1. Jennifer Greenberg Seth1
Preschool feeding practices and beliefs among low-income Hispanic WIC parents
Jennifer Greenberg Seth1
Alexandra Evans2
Karol Kaye Harris1
Jennifer Loyo1
Tara Campbell Ray1
Carol Spaulding1
Nell Gottlieb1
1 Department of Kinesiology and Health Education, University of Texas Austin
2 Michael & Susan Dell Center for Advancement of Healthy Living, University of Texas School of Public Health in Austin

2. Purpose
To examine preschooler feeding practices and beliefs among low-income Hispanic parents in Texas
To compare practices and beliefs between Spanish-speaking and English-speaking Hispanics

3. Background
Prevalence of child overweight is especially high among Hispanic children in the U.S.
Some parental feeding practices are associated with children’s food intake and weight status
Cultural context
Feeding child associated with being a good mother
Perception of “chubby” kids as healthy
Level of acculturation is related to dietary practices – to feeding practices as well?
Sources: Hoelscher et al. 2004; Faith et al.; Hodges 2003, Baughcum et al. 2001; Satter 1996; Kaiser et al. 1998; Gonzalez et al. 2006; Brewis 2003; Crawford 2001; Reifsnider et al. 2006; Romero-Gwynn & Gwynn 1997)
In Texas, highest rates of overweight in schoolchildren among Hispanic boys about 30-33%; Hispanic girls – 23-27% (Hoelscher et al 2004).
TX DSHS(2006) also found higher rates of overweight among Hispanics preschoolers in the WIC program.
2) Parental feeding practices include restrictive behaviors and pushing child to eat (Faith et al 2004; Hodges 2003; Baughcum et al 2001; Ellen Satter 1996)
Not much research with this specific population group in the US
Qualitative studies indicate that Hispanic parents may engage in feeding practices that can lead to increased caloric intake among their children.
Kaiser et al - bribery, threats to get their children to eat more
3) Hispanic culture and values may predispose parents to engage in certain practices conducive to increased caloric intake among children
Cultural context: Mexican women associate being a good mother with feeding child; chubby
4) Dietary acculturation described by Romero-Gwynn and Gwynn (1997) = less use of traditional foods, use of traditional foods in new ways and adopting the foods of the host country

4. Methods
Convenience sample of parents (of children ages 1-5) recruited from 7 WIC clinics in Texas
Participants were asked to complete several instruments including:
The Preschooler Feeding Questionnaire
Short demographic survey
Respondent’s age
Monthly income
Main language spoken at home
WIC = USDA – Special Supplemental Program for Women, Infants, and Children
Eligibility = 185% of US Federal Poverty Level
Parents were given $10 for completion of survey
IRB approval from University of Texas and Texas Department of State Health Services
Main language spoken at home was used as a proxy measure of acculturation - (Lara et al review article).

5. Preschooler Feeding Questionnaire (PFQ)
Developed by Baughcum, et al (2001)
32-items measuring parental practices and beliefs
8 subscales:
Difficulty in child feeding
Concern about child overeating or being overweight
Pushing child to eat more
Using food to calm child
Concern about child being underweight
Child’s control of feeding interactions
Structure during feeding interactions
Age inappropriate feeding
5-point Likert scale
How often do you….? “Never” to “Always”
Agree/Disagree statements
Implemented by Baughcum in WIC clinics in Kentucky and pediatricians’ offices in Cincinnati.
Baughcum found an 8-factor structure in her population
Appropriate for use with low literacy populations

6. Preschooler Feeding Questionnaire (PFQ)
Translated into Spanish and back-translated
Field tested in Texas WIC clinics for readability and comprehension
Appropriate for low-literacy populations
For our study, we translated the survey into Spanish…….

