I. ‘
:53. A\. t
.. .t

.37,
. 112

'\'« u. 2"
. 3.7:...»
: LEVI.
n;a:wa~

.u
r . .3,
. ex... 3:
. 2..)I..‘D

2

3...: I... 1.1.

1: .3 rate... :-

2:3: .. x2 ,1

.2

.i 5!?
y .

 

 

. m . r3: . . .
2. i9... .: .
2.7! . i.
1 ‘

 

 

 

Khalil;

acm>\

 

_ LIBRARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

Achieving Fat and Fiber Recommendations with Foods
Consumed From the USDA School Meal Programs

presented by
Madhuri V. Yagalla

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Human Nutrition

 

Mafior professor
Dr. Sharon . Hoerr

8-16-01

 

Date

MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

|
robin 1220055

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DatoDue.965-n15

ACHIEVING FAT AND FIBER RECOMMENDATIONS WITH
FOODS CONSUMED FROM THE USDA SCHOOL MEAL PROGRAMS

By

Madhuri V. Yagalla

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Dept. of Food Science and Human Nutrition

2001

ABSTRACT
ACHIEVING FAT AND FIBER RECOMMENDATIONS WITH FOODS
CONSUMED FROM THE USDA SCHOOL MEAL PROGRAMS
By

Madhuri V. Yagalla

Objectives: 1) to identify clusters of foods and food components consumed by USDA
school meal program participants by total fat, saturated fat and dietary fiber, 2) to assess
dietary fiber intakes of US. school children by age, gender, ethnicity, family income,
student eligibility for free or reduced price school meals and participation in USDA
school meal programs, 3) to determine % of US. school children consuming vended
foods/beverages by grade, race, gender, eligibility for subsidized meals and participation
in USDA meal programs and determine differences in Dietary Guideline (DO) and RDA
achievement between vended food consumers and nonconsumers.

Methods: The USDA School Nutrition Dietary Assessment Study (SNDA) data with a
nationally representative sample of 3,381 children in grades 1-12 from the 48 contiguous
states were examined. Food codes from the 24 hr. recalls of children were matched with
the dietary fiber updated USDA Nutrient Database for Individual Intake Survey Analysis
version 4 to obtain dietary fiber composition of foods and intakes of children and
composition of vended foods. The SNDA assessed DG achievement as fat (<3 0%
energy), saturated fat (<10% energy) and cholesterol (<3 00 mg/d) as a categorical
variable. RDA achievement (for protein, iron, calcium, magnesium, vit. A, C, B6, B12,

folate, niacin, thiamin, and riboflavin) was also a categorical variable in the SNDA data.

The American Health Foundation=s (AHF) dietary fiber intake recommendations

[age (yr.) + 5 g/d] were used to assess fiber intakes. Fuzzy K-means cluster analysis of
foods consumed by USDA meal program participants was performed. Chi-square
analyses, Tukey-Kramer t-tests, & correlations were used to assess differences between
groups of children, using SAS, SUDAAN and JMP.

Results: 1890 foods and food components low in fat and saturated fat and high in dietary
fiber were identified by the cluster analysis. These foods included, hot chocolate and
chocolate milk and ready-to-eat breakfast cereals. The % of children achieving AHF
dietary fiber recommendations ranged from over 60% of 6 yr. old males to less than 10%
of 18 yr. old females. Females 12-18 yr. of age had significantly lower mean fiber
intakes than males of the same ages (p<0.001). African-American children had
significantly lower mean dietary fiber intakes at most ages than Caucasian children.
Hispanic adolescents had significantly higher mean dietary fiber intakes than Caucasian
adolescents of the same ages (p<0.05). Dietary fiber intakes did not differ significantly
by eligibility for subsidized meals. Dietary fiber intakes of USDA school breakfast
program (14.1 vs. 17.3 g/d, p<0.001) and National School Lunch Program participants
(13.3 vs. 16.6 g/d, p<0.001) were significantly lower than that of nonparticipants. An
average of 11.4% of school children consumed vended foods/beverages. These items
contributed approx. one third of the daily energy intake of those who consumed them.
Male and female vended food consumers were significantly less likely to have achieved
age appropriate RDAS than nonconsumers of the grade and gender (p<0.001), but often

more likely to have achieved DGs than nonconsumers.

ACKNOWLEDGMENTS

This dissertation is a result of professional and personal mentoring from
Dr. Sharon Hoerr, Dr. Won Song, Dr. Jenny Bond, Dr. Howard Teitelbaum, and
Dr. Leonard Bianchi. Their patient instruction has given me, not only academic
learning, but personal growth as well. Without the unconditional love and support
of my husband, parents, son, family and friends, I would not be here today. To
them, I can only say that whatever I do, including this dissertation, is for you. I
would also like to thank the Dept. of Food Science and Human Nutrition, College
of Human Ecology, Graduate School, and College of Human Medicine at
Michigan State University and Kellogg Co. for their support of this work. Special
thanks to Ms. Sang-Jim Chung and Dr. Randy Fotiu and Ms. Debra Keast for

technical assistance.

iv

TABLE OF CONTENTS

M 33

Title

Abstract

Acknowledgements

Table of Contents

List of Tables

List of Figures

1. Introduction
Legislative Context of Proposed Study
Statement of the Problem
Research Objectives
Research Questions
Null Hypotheses
Variable Definitions

2. Review of the Literature
Physiological Effects of Breakfast
Breakfast and Nutritional Content of the Diet
School Breakfast Program and Nutrition
Breakfast and Performance
Prevalence of Breakfast Skipping
Breakfast and Obesity
Implications of Reviewed Literature for Proposed Research
Sample Design and Statistical Analysis Methods

3. Clusters of High Fiber, Low Fat Foods Selected by Children
in USDA School Meals Programs
Abstract
Introduction
Methods
Results
Discussion
Implications for Health Professionals

iv

vii

viii

NOMAth-s

10
12
13
18
20
21
22

24
24
26
28
31
32
36

Dietary Fiber Intakes of U.S. School Children Do Not Meet
American Health Foundation Recommendations
Abstract

Introduction

Methods

Results

Discussion

Conclusions

Vended Food Consumption by U.S. School Children
Abstract

Introduction

Methods

Results

Discussion

Implications

Conclusions
Recommendations for Future Studies

References

vi

40
40
42

47
48
S4
59
59
61
63
66
69
72

83
85

86

Table No.

Table 3-1:

Table 3-2:

Table 3-3:

Table 4-1:

Table 4-2:

Table 5-1:

Table 5-2:

Table 5-3:

Table 5—4:

LIST OF TABLES

'1!

Title ag

Cluster center profiles for clusters of foods consumed by
USDA school meal program participants based upon the
food=s dietary fiber, total fat, and saturated fat content 37

High dietary fiber, low fat and saturated fat cluster of

foods consumed most frequently by dietary recommendation

(DR) achievers in USDA school meal programs compared

to foods in the high fat and saturated fat, low dietary fiber

cluster consumed most frequently by school meal program

participants not achieving DR for fat, saturated fat and

dietary fiber 38
Percentage of children who achieved dietary

recommendations (DR) for dietary fiber, total fat, and

saturated fat in 24 hr. By USDA meal program

participation 39

Dietary fiber intakes (g/day) of U.S. school children by
age and gender 61

Mean dietary fiber intake (g/day) of U.S. school children
by age and ethnicity 62

Vended Foods and Beverages Consumed by U.S. School
Children Ranked by Frequency of Consumption 75

Mean nutrient contribution of foods and beverages obtained
from vending machines or school stores to the 24 hr.
intake of U.S. school children in elementary, middle and
high schools consuming them 76

Mean 24 hr. nutrient intake of U.S. school children consuming vended
foods and beverages in elementary, middle and high school grades

78
Achievement of Dietary Guidelines and Recommended
Dietary Allowances Elementary,
middle and high school male and female
vended food nonconsumers vs. consumers 80

vii

Table 5-5: Mean 24 hr. nutrient intake of U.S. school children
consuming vended foods and beverages vs. nonconsumers
in elementary, middle and high school grades.

viii

81

LIST OF FIGURES

Figure No. Title

Figure 4-1. Percent of U.S. school children who met American
Health Foundation dietary fiber recommendations
(Age + 5 g/day)

Figure 5-1. Percent of U.S. school children consuming vended foods
by grade

ix

1%”

58

74

Chapter One
Introduction

Significance

At present, more than half of children get at least one of their three major meals in
school (Burghardt, 1995). One in ten children get two out of three meals per day at
school (Burghardt, 1995). The current political atmosphere is one that questions both the
efficacy of school meal programs in achieving their original objective of "safeguarding
the health and well being of the nation's children" and the amount and distribution of the
programs' funds (USDA, 1994). Therefore, it is of vital importance to improve school
meal program ability to provide nutrition, without contributing to excess consumption of
dietary components associated with chronic disease risk. Possible methods of improving
the total fat, saturated fat, dietary fiber and other nutrient content of meals offered in these
programs to better meet the United States Department of Agriculture and Department of
Health and Human Services '( USDA-DHHS) Dietary Guidelines for Americans must be
identified (USDA-DHHS, 1989).

The USDA School Nutrition Dietary Assessment Study (SNDA), published in
October 1993, assessed demographics and dietary information on a nationally
representative sample of over 3,000 children in grades 1-12 in 545 schools across the
country (Burghardt, 1993). The findings included a partial nutritional analysis of School
Breakfast Program (SBP) and National School Lunch Program (N SLP) participants and

nonparticipants' diets with respect to the National Research Council Recommended

Dietary Allowances (RDA) and USDA-DHHS Dietary Guidelines for Americans
recommendations on total fat, saturated fat, and cholesterol intake (National Academy of
Sciences, 1989; USDA-DHHS, 1989; Burghardt et al., 1993; Burghardt et al., 1993;
Devaney et al., 1993). However, all diets were analyzed with respect to the
Recommended Dietary Allowances for selected nutrients of interest, not within the
context of food group serving schemes, as the SBP and NSLP meals were defined.
Neither the impact of vended foods on nutrient intake of children nor dietary fiber intake
were examined using SNDA data. The proposed analysis will yield timely food-based
recommendations to assist schools and school children to meet USDA-DHHS's Dietary
Guidelines. It will also provide valuable data on vended foods impact on the diets of
children.
Legislative Context of Proposed Study

The USDA's June 1995 final ruling on School Meal Initiatives For Healthy
Children requires school meal program conformance to the 1990 USDA ~DHHS Dietary
Guidelines for Americans by the 1996-97 school year (USDA, 1995). While it is
important to ensure the nutritional quality of school meals, it is equally important to make
the process of planning, preparing, serving and selecting nutritious meals cost efiicient
and user-friendly. For this reason, the USDA, in response to food service personnel
concerns, offered three options: 1) nutrient analysis based menu planning (known as
NuMenus), 2) assisted nutrient analysis based menu planning (Assisted NuMenus) and 3)
food group based menu planning (USDA, 1995).

Another issue of importance is the dietary fiber intakes of U.S. school children.

2

The USDA=S nutrient goals for the school meal programs do not include dietary fiber as
part of their nutrient goals at present (Burghardt, 1995). This was due in part to the lack
of adequate information on the dietary fiber content of foods in computerized nutrient
analysis sofiware and to the lack of appropriate pediatric dietary fiber intake
recommendations at the time of the initial USDA legislation regarding nutrient goals for
the school meal programs (USDA, 1990). However, with the release of USDA=S nutrient
database version 4 with dietary fiber content of foods and the American Health
Foundations: age-specific dietary fiber intake recommendation (age + 5 g/day) it is now
appropriate to examine dietary fiber intake of school children (USDA, 1990; Williams et
aL,1995)

The USDA, in their June 1995 final ruling on school meal programs, stated that
significant concern regarding lack of regulation of vended foods has been expressed by
parents, health professionals and school officials (USDA, 1995). The USDA currently
has no jurisdiction, however, over vending machine offerings or the use of vending
machines outside cafeterias in schools (USDA, 1995). The impact of vended foods must
be examined in order to determine whether regulation of vended foods is warranted.
Statement of the Problem

The SNDA study has data collected which can be used to answer several key
previously unanalyzed questions.

1) What specific foods currently offered to and selected by children in the school meals
programs have low total fat, and saturated fat and high dietary fiber contents?
2) What are the dietary fiber intakes of U.S. school children and how do they compare to

3

the American Health Foundation (age + 5 g/day) recommendation?
3) What impact do competitive foods (such as those sold in vending machines) have on

nutrient intake and the dietary guideline achievment of school children?
Research Objectives

This project aims to use fuzzy cluster mathematical modeling to identify specific
low fat, low saturated fat, high dietary fiber foods which increase likelihood of achieving
the Dietary Guidelines for Americans recommendations for total and saturated fat, and
the American Health F oundation=s recommendation for dietary fiber (Windham et al.,
1985). The percent of U.S. school children achieving American Health Foundation
recommendations for dietary fiber intake will be examined by gender, age and ethnicity.

Although the USDA legislation prohibits the service of "foods of minimal
nutritional content" in cafeterias during school meal hours, little data are available on the
impact of vended foods on nutrient intake of school children (USDA, 1995). This project
proposes to examine the impact of vended foods on children's nutrient intakes. Relation
of participation in school meal programs to vended food consumption will also be
examined. Healthful foods which could be offered via vending machines will be

identified.

Research Questions
Descriptive analyses were conducted to assess any gender, ethnic, or age

differences in dietary fiber intakes and vended food consumption. For children in grades

1-12 who participated in the School Nutrition Dietary Assessment Study (SNDA):

1) What cluster of low fat and saturated fat and high dietary fiber foods is associated with
achieving the dietary recommendation for:
a) fat (<30% energy)?
b) saturated fat (<10% energy)?
c) dietary fiber (age + 5 g/d)?

2) What are the 24 hr. dietary fiber intakes of U.S. school children by age, gender and
ethnicity?

3) What percent of U.S. school children achieve American Health Foundation (age + 5
g/day) dietary fiber intake recommendation?

4) How does participation in SBP or NSLP relate to vended food and beverage
consumption in schools?

5) For all students who consumed vended foods, what is the total fat, saturated fat and
dietary fiber derived from vended foods and beverages?

6) Do vended foods consumers differ from nonconsumers in 24 hr. mean intakes of the
following nutrients:
a) total fat

b) saturated fat

c) dietary fiber

Null Hypotheses:

For children in grades 1-12 who participated in the School Nutrition Dietary Assessment

Study:

1) No clusters of foods are associated with achieving the dietary recommendation for
a) fat (< 30% energy),
b) saturated fat (<10% energy),
0) dietary fiber (age + 5 g/d).

2) Mean 24 hr. dietary fiber intakes of U.S. school children do not differ by age, gender
or ethnicity.

3) U. S. school children achieve the American Health Foundation (age + Sg/day) dietary
fiber intake recommendation.

4) Participation in SBP does not relate to using vended food and beverage consumption in
schools.

5) For all students who consumed vended foods, no total fat, saturated fat or dietary fiber
was derived from these foods or beverages.

6) Vended food consumers do not differ from nonconsumers in total 24 hr. intakes of the
following nutrients:
a) total fat
b) saturated fat

0) dietary fiber

Variable Definitions

1) Dietary fat: total fat g/day; percent of energy from total fat; g/ 1000 kcal in foods.

2) Saturated fat: saturated fat g/day; percent of energy from saturated fat; g/1000 kcal in
foods.

3) Dietary fiber: total dietary fiber g/day; g/kcal in foods.

4) Vended foods or beverages: any foods or beverages purchased from a vending machine
or on campus school store.

5) Vended food consumers: those who reported any food or beverage from a vending
machine in their 24 hr. recall.

6) Vended food nonconsumers: those who did not consume any foods or beverages from

vending machines or school stores.

Chapter Two

Review of Literature

The goal for implementing national school meal programs was initially to
safeguard the health and well being of school children (USDA, 1995). This includes the
impact that the meals themselves have upon health, nutrition, growth, physiology, and
cognition of school children. Therefore, the following historical review of the literature
will focus on the impact of breakfast on various outcomes. Since meals are offered
through the school systems, it is of value to examine the specific effects of breakfast upon
school performance, as well.

Physiologic effects of breakfast

Most of the research focusing on the effect of breakfast on physiologic response
was conducted in the early 1950's. It is important to review, however, because of the
relevance of the findings to the proposed study.

Kate Daum and her colleagues conducted several controlled metabolic trials, in
the 50's, examining the effects of various types of breakfast on physiologic response
(Daum et al., 1950), utilization of nitrogen, phosphorus, calcium and iron (Daum et al.,
1951) and thiamine and nitrogen excretion (Daum et al., 1951). One study focused on
how several different breakfast compositions impacted neuromuscular tremor, maximum
work output, and choice reaction time (Daum et al., 1950). Ten university women aged

18-25 years consumed one of the following for five weeks: a) a heavy breakfast (bacon,

cereal, toast, egg, fruit, etc.) vs. a light breakfast (toast, butter and fruit), or b) a basic
breakfast (toast, cereal, fruit and milk) vs. no breakfast, or c) no breakfast vs. black coffee
only, or (1) basic breakfast with coffee. Test meals were administered between 7 and 8
am. No breakfast indicated no food or beverage intake from 8 pm. of the night before
until noon on the test day. Physiologic response data were collected between 10:30 A.M.
and noon.

