and Halke, Hamburg, Germany), and height was measured using a stadiometer (Shorr Height Measuring Board, Olney, MD). BMI (kilograms per meter squared) was computed for girls. Parental height and weight were also measured, and BMI was calculated.

Child-reported Measures 24-Hour Dietary Recalls. At baseline, each girl completed two 24-hour dietary recalls, 1 to 2 weeks apart. Details of the assessment methods have been presented elsewhere (16). Recalls were averaged across the 2 days and were analyzed for total energy intake (reliability ICC1 ⫽ 0.43), percent of energy from fat (reliability ICC ⫽ 0.13), servings of fruit (reliability ICC ⫽ 0.17), servings of juice (reliability ICC ⫽ 0.16), servings of vegetables (reliability ICC ⫽ 0.06), water (reliability ICC ⫽ 0.38), and sweetened beverage intake (reliability ICC ⫽ 0.04) (17). Vegetable consumption did not include French fries (17). Meal patterns were coded as the number of meals, snacks, and breakfasts eaten each day, averaged over the two recalls. Sweetened Beverage Preferences. Girls at the Baylor and Memphis field centers completed the sweetened beverage preference scale. This scale had acceptable internal consistency (Cronbach’s ␣ of 0.71) and 12-week test-retest reliability (ICC of 0.79) (17). Higher values represented higher preferences for sweetened beverages. Water preference was assessed with one question at the Baylor and Memphis field centers. Higher values represented higher preferences for water. Social Desirability. Social desirability, a form of response bias, was assessed with the “Lie Scale” from the Revised Children’s Manifest Anxiety Scale (18). The modified subscale consisted of eight items assessing socially desirable behaviors and was coded as “yes” or “no” to each query. A higher score represents higher social desirability. The scale was originally formed on a large sample of diverse children, 6 to 19 years of age, from 13 states in the United States and showed reliability and validity across age, sex, and race/ethnicity. In the Girls Health Enrichment Multisite Studies sample of 8- to 10-year-old AfricanAmerican girls from the Baylor and Memphis field centers, the internal consistency of the social desirability score was substantial with a Cronbach’s ␣ of 0.78 (19). Parent-reported Measures Sociodemographics. Family income, material possessions, and highest educational level of the household were reported by parents on demographic questionnaires (20). Home Barriers to Eating. Parents/caregivers of girls from the Baylor and Memphis field centers completed a 20-item questionnaire about barriers to preparing, serving, and consuming fruit, juice, vegetables, and low-fat foods. The questionnaire yielded two factors: low-fat food barriers and fruit, juice, and vegetable barriers. Internal consistency, assessed with Cronbach’s ␣, was 0.80 for both factors. The 12-week test-retest reliabilities were substantial with ICCs of 0.71 for low-fat food and 0.80 for fruit, juice, and vegetable barriers

1

Nonstandard abbreviations: ICC, intraclass correlation.


21S


(16). Higher scale values represent higher perceived barriers. Food Preparation for Daughters. Two factors, low-fat food preparation and high-fat food preparation, were obtained from the Food Preparation for Daughters’ questionnaire completed by Memphis and Minnesota parents or caregivers. Internal consistency was 0.66 for low-fat food preparation practices and 0.58 for high-fat food preparation practices. The 12-week test-retest reliabilities were 0.69 and 0.66 for low-fat and high-fat food preparation practices, respectively (16). Higher values indicated a greater number of low-fat or high-fat food preparation practices. Fruit, Juice, and Vegetable Availability. Parents or caregivers from the Baylor and Memphis field centers identified (yes/no) which of 4 juices, 17 fruits, and 18 vegetables were present in the home environment in the past week. Internal consistency for the fruit, juice, and vegetable availability scale was substantial (␣ ⫽ 0.77) (17). The 12-week testretest reliability was moderate at 0.50. Higher values indicated greater availability of fruits, juice, and vegetables in the home. Fruit, Juice, and Vegetable Accessibility. Parents or caregivers from the Baylor and Memphis field centers were asked to identify (yes/no) whether five fruits, juices, and vegetables were easily accessible to children in the home environment in the past week. Internal consistency for the fruit, juice, and vegetable accessibility scale was moderate (␣ ⫽ 0.45) (17). The 12-week test-retest reliability was poor at 0.10. Higher values represented greater accessibility of fruit, juice, and vegetables in the home. Low-Fat and Fat-free Food Availability. Home availability of high-fat foods and low-fat and fat-free food alternatives in the last 2 weeks was indicated (yes/no) with 32 food items at the Memphis and Minnesota field centers. Internal consistency was substantial for the high-fat food availability (␣ ⫽ 0.64) scale and moderate for the low-fat scale (Cronbach ␣ ⫽ 0.59). The 12-week test-retest reliabilities were moderate at 0.44 for high-fat and 0.60 for low-fat food availability (17). Higher values represented greater availability of low-fat and fat-free foods. Statistical Analyses Descriptive statistics were calculated for demographic, anthropometric, psychosocial, dietary behavior, dietary consumption, and meal pattern variables. Pearson correlation coefficients were calculated among all variables. Regression models were calculated separately for the following dependent variables: total energy intake, percent energy from fat, and fruit, juice, vegetable, sweetened beverage, and water servings per day. For each food/nutrient, separate hierarchical stepwise regression analyses (with backward deletion and p ⬎ 0.05 to delete terms) were conducted. The independent variables were modeled in blocks (models 1 to 4): 1) sociodemographics (girls’ age, family income, highest 22S

