Family Economics and Nutrition Review [Volume 10, Number 4]

              Tre
NFC_
U.S. DEPOSITORY
PROPERTY OF THE LIBRARY
The University ot lliUJLO Carolina
at Greensboro 4-95
Cecilia Wilkinson Enns, Joseph D. Goldman, and Annetta Cook
Trends in Marketing and Usage of Fat-Modified Foods:
Implications for Dietary Status and Nutrition Promotion
Nancy E. Schwenk and Joanne F. Guthrie
Supplement Use May Not Be Associated With Better
Food Intake in All Population Groups
David L. Pelletier and Anne Kendall
Research Summaries
51
54
56 Making Payments on the Internet
58 Farmers' Markets: A Survey
Regular Items
60
62 Research and Evaluation Activities in USDA
65 Cost of Food at Home
66 Consumer Prices
67 Index of Articles in 1997 Issues
68 Index of Authors in 1997 Issues
Reviewers for 1997
Dan Glickman, Secretary
U.S. Department of Agriculture
Shirley R. Watkins, Under Secretary
Food , Nutrition, and Consumer Services
Rajen Anand, Executive Director
Ce nter for Nutri tio n Policy and Promotion
Editorial Board
Mohamed Abdel-Ghany
University of Alabama
Rhona Applebaum
National Food Processors Association
Johanna Dwyer
New England Medical Center
Jean Mayer USDA Human Nutrition Research Center
on Aging at Tufts University
Jay Hirschman
Food and Nutrition Service
U.S. Department of Agriculture
Helen Jensen
Iowa State University
Janet C. King
Western Human Nutrition Research Center
U.S. Department of Agriculture
C. J, Lee
Kentucky State University
Rebecca Mullis
Georgia State University
Suzanne Murphy
University of California-Berkeley
Donald Rose
Economic Research Service
U.S. Department of Agricul ture
Ben Senauer
University of Minnesota
Laura Sims
University of Maryland
Retia Walker
University of Kentucky
Editor
Joan C. Courtless
Managing Editor
Jane W. Fleming
Family Economics and Nutrition Review is
written and published each quarter by the
Center for Nutrition Policy and Promotion,
U.S. Department of Agriculture, Washington, DC.
The Secretary of Agriculture has determined
that publication of this periodical is neces­sary
in the transadion of the pubfic busness
required by law of the Department.
This publication is not copyrighted. Contents
may be reprinted without permission, but
credit to Family Economics and Nutrition
Review would be appreciated. Use of
commercial or trade names does not imply
appr011al or constitute endorsement by USDA.
Family Economics and Nutrition Review is
indexed in the following databases: AGRICOlA,
Ageline, Economic Literature Index, ERIC,
Family Studies, PAIS, and Sociological
Abstracts.
Family Economics and Nutrition Review is
for sale by the Superintendent of Documents.
Subscription price is $10.00 per year ($12.50
for foreign addresses). Send subscription
orders and change of address to Superinten­dent
of Documents, P.O. Box 371954,
Pittsburgh, PA 15250-7954. (See subscription
form on p. 71 .)
Original manuscripts are accepted for
publication. (See "guidelines for authors" on
back inside cover). Suggestions or comments
concerring this publication should be
addressed to: Jane W. Fleming, Managing
Editor, Family Economics and Nutrition Review,
Center for Nutrition Pdicy en! Prorrdion, USDA,
112020th St., NW, Suite 200 North Lobby,
Washin~on. DC 20006. Phone (202) 6064816.
USDA prohibits discrimination in its programs
on the basis of race, color, national origin,
sex, religion, age, disability, political beliefs,
and marital or familial status. (Not aH prohibited
bases apply to all programs.) Persons with
disabilities who require alternative means for
communication of program information (Braille,
large print, audiotape, etc.) should contact
USDA's TARGET Center at (202) 720-2600
(voice and TDD). To file a complaint, write
the Secretary of Agricu~ure,U.S. Department
of Agricuture, Washington, DC 20250, or call
1-800-245-6340 (voice) or (202) 720.1127
(TOO). USDA is an equal employment
opportunity employer.