Heavy vs. light breakfast showed no significant differences in neuromuscular
tremor or choice reaction time (maximum work output was not tested for this). Basic
breakfast vs. no breakfast showed neuromuscular tremor to be greater and maximum
work output significantly less with no breakfast than with the basic breakfast. The results
for choice reaction time were conflicting. Interestingly, having black coffee only was
found to be worse than having nothing at all; maximum work output was lower,
neuromuscular tremor greater and choice reaction time longer with coffee only in the
morning (Daum et al., 1950).

In the study which examined nitrogen, phosphorus, calcium and iron in ten male
students (aged 21-35 years), Daum et al. compared to consumption of a cereal breakfast
with a bacon and egg breakfast for 4 to 6 weeks; other meals were controlled and constant
in approximate composition (Daum et al., 1951). They found no meaningful differences
in nitrogen balance, phosphorus, calcium or iron utilization between the two groups.
When thiarnin and niacin excretion was measured in the same men, they found that
thiarnin and niacin excretion was the same, regardless of type of breakfast eaten (Daum et

aL,1951)

It is important to examine the effect of breakfast on nutrient content of the diet as
a whole to understand how breakfast might impact nutrient excretion. Numerous studies
examined the impact of breakfast consumption on nutrient content of the total diet.
Several key studies from the past 15 years are reviewed here.

Breakfast and nutritional content of diet

Theresa Nicklas and colleagues examined breakfast consumption patterns in 467
10-year-old children from the Bogalusa Study sample (Nicklas et al., 1993). They found
that 16% skipped breakfast. The average total daily energy and macronutrient intake was
significantly lower for children who did not consume breakfast and for children who
consumed breakfast at home, when compared with children who consumed breakfast at
school. Percentage of energy from fat was lower in children who ate no breakfast
compared with those who did (34% vs. 37-39%). Perhaps the most interesting result was
that children who did not eat breakfast did not make up the difference in nutrient intake
through the balance of the day. Breakfast skippers were more likely to fail to achieve two
thirds of the National Research Council ’3 Recommended Dietary Allowances for vitamins
and minerals (Nicklas et al., 1993 a).

In a separate analysis of 393 lO—year old children from the Bogalusa sample,
Nicklas et a1. (1993) found that home breakfasts provided more animal protein and
sucrose than did school breakfasts (N icklas et al., 1993 b). School breakfasts, compared
to home breakfasts, contributed significantly higher percentage of total nutrient intake for
energy (26% vs. 19%), protein (30% vs. 18%), carbohydrate (30% vs. 20%) and sodium
(27% vs. 18%). Differences were attributed to foods commonly consumed at home (grits,

10

pancakes, oatmeal, cereal, ham, bologna, etc.) compared to foods consumed at school
(breads, dairy, bacon, sausage and fruits) (N icklas et al., 1993).

Karen Morgan and colleagues examined specifically the impact of ready-to-eat
(RTE) cereal breakfast consumption on nutrient intake of 657 American children aged 5-
12 years (Morgan et al., 1981). From 7-day food diaries, Morgan et al. noted that
breakfast consumption contributed significantly to daily nutrient intake. Only 1.5% of
the sample skipped breakfast. Children who had RTE cereal breakfasts 3 or more times
in the 7-day period were found to have consumed significantly less (p<0.001) fat and
cholesterol and significantly more Q)<0.001) fiber, thiamin, niacin, riboflavin, iron,
folacin, pyridoxine, and vitamins B12, A and D than those who had no RTE cereal at
breakfast. Breakfasts which did not contain RTE cereal had a higher average content of
energy, protein, fat, cholesterol and sodium (Morgan et al., 1981).

More recently, in 1993, Albertson and Tobelmann also examined impact of RTE
cereal consumption on the diets of children 7-12 years old in 4,000 households
(Albertson et al., 1993). Their results also indicated that frequent cereal eaters consumed
fewer calories from fat at breakfast, as well as for the total day. They also had higher
average daily vitamin and mineral intakes than less frequent cereal eaters and noncereal
eaters (Albertson et al., 1993).

In the Bogalusa Heart Study sample of 568 lO-year-old children from whom 24-
hour recalls were collected, no significant differences were noted in macronutrient,
cholesterol, calcium, sodium, potassium, phosphorus or vitamin D intake over the
24-hour period (Nicklas et al., 1994). However, iron, vitamins A, B6, BIZ, thiamin,

11

niacin, riboflavin, and folate intake were significantly higher (p<0.05) for children
consuming cereal breakfast compared with those who did not eat cereal (N icklas et al.,
1994)

Ruxton et a1. (1993), in their sample of 7-8 year-old Scottish children, also found
RTE cereal breakfasts contributed more vitamin BZ, niacin, vitamin B6, and folate than
breakfasts consisting of other foods (Ruxton et al., 1993). However, this study found no
significant differences in mean energy intake between cereal and noncereal breakfast
eaters (Ruxton et al., 1993).

Although the previously reviewed literature indicates that consumption of
breakfast, in general, is associated with increased micronutrient content of the diet, it does
not address the success of providing school breakfast on improving nutrient intake of
school children. The following study specifically assessed the impact of the national SBP
on school children's nutrient intake.

School Breakfast Program (SBP) and nutrition

Hanes et al. (1984) assessed specifically the impact of participation in the United
States Department of Agriculture School Breakfast Programs on 24-hour nutrient intake
in 5212 children in grades 3-12 (Hanes et al., 1984). Data from 1089 children in grades 1
and 2 were collected at home from parents. Nutrient analyses were performed using the
Highland View Hospital-Case Western Reserve University nutrient database. This study
used least squares regression analysis and controlled extensively for: age, gender,
anthropometric measurements, activity level, ethnicity, family education, and other
demographic variables (Hanes et al., 1984).

12

Hanes et al. found that children in schools where the SBP was available were
significantly (p<0.05) more likely to consume breakfast than those children who did not
have SBP availability. Children who skipped breakfast had lower nutrient adequacy
ratios for energy and all nutrients over the 24-hour period than children who ate breakfast.

When the sample data were projected for the United States school-aged population,
approximately 3 million children were estimated to Skip breakfast on any given day.
Hanes et al., further estimated that an additional 600,00 children, who would otherwise
skip breakfast, would eat breakfast if the SBP were available to them (Hanes et al., 1984).

Given this information on the impact of breakfast consumption on nutrient intake
and physiology, it is logical to question the impact of breakfast on performance. This
literature review will first examine key studies on the following aspects: mood; physical
performance; Short-term recall; and specific task performance. Later, the broader
response of cognitive performance in school, which includes the previously listed aspects,
will be examined.

Breakfast and performance

Pollitt et a1. (1981) described the effects of skipping breakfast on selected aspects
of cognition in children and showed that fasting adversely affects the accuracy of problem
solving ability (as tested by Hagan Central-Incidental Task), but beneficially impacts
immediate recall in short-term memory (Pollitt etal., 1981). However, initial studies
conducted on French adolescents failed to Show any effects of breakfast skipping
(Michaud et al., 1989).

Therefore, Claude Michaud and colleagues (1991) decided to examine specifically

13

the effect of breakfast Size on short-term memory, concentration, mood and blood glucose
in 319 free living adolescents, aged 13-20 years (Michaud et al., 1991). Michaud et al.
found that a 63% mean increase in energy intake (compared to habitual intakes) had no
effect on blood glucose or late morning mood as assessed by the Herbert visual analogue
scale. High breakfast energy intake did have a beneficial effect on average immediate
recall in short—term memory (2 minute gap between initial display of test items and test
recall). However, concentration on a word puzzle (finding the word in a group of letters)
was impaired by high calorie breakfasts (Michaud et al., 1991).

This seemed to support Pollitt's earlier conclusion that high energy breakfast
consumption positively influenced short term recall but negatively affected long term or
sustained attention activities. Variation in blood glucose can not explain these observed
differences in performance, because both Michaud et a1. and Pollitt et al. found blood
glucose to remain stable, despite differences in energy content of breakfast or breakfast
skipping (Pollitt et al., 1981; Michaud et al., 1991).

Craig points out that numerous studies, including studies conducted by Craig
himself, Show a post—lunch dip in performance (Craig, 1986). However, such a dip in
performance was not observed following breakfast consumption in a study conducted by
King and colleagues as early as 1945. This early study found that levels of both visual
and motor functioning were poorer when breakfast was not eaten than when it was.
However, the tests were performed 2-3 hours following breakfast and may not assess the
acute effects of breakfast (King et al., 1945).

Pollitt and colleagues (1981) have shown that, in the late morning, school children

14

are more likely to make errors on a picture identification task (Matching Familiar Figures
Test) when they had missed breakfast (Pollitt et al., 1981). However, breakfast was not
found to affect mean efficiency on the Continuous Performance Test (CPT). Conners and
Blouin (1982/1983) also found no impact of breakfast on CPT (Conners et al., 1982/83).

Spring et a1. (1983) reported that on a sustained dichotic shadowing task - a task
that requires selectively attending to and repeating out loud words entering one car while
ignoring those arriving at the other car - omission errors were greater, and overall
accuracy lower, after a meal that was high in carbohydrate compared to a high-protein
meal (Spring et al., 1983). For younger people (<40 years old), the difference seemed
greater at breakfast than at lunch. For older people the differential effect was found to be
noticeably larger after lunch. The shadowing task took 20 minutes to complete and a
brief (2 to 3 minute) reaction time test failed to Show any differences due to meal type or
time of meal (Spring et al., 1983).

Clarke et a1. (1990) compared the nutrition and health of 167 passing versus 158
failing 5th grade children from five randomly selected primary schools in Kingston,
Jamaica (Clarke et al., 1990). Academic performance was tested with the Wide Range
Achievement Test, which had been previously used in Jamaica. Failing children had
lower heights-for-age, weights-for-height, hemoglobin levels and poorer breakfasts than
passing children. However, this study classified breakfast intake into only three
categories: solid, fluid, or none. Only 6% of the passing children (compared to 19% of
failing children) skipped breakfast or consumed a fluid breakfast. Poor nutrition and
health parameters were found to have a negative impact school achievement, even after

15

controlling for socio-economic factors (Clarke et al., 1991 ). Many researchers have
examined the effect of breakfast consumption on total daily nutrient intake and adequacy.

Angus Craig detailed some of the reasons for discrepancies in research on
breakfast consumption and performance (Craig, 1986). The delay between consumption
of breakfast and the measurement of performance, the instrument used to assess cognitive
efficiency or performance, quantity, volume and quality of the food consumed,
individual's previous history or habit of breakfast consumption/skipping and the nature of
the individual responding to the breakfast meal (diurnal rhythm) can all affect the
observed response to breakfast (Craig, 1986).

The early part of the waking day is a period that is generally associated with the
upward phase of the circadian rhythm, when there is a rapid increase in alertness, arousal
and activation. Therefore, endogenous time-of-day effects could mask any effects that
are directly caused by eating breakfast (Craig, 1986). All of these factors must be born in
mind when evaluating and interpreting breakfast research.

Current evidence is still insufficient, in both quantity and quality, to permit firm
conclusions about the influence of breakfast on performance. The mechanisms by which
breakfast is hypothesized to impact performance is through its influence, if any, on
concentration, mood, satiety, nutritional status and general health. If breakfast does
influence these variables, what is its effect on educational performance?

Alan Meyers and colleagues examined the broader response of cognitive
performance in school, specifically the effect of participation in the SBP on academic
performance and attendance, of children in grades 3-6 in six elementary public schools in

16

Lawrence, Massachussetts (Meyers et al., 1989). Differences in scores on the annually
administered standardized test, Comprehensive Test of Basic Skills (CTBS), were used as
a measure of academic performance. While this was an improvement over the use of
teacher evaluation of changes in behavior and performance in nonblinded studies of
impact of breakfast or morning supplements on performance, this parameter has its own
limitations. Meyers' sample consisted of a Significant proportion of Hispanic children in
transitional bilingual classes, who might have been adversely assessed on a standardized
test designed for children with proficiency in English. However, the researchers
attempted to control for environmental variables like family influence, as well as genetic
variability in intelligence by using the differences in children's scores between two years.
Thus, the children served as their own individual controls (Meyers etal., 1989).

Meyers et al. found that, after controlling for other variables, participation in the
SBP positively impacted CTBS total battery (p<0.01), reading (p<0.05) and mathematics
(p<0.01) scores. Increases in CTBS scores from 1986 to 1987 were Significantly greater
for children participating in the SBP than those who did not participate in the SBP
(p<0.01). Tardiness rates decreased from 1986 to '87 for SBP participants while
increasing for nonparticipants (p<0.01). SBP participation contributed significantly to
the prediction of 1987 CTBS battery total scale scores (P<0.05), and followed only grade
in school and 1986 battery total scale score in the weight of its contribution (Meyers et
al., 1989).

Lopez and collegaues examined the effect of breakfast omission on cognitive
performance in 279 normal, wasted and stunted schoolchildren (Lopez et al., 1993).

17

Subjects ranged in age from 8 to 12 yr. Short-term visual memory, problem solving and
attention were found to be unaffected by 14 hr. fasting when compared with subjects who
consumed a standard breakfast (Lopez et al., 1993). However, several studies conducted
by Sally Grantham-McGregor and colleagues have shown an association between
breakfast skipping and poor nutritional status in Jamaican schoolchildren (Grantharn-
McGregor, 1993; Grantham-McGregor et al., 1991). Both wasted and stunted children
were found to perform more poorly on cognitive tests when breakfast was omitted.
Breakfast skipping was one of the most powerful predictors of school failure, along with
measures of socioeconomic status and attendance and illness measures (which could be
correlated with poor nutrition as well) (Clarke et al., 1991). The discrepancy between
these two researcher's results could partially be accounted for by the effect of motivation
or stimulation on cognitive performance in the testing situation.

Michaud et al. also found an association between breakfast and cognitive
performance, although not mood in a group of 319 well-nourished French adolescents
(Michaud et al., 1991). The tests used were the scale test of short-term memory and word
test of concentration and the Herbert Analogue Mood Rating Scale for mood. This was,
however, a univariate analysis not a multivariate model.

The prevalence of breakfast skipping must be assessed to understand the impact
of breakfast consumption or Skipping on outcomes including health and performance.
Child health advocates must be aware of not only the prevalence of meal Skipping but
also negative health outcomes associated with such eating patterns.

Prevalence of breakfast skipping

18

In a review of breakfast habits in children published in Nutrition and Food
Science, Carrie Ruxton and colleagues (1993) commented that children under 12 years of
age and adults over 65 are the most likely to consume breakfast, while young adults
displayed the highest tendency to omit breakfast (Ruxton et al., 1993). In this study of
136 Scottish children, only 1.5% of children were found to skip breakfast regularly
(Ruxton et al., 1993). Another study documenting the diets of 278 junior high school
home economics students in Rhode Island found that, on average, both male and female
students skipped breakfast 3 days per week (Brown et al., 1979).

Walker and colleagues examined breakfast consumption and prevalence of
skipping in 4717 South African: Black, Indian, European and African-Malay adolescents
(16-18 years old), using 24-hour recalls (Walker et al., 1982). Percent of the sample
skipping breakfast ranged from 13% in Europeans and Indians to 21% in Blacks.
However, this study failed to find significant differences in weight/height, academic class
rank or attendance in individuals who were consuming a liquid (porridge) breakfast
versus a breakfast of cooked food with or without cereal (Walker et al., 1982). This was
due, in part, to the manner in which breakfast groups were defined. There was no calorie
restriction on liquid breakfasts and porridge can be a substantially nutritive food.

Breakfast size and skipping breakfast are considered by health educators as a
major problem in nutrition of adolescents (Bull, 1988; Stewart-Truswell, 1985).
Breakfast skipping is hypothesized to be motivated by the concern for weight loss
prevalent in many adolescents and young adults. However, evidence suggests that
consuming high fiber foods, decreases not only the energy consumed at breakfast but also

19

up to three and a half hours later (Levine et al., 1989). Resnicow and colleagues found
that children who skip breakfast tend to consume greater amounts of high fat snacks
(Resnicow et al., 1991).

Regular breakfast skipping is often seen in dieters and is associated with increased
prevalence of obesity. The following study examines the effect of breakfast skipping on
weight loss efforts.

Breakfast and obesity

In a study conducted by Schlundt et al. (1993), 52 moderately Obese adult women
were stratified according to their baseline breakfast-eating habits and randomly assigned
to a weight-loss program (Schlundt et al., 1993). The no-breakfast group ate two meals
per day and the breakfast group ate three meals per day. The energy content of the two
weight-loss programs was identical. After the 12 week treatment, baseline breakfast
eaters lost 8.9 kg in the no-breakfast treatment and 6.2 kg in the breakfast treatment.
Baseline skippers lost 7.7 kg in the breakfast treatment and 6.0 kg in the no-breakfast
treatment. Substantial changes from previously established patterns seemed to improve
weight-loss. Schlundt and his group found that eating breakfast helped reduce dietary fat
intake and minimize impulsive snacking and, therefore, may be an important part of a
weight-reduction program (Schlundt et al., 1993).