household education level), 2) parent and child completed psychosocial variables, 3) meal pattern variables, and 4) BMI. Backward deletion was conducted after adding each group of variables and before adding the next group. Because of possible bias caused by the relationship between social desirability and each covariable, social desirability was entered as a potential covariate in each model. Terms for field center and field center-by-independent variable interactions were added to the multivariate models. If not significant, these terms were dropped from the analyses. Demographics and energy consumption were also included in models 2 to 4 to control for potential confounders in the models. Because each psychosocial questionnaire was not completed at each center, but all three field centers had complete data for demographics, meal patterns, and BMI, the first analyses were conducted on girls from all three centers and only included these variables. The remaining analyses included the psychosocial variables as a block for specific fields centers. Only final models with total R2 ⬎0.10 are presented. Models with R2 ⬍0.10 are left blank or deleted from the tables. All analyses were conducted using Statistical Package for Social Sciences (SPSS version 11.0 for Windows; SPSS Inc., Chicago, IL).

Results Demographic characteristics of the participating girls and descriptive statistics for the psychosocial variables and dietary behaviors are presented in Table 1. The number of snacks consumed per day was related to higher energy intakes (p ⬍ 0.05; Table 2). A significant positive correlation was found between age of girls and percentage energy from fat (p ⬍ 0.01), and there was a negative relationship between percentage energy from fat and eating breakfast (p ⬍ 0.05). Lower juice (p ⬍ 0.05) and vegetable consumption (p ⬍ 0.01) was correlated with higher BMI. Juice consumption was negatively related to fruit, juice, and vegetable accessibility (p ⬍ 0.01). Sweetened beverage consumption was significantly positively correlated with highest education in the household (p ⬍ 0.05), greater high-fat food availability (p ⬍ 0.01), and high-fat food preparation practices (p ⬍ 0.05), and negatively correlated with greater low-fat food preparation practices (p ⬍ 0.01). Positive relationships with water consumption were found for BMI (p ⬍ 0.01), eating breakfast (p ⬍ 0.05), and bottled water preference (p ⬍ 0.05). The first block of regression analyses included demographic variables, meal patterns, and BMI for girls at the three field centers (Table 3). Only three dietary consumption variables had R2 ⬎0.10 (models 2 and 3 in Table 3). Memphis (p ⬍ 0.05) and Minnesota (p ⬍ 0.001) girls consumed less energy than Baylor girls. The number of meals (p ⬍ 0.001) and snacks consumed (p ⬍ 0.01) was



Table 1. Baseline demographic characteristics, anthropometric measures, dietary consumption, psychosocial variables, and meal patterns of the participating girls (n ⫽ 150)* Variable type

Variables

Age (years)

7 years old 8 years old 9 years old 10 years old Maximum household High school graduate or less education Technical school or some college College graduate or post graduate Household income ⬍$20,000 $20,000 to $40,000 ⬎$40,000 Anthropometric BMI (kg/m2) Dietary consumption Total energy† Percent energy from fat Fruit (svg)† 100% fruit juice (svg)† Vegetables (svg)† Total FJV (svg)† Sweetened beverages (svg)† Water (svg)† Psychosocial Home availability of FJV (Baylor, Memphis) Home accessibility of FJV (Baylor, Memphis) Regular/high-fat food availability (Baylor, Memphis) Low-fat food availability† (Memphis, Minnesota) Low-fat food practices (Memphis, Minnesota) High-fat food practices† (Memphis, Minnesota) Low-fat food barriers (Baylor, Memphis) FJV barriers† (Baylor, Memphis) Sweetened beverages preference (Baylor, Memphis) Bottled water preference (Baylor, Memphis) Social desirability (Baylor, Memphis) Meal patterns Mean number of meals/day Mean number of snacks/day Ate breakfast both days

N

Percentage

1 82 33 33 35 61 49 38 56 53 155 147

0.7 55.0 22.1 22.1 24.1 42.1 33.8 25.9 38.1 36.1

149 148 147 147 148 148 145 95 95

Min

12.70

Max

45.80

Mean

22.32

SD

6.10

413.1 3022.0 1544.0 481.0 16.98 50.24 34.38 5.79 0 1.91 0.33 0.45 0 2.46 0.65 0.57 0 3.03 0.83 0.76 0 5.66 1.88 1.24 0 3.91 1.15 0.86 0 3.25 0.73 0.68 0.11 0.79 0.45 0.15 0

1

0.71

0.23

114

0.13

0.93

0.55

0.16

112

0

0.41

0.12

0.10

114

1.25

2.83

1.99

0.34

114 95 94

1.43 1 1

3 2.78 2.11

2.40 1.57 1.28

0.32 0.43 0.31

95 95 95 162 162 162

1.40 1 0.00 1 0 0

3.00 3 1.00 4.0 3.5 1

2.44 2.61 0.48 2.8 1.5 0.8

0.35 0.66 0.30 0.5 0.8 0.4

* Not every girl or parent completed every measure. † Outliers (values ⬎ 3 SD) have been removed; Dietary consumption, number of meals, and snacks were averaged across 2 days of recall.