7. Analysis Exploratory factor analysis with varimax rotation
Loadings above .40 considered
2 items deleted
8-factor structure emerged
Cronbach’s alpha / Pearson correlation coefficients calculated to determine internal reliability

8. Analysis Means and standard deviations calculated for subscales
Non-Hispanics
All Hispanics
English-speaking Hispanics
Spanish-speaking Hispanics
ANOVA with post hoc analyses, controlling for education
Significance level set at p<0.05
ANOVA to determine differences between non-Hispanics, English- and Spanish-speaking Hispanics

9. Characteristics of participants
265 participants – 30 deleted because of ineligibility
177 Hispanics reported Spanish as main language spoken at home
Mostly women (98%)
40% had monthly household income < $1000

10. Results for PFQ Factor Sample Mean SD Internal Consistency Coefficient
1-Parental worry about weight status
2.89
1.20
.87
2-Difficulty in child feeding
2.30
.75
.80
3-Concern about child overeating
1.62
.77
.72
4-Using food to calm child
1.98
.68
5-Pushing child to eat more
2.39
.90
.64
6-Use of Rewards
3.12
.51
7-Child’s control of feeding
2.69
.86
.40*
8-Unstructured mealtimes
2.84
.93
.20*
Mean range: 1 = never – 5 = always
Factors:
Worry about both overweight and underweight
Difficulty in child feeding includes child is a picky eater, parents have to make special meals, struggle to get child to eat
Concern about overeating includes restrictive behavior, getting upset if child eats too much
Using food to calm child - when child is bored, tired, having a tantrum
Strategies to push child to eat more - finish all food on plate, punish child to get him to eat more, bribe
Use of rewards - parent sits down with the child, make something else if child doesn’t like what is served, offer dessert to get child to eat more
7) Child’s control of feeding- choose food from what is served, eat snacks whenever
8) Unstructured mealtime - TV at meals, no set meal and snack time
ANOVA between 3 groups (non-Hispanics, English-speaking Hispanics, Spanish-speaking Hispanics) with post hoc analyses. No significant differences between non-Hispanics and English-speaking Hispanics.

11. Comparisons between English and Spanish-Speaking Hispanics
English Speakers
Spanish Speakers
F and p values
1-Parental worry about weight status
2.05 (.89)
3.38 (1.08)
70.32**
2-Difficulty in child feeding
2.13 (.78)
2.39 (.72)
4.81*
3-Concern about child overeating
1.43 (.67)
1.72 (.81)
5.76*
4-Using food to calm child
1.83 (.72)
2.06 (.83)
3.63
5-Pushing child to eat more
2.18 (.92)
2.51 (.87)
5.57*
Looking at the top 5 factors with acceptable alphas - significant differences between English- and Spanish-speaking Hispanics.

12. Conclusions and Implications
Acculturation is an important variable to consider
Less acculturated parents engaged in obesity promoting behaviors more often
They were also more concerned
Do these results conflict with other data?
Implications for interventions
It is important to address feeding practices in obesity prevention programs directed at preschoolers.
Use of rewards was feeding practice most often reported
Unstructured meal times also often
Less concern of child overeating and more likely to push children to eat more
In other studies these practices have been associated with child overweight.
Differences between Spanish and English speakers:
Spanish speakers more likely to be worried about children’s weight
Push children to eat
Previous research on acculturation and weight:
More overweight among Hispanics who are more acculturated
Importance of the food environment
increase in weight with acculturation may be related to abandoning the traditional “Latin diet” in favor of more processed foods with refined grains and sugars and large portion sizes.
Foods that are abundant and accessible in Mexico may not be affordable to low-income families in the US
Common feeding practices that are not associated with overweight in Mexico may put children in prison in the context of the American food environment
Implications:
Remember that there is diversity within the US Hispanic population
Focus on feeding practices in the context of the new food environment for immigrants
Sensitivity to cultural norms
May be an environmental influence on the relationship between child feeding and overweight
Likert scale may not be appropriate for these questions – “no one would report ‘always” engaging in the behaviors that are on the PFQ.

13. Acknowledgements
Thanks to the WIC clinics for their help in implementation and to all the WIC clients who completed the surveys!