Center for Nutrition Policy and Promotion
Feature Articles
2
: . 16
32
Trends in Food and Nutrient Intakes by Adults:
NFCS 19n-78, CSFII1989-91, and CSFII1994-95
Cecilia Wilkinson Enns, Joseph D. Goldman, and Annetta Cook
Trends in Marketing and Usage of Fat-Modified Foods:
Implications for Dietary Status and Nutrition Promotion
Nancy E. Schwenk and Joanne F. Guthrie
Supplement Use May Not Be Associated With Better
Food Intake in All Population Groups
David L Pelletier and Anne Kendall
Research Brief
45 Snacking Habits of Different Income Groups
Shanthy A. Bowman
Research Summaries
51
54
56
Expenditures on Necessities by the Elderly and Nonelderly
During the 1980's
Price and Income Affect Nutrients Consumed From Meats
Making Payments on the Internet
58 Farmers' Markets: A Survey
Regular Items
60 Charts From Federal Data Sources
62 Research and Evaluation Activities in USDA
65 Cost of Food at Home
66 Consumer Prices
67 Index of Articles in 1997 Issues
68 Index of Authors in 1997 Issues
70 Reviewers for 1997
VoluiM 10, Number 4
1997
2
Feature Articles
Trends in Food and Nutrient
Intakes by Adults: NFCS
1977-78, CSFII1989-91,
and CSFII 1994-95
Cecilia Wilkinson Enns
Joseph D. Goldman
Annetta Cook
Beltsville Human Nutrition Research Center
Agricultural Research Service
Evaluations of diet quality and tracking changes in the diet over time have
many useful applications, including policy formation, program planning, and
targeting appropriate nutrition education messages. CSFII 1994-95 data
on food and nutrient intakes by adults were used to examine diet quality in
1994-95 and changes since 1977-78. The largest changes were decreased
consumption of whole milk and increased consumption of grain products,
especially grain mixtures; bananas; meat, poultry, and fish mixtures; beer
and ale; fruit drinks and ades; and soft drinks. In general, the nutrients that
were below the RDA in 1994-95 are the same nutrients that were below the
RDA in 1977-78. In 1994-95, intakes of magnesium and zinc were below
the RDA for both women and men. Women's intakes were also below the
RDA in vitamin 86, vitamin E, and calcium. Future increases in whole grains,
fruits, dark green vegetables, legumes, nonfat or lowfat dairy products, and
lean meats and decreases in fats and sugars are desirable.
[!] n accordance with the
National Nutrition Monitor­ing
and Related Research
(NNMRR) Act of 1990
(P.L. 101-445), the NNMRR Program
(NNMRRP) monitors the nutritional
status of the U.S. population. As a
cornerstone of the NNMRRP, USDA
surveys such as the Continuing Survey
of Food Intakes by Individuals (CSFII)
provide up-to-date information on food
intakes by Americans for use in policy
formation, regulation, program planning
and evaluation, education, and research.
For example, CSFII data have been used
to evaluate the impact of food fortifica­tion
on nutrient intakes, to estimate
exposure to pesticide residues and other
contaminants from foods, and to target
nutrition programs to those who need
them most. CSFII data are essential for
monitoring changes over time in the
food choices Americans make and the
adequacy of their diet.
Family Economics and Nutrition Review
This article contains information from
the first 2 years of the 1994-96 Continuing
Survey of Food Intakes by Individuals
(CSFII), conducted by the Agricultural
Research Service (ARS) of the U.S.
Department of Agriculture (USDA).
Some comparisons are made between
intakes in the CSFII 1994-95 and intakes
in the Nationwide Food Consumption
Survey (NFCS) 1977-78 and the CSFII
1989-91. Methods and comparability
of the data from the three surveys are
discussed briefly in the "Design and
Methods" section below.
Mean intakes and percentages of
individuals using foods from 12 food
groups and 29 selected subgroups are
presented for women (tables I and 2,
pp. 4 and 5) and men (tables 3 and 4,
pp. 6 and 7) age 20 years and over who
participated in the NFCS 1977-78,
CSFII 1989-91, and CSFII 1994-95.