Fricker and colleagues (1990) also examined circadian rhythm of energy intake
and obesity status in 1312 adults seeking medical advice for weight loss (Fricker et al.,
1990). The categories of obesity (non-obese, mildly obese and massively obese) were
defined according to National Health and Nutrition Examination Survey 11 criteria, while

20

dietary intake was assessed with a diet history. The 24 hour energy intake was found to
increase with degree of obesity. Quantity and quality of breakfast consumed was found
to have an inverse association with degree of obesity, while lunch was the opposite
(p<0.001). The relationships of obesity with breakfast and lunch persisted after
correction for nutrient intake. No relationships were observed with dinner or snacks
(Fricker et al., 1990).
Implications of reviewed literature for proposed research

Given the relation of school meals to over all nutrient intake, health and
performance of school children, it is important for school meals to conform to health and
nutrient goals including the USDA-DHHS Dietary Guidelines and RDAs (USDA-DHHS,
1989; National Academy of Sciences, 1989). It is also important to examine dietary fiber
intake of school children with the newly available computerized nutrient databases
containing dietary fiber information (USDA, 1990). Vended foods are being offered in
competition with school meals and, therefore, also deserve attention as part of efforts to

ensure diet quality for U.S. school children.

21

Sample Design and Method of Statistical Analysis — Standard Error

 

Calculation

The School Nutrition Dietary Assessment Study used a multistage
stratified sampling plan. The Study sampled school districts, schools within the
sampled school districts and students within the selected schools. First, districts
were separated into four distinct types. Within each district type, the districts
were separated into 10 census regions. Then districts were selected using
probability proportional to size.

Each sampled district was assigned to one of two types of data collection, in-
person visits (IP); or meals offered (MO). Those in the IP data collection sampled
students while those in MO did not sample students. Three schools were sampled
from the selected districts in the IP collection using probabilities proportional to
size.

Ten students were sampled from each of the three chosen schools, totaling 30
students per district. In small districts for eg. those with only one school, 30
students were chosen at random from the school. In districts with two schools, 15
students were chosen at random from each school.

When the data are collected as part of a complex multistage plan a method
is needed to produce unbiased estimates of the sampling variance. Most statistical
programs such as SAS and SPSS base their methods of estimating standard error
on the assumption that the data was collected using a simple random sampling

plan. The standard error, used in creating confidence intervals of means,

22

regression coefficients, etc and in inferential statistics, will be underestimated if
the sampling design is not taken into consideration in the analysis.

The statistical analysis program SUDAAN utilizes design statements, plus
an optional DESIGN parameter on the PROC (procedure) statement to describe
the input data file. Three types of sample designs with equal or unequal
probability and with or without replacement can be accommodated by SUDAAN.
Variance estimation for the statistics computed by SUDAAN procedures is based
on either the Taylor series linearization method (equivalent to GEE in regression
procedures) or replication methods such as BR and J ackknife.

For this analysis, Taylor linearization with replacement (WR) at the first
stage, subsequent stages and equal probabilities of selection at both the first and
subsequent stages was used. For a WR design, SUDAAN uses the between-PSU
within-stratum variance component to estimate the variances. This corrects for
the stratified multistage sampling design used for the School Nutrition Dietary

Assessment Study data collection when calculating the standard error measures.

23

Chapter Three

Clusters of High Fiber, Low Fat Foods Selected by Children
in USDA School Meals Programs
ABSTRACT
Objectives: To identify clusters of foods and food components consumed by USDA
school meal program participants by total fat, saturated fat, and dietary fiber content and
determine percent of program participants vs. nonparticipants achieving dietary
recommendations (DR) for total fat (<30% energy), saturated fat (<10% energy) and
dietary fiber (age + 5 g/d).
Methods: The USDA School Nutrition Assessment Study data set is a nationally
representative sample of 3,381 students from 329 schools across the country, grades 1-12
and food composition data on 38,380 foods and food components selected by school
children. Dietary fiber content of these items was estimated by USDA Nutrient Analysis
of Individual Intake Surveys, version 4. High fiber, low fat items consumed by children
were identified using fuzzy K-means cluster analysis from the Statistical Analysis
Software, version 6.03. Percentage of children achieving DR was determined by
participation in the school breakfast program (SBP), and/or national school lunch
program (N SLP).
Results: Five main clusters of foods consumed by school children were identified.
Cluster #1 contained 1,890 foods high in dietary fiber and low in fat and saturated fat,
including sugar free hot chocolate, ready-to-eat breakfast cereal, whole wheat bagels,
baked snack chips, etc. Three clusters contained foods high in fat and low in fiber such
as pizza, cheese, ice cream, hamburgers, etc. and one cluster had items low in all three
nutrients. Significantly more children who did not participate in USDA school meal
programs achieved DR for fat, saturated fat and dietary fiber (5.7%, p<0.001), than those
who participated in the SBP only (2.9%), NSLP only (1.8%) or both SBP and NSLP
(2.8%). Conclusions: Many high fiber, low fat foods popular with children are already

24

available in school meals programs. Cluster analysis is a powerful, flexible tool to
identify such foods from existing school offerings. Increasing the number of times these
foods are offered in meals programs may be one way to increase DR achievement by

school meal program participants.

25

Clusters of High Fiber, Low Fat Foods Consumed by Children
in USDA School Meals Programs
INTRODUCTION
Presently over half of U.S. children get at least one of three major meals per day
in school. Ten percent of children receive two out of three meals per day at school
(N icklas ct a1. 1993). Therefore, it is of vital importance to evaluate the school meal
program efficacy in offering children choices to encourage achievement of the United
States Department of Agriculture=s (USDA) and Department of Health and Human
Services: (DHH S) Dietary Guidelines for Americans recommendations for total fat
(<3 0% energy) and saturated fat (<10% energy) and dietary fiber. The American Health
Foundation=s recommendation for daily dietary fiber intake (g/day) is equal to at least the
child's age plus five grams of fiber (Williams et a1. 1995). These three nutrients play key
roles in the prevention of coronary artery disease, some cancers, and gastrointestinal
problems such constipation (Osganian ct a1. 1995).
The Child Nutrition Act was first ratified in 1946, establishing the National
School Lunch Program in an effort to meet the nutrition needs of underprivileged
children in the U.S. (Dwyer 1995). In 1966, this act was amended to begin the service of
breakfast (School Breakfast Program) in schools in particularly poverty stricken areas.
Congress recently modified this legislation as part of the Welfare Reform Act of 1996 in
an attempt to streamline the program’s funding and regulations (Johnson 1996).

26

Dietitians and other health professionals are lobbying to update dietary recommendations
for the NSLP and SBP to be consistent with chronic disease prevention rather than
merely the original goal of prevention of undemutrition (USDA 1995).

In June of 1995, the USDA proposed to change the existing food group-based
menu planning system used in school meals programs to a nutrient-based system in order
to promote achievement of dietary guidelines for selected nutrients such as fat, saturated
fat, calcium, iron, etc. (USDA 1995). While the USDA School Meals Initiative for
Healthy Children was greeted with approval from parents and health professionals, it
inspired nearly unanimous concern and disapproval from school food service personnel.
School food service personnel were concerned primarily about the training and cost of
computer technology required to change the menu planning system to meet the dietary
guidelines (USDA 1995). USDA later permitted the schools to keep existing menu
planning or nutrient based planning to meet DR (USDA 1997). Federal policy mandates
that school meals meet dietary recommendations (DR) by 1997/98 (US DHHS 1990).
One of the few suggestions given to schools by the USDA to help them meet these goals is
to make healthful foods widely available (USDA 1997).

The goal of our study was to identify high fiber, low fat and saturated fat foods
consumed by children in the USDA school meal programs. In order to analyze
simultaneously the content of several nutrients in items consumed by children in USDA
school meals programs, we chose a statistical analysis tool not widely used in dietary
analysis, Fuzzy Clustering. Fuzzy Clustering is a flexible, powerful analysis tool
permitting examination of multiple nutrients in the natural distribution, without

27

investigators setting out points a priori for the number of groups of foods or high/low
nutrient contents (Aldenderfer et al. 1984). The fuzzy K-means algorithm identifies
natural clusters of foods from the distribution of foods in several dimensions; here we
chose total fat, saturated fat and fiber (Aldenderfer et al. 1984). Not only are the foods
identified by this study ones already offered by school food service system menus, but
ones that children consumed voluntarily. Increasing the availability of these identified
foods could, then, be one practical means by which school food service systems could
move toward achieving the USDA dietary recommendations with little additional cost or
training.

METHODS

Sample and Data set

 

We used the USDA School Nutrition Dietary Assessment Study (SNDA) data set
collected in 1991-1992, a nationally representative sample of 3,381 children in grades l-12
(approx. 250 students per grade) from 329 schools across the U.S. (Burghardt et al. 1993).

Students not attending school on the day of data collection, those enrolled in programs
offering self-contained classes for students with disabilities, students from Alaska or
Hawaii, and students enrolled in kindergarten were not included in the sample (Burghardt
ct a1 1993). Food intake data for a 24 hour period were collected by trained interviewers
administering questionnaires. To collect dietary intake data for children in grades 1-3,
parents were interviewed along with the child. Nutritional composition data were
included on foods consumed by children from the USDA National School Lunch Program
(N SLP) and School Breakfast Programs (SBP) at each school, as well as foods purchased

28

in vending machines or in campus snack shops (Burghardt et al. 1993).

Although the SNDA study also included dietary information on students who did
not participate in USDA school meal programs, we examined only foods consumed by
program participants for the cluster analysis. For the assessment of DR achievement we
included all SNDA children. All data were coded to maintain anonymity of participants.
Analysis by ethnicity was not possible due to errors in the coding for the variable Aracez
in the SNDA data set.

M

The initial USDA analysis of the SNDA data did not address dietary fiber, due to
incomplete information on dietary fiber content of foods in USDA computerized nutrient
databases and lack of established recommendations for pediatric dietary fiber intake at the
time of data collection (Burghardt et al. 1993). However, a subsequent release of the
USDA Nutrient Database for Analysis of Individual Intake Surveys version 4 in 1991
made our analysis possible (USDA 1990). SNDA data were obtained from USDA on
IBM mainframe magnetic reel tapes in Statistical Analysis for the Sciences (SAS) ASCII
files with accompanying SAS programs to label data. Five relevant files with
participants: 24 hour nutrient intakes, diet recalls, demographic information and food
composition were downloaded to SAS version 6.03 personal computer files (SAS Institute
1988). Food composition of items from school sources such as cafeteria, vending
machines, and on campus snack shops, consumed by children was determined from
product labels, recipe analyses, food service nutrient profile databases by the SNDA study.

Foods brought from home were not included in our analysis. Foods listed more than once

29

were entered only once, however, similar foods were kept separate. For example, milk 2%
fat and milk 1% fat were counted as different foods. The large number of foods in the
database is due to separation of mixed dishes such as spaghetti into component foods (i.e.,
spaghetti sauce, pasta, cheese, ground meat, etc). Food codes from all SNDA children's
diets were matched with USDA dietary fiber updated Nutrient Database for Individual
Intake Surveys version 4 to determine dietary fiber content of foods, as well as to calculate
total 24 hr. dietary fiber intake of children (USDA 1990).

Fuzzy cluster analysis was used to determine which foods selected and consumed
by children in USDA school meal programs were low in fat, saturated fat and high in
dietary fiber. Then, items in the high fiber, low fat cluster most popular with DR achievers
was determined and compared to the high fat, low fiber items most popular with those
who did not achieve DR. Popularity was defined as the number of children consuming
these items. Grams/1000 kcal of total fat, saturated fat and dietary fiber in each food were
used for clustering foods. Extreme values for these three nutrients in the data (> 2
standard deviations from the mean) were removed fi'om the analysis to prevent skewing of
the clusters (Aldenderfer et al. 1984). Fuzzy cluster analysis by the SAS F astclus
procedure utilizes a Fuzzy K-means algorithm to determine the appropriate number of
groups of foods based upon their content of multiple nutrients and calculates a cluster
center nutrient profile for each group (SAS Institute 1988). This clustering is based upon
Euclidean distances between observations (those distances within groups being less than
distances between groups). The program, then, generates an association value ranging
between 0 and l for each food in the cluster compared to the cluster center as a goodness

30

of fit measure (SAS Institute 1988).

Variables describing participants used for analysis included age, participation in
school meal programs, percent of energy from total fat and saturated fat and total grams of
dietary fiber consumed in 24 hr. Descriptive analyses of percent of children achieving DR
for nutrients of interest (total fat, saturated fat and dietary fiber) based upon USDA meal
program participation were run using SAS version 6.03 (SAS Institute 1988).

RESULTS

The cluster analysis of foods consumed from school sources by children yielded
five main clusters (of greater than 1,000 foods per cluster) based upon nutrient density
(g/1000 kcal) for dietary fiber, total fat and saturated fat. Clusters with less than 1,000
items are not reported here. The results shown in Table 1 indicate that approximately
1,890 low total fat, low saturated fat, and high fiber foods were identified (Cluster #1).
Table 2 shows the high fat, low fiber items in Clusters # 2, 4 & 5 compared to Cluster #1.
These items were determined by assessing which foods in the low fat, high fiber cluster
(#1) were consumed most frequently by children achieving the DR and which high fat,
low fiber foods (clusters 2, 4, & 5) were most frequently consumed by children not
achieving the DR. The high fat, low fiber foods included hot dogs, hamburgers, ice cream,
high fat desserts, cheese, etc., while the high fiber, low fat foods included some surprises
as sugarfree hot chocolate (with skim milk and soluble gum thickener) and baked snack
chips along with the expected high fiber ready-to-eat breakfast cereals and whole wheat
bread. Cluster # 3 consisted of foods low in all three nutrients such as beverages, some
condiments, etc. The largest number of foods and food components fell into Cluster #3.

31

A complete list of foods in each cluster is available to interested parties on disk by
contacting the first author.

The percentage of children achieving DR for total fat, saturated fat and dietary
fiber in 24 hr based upon USDA meal program participation is shown in Table 3. Of those
children who did not participate in either the USDA SBP or NSLP, 5.7% achieved DR for
the nutrients of interest. This was significantly more (p<0.001) than those who only
participated in the SBP but not NSLP (of whom only 2.9% achieved DR) as well as those
who participated in the NSLP but not SBP (of whom only 1.8% achieved DR). A
significantly greater percentage of children who did not participate in either the SBP or
NSLP achieved DR compared to children who participated in both programs (5.7% vs.
2.8%, p<0.001).

DISCUSSION

The USDA=S analysis of the SNDA found that the mean percentage of energy
from fat was 34% when the nutrient intakes of all study participants was included, while
saturated fat constituted 13% of energy (Devaney et al. 1995). However, fat intake of
school meal program participants alone averaged 37% of energy with 14% energy from
saturated fat (Devaney et a1 1995). This high average fat intake could reflect the lack of
enough food choices which are appealing, low fat sources of dietary fiber in the USDA
school meals programs, or lack of popularity of those items when children make their own
food selections.

The Child and Adolescent Trial for Cardiovascular Health (CATCH) included the
Eat Smart Program which targeted school food service staff and aimed to lower fat,

32

saturated fat and sodium content of school meals (Osganian et al. 1996). There was a
significantly greater mean reduction in the percentage of energy from fat (-4.1%; p<0.001)
and saturated fat (-1.3%; p=0.003) in response to USDA=S mandate to meet DR ), in the
56 schools where the intervention was conducted over a two and half year period
compared to control schools (Osganian et a1. 1996). This indicates that nutrition and
education for food service staff about low fat food preparation leads to improved
compliance of school meals with DR. However, the CATCH study did not address dietary
fiber content of foods (Osganian et a1. 1996).

Our study found that hot chocolate and chocolate milk, with carageenan thickener
were popular sources of dietary fiber in the SNDA children=s diets. The other major
sources of dietary fiber in the SNDA children=s diets were whole grain products such as
baked snack chips and high fiber ready-to-eat breakfast cereal. Beans and vegetables were
significant contributors to dietary fiber, but far less popular in the self-selected diets of
these school children. Interestingly, from the Bogalusa study, Nicklas et al. also found
that the milk group (including mixed dairy products) was a significant contributor of
dietary fiber especially in the form of carageenan found in low fat chocolate milk (Nicklas
et al. 1995). The two categories of food contributing most to dietary fiber intake in the
Nicklas et al. study were vegetables/soups and breads/ grains.