23S


Table 2. Pearson correlations between nutrient and food intake and demographic characteristics, anthropometric measures, and psychosocial and meal pattern variables

Demographics Age of girl Highest education: all Total income Material possessions–8 items Anthropometric BMI (girl) Meal patterns Number of meals Number of snacks Ate breakfast both days Home environment measures Home availability of FJV

Energy

Percent energy as fat

Fruit Juice Vegetable servings servings servings

⫺0.149 147 0.096 145 0.049 143 ⫺0.042 142

0.244* 149 ⫺0.076 147 ⫺0.049 145 0.040 144

⫺0.154 148 0.116 146 ⫺0.063 144 ⫺0.005 143

Corr 0.024 N 146

0.111 148

⫺0.148 ⫺0.172† 147 146

Corr 0.185 N 147 Corr 0.252† N 147 Corr 0.211 N 147

Corr N Corr N Corr N Corr N

Corr N Home accessibility of FJV Corr N Regular/high-fat food availability Corr N Home/low-fat food availability Corr N Mean home low-fat food practice Corr N Mean home high-fat food practice Corr N Psychosocial variables Mean low-fat food barriers Corr N Mean FJV barriers Corr N Sweetened beverage preferences Corr N Bottled water preference Corr N Social desirability Corr N

⫺0.104 147 ⫺0.047 145 ⫺0.050 143 ⫺0.093 142

⫺0.126 148 0.183† 146 0.110 144 0.155 143

0.037 145 ⫺0.083 143 0.002 142 ⫺0.101 140

⫺0.211* 146

0.081 147

0.239* 145

⫺0.112 0.009 0.065 149 148 147 ⫺0.029 0.017 0.071 149 148 147 ⫺0.166† 0.151 0.090 149 148 147

⫺0.003 147 0.030 147 ⫺0.033 147

⫺0.094 148 ⫺0.085 148 0.140 148

0.129 145 0.141 145 0.176† 145

⫺0.009 93 ⫺0.154 93 0.124 114 ⫺0.107 112 ⫺0.054 114 0.134 114

0.016 95 ⫺0.059 95 0.142 114 ⫺0.144 112 ⫺0.225 114 0.242 114

⫺0.156 95 0.027 95 ⫺0.067 113 0.023 111 0.068 113 ⫺0.057 113

⫺0.163 95 ⫺0.285* 95 0.029 112 0.072 110 0.144 112 0.053 112

⫺0.017 93 ⫺0.113 93 0.008 113 ⫺0.050 111 0.005 113 0.150 113

⫺0.058 94 ⫺0.094 94 0.227* 114 ⫺0.141 112 ⫺0.196† 114 0.197† 114

⫺0.106 92 ⫺0.096 92 ⫺0.042 112 0.070 110 ⫺0.008 112 ⫺0.152 112

0.041 93 0.057 92 ⫺0.081 93 0.091 93 0.171 93

0.141 95 ⫺0.018 94 ⫺0.118 95 0.056 95 ⫺0.007 95

0.158 95 0.017 94 ⫺0.152 95 0.190 95 0.105 95

⫺0.035 95 0.150 94 0.000 95 0.062 95 0.077 95

0.152 93 0.088 92 0.104 93 ⫺0.011 93 0.085 93

⫺0.023 94 0.042 93 0.121 94 0.009 94 ⫺0.064 94

⫺0.114 92 ⫺0.054 91 0.024 92 0.251† 92 ⫺0.168 92

* Correlation is significant at the 0.01 level (two-tailed). † Correlation is significant at the 0.05 level (two-tailed).

24S

⫺0.068 147 ⫺0.083 145 ⫺0.018 143 ⫺0.127 142

Sweetened beverage Water servings servings



Table 3. Results from hierarchical block regression (and backward deletion) for dietary consumption (energy intake, percent fat, fruit, juice, vegetables, sweetened beverages, water) with independent variables demographic (age, family income, highest household education, material possessions, field center), meal patterns (no. of meals, no. of snacks, ate breakfast both days), and BMI among 8- to 10-year-old African-American girls at the Memphis (n ⫽ 60), Minnesota (n ⫽ 54), and Baylor (n ⫽ 35) field centers Model 2 meal patterns (plus demographics) Dependent variables Energy intake Field center: Memphis* Field center: Minnesota* Number of meals Number of snacks

Model 3 BMI (plus demographics)

Std.

t

Sig.