Some trends in food intakes that can be
traced from the NFCS 1977-78 through
the CSFII 1989-91 to the CSFII 1994-
95 are highlighted in table 5, p. 8 (see
section on "Design and Methods" regard­ing
the definition of a trend). The effect
of analyzing food group intakes with
grrun mixtures and meat mixtures bro­ken
down is shown in tables 6 and 7,
pp. 8 and 9 and figure 1, p. 10.
Intakes of energy and nutrients as per­centages
of the Recommended Dietary
Allowances (RDA) are presented in
table 8, p. 11. Figure 2, p. 12, shows
trends in intakes of fat and carbohydrate
from 1977-78 to 1994-95. Saturated fat,
cholesterol, and dietary fiber intakes are
discussed.
1997 Vol.JONo. 4
Design and Methods
The recently completed CSFII 1994-96,
popularly referred to as "What We Eat
in America," was conducted by Westat,
Inc. (Rockville, MD) under contract to
USDA's Food Surveys Research Group
(FSRG), Agricultural Research Service
(ARS). In each of the 3 survey years, a
nationally representative sample of non­institutionalized
individuals residing in
the United States provided, through in­person
interviews using a 1-day dietary
recall, food intakes on 2 nonconsecutive
days and health-related information.
Some estimates from the NFCS 1977-78
and the CSFII 1989-91 are compared
with estimates from the first 2 years of
the CSFII 1994-96. Some differences
exist between the three surveys in sam­pling
and methodology. Detailed infor­mation
on the methodology of each
survey is available elsewhere ( 13-17).
In 1994-96, the target population was
noninstitutionalized individuals in all
50 States; in 1977-78 and 1989-91, the
target population covered only the 48
conterminous States. The 1994-96 survey
included an oversampling of the low­income
population; earlier surveys
included separate supplemental samples
for the low-income population. In the
summer, fall, and winter quarters of
the NFCS 1977-78, in households with
more than one individual19 years or
over, one-half of the individuals in that
age group were asked to participate. In
the CSFII 1988-91 , all individuals in
sample households were eligible. CSFII
1994-96 data were collected from selected
individuals within each household. The
1994-96 survey provides a proportion­ately
larger sample in selected sex-age
categories, specifically young children
and elderly people.
In the NFCS 1977-78 and the CSFII
1989-91, dietary information was col­lected
on 3 consecutive days using a
1-day dietary recall and a 2-day dietary
record. In the CSFII 1994-96, dietary
data were collected by means of 1-day
dietary recalls on 2 nonconsecutive days
(3 to 10 days apart). The 1-day recall
was modified for the CSFII 1994-96 to
include multiple passes through the list
of all foods and beverages recalled by
the respondent in order to maximize the
amount of information collected. The
multiple-pass approach and its develop­ment
are described in detail elsewhere
(3, 5, 13). The procedure has respondents
provide a list of all foods eaten the
previous day, using any recall strategy
they desire. Then interviewers get a
more detailed list by probing for addi­tions
to food and giving respondents an
opportunity to recall food items initially
forgotten. In a third pass, the interviewers
review with the respondents the list of
food they reported to stimulate more
reports of food and eating occasions.
In order to track trends over time from
surveys with different numbers of days
of dietary information, tables and figures
comparing food and nutrient intake
estimates among the surveys are based
on only the first day's data collected
from each individual. All estimates
are weighted to be nationally
representative.
In addition to the changes outlined
above, the USDA nutrient data base
was updated for each survey to reflect
changes in foods on the market and also
to incorporate improved nutrient values.