Substitution of foods high in dietary fiber and complex carbohydrates for foods
high in total fat and saturated fat is important in the prevention and treatment of
constipation, obesity, hyperlipidemia, diabetes mellitus and some cancers in children and
adults (Nicklas et al. 1995). Nicklas et al. argue that a shift toward greater consumption

33

of whole grains, vegetables, legumes and fruits will be necessary to reach this goal
(Nicklas et al. 1995). In 1991, the National Cancer Institute began the 5 A Day for Better
Health Program to encourage Americans to consume more fruits and vegetables to
increase their likelihood of obtaining not only adequate dietary fiber but essential
micronutrients such as antioxidant vitamins (Nicklas et al. 1996). Such food-based
recommendations appeal to consumers more than nutrient based recommendations, as
nutritionists and other health professionals are well aware (N icklas et a1. 1996).

Only a very small percentage of children in the SNDA sample achieved the dietary
recommendations for total fat, saturated fat and dietary fiber (5.7% of those who
participated in neither the NSLP or the SBP). Participation in the school meals programs
was associated with a significantly lower percentage of children achieving DR. More
children were unable to achieve the AHF recommendation for dietary fiber than the
recommendations for fat and saturated fat. Our study categorized children into DR
achievers or nonachievers based upon the 24-hour recalls from the SNDA sample. Due to
day to day variation of intake, it is possible that achievement of DR over time may be
different from the low percentage we found.

The Bogalusa Heart Study found that 15-22% of school age children achieved DR
for total fat (<30% energy) between 1973 and 1988 respectively. The percentage of
children achieving saturated fat recommendations (<10% energy) ranged from 8% in 1973
to 22% in 1988 in the Bogalusa study (N icklas et al. 1996). The average dietary fiber
intake did not vary in the Bogalusa trials from 1973 to 1988, remaining at approximately
12g/day (Nicklas et al. 1996).

34

Wirfat et al. used fuzzy cluster analysis to group people according to specific foods
consumed, for example meat and pastry consuming groups, and related those groupings to
nutrient content of the diet (Wirfat et al. 1997). For example, individuals whose dict
contained most foods from the meat and pastry group were found to have a higher average
total daily fat intake than foods in other groups (Wirfat 1997). In the current analysis, we
looked at the opposite relation. We identified categories or clusters of low fat, high
dietary fiber foods from the diets of children achieving DR for fat and fiber.

We found Fuzzy Cluster analysis to be a valuable tool for finding patterns of foods
which can be used for framing food-based recommendations rather than nutrient
recommendations. However, cluster analysis is limited by its sensitivity to extreme values
in a data set. Therefore, data must be cleaned carefully prior to employing this technique.
For example, the SNDA data set had errors showing individuals to have unusual nutrient
intakes, for example 154 g of dietary fiber intake in a 24 hr. period due to a data entry
error of a misplaced decimal on portion size of chilli consumed. Each dict record and food
composition profile had to be reviewed for such errors by a trained dietitian to remove
such cases from the analysis.

Dietitians often teach that foods are not inherently good or bad. However, the
encouragement of consumption of Specific foods promoting health (e. g. fi'uits and
vegetables as promoted by the National Cancer Institute) is an effective manner of
translating nutritional science into advice easily accessible to target audiences (National
Cancer Institute 1994). The US Dietary Guidelines are difficult for some food service
directors to recall and apply. For example, Hurd et al. reported that while 76% of the food

35

service directors in Texas reported awareness of the Dietary Guidelines, only 59% could
name at least two (Hurd et a1. 1996). The school cafeteria can be a good place to display
point-of-selection educational graphics showing the foods contributing to achievement of
DR for optimum health (Nicklas et al. 1995).
IMPLICATIONS FOR HEALTH PROFESSIONALS

Many (n=l,890) low fat, high dietary fiber foods popular with school children are
already available within USDA school meal programs offerings. Perhaps, an easily
applicable measure to help children increase their likelihood of achieving dietary
recommendations for fat and fiber is to offer such foods more frequently. Another
important step would be to identify these food choices, at the point of selection, with
colorful graphic messages stating that they are foods likely to promote optimum health.
Nutritionists can help train food service staff to plan creative menu combinations
incorporating such existing low fat, high fiber foods and promote selection by appropriate
nutrition education point-of-selection nutrition education. Cluster analysis is a powerful

technique for identifying food patterns that promote DR and optimum health.

36

Table 1. Cluster center profiles for clusters of foods consumed by USDA school meal
program participants based upon the food’s dietary fiber, total fat and saturated fat content.

 

ClusterType No. of Foods Dietary Fiber Total Fat Saturated Fat
g/1,000 kcal g1,000 kcal g/ 1 ,000 kcal

 

1 High fiber, low fat 1,890 4.7 1.3 0.3
2 Low fiber, mod.fat 4,874 0.6 10.9 4.5
3 Low fiber, low fat 29,467 0.5 1.9 0.7
4 Low fiber, high fat 1,475 1.4 17.4 4.9
5 Low fiber, high fat 1,043 1.1 24.9 8.3

 

37

 

Table 2. High dietary fiber, low fat & saturated fat cluster of foods consumed most
frequently by dietary recommendation (DR) achievers in USDA school meal
programs compared to foods in the high fat & saturated fat, low dietary fiber
cluster consumed most frequently by school meal program participants not
achieving DR for fat, saturated fat and dietary fiber.

 

Cluster #1 (high dietary fiber/ low fat) Cluster # 5 (low fiber/ high fat)

 

Sugar Free Hot Chocolate Cheese
Chocolate milk Hot Dogs

Baked Snack Chips Hamburgers
Ready-to-Eat Breakfast Cereals Ice Cream
Whole Wheat Bagels Pizza

Whole Fresh Apples Butter/Margarine
Carrots Macaroni and Cheese
Baked Beans Salad dressings
Baked Potato Strings Gravy

Pinto Beans Mayonnaise
Chili High Fat Dessert Items
Green Peas Lasagna

 

* frequency was determined by counting the number of DR achievers and nonachievers
consuming these foods

38

Table 3. Percentage of children who achieved dietary recommendations (DR). for dietary
fiber, total fat, and saturated fat in 24 hr., by USDA meal program participation

 

 

Group Unweighted n Percent
achieving DR
National School Lunch Program only 1744 1.8%
School Breakfast Program only 319 2.9%
Participants in both SBP and NSLP 145 2.8%
Students not participating in either program 1637 5.7%"

 

* Less than 30 % energy from total fat, less than 10% energy fiom saturated fat and equal
to or greater than age (yr.) plus five grams per day of dietary fiber.

** p<0.001 different from all other groups, based on weighted tabulations fi'om the SNDA
data.

39

Chapter Four

Dietary Fiber Intakes of U.S. School Children Do Not Meet
American Health Foundation Recommendations

ABSTRACT

Objectives: To assess dietary fiber intakes of U.S. school children by age, gender,
ethnicity, family income, student eligibility for free or reduced price school meals and
participation in USDA school meal programs.

Methods: The USDA School Nutrition Dietary Assessment Study (SNDA) data with a
nationally representative sample by grade of 3,381 children in grades 1-12 were used.
Dietary fiber intakes were calculated using the dietary fiber data of updated USDA
Nutrient Database for Analysis of Population Based Surveys, version 4 released in 1990.
Descriptive analyses, Tukey-Kramer t-tests, Chi-square analyses and correlations were
run using SAS version 6.03 and JMP from the SAS Institute and SUDAAN. American
Health Foundation (AHF) recommendation [age (yr.) + 5 g/day] was used as the
reference standard for dietary fiber intake. All analyses were weighted to make the
results nationally representative.

Results: Dietary fiber intakes of U.S. school children, particularly adolescents, do not
meet AHF recommendations. Percent of U.S. school children achieving AHF
recommendation ranged from 61.7% of males age 6 to 9.1% of females age 18. Fewer
females age 3 12 yr. achieved AHF recommendations than males of the same age. Black

children had significantly lower mean dietary fiber intakes than white children at nearly

40

all ages, while Hispanic adolescents had significantly higher mean intakes than white
adolescents at most ages. Annual family income was only weakly correlated with dietary
fiber intakes (r=0.06, p<0.01) and dietary fiber intakes did not differ significantly by
student eligibility for free or reduced price school meals. Signficantly fewer students
who participated in School Breakfast Program (52% vs. 61%, p<0.05) or National School
Lunch Program (42% vs. 48%, p<0.05) achieved AHF daily dietary fiber intake
recommendations than children who did not participate in USDA school meal programs.
Conclusions: Dietary fiber intakes of children, especially adolescent females fail to meet
AHF recommendations. Given the availability of dietary fiber content of foods and
guidelines for pediatric dietary fiber intake, dietary fiber should be included in the
nutrient goals for USDA school meal programs. Physicians and other health
professionals should educate parents and children on the importance of consistently

making high fiber food choices to reduce risk of chronic diseases in adulthood.

41

Dietary Fiber Intakes of U.S. School Children Do Not Meet
American Health Foundation Recommendations

INTRODUCTION

Optimizing the diet of children is important in order to meet the following goals:
1) to support adequate grth and development (American Academy of Pediatrics 1983;
Dwyer 1987; American Academy of Pediatrics 1981, American Heart Association 1983),
2) to reduce the risk of diet-related chronic diseases such as coronary artery disease
(Williams 1992; Wynder et al. 1989, National Cholesterol Education Program 1991),
some cancers (Williams et al. 1993, National Cancer Institute 1987), stroke, adult-onset
diabetes, obesity and hypertension (Public Health Service 1991) , and 3) to establish
healthy eating habits in early life (Kleinman ct a1. 1996). Recommendations for micro
nutrient, macro nutrient, and energy intake for children of various ages have been
established by the National Academy of Sciences and recently updated for selected micro
nutrients (National Academy of Sciences 1989). These Recommended Dietary
Allowances are the levels of intake of essential nutrients that are deemed adequate to
meet the known nutrient needs of practically all healthy persons (National Academy of
Sciences 1989). The current paradigm of optimizing diets seeks to not only ensure
nutritional adequacy and good health for the present but also to minimize future risk of
diet-related chronic disease (Public Health Service 1991).

42

With this aim of disease prevention, initial recommendations for dietary fiber
intakes in children ranged from 0.5 g/kg of body weight to 10 g/1000 kcal (American
Academy of Pediatrics 1993). The 0.5 g/kg body weight recommendation, put forth by
the American Academy of Pediatrics, resulted in fiber recommendations for obese
adolescents in excess of 40 g/day, placing individuals with borderline or low mineral
intake at risk for nutrient deficiency (American Academy of Pediatrics 1993). In 1994,
the American Health Foundation’s age-Specific recommendation for dietary fiber intake
in children, age plus five grams per day, was accepted as safe and adequate for health
promotion and disease prevention by nutritional and medical scientists (Williams et a1.
1995).

Dietary fiber has important health benefits in childhood, including promoting
bowel regularity (F ulgoni et al. 1991). Adequate dietary fiber intake throughout life has
been related to reduced risk of some adult onset cancers, such as breast, colon, pancreas,
ovary, endometrium and prostate cancer (Williams et al. 1993, National Cancer Institute
1987). Dietary fiber moderates blood glucose concentration in adult onset diabetes,
reduces blood cholesterol and even helps maintain a healthy weight for height by
inducing satiety without increasing calories content (Public Health Service 1991).
Because the methodology for analyzing dietary fiber was only developed within the past
decade, nutrient databases used for population dietary survey analysis did not have
complete, accurate and consistent information on dietary fiber until the early 1990S
(Slann 1987, United States Department of Agriculture 1990). For this reason, relatively
little information about the dietary fiber intakes of children is available in the literature.

43

The United States Department of Agriculture (USDA) collected extensive data on a
nationally representative sample of children’s diets and nutrient composition of foods
offered in the National School Lunch Program (N SLP) and School Breakfast Program
(SBP) as part of their School Nutrition Dietary Assessment Study (SNDA) in 1991-92
(Burghardt 1995). However, they were unable to examine dietary fiber intakes of their
study participants due to inadequacy of dietary fiber information in computerized
databases and the lack of dietary fiber intake guidelines at the time of data collection
(Burghardt 1995). The objectives of our study were to assess dietary fiber intakes of U.S.
school children using dietary fiber updated databases to examine SNDA data.
Secondarily, we examined differences in dietary fiber intake by age, gender, ethnicity,
family income, student eligibility for free or reduced price meals, and participation in
USDA school meal programs.

METHODS

Sample and Data set

 

The SNDA data set (1991-92) includes 24 hour recall dietary data collected fi'om
a nationally representative sample of 3,381 children in grades 1-12 in a nationally
representative subsample of 329 from a total of 545 schools fi'om the 48 contiguous states
from which school level data were collected (Burghardt 1995). The nationally
representative children were selected in a multistage sampling design. Food intake recall
data for a 24 hour period, including a school day, were collected by trained interviewers
administering questionnaires. Children in grades 1-3 were interviewed with their parents

to obtain dietary intake data. Children in grades 4-12 reported their own intakes. Student

44

 

level data were weighted to adjust for selection bias. All data were coded to maintain
anonymity of participants (Burghardt 1995).

Food composition data on all foods from the USDA National School Lunch
Program (N SLP) and School Breakfast Programs (SBP) were coded by nutritionists from
the Minnesota Nutrition Coordinating Center. Lists of foods purchased from vending
machines, A La Carte programs, and on campus snack shops at each school were
collected from school food service administrators (Burghardt 1995) . The race of the
student was identified and recorded by the interviewer collecting the data, rather than
student self report. Family income was obtained from a questionnaire assessing
demographic information and student/parent attitudes toward USDA meal programs
mailed to the parents of each child in the study. School administrators reported each
student=s eligibility for free (annual family income <135% Poverty Index) or reduced
price (family income between 135% and 185% of Poverty Index) school meals
(Burghardt 1995). Student participation in the school meal programs was determined
from students: 24 hr. diet recalls. The source from which each item was obtained was
recorded as part of the diets recalls. Students were considered SBP participants if they
consumed 2 of the 4 food group components required for USDA reimbursable breakfasts.

NSLP participants were defined as those students who consumed at least 3 of the 5 meal
pattern components required for a USDA reimbursable lunch (Burghardt 1995).
An_alys_is

SNDA data were obtained from the USDA on IBM mainframe magnetic reel

tapes in Statistical Analysis for the Sciences (SAS) ASCII files with accompanying SAS

45

programs to label data. Five relevant data files (stanal.dat, strecipedat, stchar.dat,
stfcomp.dat, and scvend.dat) were downloaded to SAS version 6.03 personal computer
files (SAS Institute 1998). Food codes from the SNDA participants: diets were
matched with USDA dietary fiber updated Nutrient Database for Analysis of Population
Based Dietary Surveys version 4 to determine dietary fiber content of foods to calculate
24 hour dietary fiber intake of children (USDA 1990) . These 24 hour dietary fiber
intakes were compared to the age-specific American Health Foundation dietary fiber
guideline [age (yr.) plus five grams per day] (Williams 1995). For the analyses by age,
children aged 5, l9, and 20 yr. (n = 20 total) were not included due to insufficient power
(small cell number). For the comparisons by ethnicity only white (n=2188), black
(n=544) and hispanic (n=459) students: diets were included due to small samples with
inadequate power in the mixed race and other categories (n=190 total). Comparisons by
ethnicity were performed with students collapsed across gender because the gender
distribution of students within each ethnic category did not differ significantly.
Variables used for analyses included total grams per day of dietary fiber in each
student=s 24 hr. intake, participation in school meals programs, age, race, annual family
income, and eligibility for free or reduced price meals. Family income was a continuous
variable in the SNDA data. Relationship of family income to dietary fiber intake was
assessed with a simple regression. Chi-square analyses including all children, regardless
of race or age, were used for comparions by USDA meal program participation and
eligibility for free or reduced price meals. Weighted descriptive analyses of mean fiber
intakes and differences in fiber intakes by various factors were conducted using SAS

46

version 6.03 and J MP from the SAS Institute (SAS Institute 1998) and SUDAAN .
Differences in mean dietary fiber intakes of various groups were tested using Tukey-
Kramer t-tests. This technique utilizes a conservative family based alpha level for
multiple comparisons resulting in a reduced power but greater assurance of finding true
differences between groups (Tukey 1960). This was preferable to a multiple regression
because dietary data available were only for one 24 hr. recall rather than diet over time.
RESULTS

The percent of children achieving the American Health Foundation
recommendation of age plus five grams or more of dietary fiber per day is represented in
Figure 1. Note the steady decline in fiber recommendation achievement with increasing
age, especially in adolescent females. A significantly smaller percent of females met the
AHF fiber recommendations compared to males (36% vs. 49%, p<0.001). The 24 hour
mean dietary fiber intakes by age and gender are shown in Table 1.