⫺0.25 ⫺0.37 0.28 0.23

⫺2.57 ⫺3.69 3.66 2.97 R 2 ⫽ 0.184

0.011 0.000 0.000 0.004

Vegetables Field center: Memphis* Field center: Minnesota* BMI Household education Sweetened beverages Field center: Memphis* Field center: Minnesota* Number of snacks Field center: Memphis-snacks interaction* Field center: Minnesota-snacks interaction*

0.05 0.15 0.64 ⫺0.45 ⫺0.85

0.25 0.70 3.93 ⫺2.30 ⫺3.54 R 2 ⫽ 0.252

Std.

t

Sig.

⫺0.28 ⫺0.41 ⫺0.26 0.17

⫺2.66 ⫺3.69 ⫺3.17 2.04 2 R ⫽ 0.136

0.021 0.001 0.002 0.043

0.799 0.483 0.000 0.023 0.001

* Baylor was referent.

positively associated with energy consumption (model 2). Model 2 explained 18.4% of the total variance in energy consumption. Controlling for field center differences, lower BMI (p ⬍ 0.01) and higher household education (p ⬍ 0.05) were associated with greater vegetable consumption, explaining 13.6% of the total variance (model 3). Baylor girls reported higher consumption of vegetables than Memphis (p ⬍ 0.05) and Minnesota girls (p ⬍ 0.01). For sweetened beverage consumption, the interaction terms of field centers by number of snacks consumed (Memphis, p ⬍ 0.05; Minnesota, p ⬍ 0.01) were significant, and thus, were retained as factors; the final model explained 25.2% of the total variance (model 2). A greater number of snacks was associated with higher levels of reported sweetened beverage consumption for the Baylor girls relative to low correlations in the other centers.

The results from the hierarchical block regression analysis, with the variables available from the Memphis and Minnesota field centers, are presented in Table 4. Only models for energy intake and percentage energy from fat had R2 ⬎0.10. Model 3 explained 14% of the total variability. Higher number of meals per day (p ⬍ 0.05) was positively associated with energy intakes, and the interaction term of field center by high-fat food preparation practice (p ⬍ 0.01) was significant in model 3. For Memphis participants, greater high-fat food preparation practice scores were related to higher total energy intake, whereas Minnesota girls reported little difference in energy intake at any point of the high-fat food scale. Age (p ⬍ 0.05) and field center (p ⬍ 0.01) were significant correlates of percentage energy for fat (model 2), whereas low-fat food preparation practices was a negative correlate (p ⬍ 0.05; R2 ⫽ 0.17; model 2).


25S


Table 4. Results from hierarchical block regression analyses (and backward deletion) for dietary consumption (energy intake, percent fat, fruit, juice, vegetables, sweetened beverages, water) with independent variables demographic (age, family income, highest household education, material possessions, field center), psychosocial (low-fat and high-fat food availability and practices), meal pattern (no. of meals, no. of snacks, ate breakfast both days), and anthropometric (BMI) among 8- to 10-year-old African-American girls at Memphis (n ⫽ 60) and Minnesota (n ⫽ 54) field centers Model 2 psychosocial (low-fat and high-fat food availability and practices) (plus demographics) Dependent variables

Std.

t

Model 3 meal patterns (plus demographics) Sig.

Energy intake Field center* Mean high-fat food practice Field center ⫻ high-fat food practice Number of meals Percent calories from fat Age Field center* Mean low-fat food practice

0.22 0.24 ⫺0.19

2.57 2.79 ⫺2.20 R 2 ⫽ 0.171

Std.

t

Sig.

⫺1.33 ⫺0.04 1.41 0.29

⫺1.94 ⫺0.35 2.01 3.22 R 2 ⫽ 0.141

0.055 0.726 0.047 0.002

0.011 0.006 0.030

* Baylor was referent.

The results from the Baylor and Memphis models are found in Table 5. All dietary variables except fruit consumption had R2 ⬎0.10. Field center (Memphis reported lower energy; p ⬍ 0.05), numbers of meals (p ⬍ 0.01), and snacks consumed (p ⬍ 0.01) were significantly associated with higher energy intake, accounting for 24% of the total variance in the model (model 3). For percentage energy from fat, the model explained 28% of the total variance. Three interaction terms were significant: field center by “like water a lot” (p ⬍ 0.01), field center by number of meals (p ⬍ 0.05), and field center by eating breakfast (p ⬍ 0.05; model 3). For girls from the Baylor field center, lower preference for water was correlated with higher percentage energy from fat. There was little association between percentage energy from fat and water preference for girls from Memphis. Percent energy from fat was higher for Baylor girls than Memphis girls, regardless of the number of meals or breakfasts eaten. Higher education (p ⬍ 0.05) was positively associated with juice consumption (model 2). Fruit, juice, and vegetable accessibility was negatively associated with juice consumption (p ⬍ 0.02). Two interaction terms were significant correlates of juice consumption: field center by fruit, juice, 26S