3
Table 1. Mean intakes per individual from selected1 food groups and standard errors of the mean, women 20
years and over, 1 day, NFCS 1977-78, CSFII 1989-91, and CSFII 1994-95
1977-78 1989-91 1994-95
Food group2 Mean (SEM) Mean (SEM) Mean (SEM)
Grams
Total grain products 177 ( 2.4) 234 ( 5.8) 255 ( 5.3)
Yeast breads and rolls 48 ( .5) 45 ( .9) 45 ( 1.0)
Cereals and pasta 42 ( 3.5) 63 ( 3.2) 64 ( 2.6)
Ready-to-eat cereals 7 ( .2) 11 ( .5) 13 ( .5)
Mixtures mainly grain 42 ( 1.8) 71 ( 3.0) 87 ( 4.3)
Total vegetables 205 ( 2.4) 187 ( 3.6) 189 ( 3.8)
Dark-green vegetables 10 ( .7) 12 ( .8) 14 ( 1.2)
Deep-yellow vegetables 10 ( .3) 9 ( .7) 10 ( .9)
Tomatoes 27 ( 1.0) 25 ( l.l) 28 ( 1.3)
Total fruits 142 ( 4.4) 150 ( 4.7) 156 ( 5.0)
Citrus juices 54 ( 2.6) 53 ( 2.7) 52 ( 2.9)
Bananas 8 ( .4) 13 ( .7) 17 ( .8)
Noncitrus juices and nectars 12 ( .8) 20 ( 2.2) 16 ( 1.7)
Total milk and milk products 203 ( 4.2) 206 ( 5.6) 202 ( 6.0)
Total fluid milk3 151 ( 3.5) 156 ( 5.1) 135 ( 4.2)
Whole milk 67 ( 4.3) 46 ( 2.9) 35 ( 2.9)
Lowfatmilk 34 ( 1.2) 74 ( 4.2) 56 ( 3.5)
Skim milk 13 ( .7) 34 ( 2.4) 41 ( 2.7)
Milk desserts 16 ( .6) 18 ( .9) 25 ( 1.8)
Cheese 16 ( .7) 12 ( .7) 15 ( 1.0)
Total meat, poultry, fish 184 ( 1.9) 167 ( 3.3) 168 ( 3.1)
Beef 47 ( 1.5) 22 ( l.l) 19 ( .9)
Pork 18 ( .5) 10 ( .5) 10 ( .6)
Mixtures mainly meat, poultry, fish 60 ( 2.0) 78 ( 3.2) 85 ( 3.4)
Eggs 24 ( .9) 16 ( .9) 16 ( .8)
Legumes 18 ( 1.4) 17 ( 1.3) 19 ( 1.2)
Nuts and seeds 2 ( .2) 3 ( .3) 3 ( .3)
Total fats and oils 13 ( .6) 16 ' ( .5) 16 ( .6)
Table fats 6 ( .2) 5 ( .2) 3 ( .2)
Salad dressings 6 ( .3) 9 ( .5) 10 ( .5)
Total sugars and sweets 17 ( .6) 17 ( .8) 19 ( 1.0)
Sugars 5 ( .2) 4 ( .2) 4 ( .2)
Candy 2 ( .2) 4 ( .3) 5 ( .5)
Total alcoholic beverages 32 ( 2.1) 40 ( 4.7) 59 ( 5.6)
Beer and ale 19 ( 1.7) 27 ( 4.2) 38 ( 4.6)
Total nonalcoholic beverages 698 (18.1) 753 (14.5) 854 (20.4)
Coffee 376 (14.5) 327 (13.0) 332 (13.9)
Total fruit drinks and ades 29 ( 2.1) 46 ( 2.7) 58 ( 4.4)
Total carbonated soft drinks 137 ( 4.0) 238 ( 7.6) 293 (13.3)
Regular carbonated soft drinks 101 ( 4.5) 140 ( 6.4) 178 (10.8)
Low-calorie carbonated soft drinks 34 ( 3.4) 97 ( 5.5) 115 ( 8.9)
1 Because only selected food groups are included here, intakes from subgroups may not sum to total s.
2For descriptions of the foods included in the food groups named in this article, see "Table Notes" in reference 12.
3
The proportion of total fluid milk intake that could be classified as whole, lowfat, or skim on the basis of information obtained from respondents was much
higher in 1994-95 (nearly 98 percent) than in 1977-78 (75 percent or less).