Interestingly, mean fiber intakes were significantly but only weakly correlated to
family income (r=0.06, p<0.001). Children eligible for free or reduced price school
meals were not significantly more likely to achieve dietary fiber recommendations than
those ineligible (43% of those eligible vs. 42% of those ineligible for low cost or free
meals). Significantly fewer children who participated in the USDA School Breakfast
Program (SBP) achieved AHF fiber intake recommendations than children who had a
breakfast other than USDA SBP (52% vs. 61%, p<0.05). Significantly fewer children
who participated in USDA National School Lunch Program (N SLP) achieved AHF fiber
intake recommendations than nonparticipants (42% vs. 48%, p<0.05). Low fat food

47

sources of dietary fiber in the SNDA children’s diets are described elsewhere (Yagalla et
a1. 1998) . Table 2 shows differences in mean fiber intakes between white, black and
hispanic children. Black children had significantly lower mean fiber intakes than white
children at nearly all ages. Hispanic adolescents, however, had Significantly higher mean
fiber intakes than white adolescents. These ethnic differences in fiber intake do not seem
likely to be attributable to differences in family income, which was only very weakly
correlated with fiber intake.
DISCUSSION

AS illustrated in Figure 1, most children under the age of 10 in the SNDA sample
achieved the American Health Foundation age plus five grams of dietary fiber per day,
however, this declined as children entered adolescence. Adolescent females were
significantly less likely to achieve AHF dietary fiber recommendations than adolescent
males, and the drop occurred 2 years earlier in females. Adolescents have more freedom
in food selection, including going off campus to fast food and other restaurants during
high school years than younger children. This increased freedom of choice coupled with
a lack of awareness of dietary impact on chronic disease prevention and food sources for
dietary fiber may be contributing to the lower fiber intakes in older children (Murphy
1994) . Adolescent females are perhaps most at risk for low fiber intakes due, in part, to
the high prevalence of restrictive eating, dieting and meal skipping (Murphy 1994) .
Burghardt and colleagues in the initial SNDA analysis found that 22% of 15-18 yr. old
females skipped breakfast and 12% ate no lunch (Burghardt 1995).

The chronic pattern of low fiber intake after elementary school in youth is of

48

concern due to the elevated risk for chronic diseases in adulthood such as coronary artery
disease, some cancers, and hypertension (Williams et al. 1993, National Cancer Institute
1987). McClung and colleagues, in a study of dietary fiber intake and gastrointestinal
function in children aged 2 to 12 yr., found that approximately half of their subjects
consumed less than age + five g/day (McClung et al. 1995). They noted that those
children with low fiber intakes were significantly more likely to experience constipation
than children consuming adequate fiber and water (McClung ct al. 1995). Despite
physician instruction to children and parents to increase fiber intakes, McClung et al.,
reported that children with constipation continued to have less than half the mean fiber
intakes recommended (McClung et al. 1995).

Significantly fewer participants in USDA meal programs achieved AHF fiber
intake recommendations than children who did not participate. Although this could be
due to gender, age, or other differences between program participants and
nonparticipants, another reason for this could be an inadequate number of high fiber
offerings in USDA programs. For example, Burghardt and colleagues, in the original
analysis of the SNDA data, found that although high schools and middle schools were
more likely to offer fi'esh hits and vegetables than elementary schools (approximately
35% vs. 12% of elementary schools, p<0.001), they were also more likely to offer low
fiber breads such as hamburger and hotdog buns and white rolls (50% vs. 25%, p<0.001)
and more likely to offer low fiber entrees such as hamburgers, hotdogs, and cheese
sandwiches than elementary schools (25% vs. 10%, p<0.001)(Burghardt et al. 1995).
Hot chocolate and chocolate milk have been found to be popular sources of dietary fiber

49

for children of all ages, however, these beverages were only offered on 42% of school
days (Burghardt et al. 1995).

The SNDA participants’ mean and median dietary fiber intakes are higher than
those found in the Bogalusa Heart Study (Nicklas et al. 1995). The Bogalusa Heart Study
was a prospective study of coronary artery disease risk factors in children. The dietary
recalls collected from 1976-1987 were reanalyzed for dietary fiber using USDA
computerized nutrient database version 4, the same database used for our analysis. The
mean dietary fiber intakes often year olds in the SNDA sample was 15.2 g/day compared
to mean intakes ranging 11.2 g/day in 1978 to 12.5 g/day in 1987 for ten year olds in the
Bogalusa sample. Thirteen year olds’ dietary fiber intakes in the SNDA averaged 15.8
g/day compared to Bogalusa sample intakes ranging from 11.5 g/day in 1976 to 14.0 in
1987 (Nicklas et al. 1995). There appears to be a trend toward increased dietary fiber
intake over time from 1976 to 1987 values within the Bogalusa sample compared to the
SNDA 1992 levels.

We found that dietary fiber intakes in SNDA adolescent females aged 12-18 were
significantly lower than those of adolescent males ages 12-18. The Bogalusa study also
reported similar results with females having significantly lower fiber intakes than males
(p<0.05) (Nicklas et al. 1995) . When grams of fiber per 1000 calories was used to adjust
for energy intakes, Nicklas and colleagues found the gender difference was no longer
significant (Nicklas et a1. 1995). However, neither the American Health Foundation
fiber recommendations for children nor the USDA Food Guide Pyramid fiber
recommendations for adults are adjusted for differences in energy intake or gender

50

(USDA 1992) , so we chose to use grams of dietary fiber consumed per day rather than
grams of fiber per 1000 kcal.

Both the mean (15.6 g/d) and median (14.0 g/d) dietary fiber intakes of SNDA
children aged 6-11 yr. were higher than children aged 6-11 yr. in the Nationwide Food
Consumption Survey (NFCS) 1977-78 (12.1g/day) and 1987-88 (11.5 g/day) (Saldanha
1995). Similarly, for SNDA males aged 12-18 yr. mean dietary fiber intake was 21.0 g/d
and median 18.3 g/d, higher than the mean of 15.2 g/d in the NF CS 1977-78 males of this
age and 14.0 g/d in NFCS 1987-88 males aged 12-18 yr. Females aged 12-18 yr. in the
SNDA study also had higher mean (12.6 g/d) and median (12.5 g/d) fiber intakes than
NFCS 1977-78 females (mean 11.0 g/d) and NFCS 1987-88 females (10.6 g/d). The
NFCS collected two day diet records and one 24 hr. recall per child, which were averaged
to assess nutrient intake. However, Saldanha and colleagues used USDA Nutrient
Database for Individual Food Intake Surveys version 2 released in 1985 to assess the
dietary fiber intakes of these children (Saldanha 1995). Dietary fiber content of many
foods was still missing in that version, and could have led to underestimation of total
daily fiber intakes. Version 4 of the USDA nutrient database is more than 98% complete
for dietary fiber content of foods (USDA 1990). The higher mean intakes found in the
SNDA sample could also reflect a greater variety of high fiber foods available in the
current food supply than earlier. For example, many new high fiber ready-to-eat
breakfast cereals have been released in the past decade. High fiber cereals were found to
be popular sources of dietary fiber in children’s diets in our study as well as those
conducted by other researchers (Saldanha 1995, Nicklas et a1 1995). Other low fat food

51

sources of dietary fiber consumed by the SNDA participants are enumerated elsewhere
Wagalla 1998).

Levine and Guthrie, in a study of nutrient intakes and eating patterns of teenagers
in the Continuing Survey of Food Intakes of Individuals (CSFIl) 1989-1991, found that
mean fiber intake ranged from 11 to 16 grams per day in children ages 13-18 (Levine et
al. 1997). Dietary fiber Intakes of females were particularly low in this analysis, with
white females aged 13-18 averaging 11.2 g/day and black females averaging 11.0 g/day.
Hispanic and other ethnicity children were not included in the CSFII analysis due to
insufficient sample size (Levine et al. 1997) . The CSFII analysis was performed with the
USDA fiber updated version 4 database.

The dietary fiber intakes of white, black and hispanic children in the SNDA study
differed significantly as shown in Table 3. We found that black children, at most ages,
consumed significantly less fiber than white children. This differs from the Bogalusa
study, which found black females ages 10-17 yr., consumed significantly more fiber than
white females (p<0.01)(Nicklas et al. 1995). Hispanic children were not included in the
Bogalusa study. Hispanic adolescents in the SNDA sample consumed significantly more
fiber at most ages than white adolescents. This could be due to the culturally prevalent
practice of using beans and bean products in Hispanic cuisine. Family income, which has
been found in earlier studies of adults to account for the variance in etlmie and racial
differences in diets, was only weakly correlated to fiber intake in our study.

Our study is the only analysis of a large nationally representative sample of
children ages 6-18 using the most up to date dietary fiber databases and intake

52

recommendations, therefore, presents valuable and valid information on differences in
mean dietary fiber intakes by age, gender and ethnicity. However, the SNDA study
dietary information is based on a 24 hour recall, not an assessment of diet over time.
Therefore, individual achievement of dietary fiber recommendations over time may vary
due to day to day variability in food intake. Multiple regression analysis of factors
contributing to differences in dietary fiber intake would be valuable, but only if dietary
information was based on a food frequency questionnaire or other assessment of diet over
time rather than a single 24 hour intake. Since the AHF recommendations themselves
were only established in 1994, it is important to assess achievement with the available
SNDA data. Our analysis by ethnicity (race) would have been stronger if the variable
race in the SNDA data set was by determined by student self report rather than being
coded by observers.

Our study examines the distribution of dietary fiber intakes in the normal diets of
U.S. school children. Most studies documenting the impact of dietary fiber on promoting
bowel regularity (McClung et a1. 1995), weight reduction in obese children (Ryttig et al.
1989, Rigaud et al. 1990, Astrup et a1 1990) and serum cholesterol reduction in children
at risk for coronary artery disease (Williams et al. 1991, Dennison et al. 1993,
Blumenschein 1991) have used fiber supplements such as psyllium or oat bran, rather
than increasing the consumption of fiber rich foods in the daily diet. The risk of chronic
disease is best reduced by incorporating healthy eating habits in daily life, rather than
supplement usage. Physicians and other health professionals should educate children and
their families on the benefits of regularly consuming low fat, high fiber diets to reduce

53

risk of chronic diseases. Another key place for such nutrition education is school
cafeterias, where low fat, high fiber foods can be promoted with educational graphics
explaining their health benefits. Nutrition education should begin during early childhood
to promote healthier eating habits in adult life, but our data clearly show the need to
reernphasize these concepts during adolescence. Drewnowski and others have
documented that the foods adults prefer most are those they were offered most and
consumed most frequently as children (Drewnowski 1997). Therefore, health
professionals Should concentrate their disease preventive efforts on children and
adolescents.
CONCLUSIONS

Dietary fiber intakes of children, especially adolescent and females fail to meet
American Health Foundation recommendations. Achievement of AHF dietary fiber
recommendations falls from over 60% in 6 year olds to less than 20% in 18 year olds.
Black children have lower mean dietary fiber intakes than white children, while Hispanic
adolescents have higher mean fiber intakes than white adolescents. Participants in USDA
school meal programs have lower average daily fiber intakes than children who do not
participate is school meal programs. Given the new AHF fiber intake recommendations
and the availability of data on the dietary fiber content of foods, the USDA should
incorporate dietary fiber as one of the nutrient goals for their school meal programs.
School meal programs should offer low fat, high fiber food choices such as legume based
entrees as meat alternates, whole grain breads and cereals, and fresh fruits and vegetables
often to promote the likelihood that program participants will meet AHF dietary fiber

54

goals.

Physicians and other health care professionals should emphasize to both parents
and children the importance of consistently making high fiber food choices as a part of a
healthy lifestyle. Researchers have documented that parents and children are not able to
translate nutrient goals to foods in their usual diet (McClung et al. 1995). Therefore,
physicians and dietitians should make food-based recommendations such Aeat more
whole grains, , beans legumes, and fresh fruits and vegetables; rather than nutrient-based
recommendations such as Aincrease your fiber intakez. Low fat, high fiber foods
voluntarily selected and consumed by children have been identified fi'om SNDA data and

can be used effectively as part of such educational efforts (Y agalla et a1. 1998).

55

Table 1. Dietary fiber intakes (g/day) of U.S. School children by age and gender.

 

 

Age (yr.) Females Males

Mean : SE n Mean : SE n
6 13.7 : 0.84 52 13.9 : 1.34 60
7 15.1 : 0.71 134 14.4 : 0.78 135
8 13.9 : 0.54 158 16.2 : 0.69 134
9 15.0 : 0.52 186 16.7 : 0.78 200
10 16.2 : 0.72 176 17.1 : 0.78 175
11 17.3 : 0.83 155 18.6 : 0.92 143
12 14.0 : 0.69 123 18.1 : 0.65* 162
13 14.8 : 0.67 148 20.6 : 0.89“ 137
14 14.9 : 0.89 126 22.9 : 1.26* 137
15 14.3:0.79 131 19.9: 1.08* 117
16 15.5: 1.16 120 21.6: 1.05* 120
17 15.1 : 1.14 110 23.7 :1.74* 127
18 14.8 : 1.99 44 20.1 : 1.36* 50

 

SOURCE: weighted tabulations from USDA School Nutrition Dietary Assessment Study
data, 11 shown are unweighted, but analyses are done on weighted means.

* statistically significant different with a two-tailed t-test, p<0.05.

56

Table 2. Mean dietary fiber intake (g/day) of U.S. school children by age and ethnicity

 

 

Age White Black
Hispanic
yr. Mean : SE n Mean : SE n Mean : SE n
6 14.7 + 1.34 62 13.1 : 0.92“ 24 11.9 :1.24* 22
7 15.5:0.60 159 13.5:0.83* 38 l6.6:2.11** 41
8 15.2 : 0.52 183 15.9 : 0.97 44 14.5 : 1.27* 41
9 16.1 : 0.61 259 15.2 :1.07* 60 15.7 :1.11 53
10 l6.7:0.67 243 15.1:l.15* 48 l8.5:l.9l** 47
11 18.5 : 0.77 210 16.8 :1.29* 32 17.0: 1.63* 39
12 16.5 :0.65 194 15.5 :1.13* 44 16.1 : 1.38 28
13 18.3 : 0.85 174 15.6 : 0.93* 53 17.0 : 1.98”“ 39
14 18.6 : 1.05 173 16.0 :1.45* 38 23.6 : 2.43" 39
15 16.5:0.80 160 18.6: 1.95" 47 16.6:1.18 36
16 17.2 : 0.93 151 21.2 : 2.37" 47 19.5 : 2.48M 31
17 20.1 : 1.49 145 17.2 :1.24* 47 21.4 : 3.07" 33
18 16.7:1.22 65 l6.5:3.10 16 17.6:1.43** 7

 

SOURCE: weighted tabulations from USDA School Nutrition Dietary Assessment Study
data, n shown are unweighted, but analyses are based on weighted means; students are
collapsed across gender for this analysis because gender distribution within each ethnic
category did not differ significantly.

* mean is significantly (p<0.05, two tailed t-test) lower than mean intake in white
children of same age

** mean is significantly (p<0.05, two tailed t-test) higher than mean intake in white
children of same age

57

Ems? H. 389: cm C.m. mega 0358: 2:0 Beak/8010mm E85.
mocsamaos “LARSON mean HmocBBasmmaosm Ammo + m m\&

we

 

mc 11
o\o
we 1
nc l
.5 i

 

 

 

 

 

 

 

 

 

 

 

 

a N m 9.33.33.33.53;

one Q:

 

 

 

 

 

 

 

h.

 

 

D 25.8..
I mesaem

 

 

 

 

Chapter Five

Vended Food Consumption by U.S. School Children
ABSTRACT
Objectives: 1) to determine percent of U.S. children consuming vended foods/beverages
on a typical school day by grade, race, gender, eligibility for subsidized meals, and
participation in USDA meal programs, 2) to assess the nutrient contribution from vended
items, 3) to determine differences in nutrient intakes and Dietary Guideline (DG) and
RDA achievement between vended item consumers and nonconsumers.
Methods: The USDA School Nutrition Dietary Assessment Study data with a nationally
representative sample of 3,381 children in grades 1-12 from the 48 contenninant states
was examines. Food codes from the 24 hr. dietary recall data were matched to the USDA
computerized nutrient database for population survey analysis version 4 to determine
nutrient contribution of vended foods/beverages. The SNDA assessed DG achievement
for fat (<30% energy), saturated fat (<10% energy) and cholesterol (<300 mg/d) as a
categorical variable. RDA (for protein, iron, calcium, magnesium, vit A, C, B6, B12,
folate, niacin, thiarnin and riboflavin) achievement was also determined as a categorical
variable as part of the SNDA. Descriptive analyses were performed using SAS version
6.03 and J MP statistical software. Chi square analyses and t-tests were performed to
assess differences between groups.
Results: 11.4% of U.S. school children consumed vended foods/beverages and these

items contributed an average of one third of the daily energy intake of those who

59

consumed them with energy ranging from 0 to 2170 kcal. Males and females consumed
vended items equally (l 1.1 vs. 11.1%) and children of all ethnicities consumed vended
foods equally (11.9% Hispanic, 12.5% Asian, 10.9% Caucasian and 12.2% Afiican
American). Student participation in the USDA School Breakfast Program reduced the
percent of children consuming vended items compared to nonparticipants, however, no
relation was observed with participation in the USDA National School Lunch Program.
Fewer children who consumed vended items achieved DG than nonconsumers, but more
vended food consumers achieved age-specific RDAs than nonconsumers.