and vegetable availability (p ⬍ 0.05) and field center by fruit, juice, and vegetable barriers (p ⬍ 0.05). Model 2 accounted for 24% of the total variability in juice consumption. At mean levels of fruit, juice, and vegetable availability, juice consumption was higher for Memphis girls compared with Baylor girls. In contrast, at mean levels of fruit, juice, and vegetable barriers, juice consumption was higher for Baylor girls. There were only two significant correlates of vegetable consumption, accounting for 20% of the model variance (model 4). Consumption of vegetables was lower for Memphis vs. Baylor girls (p ⬍ 0.05) and was higher at lower BMI levels (p ⬍ 0.05). Twenty-seven percent of the variance in sweetened beverage consumption was explained by model 3. The interaction term of field center by number of snacks (p ⬍ 0.01) was significant. As number of snacks increased, sweetened beverage intake increased for Baylor girls but not Memphis participants. The model for water consumption (model 2, Table 5) accounted for 21% of the total variance. Low-fat food barriers (p ⬍ 0.05) were a negative correlate of water



Table 5. Results from hierarchical block regression (and backward deletion) with dietary consumption (energy intake, percent fat, fruit, juice, vegetables, sweetened beverages, water) with independent variables demographic (age, family income, highest household education, material possessions, field center), psychosocial (availability, accessibility, barriers, and preferences), meal patterns (no. of meals, no. of snacks, ate breakfast both days), and anthropometric (BMI) among 8- to 10-year-old African-American girls at the Memphis (n ⫽ 60) and Baylor (n ⫽ 35) field centers Model 2 psychosocial (availability, accessibility, barriers, and preferences) (plus demographics) Dependent variables

Std.

t

Sig.

Energy Intake Field center* Number of meals

Model 4 BMI (plus demographics)

Std.

Std.

t

Sig.

⫺2.61 ⫺0.97 1.44 ⫺0.31 0.26 2.07 ⫺0.67 R2

Sig.

0.23 0.10 ⫺0.49 ⫺0.26 0.38 0.94 ⫺0.99

2.22 0.19 ⫺2.36 ⫺2.39 2.43 2.36 ⫺2.34 R 2 ⫽ 0.238

⫺2.96 0.004 ⫺2.28 0.025 2.91 0.005 ⫺1.55 0.126 1.14 0.256 2.32 0.023 ⫺2.07 0.042 ⫽ 0.280

0.029 0.850 0.020 0.019 0.017 0.021 0.022

Vegetables Field center* BMI

⫺0.25 ⫺2.48 0.015 ⫺0.26 ⫺2.65 0.010 R 2 ⫽ 0.198

Sweetened beverages Field center* Number of snacks Field center ⫻ snacks* Water Field center* Mean low-fat food barriers Sweetened beverages preference Likes water a lot† Field center ⫻ sweetened beverage preference*

t

⫺0.23 ⫺2.42 0.017 0.31 3.28 0.001 R 2 ⫽ 0.241

Percent energy from fat Field center* Likes water a lot* Field center ⫻ likes water a lot* Number of meals Ate breakfast every day Field center ⫻ number of meal* Field center ⫻ ate breakfast* Juice Highest education: all Field center* Mean availability of FJV Mean accessibility of FJV Mean FJV barriers Field center ⫻ FJV availability* Field center ⫻ FJV barriers*

Model 3 meal patterns (plus demographics)

0.05 0.24 0.809 0.54 3.73 0.000 ⫺0.46 ⫺2.19 0.031 R 2 ⫽ 0.273 ⫺1.77 ⫺0.23 ⫺0.24 0.38 2.01

⫺2.50 ⫺2.24 ⫺1.70 2.45

0.014 0.028 0.093 0.016

2.88 0.005 R 2 ⫽ 0.206

* Baylor was referent. † Water referent: doesn’t like water.


27S


consumption (p ⬍ 0.05). A higher preference for water was positively associated with water consumption (p ⬍ 0.05). The relationship between water consumption and sweetened beverage preference varied by field center (p ⬍ 0.05). At average levels of sweetened beverage preference, Memphis girls reported higher water consumption than Baylor girls. The social desirability measure was not significant in any model with any model.

Discussion This study explored the relationships among anthropometric, demographic, and psychosocial variables and meal behaviors in 8- to 10-year-old African-American girls. It examined whether girls’ BMI; parent-reported fruit, juice, vegetable, and low-fat food availability and accessibility, barriers to eating fruit, juice, and vegetables, and low- and high-fat food preparation practices; and girl-reported preferences for sweetened beverages and water were related to girls’ energy intake, percent energy from fat, fruit, juice, vegetable, water, and sweetened beverage consumption. Complex relationships were found that differed across geographic regions. The energy balance equation suggests that excess adiposity is the result of excess energy intake over expenditure; therefore, correlates of energy intake should have implications for obesity prevention efforts. After controlling for demographic and location of field center confounders, the number of meals and snacks consumed were related to energy intake in all three regression models, accounting for 14% to 24% of the variance. For Memphis girls in the Memphis-Minnesota model, parent-reported high-fat food preparation practices were significant and accounted for 14% of the variance (model 2, Table 4). Low-fat food preparation practices were negatively related to percentage energy from fat for the Minnesota and Memphis girls, accounting for 28% of the variance. Surprisingly, neither BMI nor education was related to energy intake. Data on the relationships between intake of energy and fat-related food practices among children are limited, making conclusions difficult. These results suggest regional differences in food preparation practices. Data from this study are consistent with results from 10- to 14-year-old African-American boys in the Houston area, where total high-fat food practices was significantly correlated with total energy intake (r ⫽ 0.46, p ⬍ 0.001) (25). The relationship between energy intake, family food preparation practices, and number of meals and snacks should be explored further, because these behaviors may be important targets for future obesity prevention interventions. Dietary fat intake has been directly related to BMI in children (26,27). After controlling for demographic confounders and field center, variables related to percentage of energy from fat included girls’ age (positive), greater fre28S