4 Family Economics and Nutrition Review
Table 2. Mean percentages of individuals using items from selected food groups and standard errors of the
percentage, women 20 years and over, 1 day, NFCS 1977-78, CSFII 1989-91, and CSFII 1994-95
1977-78 1989-91 1994-95
Food group1 Mean (SEP) Mean (SEP) Mean (SEP)
Percent
Total grain products 94.8 (0.24) 96.3 (0.34) 96.8 (0.30)
Yeast breads and rolls 76.4 ( .46) 69.6 (1.00) 66.9 ( .81)
Cereals and pasta 35.6 ( .52) 41.3 (1.11) 43.8 ( 1.0 I)
Ready-to-eat cereals 18.4 ( .41) 23.3 ( .70) 24.0 ( .86)
Mixtures mainly grain 16.8 ( .41) 26.6 ( .88) 31.2 (1.08)
Total vegetables 86.3 ( .37) 82.7 ( .73) 84.0 ( .79)
Dark-green vegetables 7.8 ( .29) 11.4 ( .68) 12.4 ( .74)
Deep-yellow vegetables 10.2 ( .33) 11.1 ( .64) 15.3 ( .84)
Tomatoes 25.5 ( .47) 30.5 ( .78) 39.0 (1.11)
Total fruits 54.2 ( .54) 54.5 (1.18) 54.8 (1.12)
Citrus juices 26.7 ( .48) 22.2 ( .98) 20.8 ( .87)
Bananas 7.1 ( .27) 12.1 ( .67) 14.9 ( .64)
Noncitrus juices and nectars 4.9 ( .24) 6.7 ( .56) 5.1 ( .45)
Total milk and milk products 77.1 ( .45) 76.3 ( .88) 77.1 ( .87)
Total fluid milk2 59.1 ( .53) 57.1 (1.11) 51.8 (1.16)
Whole milk 25.2 ( .47) 19.6 ( .85) 14.7 (1.23)
Lowfatmilk 12.2 ( .35) 26.3 (1.04) 22.8 ( .90)
Skim milk 5.2 ( .24) 11.8 ( .65) 14.4 ( .81)
Milk desserts 13.1 ( .36) 13.1 ( .60) 17.0 ( .83)
Cheese 26.4 ( .48) 27.1 (1.03) 32.1 (1.08)
Total meat, poultry, fish 92.1 ( .29) 88.0 ( .67.) 85.3 ( .70)
Beef 33.4 ( .51) 20.5 ( .87) 19.4 ( .74)
Pork 25.1 ( .46) 15.8 ( .65) 16.6 ( .79)
Mixtures mainly meat, poultry, fish 27.7 ( .48) 32.7 (1.18) 34.0 (1.23)
Eggs 30.9 ( .50) 19.2 ( .76) 18.5 ( .74)
Legumes 10.6 ( .33) 11.3 ( .58) 14.1 ( .62)
Nuts and seeds 7.2 ( .28) 8.5 ( .51) 8.2 ( .49)
Total fats and oils 63.2 ( .52) 64.4 ( .79) 60.9 (1.21)
Table fats 45.4 ( .54) 40.2 (1.02) 33.6 (1.09)
Salad dressings 27.9 ( .49) 32.8 ( .89) 32.1 (1.21)
Total sugars and sweets 53.0 ( .54) 51.4 (1.12) 56.8 ( .95)
Sugars 38.5 ( .52) 34.8 (1.03) 36.1 ( .83)
Candy 4.6 ( .23) 9.3 ( .65) 12.2 ( .74)
Total alcoholic beverages 9.7 ( .33) 10.1 ( .88) 11.8 ( .91)
Beer and ale 3.2 ( .19) 4.0 ( .48) 4.3 ( .38)
Total nonalcoholic beverages 91.8 ( .30) 89.1 ( .85) 90.7 ( .54)
Coffee 66.9 ( .51) 55.5 (1.31) 54.2 (1.17)
Total fruit drinks and ades 8.8 ( .30) 12.2 ( .65) 13.1 ( .58)
Total carbonated soft drinks 32.2 ( .50) 44.3 (1.07) 48.7 (1.32)
Regular carbonated soft drinks 24.5 ( .48) 28.2 ( .97) 31.7 (1.24)
Low-calorie carbonated soft drinks 8.1 ( .30) 18.2 ( .78) 18.5 ( .93)
1 For descriptions of the foods included in the food groups named in this article, see "Table Notes" in reference 12.
2The proportion of total fluid milk intake that could be classified as whole, lowfat, or skim on the basis of information obtained from respondents was much
higher in 1994-95 (nearly 98 percent) than in 1977-78 (75 percent or less).