Conclusions: At present no federal guidelines on vending machine use during meal hours
in schools or nutritional quality of the items offered in them exist. Given the range of
nutrients being derived fi'om vended items and the percent of children consuming them,
dietitians should incorporate vended foods as part of the ADA’s Child Nutrition and
Health Campaign. Health professionals should lobby for legislation governing vending
machine use in schools and encourage local and school officials to offer healthful options

in vending machines.

60

Vended Food Consumption by U.S. School Children

INTRODUCTION

The major sources of foods in the diets of school aged children include foods: 1)
eaten at or packed from home, 2) purchased at school as part of the United States
Department Of Agriculture’s (USDA) school meal programs or A La Carte offerings,
3) purchased from restaurants, 4) given to children by others, and 5) purchased from
vending machines and school snack shops (Burghardt et al. 1995). Although the USDA
and others have studied the impact of participation in USDA school meal programs on
the diets of children, very little information is available in the literature concerning the
impact of competitive foods such as those sold via vending machines and in school stores
(Dwyer 1995).

At present there are no federal guidelines regulating what foods and beverages
are offered in vending machines in schools. Rather, this is a matter regulated locally
within each school and school district (Dwyer 1995). In 1970, Congress authorized the
USDA to regulate the sale of foods that compete with National School Lunch Program
(N SLP) lunches. Initially, USDA regulations limited the foods sold in schools to those
that either contributed to one of the required meal components or that were served as an
additional item with the school meal (Burghardt 1995). However, in 1972, Congress
amended the USDA authority over competitive foods by exempting from restriction the
sale of any food whose proceeds accrued to the school or to school organizations

(Burghardt et al. 1995).

61

In 1977, Congress further modified this exemption by authorizing that the USDA
establish regulations approving foods that could be sold in school in competition with the
meal program. Such regulations were intended to restrict the sale of foods of poor
nutritional quality (such as sweetened soft drinks, candy and fiied chips) from being sold
in school vending machines, stores and A La Carte programs. However, by the mid-
19803, lawsuit judgements forced the USDA to reconsider its regulations on competitive
foods (including those available in vending machines) and to loosen the restrictions on
what foods could be sold (Burghardt ct al. 1995).

Due to the scarcity of information in the scientific literature on the actual
contribution of vended foods to children's diets, it is difficult to say whether such foods
are truly detrimental to child nutrition. Health professionals have long expressed concern
about the apparent contradiction between classroom promotion of healthful diets,
followed by lunch or snacks of chips and candy (Hinkle 1982). One study found that
32% of a junior high school population consumed vended foods as either an entire lunch
or part of one (Ho et al. 1991). However, healthful foods and beverages (e. g. sandwiches,
low-fat yogurt, and milk) could also be supplied via vending machines as a viable
alternative to the foods offered in the school meals programs (Wilbur et al. 1981). A few
studies have found that when popular, healthful foods were available in vending
machines, they sold well (Hruban 1997, United Fresh Fruit and Vegetable Association
1978, Hoerr et al. 1993). Therefore, vended food consumption by U.S. school children
deserves investigation.

The USDA School Nutrition Dietary Assessment Study (SNDA) collected data on

62

foods purchased by children from vending machines and stores in schools but did not
examine the relation of consumption of these foods and beverages on the nutrient intakes
of school children (Burghardt 1995). In the publication of the SNDA results, the study
authors expressed the need to examine the issue of vended food nutrient contribution
more closely (Dwyer 1995). A potential scientific basis for legislation regulating foods
and beverages in vending machine in schools could be obtained from examining the
SNDA study vended foods data.

The objectives of our study were, therefore, 1) to examine the percentage of
school children consuming vended foods by grade, race, gender, eligibility for free or
reduced price school meals, participation in USDA school meal programs, and location of
the vending machine in school using the SNDA data; 2) to assess the average nutrient
contribution from vended foods in the diets of school children who consumed them; and
3) to determine differences in 24 hour nutrient intakes and achievement of USDA Dietary
Guidelines for Americans (DG) for fat (<30% energy), saturated fat (<10% energy) and
cholesterol (<300 mg/d) and National Academy of Sciences Recommended Dietary
Allowances (RDA) for protein, calcium, iron and other select nutrients between vended
food consumers and nonconsumers in elementary, middle and high schools (U S DHHS
1989, National Research Council 1989).

METHODS

Sample and Data set

 

The SNDA data set includes information collected in the winter and spring of
1991-1992 on a nationally representative sample of 3,381 children in grades 1-12 from

63

545 schools across the country (Burghardt 1995). The sample was a multistage, stratified
sample of districts, schools within sampled districts, and students within a subsample
(329/545) of the schools. The sample of schools was designed to be representative of all
schools in the 48 contiguous states plus the District of Columbia. The sample of students
was designed to be representative of all students in grades 1 through 12 who were
attending school on a typical school day. Students in kindergarten, pre-kindergarten, or
reading readiness (pre-first grade) classes, as well as students enrolled in special
education programs with self contained classrooms were excluded from the SNDA
sample (Burghardt 1995).

Food recall data for a 24 hour period including a school day were collected by
trained interviewers administering questionnaires. Parents of children in grades 1-3 were
interviewed along with the children to obtain dietary intake data. All children in grades
4-12 reported their own intakes. All data were coded to maintain anonymity of
participants (Burghardt 1995). Food composition data on all foods children consumed
from the USDA NSLP and School Breakfast Programs (SBP) at each school were coded
by nutritionists at the Minnesota Nutrition Coordinating Center (Burghardt 1995) .
Although list of foods children consumed from A La Carte offerings, vending machines
and school stores were collected, the initial SNDA study did not analyze the nutrient
content of these foods. However, each food in each study participantzs diet recall was
coded to identify the source from which it was obtained. The foods obtained from both
vending machines and on campus stores were given the same code, so it was not possible
to separate these two sources. The location of the vending machine in schools was coded

64

as a categorical variable (either in or out of the school cafeteria) (Burghardt 1995).
$121118

The initial analysis of the SNDA conducted by the Mathematica Policy Research,
Inc. did not examine the impact of vended food consumption on the nutrient intakes of
children (Burghardt 1995). SNDA data were obtained from the USDA on IBM
mainframe magnetic reel tapes in Statistical Analysis for the Sciences (SAS) ASCII files
with accompanying SAS programs to label data. Five relevant files (stanal.dat,
stfcomp.dat, strecipe.dat, scvend.dat, stchar.dat) were downloaded to SAS version 6.03
personal computer files (SAS Institute 1998). For analysis by grade, participation in
school meal programs or eligibility for free or reduced price meals all children (n=3,381)
were included. Only Caucasian (n=2188), African American (n=544), Hispanic (n=459)
and Asian (n=167) students’ diets were included in cross cultural comparisons due to
small samples with inadequate power in the mixed race and other categories (n=l90
total). Food codes from the SNDA participants’ diets were matched with USDA dietary
fiber updated Nutrient Database for Analysis of Population Based Dietary Surveys
version 4 to determine nutrient content of foods/beverages from vending machines and
school stores only (not A La Carte foods), as well as to calculate 24 hour total nutrient
intake of children (USDA 1990).

Variables used for analyses included grade in school, total grams per day of fat,
saturated fat, and dietary fiber and total milligrams per day of cholesterol in each
student’s 24 hr. intake, participation in school meals programs, race, and eligibility for
free (student’s family income was <135% of the Poverty Index or P1) or reduced price

65

(student’s family income was >135% of PI but < 185% of PI) school meals.

A student was classified as an SBP participant if they consumed 2 out of 4 of the
USDA required meal components and an NSLP participant if they consumed at least 3
out of the 5 USDA required meal components. DG (for fat, saturated fat and cholesterol)
and RDA (for protein, vitamin A, vitamin C, thiamin, riboflavin, niacin, vitamin B-6,
folate, vitamin B-12, calcium, iron, phosphorus, magnesium, zinc, and total energy)
achievement in 24 hr. intakes was assessed as part of the original SNDA study and
entered as a categorical variable (I: achieved all DGs or RDAS and 0= did not achieve all
DGS or RDAs) (Burghardt et al. 1993).

Students were classified as vended foods consumers if their diet recalls included
any foods or beverages from vending machines or school stores. Descriptive analyses of
differences in nutrient intakes by vended food consumption were conducted using SAS
version 6.03 and JMP from the SAS Institute (SAS Institute 1998, SAS Institute 1998).
All student level data were weighted according to SNDA protocol to make comparisons
nationally representative and adjust for selection bias (Burghardt 1995). Statistical tests
employed to test significance of differences between groups included, Chi-square
analyses and Student t-tests (SAS Institute 1998).

RESULTS

Approximately 668 (1.2%) of the 53,494 foods and food components consumed
by the SNDA study participants were purchased from vending machines or stores in
schools . The foods consumed from vending machines and stores ranged from soft drinks
to refiigerated items such as sandwiches and are reported in the SNDA results (Burghardt

66

et al. 1993). We found that foods or beverages from vending machines or school stores
were consumed by 1 1.4% (n=387/3,381) in the unweighted sample) of U.S. school
children. The percentage of children consuming vended foods/beverages by grade is
represented in Figure 1. Note the increased consumption of vended foods during middle
school years and the decline during high school years when many schools allow off-
campus privileges. Hispanic (11.9%) and Asian children (12.5%) were significantly
more likely to consume vended foods than Caucasian (10.9%) and African American
(12.2%) children (p<0.05).

Interestingly, vended food consumption did not differ significantly between males
and females (11.1% vs. 11.1%). Eligibility for free or reduced price school meals did not
significantly affect vended food consumption, in fact the trend was toward increased
consumption among those who qualified for free or reduced price meals (12.6% of those
not eligible vs. 7.6% of those eligible). Students who participated in the USDA SBP
were Significantly less likely to consume vended foods than nonparticipants (3.5% vs.
11.6%, p<0.001). However, participants in the NSLP were significantly less likely to
consume vended foods (15.2% of the nonparticipants vs 7.3% of the participants).
Whether the vending machines were located inside the school cafeteria or elsewhere in
the school also did not affect prevalence of vended food consumption by children.

Table 1 shows the list of foods and beverages consumed by U.S. school children
ranked by frequency of items consumed. Soft drinks, candy, fiied chips, cakes and
cookies were the most popular items (n=474 total). Healthful items such as sandwiches,
milk, bread, crackers, fruit and juices ranked far lower (n=104 total). Miscellaneous

67

beverages included items such as seltzer water, tea, hot chocolate and coffee and
miscellaneous foods included mostly condiments, and a few nuts, seeds, etc.

Table 2 shows the average nutrient contribution of vended foods/beverages to
the diet of U.S. school children in elementary, middle and high schools (unweighted
n=3 87) who consumed vended foods. Note that the range indicates that some children
are consuming entire meals worth of food energy and macronutrients in the form of
vended foods/beverages, while others seem to be using these foods/beverages as adjuncts
to their diet. The lower end of the range (zero food energy and macronutrients)
represents sugar-free soft drinks consumed.

The 24 hour intake of food energy from various nutrients in the diets of vended
food consumers vs. nonconsumers by gender and grade category is shown in Table 3.
Within most gender and grade categories except middle and high school males, vended
food consumers had lower mean intakes of most micronutrients (e. g. vitamins A & C,
calcium, and iron) while energy and macronutrient intake was often higher compared to
their counterparts who did not consume vended foods or beverages. These differences in
nutrient intakes among vended food consumers were statistically significant for some
nutrients, although not for all.

Table 4 shows the percent of male and females achieving USDA DG for fat,
saturated fat, and cholesterol in a 24 hr. period was significantly different for vended
foods consumers vs. nonconsumers in each grade category except middle and high school
males. Note that the trend was for a smaller percentage of those who did consume
vended foods achieving fat, saturated fat and cholesterol recommendations compared to

68

nonconsumers. However, vended foods consumers in some gender and grade categories
were significantly more likely than nonconsumers to achieve age appropriate RDAs for
protein, iron, calcium and other vitamins and minerals.

DISCUSSION

Given that one in ten U.S. school children consumes vended foods and beverages
in a typical school day, it is imperative that dietitians and other health professionals
address the nutrient content of these items as part of the American Dietetic Association=s
Child Nutrition and Health Campaign (Stedronsky 1998). The age, gender, and ethnicity
patterns in vended food consumption in U.S. school children is of particular interest.
Specifically, the peak of vended food consumption in high school years could reflect both
increased independence in food selection and access to money. Whatever the reasons,
middle and high school group appear to be the most appropriate target for educational
efforts encouraging healthy vending selections. Such educational messages Should be
part of the Child Nutrition and Health Campaign’s objective to publicize the importance
of healthful snacks for good nutrition for children (Stedronsky 1998).

It is important to identify the reason for vended food consumption in school
children since its prevalence seems universal rather than limited to a specific gender or
ethnic group. The most common reason for student nonparticipation in USDA meal
programs, and perhaps use of alternative sources of food such as vending machines was
dislike for the foods offered (Burghardt 1993). Interestingly, there appear to be no gender
differences in the prevalence of vended food consumption in U.S. school children.
Surprisingly, an individual’s qualification for free or reduced price school meals did not

69

influence their likelihood of consuming vended foods. Vended foods and beverages in
schools may be priced so inexpensively that they are becoming an economic alternative
to reduced price or full priced school meals. These issues warrant further research.

Burghardt and colleagues, in their technical report of the SNDA findings reported
that only 16 percent of elementary schools, compared with 52 % of middle schools and
88% of high schools in the U.S. (based on weighted tabulations) have at least one
vending machine in school (Burghardt et al. 1993). Vending machines are available for
student use at all times including during school hours in 16% of schools. Beverages
were offered in nearly all the school vending machines, but soft drinks constituted more
than 2/3 of the beverages. Baked desserts (such as creme filled cakes), snack chips and
candy constituted the majority of the other offerings in school vending machines while
only 1% of high school machines (none of the elementary school or middle school
machines) offered fresh fruit and none offered yogurt (Burghardt et al. 1993). This is
alarming when considering the range of energy and macronutrients that U.S. school
children are deriving fiom vended foods and beverages. The items most frequently
purchased from vending machines and school stores identified in our study reflect the
limited range of foods and beverages offered in school vending machine and stores. If
healthful options are not available for e. g. sandwiches and yogurt in refrigerated vending
machines, children are at risk for poor nutrition by choosing vended items over a USDA
meal or a meal packed from home.

We found that the average food energy derived from vended foods and beverages
in school children consuming them was about one third of the their daily intake. One

70

study of vended snack consumption in college students found that one in ten students
received 20% of their calories from vended snacks (Aljadir et al. 1981). When we
examined the achievement of the age-appropriate dietary recommendations (DGs and
RDAs) based on vended food consumption, we found surprising contradictions. Vended
food consumers tended to be less likely to achieve the DG for fat, saturated fat and
cholesterol than nonconsumers. This can be explained by the high prevalence of
sweetened and fried dessert and chips consumption in vending consumers. The RDAS,
which look for actual amounts of various macro and micro nutrients consumed, reflected
differences in diet quality of vended food consumers compared to nonconsumers. Some
individuals who consumed Another way to examine the health impact of vended foods
and beverages would be to do food group analysis.

One study of unrefrigerated vended snacks at the National Institute of Health
found an increase in total snack sales when low-calorie snacks were marketed in a
Ahealthy snack machine placed next to a machine vending standard snacks (Wilbur et al.
1981). Hoerr and colleagues also found that sales of vended snacks increased among
university students when colorful graphics identifying nutrient-dense snacks were placed
on vending machines, provided the snacks available were ones popular with students, e. g.
pretzels (Hoerr et a1. 1993). A popular snack item was defined as one ranking among the
vending industry’s top 20 best-selling vended snacks nationally (Hoerr et al. 1993).
Other studies using refiigerated vending machines found that if nutrient-dense snacks
were offered, students bought them and vended snack sales did not decrease (Aljadir et al.
1981, French et al. 1997).

71

Researchers at the University of Minnesota’s School of Public Health recently
examined pricing strategies to promote low-fat snack choices through vending machines
(French et al. 1997). They found that the sale of low-fat items (those containing 3 grams
of fat or less per serving) increased from 25.7% to 45.8% of total sales when the prices of
these items were reduced by 50%. This increase in low-fat snack sales did not last,
however, beyond the price reduction intervention period, falling to 22% of total sales
when prices returned to pre-intervention levels (French et al. 1997). The low-fat items in
this study were not identified with any nutrition information, only bright yellow price
labels. This suggests that environmental approaches to promoting low-fat food choices,
such as competitive pricing, may promote low-fat food purchase and consumption in the
population as a whole, even independent of point-of-selection nutrition education (French
et al. 1997). French and colleagues suggest that small price reductions in low-fat items
coupled with modest price increases for hi gh-fat items could result in a net revenue gain
and a net profit for vendors (French et a1. 1997). A community-based survey of over 800
men and women in Minnesota found that requiring low-fat, healthful foods to be
available in school cafeterias and eliminating high-fat foods from vending machines were
the most favorably evaluated public health policies (Schmidt et a1. 1989).
IMPLICATIONS

Dietitians and other health professionals should advocate with policy makers for
legislation regulating the nutritional quality of foods and beverages offered in school
vending machines and stores, especially during meal hours. The manner in which the
nutritional quality of these items is assessed is also important. It is appropriate to restrict

72

access to vending machines during the school meal hours if their only offerings are foods
and beverages of poor nutritional quality that could place children at risk for malnutrition.
However, an alternative means of ensuring the diet quality of school children is to
promote the sale of healthful foods and beverages, such as fresh fruit, yogurt, low-fat
milk, juice and sandwiches in school vending machines.