quency of low-fat food preparation practices (negative; model 2, Table 4), and interaction terms with field center, number of meals and breakfasts eaten, and water preference (model 3, Table 5). In previous work with 10- to 14-yearold African-American boys, low-fat practices were inversely correlated with percentage of energy from fat (r ⫽ ⫺0.19, p ⬍ 0.05) (25). Low-fat practices was significantly negatively correlated with consumption of fat as a percentage of total energy (r ⫽ ⫺0.15, p ⬍ 0.01), extracted from food records as opposed to questionnaires, from fourth to sixth grade children (28). Among a group of 9- to 12-yearold children, including 20% African Americans, child-reported low-fat food practices were significantly related to parent-reported low-fat food practices (29). The interaction terms in the model for percentage energy from fat (model 3, Table 5) are difficult to explain. The results suggest that perhaps there are regional differences in breakfast meals (i.e., higher in fat) consumed by Baylor girls compared with Memphis girls. Although dietary fat consumption as a percentage of energy was similar for Baylor and Memphis girls consuming the mean number of meals, it was higher for Baylor girls and lower for Memphis girls with fewer meals. The relationship between water preference and percentage energy from fat is difficult to explain. Perhaps Baylor girls who reported lower preference for water consumed beverages with more fat at mealtimes (e.g., milk). Further understanding of meal composition is important and could provide targets for future interventions. Girls whose parents reported higher education reported greater juice consumption (model 2, Table 5), and higher parent-reported fruit, juice, and vegetable accessibility was associated with lower juice consumption. The interaction terms revealed that the level of fruit, juice, and vegetable barriers and availability made no difference in juice consumption among Memphis girls. However, among Baylor girls, juice consumption was higher for lower levels of fruit, juice, and vegetable availability but higher for those with greater fruit, juice, and vegetable barriers. In previous work, fruit, juice, and vegetable availability and accessibility were positively related to fruit, juice, and vegetable consumption (13,14). The negative relationship may be related to consumption of fruit juice at school breakfasts. After controlling for demographic confounders, energy intake, and field center, girls with lower BMI and parents with higher education reported higher vegetable consumption; however, vegetable consumption differed by field center (model 3, Table 3; model 4, Table 5). The relationship between lower BMI and higher vegetable consumption is of great interest. The finding supports the rationale of the national 5-A-Day campaign: eating a greater amount of fibrous fruits and vegetables may displace foods higher in fat and energy and potentially contribute to lower body weight (21). A few previous studies identified a positive



relationship between fruit and vegetable consumption and leaner body mass (22,23). Among women with multiethnic backgrounds, vegetable consumption calculated from a food frequency questionnaire was negatively related to BMI (24). BMI was used as an independent variable in these analyses, but these relationships may reflect the reverse relationship. Our results also suggest that vegetable intake may be influenced by education and location of study as seasonal changes between and within field centers could affect vegetable availability, pricing, and consumption. The results also suggest that regional and educational level may influence vegetable intake. This is an important area that needs further research, particularly with longitudinal studies that include dietary assessment and measures of BMI in diverse populations. Surprisingly, no significant relationships were found for fruit consumption. Recent studies have documented that fruit consumption was positively related to home availability of fruit, juice, and vegetables and negatively related to fruit, juice, and vegetable barriers (11,12,14). Fruit consumption has been found to be higher for those reporting lower BMI (30). Further exploration of this relationship is important for obesity prevention efforts. Perhaps the reliability from 2 days of assessment did not allow detection of these relationships. Higher levels of sweetened beverage intake have been related to higher BMI (4) and increased energy intake (31). After controlling for demographic confounders and field center, the major variables associated with sweetened beverage consumption were interaction terms of field center by number of snack meals (model 2, Table 3; model 3, Table 5). Greater snacking was associated with higher sweetened beverage consumption for Baylor girls, but there were smaller differences by number of snacks for Memphis girls. These field center differences could reflect regional differences in beverages consumed as snacks. However, common snacks included sweetened beverages along with high-fat and high-salt food items, and snacking rates are generally high among youth (32). Because of the relationship between sweetened beverages, energy intake, and obesity (4), this is an area that deserves further research. It would be important to determine whether reducing snacks, altering beverage choice, or both are appropriate targets for obesity prevention programs. Some studies suggest that people who drink more water should feel greater satiety, thereby eating less and having a lower BMI (33). After controlling for demographic confounders and field center, water consumption was significantly negatively related to low-fat food barriers and positively related to water preference, with a field center by sweetened beverage preference interaction term (model 2, Table 5). For Baylor and Memphis girls, lower water consumption was related to greater low-fat barriers. Perhaps

families that perceive more barriers to eating a healthier diet (e.g., obtaining low-fat foods) do not encourage water consumption. No previous reports are available for comparison with our findings. In previous work with fruit, juice, and vegetable consumption, higher preferences were related to greater consumption (10,14). Surprisingly, Memphis girls with higher sweetened beverage preferences reported the highest water consumption. This relationship is difficult to explain but may reflect reduced sweetened beverage availability in the home. Further research to explore ways to increase water consumption and reduce preference for sweetened beverages among youth are needed. Several limitations of this paper should be noted. First, all data were from parental and girl self-report and thereby subject to possible attention, comprehension, memory, and recording errors. However, dietary consumption was reported by girls, and the majority of the psychosocial questionnaires were reported by the parent, thereby minimizing common response bias accounting for the correlations with diet. Second, 2 days of 24-hour food recalls are known to estimate usual fruit, juice, and vegetable intake with low reliability. The poor reliability of the dietary intake measures minimizes the ability to detect relationships with other variables and likely accounts for the small number of statistically significant correlations between the psychosocial scales and dietary intake variables. Small sample size may also have hampered our ability to detect statistically significant relationships. Although numerous significant associations were found, many were low in magnitude (⬍0.20) and need to be interpreted cautiously. Finally, several questionnaires had reliabilities less than the acceptable 0.70, which could reduce our ability to detect significant relationships. The strengths of this study include the evaluation of innovative measures to assess family-based attitudes and practices related to dietary behaviors. These data were obtained from 8- to 10-year-old African-American girls at risk for weight gain and provide a unique look at an understudied group. With the burgeoning rates of obesity in minority populations, research to identify influences on dietary behaviors is critical to forming appropriate prevention approaches. In conclusion, dietary behavior is complex and differs by geographic areas. Low-fat food preparation practices in the home seemed to be an important influence on the percentage of energy consumed from fat. Greater vegetable consumption was associated with lower BMI. The results of this study with 8- to 10-year-old African-American girls identified several relationships that should be explored further. In particular, the relationships between vegetable consumption and BMI and between consumption of energy and fat and home low-fat food preparation practices deserve further study with other populations.


29S


Acknowledgments This work is a publication of the USDA/ARS Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine and Texas Children’s Hospital, Houston, TX. This research was funded grants U01 HL62662, U01 HL62663, U01 HL62668, U01 HL62732, and U01 HL65160 from the National Heart, Lung, and Blood Institute, NIH. This project has also been funded, in part, by federal funds from the USDA/ARS under Cooperative Agreement 58-6250-6001. The contents of this publication do not necessarily reflect the views or policies of the USDA, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. References 1. Ogden CL, Flegal KM, Carroll MD, Johnson CL. Prevalence and trends in overweight among US children and adolescents, 1999 –2000. J Am Med Assoc. 2002;288:1728 –32. 2. Klesges RC, Klesges LM, Eck LH, Shelton ML. A longitudinal analysis of accelerated weight gain in preschool children. Pediatrics. 1995;95:126 –32. 3. Davidson KK, Birch LL. Child and parent characteristics as predictors of change in girls’ body mass index. Int J Obes Relat Metab Dis. 2001;25:1834 – 42. 4. Ludwig DS, Peterson KE, Gortmaker SL. Relation between consumption of sugar-sweetened drinks and childhood obesity: a prospective, observational analysis. Lancet. 2001;357: 505– 8. 5. Amosa T, Rush E, Plank L. Frequency of eating occasions reported by young New Zealand Polynesian and European women. Pac Health Dialog. 2001;8:59 – 65. 6. Wahlqvist ML, Kouris-Blazos A, Wattanapenpaiboon N. The significance of eating patterns: an elderly Greek case study. Appetite. 1999;32:23–32. 7. Summerbell CD, Moody RC, Shanks J, Stock MJ, Geissler C. Relationship between feeding pattern and body mass index in 220 free-living people in four age groups. Eur J Clin Nutr. 1996;50:513–9. 8. Ohlin A, Rossner S. Factors related to body weight changes during and after pregnancy: the Stockholm Pregnancy and Weight Development Study. Obes Res. 1996;4:271– 6. 9. Baranowski T, Lin LS, Wetter DW, Resnicow K, Hearn MD. Theory as mediating variables: why aren’t community interventions working as desired? Ann Epidemiol. 1997;7: S89 –95. 10. Baranowski T, Cullen KW, Baranowski J. Psychosocial correlates of dietary intake: advancing intervention. Annu Rev Nutr. 1999;19:17– 40. 11. Cullen KW, Baranowski T, Rittenberry L, Cosart C, Hebert D, de Moor C. Socio-environmental influences on children’s fruit, juice, and vegetable consumption as reported by parents: reliability and validity of measures. Publ Health Nutr. 2000;3:345–56. 12. Cullen KW, Baranowski T, Rittenberry L, Cosart C, Hebert D, de Moor C. Child-reported social-environmental 30S

13.