The research available on the sale of such healthful items from vending machines
indicates that substantial market value exists among children and adolescents. Schools
and local officials can gain revenue and encourage good nutrition by investing in vending
machines with healthful offerings for their students’ use. Dietitians and other health
professionals must bring the issue of vended foods to the attention of policy makers

working to improve and protect the diet quality of U.S. school children.

73

Emcee H. @983 0». Cum. moroa GEES: Oosmcsdsm
x538 mooam w% 08%

3b

 

 

 

 

 

 

 

74

Table 1. Vended Foods and Beverages Consumed by U.S. School Children Ranked by
Frequency of Consumption

 

 

Food/Beverage Consumed Number of Items
Carbonated Soft Drinks 199
Candy 129
Fried Chips 85
Cakes, cookies, bars 61
Sandwiches/pizza/burrito/calzones 39
Fruit juice/juice drinks 38
Crackers/pretzels/popcom 30
Milk 15
Breads 14
Ice Cream 6
Fresh Fruit 5
Yogurt 1
Miscellaneous Foods 81
Miscellaneous Beverages 12

 

SOURCE: tabulations from the USDA School Nutrition Dietary Assessment
Study data

75

Table 2. Nutrient contribution of foods and beverages from vending machines or school
stores to the 24 hr. intake of U.S. school children in elementary, middle and high
schools consuming them (unweighted n=387/3,381)

 

 

Variable Mean : SE Range
Elementary grades (1 -5) n=47

Energy (Kcal) 199 : 25.7 20 - 912
Carbohydrate (g) 29 : 3.5 0 - 103
Total Fat (g) 7.8 : 1.2 0 - 59
Saturated Fat (g) 2.7 : 0.5 0 - 19
Protein (g) 3.8 : 1.0 0 - 30.2
Dietary Fiber (g) 1.0 : 0.2 0 - 8.2
Cholesterol (mg) 9.6 : 2.9 0 - 92
Vitamin A (meg) 21.4 : 5.7 0 - 152
Vitamin C (mg) 7.5 : 3.6 0 - 171
Calcium (mg) 43.4 : 11.0 0 - 1370
Iron (mg) 0.7 : 0.1 0 - 10.5
Middle School grades (6-8) n=90

Energy (Kcal) 353 : 51.1 5 - 1551
Carbohydrate (g) 57 : 5.6 1 - 209
Total Fat (g) 11.8 : 2.5 0 - 86.7
Saturated Fat (g) 4.6 : 1.0 0 - 31.7
Protein (g) 6.9 : 2.1 0 - 68.7
Dietary Fiber (g) 1.8 : 0.5 0 - 15.4
Cholesterol (mg) 16.8 : 5.4 0 - 191
Vitamin A (meg) 30.5 : 9.1 0 - 675
Vitamin C (mg) 10.7 : 4.1 0 - 128
Calcium (mg) 93.5 : 26.0 0 - 1370
Iron (mg) 1.5 : 0.3 0 - 11.6
High School grades (9-12) n=250

Energy (Kcal) 343 : 23.6 0 - 2170
Carbohydrate (g) 57 : 3.1 0 - 374
Total Fat (g) 11.4 :12 0 - 95

76

Saturated Fat (g)
Protein (g)
Dietary Fiber (g)
Cholesterol (mg)
Vitamin A (meg)
Vitamin C (mg)
Calcium (mg)
Iron (mg)

4.1 i 0.5
5.5 i 0.9
1.5 i 0.1

14.6 i 3.0

29.1 i 6.3

12.8 i 2.9

70.7 i 11.6
1.2: 0.1

0-53
0-79
0-14
0-255
0-806
0-481
0-1594
0-18.l

 

77

...00.0 m. .Som: NA 3. 00303 .0308 0.. GM. 00000. 03.03: 00:38.5 3000.. 30% ~50 0208.000
Annumq. 5.205.388 <0. 00000000803 A=HN00£ .0 0.03033. 8.00.0 0.2.

 

 

Em: 00:00. @0000.
$500.0 2.0.00 x H mm 0,080.8 x H mm
20000000803 <0. 00000303 200803303 <0. 00:05:03

5050035. «3000 2-3 a n gm. 0 H mm 0 u ....N 0 u .0
9.28. 080.. Bun H 00.0 ~30 H .000 ~00~ H 0.... ~50 H .3...
03.00.00.088 Am. 08 H 0:0. 0?. H 3.0 30 H 0.0 ~00 H 3.0
...08. mm. Am. 0. H ..m 0.. H 0.0 00 H ..0 mm H 0..
magma; ...». Am. 00 H 0.0 mm H N. 00 H 0.0 mm H 0..
.0385 Am. 00 H ..m 00 H 0.0 3 H ..a E H 0.0
9035. 3.0018 .0 H 0.0 .m H 0.0 .m H 0.0 .A H 0.0
0.00.0083. 30m. ~00 H 0.0 ~00 H ~90. ~00 H 0.0 Nu. H 00b
$83.: > 900$ .00. H 00.0 ..00 H 000.. 03 H N00 .30 H 0.0
$83.: 0 cam. .00 H A... .0. H 0.0 SN H A0 .8 H .0.~
00.0.03 cum. :3 H ~00 .000 H 3.. 000 H .09 wow H 5.0
.8: 3.8 .0H 0.0 .0H ..0 .AH 0.0 .0H ..0
2:0...» @0300. «.0000 8-8 0 n .30 0 n Am 0 u 000 0 N Am
mnaamv. 080.. N000 H 00.0 030 H ~30 NNMN H 3.0 ~30 H 3.0.0
00303088 Am. 0.3 H 0.0 h... H 00.. 00. H 0.0 000 H .00
4.03.0028 ..NH N0 .00H .... 00H Po 0. H 0.0

$8880 ...0. Am. AN H ... mm H 0.0 0.0. H ... 00 H 0.0

78

.0800... Am.
0.00090 2.0010.
00.0.0083. cam.
$003.: > 3.08
3008.: 0 cam.
00.0.03 30m.
.00: 3.0.

Em: 00.000. «0.00.00 A0. 8

E00000. 080..
000.003.0008 Am.
...000. 0.00 A00
008.080 0.00 Am.
00000... E.
0.00030 .0300 Am.
00.0.0083. 00m.
$008.: > 3.0%
$008... 0 30m.
00.0.5.0 cam.
.00: A80.

.000 0.0

.00 0.0
0000 .00
.00. 000.0
.000 0.0
.0000 0.0

000 0.0

:quw

0080.80
0.00 .00
.0.0 0.0
000 0..
.000 0.0
0.0 0.0
0000 .00
.0000 00.0
.000 ...0
.0.00 0.0
000 ..0

.0.0 0.0
000 ..0
0000 00.0

.0000.00.0
.0.0 00.0

.000 0.00.0
0.0 0.0

0n..0

0000 0 0.0
000 0 00.0
.00 0 0.0

00 0 0..
.00 0 0.0
00 0 ..0
000 0 00.0

.000 0 000.0
000 0 00.0

.000 0 00..

00 0 ..0

00 0 0.0
.0 0 0.0
000 0 .00
000 0 00.0
.00 0 0.0
.000 0 00.0

.0 0 0.0

:n...&

0.00 0 00.0
000 0 0.0
00 0 0..
0. 0 ..0
0. 0 0.0
.0 0 0.0
000 0 .00
00. 0 00.0
.00 0 0.0
000 0 00.0
.0 0 0.0

000 0.0

.00 ..0
0000 00.0
0000 00.0
.0. 0 .0.0
000 0.00.0
.00 ..0

=H.0.

00000..0..0
0000 00.0
000 0.0
000 0.0
000 0.0
.00 ..0
0000 00.0
0000.0..0
.000 .0..
0000 00.0
.00 ...

 

mOCw0m. 20.0300. 00.00.00.000 .0003 0.0 00.000. 2:30.00 0.00030 000000003000 080.00 0000

... 30000 2.9.: 002.00 000. 0000.0 00.00030 0.030.002? 0.0.0.0003 3. 00.800-500.00 0-800. 0A0.0m.

79

Table 4. Achievement of Dietary Guidelines and Recommended Dietary Allowances* in

Elementary, Middle and High School Male and Female Vended Food

Nonconsumers (n=2994) vs. Consumers (n=3 87, unweighted).

 

 

Males (%) Females (%)
Variable Nonconsumers vs. Consumers Nonconsumers vs. Consumers
Elementary grades (1-5) =781 n=28 n=782 n=19
Dietary Guideline
Achievement 14.0 14.2 16.5 16.7
RDA AchievementM 39.0 35.2 27.2 16.9
Middle School grades (6-8) n=3 79 =48 n=353 =42
Dietary Guideline
Achievement 12.2 6.7 14.5 17.1
RDA Achievement" 31.6 40.7 13.8 12.1
High School grades (9-12) n=353 n=119 n=345 n=l31
Dietary Guideline
Achievement 7.8 6.8 18.7 15.7
RDA Achievement" 27.1 35.5 11.4 11.8

 

SOURCE: weighted tabulations from the USDA School Nutrition Dietary Assessment

Study data

* Dietary Guidelines used were total fat (<30% energy), saturated fat (<10% energy) and
cholesterol (<3OO mg/d); RDAs were age-appropriate values for protein, calcium,

iron, etc.

** RDA achievement for both male and female vended food consumers was significantly
lower than nonconsumers in all grade categories by Chi-square analysis, p<0.05.

8O

- Table 5. Mean 24 hr. nutrient intake of U.S. school children consuming vended foods and
beverages(n=387, unweighted) vs. nonconsumers (n=2994) in elementary,
middle and high school grades.

 

 

Variable Nonconsumers . Consumers
Elementary grades (1-5) 11 = 1522 n = 47
Energy (kcal) 2223 i 27.9 2246 i 91.2
Carbohydrate (g) 288 i 3.7 308 i 16.0
Total Fat (g) 86 _‘t 1.3 84 i 4.5
Saturated Fat (g) 32 i 0.5 32 i 1.8
Protein (g) 83 i 1.2 78 i 4.2
Dietary Fiber (g) 16 i 0.3 17 i 2.2
Cholesterol (mg) 271 i 6.4 248 i 20.4
Vitamin A (mcg) 957 i 24.3 961 i215.3
Vitamin C (mg) 131 i 3.4 119 i 20.3
Calcium (mg) 1076 i 16.0 980 i 59.5
Iron (mg) 15 i 0.3 15 i 1.2
Middle School grades (6-8) 11 = 732 n = 90
Energy (kcal) 2595 i 45.9 2856 i 153.2
Carbohydrate (g) 341 i 6.5 370 i 19.3
Total Fat (g) 99 i 2.1 112: 7.2
Saturated Fat (g) 37 i 0.9 43 i 3.1
Protein (g) 94 i 1.7 103 i 6.6
Dietary Fiber (g) 17 i 0.4 17 i 1.1
Cholesterol (mg) 304 i 8.6 340 i” 26.2
Vitamin A (mcg) 1092: 47.8 1033 : 105.1
Vitamin C (mg) 141 i 5.0 146 i 13.8
Calcium (mg) 1206 j: 29.3 1236 i 97.6
Iron (mg) 17 j; 0.4 18 i 1.6
High School grades (9-12) n = 698 n = 250
Energy (kcal) 2731 i 66.8 3263 : 120.9
Carbohydrate (g) 347 i 7.7 432 i 15.9
Total Fat (g) 106 i 3.4 126 i 5.1
Saturated Fat (g) 40 i 1.3 47 i 1.9
Protein (g) 104 i 3.0 112 i 4.9

81

Dietary Fiber (g) 18 i 0.5 20 i 1.0

Cholesterol (mg) 323 i 11.1 366 i 22.4
Vitamin A (mcg) 11201 62.7 1149 j; 125.2
Vitamin C (mg) 157 i 7.0 170 i 14.5
Calcium (mg) 1231 i 38.4 1295 i 61.1
Iron (mg) 19 i 0.7 20 i 1.0

 

SOURCE: weighted tabulations from the School Nutrition Dietary Assessment Study
data

* means within gender and grade category significantly different by Tukey-Kramer t-test,
p<0.05.

82

Chapter Six

Conclusions

Over 1,800 low fat and saturated fat and high dietary fiber foods being voluntarily
selected and consumed by children were identified from existing USDA meal
program offerings. One of the ways in which school food service personnel can
meet USDA=S Recommended Dietary Allowance and Dietary Guideline goals
for school meals is to offer such foods more frequently. At present, popular low
fat sources of dietary fiber such as low fat hot chocolate or chocolate milk and
fresh fruits and vegetables are being offered less than 50% of school days.
Dietary fiber intakes of U.S. school children do not meet American Health
Foundation (age + 5 g/day) recommendations. Adolescent females are at
particular risk for inadequate dietary fiber intakes (only 10% of 18 year females
achieved AHF recommendations). Black children had significantly lower mean
dietary fiber than white children at most ages, while Hispanic adolescents had
significantly higher mean dietary fiber intakes than whites of the same age.
Dietary fiber content of foods is now available in computerized nutrient database
and the American Health Foundation has set pediatric intake recommendations.
However, U.S. school children, even those participating in USDA school meal
programs have inadequate dietary fiber intakes as shown with the SNDA data.
Therefore, the USDA should incorporate dietary fiber into the nutrient goals for
school meal programs to encourage increased dietary fiber intakes in meal

program participants. Pediatricians and other health professionals should educate

83

parents and children on consistently making high dietary fiber food choices using
food based recommendations such as Aeat more whole grain bread, fresh fruits
and vegetables and legumes.

One in ten U.S. school children consume vended foods and beverages in a typical
school day. More Asian and Hispanic children consumed vended items than
children of other ethnicities. Children participating in USDA school breakfast
program consumed vended foods less frequently than those who did not
participate. However, no relation was observed with the USDA National School
Lunch Program. Children who qualified for free or reduced price school meals
were equally likely to consume vended foods as children who did not qualify for
subsidized meals. Equal proportions of males and females consumed vended
foods.

Vended foods contributed an average of one third of the daily intake of energy in
the diets of children who consumed them. A smaller percent of children who
consumed vended foods achieved age-appropriate RDAs compared to children
who did not consume vended foods in all grade categories.

Dietary Guideline achievement often gave misleading results for analysis of the
impact of vended foods to children=s diet quality. Some children who consumed
sweetened sofi drinks, which contributed energy but no other nutrients were found
to have achieved DGs, because DG examine the nutrient density from fat (<30%
energy from fat) and saturated fat (<10% energy) in the diet. Therefore, RDA
assessment or food group analysis would be a more appropriate method of

assessing the relation of vended items to diet quality.

84

Chapter Seven

Recommendations for Future Studies

1) Dietary fiber recommendations are for average intake not intake on a given day.
Therefore, assessment of dietary fiber intake of U.S. school children using food
diaries or other multiple day records, rather than 24 hr. recalls would help assess
the achievement of the American Health Foundation intake recommendations.
This information would provide additional rationale for including dietary fiber in
the USDA=S nutrient goals for the school meal programs.

2) Dietary fiber analysis by meal comparing school meals with those brought from home
or other sources would be helpful in determining how the USDA school meals are
doing in providing dietary fiber to children.

3) Surveys examining local and district level policy regarding vending machine use and
contents of vending machines in schools would help frame future legislation
regulating vended foods.

4) Vended food consumption should also be assessed using food diaries or other multiple
day intakes rather than 24 hr. recalls to determine the average impact of such
items on diet quality of children. Factors associated with vended food use by
children should be further examined. Reasons for greater vended food
consumption by middle school children and Asian and Hispanic children should
be explored further in order to develop appropriately targeted educational

programs for healthful snacking.

85

References

Albertson AM, Tobelmann RC. Impact of ready-to-eat cereal consumption on the diets of
children 7-12 years old. Cereal Foods World 1993;38:428-434.

Aldenderfer MS, Blashfield RK. Cluster Analysis Series: Quantitative Applications in
the Social Sciences 44. Newbury Park, CA; Sage Publications, 1984.

Aljadir LP, Biggs WM, Misko JA. Consumption of Foods From Vending Machines at the
University of Delaware. J of Am College Health Assoc 1981 ;30:149-150.

American Academy of Pediatrics, Committee on Nutrition. Toward a prudent diet for
children. Pediatrics. 1983;71:78-80.