14.

15. 16.

17.

18.

19.

20. 21.

22.

23.

24.

25.

26.

27.


influences on children’s fruit, juice, and vegetable consumption: reliability and validity of measures. Health Educ Res. 2001;16:187–200. Hearn M, Baranowski T, Baranowski J, et al. Environmental influences on dietary behavior among children: availability and accessibility of fruits and vegetables enable consumption. J Health Educ. 1998;29:26 –32. Resnicow K, Baranowski T, Hearn M, et al. Social-cognitive predictors of fruit and vegetable intake in children. Health Psychol. 1997;16:272– 6. Rochon J. Common elements of the GEMS pilot studies. Ethnicity Dis. 2003;13(Suppl 1):S6 –14. Cullen KW, Klesges LM, Sherwood NE, et al. Measurement characteristics of diet-related psychosocial questionnaires among African-American parents and their 8- to 10-year-old daughters: results from the Girls’ Health Enrichment Multi-Site Studies. Prev Med. 2004;38(Suppl): S34 – 42. Cullen KW, Himes JH, Baranowski T, et al. Validity and reliability of a behavior-based food coding system for measuring fruit, 100% fruit juice, vegetable, and sweetened beverage consumption: results from the Girls’ Health Enrichment Multisite Studies. Prev Med. 2004;38(Suppl):S24 –33. Reynolds C, Paget K. National normative and reliability data for the Revised Children’s Manifest Anxiety Scale. School Psychol Rev. 1983;12:324 –36. Klesges LM, Baranowski T, Beech B, et al. Social desirability bias in self-reported dietary, physical activity and weight concerns measures in 8- to 10-year-old AfricanAmerican girls: results from the Girls’ Health Enrichment Multisite Studies (GEMS). Prev Med. 2004;38(Suppl): S78 – 87. Baranowski T, Cullen K, et al. Psychometrics of material possessions as indicators of SES. Personal communication. Havas S, Heimendinger J, Reynolds K, et al. 5-a-day for better health: a new research initiative. J Am Dietet Assoc. 1994;94:32– 6. Epstein LH, Gordy CC, Raynor HA, Beddome M, Kilanowski CK, Paluch R. Increasing fruit and vegetable intake and decreasing fat and sugar intake in families at risk for childhood obesity. Obes Res. 2001;9:171– 8. Lloyd T, Chinchilli VM, Rollings N, et al. Fruit consumption, fitness, and cardiovascular health in female adolescents: the Penn State Young Women’s Health Study. Am J Clin Nutr. 1998;67:624 –30. Maskarinec G, Novotny R, Tasaki K. Dietary patterns are associated with body mass index in multiethnic women. J Nutr. 2000;130:3068 –72. Cullen KW, Bishop RG, de Moor C. Fat practices and consumption among African-American adolescent boy scouts: the impact of meal source. Ethnicity Dis. 2002;12:193– 8. Robertson S, Cullen KW, Baranowski T, Baranowski JH, Hu S, de Moor C. Dietary intake, but not physical activity, is related to adiposity in preschool children. J Am Dietet Assoc. 1999;99:938 – 43. Lee Y, Mitshell DC, Smicklas-Wright H, Birch LL. Diet quality, nutrient intake, weight status, and feeding environ-


ments of girls meeting or exceeding recommendations for total dietary fat of the American Academy of Pediatrics. Pediatrics. Available at: http://www.pediatrics.org/cgi/content/full/107/ 6/e95. 28. Cullen K, Lara K, de Moor C. Children’s dietary fat intake and fat practices vary by meal and day. J Am Dietet Assoc. 2002;102:1773– 8. 29. Cullen KW, Lara KM, de Moor C. Familial concordance of dietary fat practices and intake. Family Community Health. 2002;25:65–75. 30. Lin BH, Morrison RM. Higher fruit consumption linked with lower BMI. Food Rev. 2003;25:28 –32

31. Cullen KW, Ash DM, Warneke C, de Moor C. Intake of soft drinks, fruit-flavored beverages, and fruits and vegetables by children in grades 4 through 6. Am J Publ Health. 2002; 92:1475– 8. 32. Zizza C, Siega-Riz AM, Popkin BM. Significant increase in young adults’ snacking between 1977–1978 and 1994 – 1996 represents a cause for concern! Prev Med. 2001;32: 303–10. 33. Lappalainen R, Mennen L, van Weert L, Mykkanen H. Drinking water with a meal: a simple method of coping with feelings of hunger, satiety and desire to eat. Eur J Clin Nutr. 1993;47:815–9.


31S