American Academy of Pediatrics, Committee on Nutrition. Carbohydrate and dietary
fiber: In: Barness L, ed. Pediatric Nutrition Handbook. 3rd ed. Elk Grove Village,
IL: American Academy of Pediatrics; 1993: 100-106.

American Heart Association. Diet in the healthy child. Task Force in Cardiovascular
Disease in the Young. Circulation. 1983;67:1411-1414.

Astrup A, Vrist E, Quaade F. Dietary fiber added to a very low calorie diet reduces
hunger and alleviates constipation. Int J Obes 1990;14:105-112.

Blumenschein S, Torres E, Kushmaul E, Crawford .1, Fixler D. Effect of oat bran/soy
protein in hypercholesterolemic children. Ann NY Acad Sci 1991;623:413-415.

Brown PT, Bergan JG, Murgo CF. Current trends in food habits and dietary intakes of
home economics students in three junior high schools in Rhode Island. Home Ec
Research J 1979;7z324-333.

Bull NL. Studies of the dietary habits, food consumption and nutrient intake of
adolescents and young adults. World Rev Nutr Diet 1988;57:24-74.

Burghardt J, Devaney B. The School Nutrition Dietary Assessment Study: Summary of
Findings. Alexandria, VA: U.S. Department of Agriculture, Food and Nutrition
Service, Office of Analysis and Evaluation, Oct. 1993.

Burghardt J, Gordon A, Chapman N, Gleason P, Fraker T. The School Nutrition Dietary
Assessment Study: School Food Service, Meals Offered and Dietary Intakes.
Alexandria, VA: U.S. Department of Agriculture, Food and Nutrition Service,
Office of Analysis and Evaluation, Oct. 1993.

86

Burghardt JA, Devaney BL. Background of the School Nutrition Dietary Assessment
Study. Am J Clin Nutr 1995;61(suppl):]78S-181S.

Burghardt J A. School Nutrition Dietary Assessment Study: overview of the study design.
Am J Clin Nutr 1995;61(supp1);182S-6S.

Burghardt J A, Gordon AR, Fraker TM. Meals offered in the National School Lunch
Program and the School Breakfast Program. Am J Clin Nutr
l995;61(supp1):187S-98S.

Clarke NMA, Grantham-McGregor SM, Powell C. Nutrition and health predictors of
school failure in Jamaican children. Ecology Food Nutr 1991;26:47-57.

Conners CK, Blouin AG. J Psych Research 1982/1983;17:193-201.

Craig A. Acute effects of meals on perceptual and cognitive efficiency. Nutr Rev
l986;May(Suppl):]63-171.

Daum K, Tuttle WW, Martin C, Myers L. Effects of various types of breakfasts on
physiologic response. J Am Diet Assoc 1950; 26:503-509.

Daum K, Tuttle WW, Kisgen R, Imig CJ, Martin C, Agle M. Utilization of nitrogen,
phosphorus, calcium, and iron by men on different breakfasts. J Am Diet Assoc
1951;27:475- 479.

Daum K, Tuttle WW, Kisgen R, Imig CJ, Bongey R, Martin C. Effect of different types
of breakfasts on thiamine and niacin excreted by men. J Am Diet Assoc
1951;27:298-301.

Dennison BA, Levine DM. Randomized, double-blind, placebo-controlled, two-period
crossover clinical trial of psyllium fiber in children with hypercholesterolernia.
Pediatrics 1 993; 123:24-9.

Devaney B, Gordon A, Burghardt J. The School Nutrition Dietary Assessment Study:
Dietary Intakes of Program Participants and Nonparticipants. Alexandria, VA:
U.S. Department of Agriculture, Food and Nutrition Service, Office of Analysis
and Evaluation, Oct. 1993.

Devaney BL, Gordon AR, Burghardt JA. Dietary intakes of students. Am J Clin Nutr
1995;61(suppl): 205S-1ZS.

Drewnowski A. Taste preferences and food intake. Ann Rev Nutr 1997;17:237-53.

87

Dwyer J. Diets for children and adolescents that meet dietary goals. Am J Dis Child.
1987;134:1073-1080.

Dwyer J. The School Nutrition Dietary Assessment Study. Am J Clin Nutr 1995;
61(suppl):173S-7S.

French SA, Jeffrey RW, Story M, Hanna P, Snyder MP. A Pricing Strategy to Promote
Low-Fat Snack Choices through Vending Machines. Am J Public Health
1997;87:849-851.

Fricker J, Giroux S, Fumeron F, Appelbaum M. Circadian rhythm of energy intake and
corpulence status in adults. Int J Obesity 1990;14:387-393.

Fulgoni VL, Mackey MA. Total dietary fiber in children=s diets. Ann NY A cad Sci.
1991:623: 369-379.

Grantham-McGregor SM. Assessments of the effects of nutrition on mental development
and behavior in Jamaican studies. Am J Clin Nutr 1993;57: 303S-9S.

Grantham-McGregor SM, Powell CA, Walker SP, Himes JH. Nutritional
supplementation, psychosocial stimulation, and mental development of stunted
children: the Jamaican Study. The Lancet 1991;338:1-5.

Gropper SS, Acosta PB. Brief Communication: the therapeutic effect of fiber in treating
obesity. J Am Coll Nutr 1987;6z533-535.

Hanes S, Vermeesch J, Gale S. The national evaluation of school nutrition programs:
program impact on dietary intake. Am J Clin Nutr 1984;40:390-413.

Hinkle MM. A mixed message: The school vending machine. J School Health.
l982;62(1):20-24.

Ho CS, Gould RA, Jensen LN, Kiser SJ, Mozar A, Jensen JB. Evaluation of the nutrient
content of school, sack and vending lunches of junior high school students. School

Food Ser Res Rev. 1991;15(2):85-89.

Hoerr SM, Louden VA. Can Nutrition Information Increase Sales of Healthful Vended
Snacks? J Sch Health. 1993;63(9):386-390.

Hruban J A. Selection of snack foods fiom vending machines by high school students. J
School Health. l987;57(l):33-37.

88

Hurd SL, Friedman BJ, Cise N. Evaluation of implementation of the US Dietary
Guidelines into the child nutrition programs in Texas. J Am Diet Assoc
1996;96:904-906.

JMP Statistics and Graphics Guide. Release 3. Cary, NC: SAS Institute; 1998.

Johnson B. Welfare reform law specifies changes and budget cuts for Food Stamp and
child nutrition programs. J Am Diet Assoc 1996;96:999-1000.

Kleinman RE, F inb'erg LF, Klish WJ, Lauer RN. Dietary Guidelines for Children: U.S.
Recommendations. J Nutr. 1996;126:1028S-103OS.

Kramer CY. Extension of Multiple Range Test to Group Means with Unequal Number of
Replications. Biometrics. 1956;57:649-55.

Levine AS, Tallman JR, Grace MK, Parker SA, Billington CJ. Effect of breakfast cereals
on short-term food intake. Am J Clin Nutr 1989;50:1303-1307.

Levine E, Guthrie J F. Nutrient Intakes and Eating Patterns of Teenagers. F am Econ and
Nutr Rev 1997;10:20-33.

Lopez 1, de Andraca I, Perales CG, Heresi E, Castillo M, Colombo M. Breakfast
omission and cognitive performance of normal, wasted and stunted
schoolchildren. Eur J Clin Nutr 1993;47: 533-542.

McClung HJ, Boyne L, Heitlinger L. Constipation and Dietary Fiber Intake in Children.
Pediatrics 1995;96:999-1001 .

Meyers AF, Sampson AE, Weitzrnan M, Rogers BL, Kayne H. School breakfast program
and school performance. Am J Dis Child 1989;143:1234-1239.

Michaud C, Musse N, Kahn JP et al. Food behavior of adolescents living in Nancy
(France). Comparison with the Recommended Dietary Allowances for French

population. Epid Pub Health 1989;37:149-159.

Michaud C, Musse N, Nicolas JP, Mejean L. Effects of breakfast size on short-term
memory, concentration, mood and blood glucose. J Adol Health 1991;12:53-57.

Morgan KJ, Zabik ME, Leveille GA. The role of breakfast in nutrient intake of 5 to 12-
year-old children. Am J Clin Nutr 1981;34:1418-1427.

Murphy AS, Youatt JP, Hoerr SL, Sawyer CA, Andrews SL. Nutrition education needs
and learning preferences of Michigan students in grades 5,8, and 11. J Sch Health

89

1994:64 (7)2273-278.

National Academy of Sciences. Recommended Dietary Allowances. 10‘h ed. Washington,
DC: National Academy Press; 1989.

National Cancer Institute. Diet, Nutrition and Cancer Prevention. Bethesda, MD: US
Department of Health and Human Services; 1987. Publication (NIH) 87-2878.

National Cancer Institute. Eat More Fruits and Vegetables - 5 A Day for Better Health.
Bethesda, Md: US Dept of Health and Human Services, Public Health Service,
National Institutes of Health publication 94-3822; 1994.

National Cholesterol Education Program. Report of the Expert Panel on Blood
Cholesterol Levels in Children and Adolescents. Bethesda, MD: National
Institutes of Health, National Heart, Lung, and Blood Institute; 1991.

National Research Council. Recommended Dietary Allowances. 10‘h ed. Washington,
DC: National Academy Press; 1989.

Nicklas TA, Bao W, Webber LS, Berenson GS. Breakfast consumption affects adequacy
of total daily intake in children. J Am Diet Assoc 1993;93:886-891.

Nicklas TA, Bao W, Berenson GS. Nutrient contribution of the breakfast meal classified
by source in 10-year-old children: Home versus school. Res Rev 1993;17:125-132.

Nicklas TA, Myers L, Berenson GS. Impact of ready-to-eat cereal consumption on total
dietary intake of children: the Bogalusa Heart Study. J Am Diet Assoc
1994;94:316-318.

Nicklas TA, Farris RP, Myers L, Berenson GS. Dietary Fiber Intake of Children and
Young Adults: The Bogalusa Heart Study. J Am Diet Assoc 1995; 95: 209-214.

Nicklas TA, Myers L, Berenson GS. Dietary fiber intake of children: the Bogalusa Heart
Study. Pediatrics 1995;96:988-994.

Nicklas TA, Myers L, Farris RP, Srinivasan SR, Berenson GS. Nutritional Quality of a
High Carbohydrate Diet as Consumed by Children: The Bogalusa Heart Study. J
Nutr 1996; 126: 1382-1388.

Nishimune T, Sumimoto T, Konishi Y, et al. Dietary fiber intake of Japanese younger

generations and the recommended daily allowance. J Nutr Sci Vitaminol (Tokyo).
1993;39:263-278.

90

Osganian SK, Nicklas T, Stone E, Nichaman M, Ebzery MK, Lytle L, Nader PR.
Perspectives on the School Nutrition Dietary Assessment Study from the Child
and Adolescent Trial for Cardiovascular Health. Am J Clin Nutr
1995;61(suppl):241 S-48.

Osganian SK, Ebzery MK, Montgomery DH, Nicklas TA, Evans MA, Mitchell PD, Lytle
LA, Snyder MP, Stone EJ, Zive MM, Bachman KJ, Rice R, Parcel GS. Changes
in the nutrient content of school lunches: results from the CATCH Eat Smart
Food Service Intervention. Prev Med 1996:25: 400-12.

Pollitt E, Leibel RL, Grenfield D. Brief fasting, stress and cognition in children. Am J
Clin Nutr 1981 ;34: 1526-33.

Public Health Service. Healthy People 2000. National Health Promotion and Disease
Prevention Objectives. Washington, DC: US Department of Health and Human
Services, 1991. DHHS Publication (PHS) 91-50212. Full report, with
commentary.

Resnicow K. The relationship between breakfast habits and plasma cholesterol levels in
schoolchildren. J School Health 1991;61:81-85.

Rigaud D, Ryttig KR, Angel LA, Apfelbaum M. Overweight treated with energy
restriction and a dietary fibre supplement: a 6-month randomized, double-blind,
placebo-controlled trial. Int J Obes 1990;14:763-769.

Ruxton CHS, Kirk TR, Belton NR, Holmes MAM. Breakfast habits in children. Nutr
Food Sci 1993;4:17-20.

Ryttig KR, Tellnes G, Haegh L, Boe E, Fagerthun H. A dietary fibre supplement and
weight maintenance after weight reduction: a randomized, double-blind, placebo-

controlled long-term trial. Int J Obes 1989;13:165-171.

Saldanha LG. Fiber in the Diet of US Children: Results of National Surveys.
Pediatrics 1995;96:994-997.

SAS/STAT User=s Guide. Release 6.03ed. Cary, NC: SAS Institute; 1988.

Schlundt DG, Hill JO, Sbrocco T, Pope-Cordle J, Sharp T. The role of breakfast in the
treatment of obesity: a randomized clinical trial. Am J Clin Nutr 1992;55:645-651.

Schmid TL, Jeffrey RW, Forster J L, Rooney B, McBride C. Public support for policy
initiatives regulating high-fat food use in Minnesota: a multicommunity survey.

Prev Med 1989;18:791-805.

91

Slann J L. Dietary fiber: classification, chemical analyses, and food sources. J Am Diet
Assoc. 1987;79:83-91.

Spring B, Maller O, Wurtman J, Digman L, Cozolino L. J Psych Res 1982/1983;l 7:155-
167.

Stedronsky FM. Child Nutrition and Health Campaign: A member update. J Am Diet
Assoc 1998;98:758-759.

Stewart-Truswell A. ABC of nutrition: children and adolescents. Br Med J 1985;291:397-
9.

Tukey JW. Conclusions vs. Decisions. Technometrics 1960;2z423-433.

US Department of Health and Human Services, Public Health Service. Healthy People
2000: national health promotion and disease prevention objectives. Washington,
DC: U.S. Government Printing Office, 1990 [DHHS publication no. (PHS) 91-
50212}

United States Department of Agriculture Nutrient Database for Individual Food Intake
Surveys. Release 4. Washington, DC: US Dept. of Agriculture, 1990.

United States Department of Agriculture. The Food Guide Pyramid. Hyattsville, MD:
Human Nutrition Information Service; 1992. Publication HG252.

United States Department of Agriculture. National School Lunch Program and School
Breakfast Program: Nutrition Objectives for School Meals. Federal Register
1994;59(1 1 1);30218-30240.

United States Department of Agriculture. National School Lunch Program and School
Breakfast Program: Compliance With the Dietary Guidelines for Americans and
Food-Based Menu Systems. Federal Register 1995;60(18):5514-5528.

United States Department of Agriculture. Child Nutrition Programs: School Meal
Initiatives for Healthy Children; Final Rule. Federal Register
1995;60(113):31188-31222.

United States Department of Agriculture. Guidelines for School Health Programs to
Promote Lifelong Healthy Eating. J Sch Health 1997; 67 (1): 9-26.

US Department of Health and Human Services, US Department of Agriculture. Dietary
guidelines for Americans. 3"d ed. Washington, DC: US Government Printing

92

Office, 1989.

Vending machines, apples and profit. Nutr Notes 76. United Fresh Fruit and Vegetable
Association: 1978; summerz3.

Walker ARP, Walker BF, Jones J, Ncongwane J. Breakfast habits of adolescents in four
South African populations. Am J Clin Nutr 1982;36:650-656.

Wilbur CS, Zifferblatt SM, Pinsky JL, Zifferblatt S. Healthy Vending: A Cooperative
Pilot Research Program to Stimulate Good Health in the Marketplace. Prev Med
1981;10:85-93.

Williams CL, Spark A, Haley N, Axelrad C, Strobino B. Effectiveness of a psyllium-
enriched step I diet in hypercholesterolemic children. Circulation 1991;84:11-6.

Williams CL. Intervention on childhood. In: Ockene IS, Ockene JK, eds. Preventive
Cardiology-A Behavioral Approach. Boston: Little Brown and Co; 1992:433-
467.

Williams CL, Bollella M, Williams GM. Cancer prevention beginning in childhood. In:
Devita VT, ed. Cancer Prevention. Philadelphia: J B Lippincott Co; 1993:1-13.

Williams CL, Bollella M, Wynder EL. A New Recommendation for Dietary Fiber in
Childhood. Pediatrics 1995; 96:985-988.

Wirfat AKE, Jeffrey R. Using Cluster Analysis to Examine Dietary Patterns: Nutrient
Intakes, Gender, and Weight Status Differ Across Food Pattern Clusters. J Am
Diet Assoc 1997: 97: 272-279.

Wynder EL, Berenson GS, Strong WB, et al. Coronary artery disease prevention:
cholesterol- a pediatric perspective. Prev Med. 1989;18:323-409.

Yagalla MV, Hoerr SM, Song WO, Bianchi L. Clusters of Foods Selected by Children in
USDA School Meal Programs Which Contributed to Achievement of Dietary
Recommendations for Total Fat, Saturated Fat and Dietary Fiber. Am J Clin Nutr.

Submitted for publication.

93

IGAN STATE USP! 1

11111111

 

llllllllllllllglll

\

l

1111

l

 
  

     
 

3 1293 02177