1997 Vol. 10 No.4 5
Table 3. Mean intakes per individual from selected1 food groups and standard errors of the mean, men
20 years and over, 1 day, NFCS 1977-78, CSFII 1989-91, and CSFII 1994-95
1977-78 1989-91 1994-95
Food group 2 Mean (SEM) Mean (SEM) Mean (SEM)
Grams
Total grain products 252 ( 4.3) 324 ( 7.5) 361 ( 9.7)
Yeast breads and rolls 75 ( 1.0) 61 ( 1.3) 63 ( 2.0)
Cereals and pasta 52 ( 3.6) 87 ( 4.6) 89 ( 4.5)
Ready-to-eat cereals 9 ( .3) 14 ( .8) 16 ( .8)
Mixtures mainly grain 56 ( 3.1) 104 ( 5.1) 128 ( 6.7)
Total vegetables 251 ( 5.5) 222 ( 4.5) 242 ( 4.4)
Dark-green vegetables 10 ( 1.0) 13 ( 1.3) 14 ( 1.4)
Deep-yellow vegetables 10 ( .4) 10 ( .8) 8 ( .6)
Tomatoes 30 ( .9) 31 ( 1.4) 37 ( 2.0)
Total fruits 142 ( 4.6) 150 ( 6.4) 172 ( 5.8)
Citrus juices 51 ( 2.8) 59 ( 4.2) 65 ( 4.6)
Bananas 10 ( .7) 11 ( .7) 19 ( 1.0)
Noncitrus juices and nectars 11 ( 1.0) 16 ( 1.9) 19 ( 2.4)
Total milk and milk products 276 ( 6.5) 254 ( 9.6) 256 ( 9.7)
Total fluid miJk3 214 ( 6.4) 193 ( 8.1) 178 ( 6.6)
Whole milk 102 ( 5.8) 66 ( 5.2) 54 ( 3.8)
Lowfatmilk 45 ( 2.1) 92 ( 5.5) 85 ( 5.6)
Skim milk 10 ( 1.0) 32 ( 3.8) 35 ( 3.6)
Milk desserts 23 ( 1.2) 24 ( 2.0) 33 ( 2.4)
Cheese 17 ( .6) 16 ( .9) 18 ( 0.9)
Total meat, poultry, fish 280 ( 2.7) 260 ( 5.4) 275 ( 5.3)
Beef 72 ( 2.4) 36 ( 1.9) 38 ( 2.4)
Pork 28 ( .8) 15 ( 1.2) 15 ( 1.0)
Mixtures mainly meat, poultry, fish 92 ( 3.7) 124 ( 5.4) 137 ( 5.0)
Eggs 38 ( 1.0) 26 ( 1.5) 23 ( 1.1)
Legumes 26 ( 1.5) 30 ( 2.7) 31 ( 3.0)
Nuts and seeds 4 ( .3) 4 ( .4) 4 ( .4)
Total fats and oils 17 .8) 18 .8) 18 .9)
Table fats 8 .4) 7 .4) 5 .4)
Salad dressings 7 ( .4) 10 ( .5) 11 ( .6)
Total sugars and sweets 24 ( 1.0) 19 ( 1.1) 24 ( 1.4)
Sugars 6 ( .2) 6 ( .4) 4 ( .3)
Candy 2 ( .3) 3 ( .4) 6 ( .5)
Total alcoholic beverages 138 ( 8.4) 162 (10.3) 238 (23.6)
Beer and ale 121 ( 8.0) 145 (10.2) 216 (22.9)
Total nonalcoholic beverages 752 (12.9) 900 (21.8) 1,066 (21.7)
Coffee 421 (13.0) 408 ( 15.6) 408 (13.4)
Total fruit drinks and ades 35 ( 2.3) 60 ( 5.1) 86 ( 6.5)
Total carbonated soft drinks 154 ( 5.6) 292 (12.3) 404 (16.1)
Regular carbonated soft drinks 136 ( 6.5) 212 (10.4) 312 (15.8)
Low-calorie carbonated soft drinks 15 ( 1.5) 79 ( 6.2) 92 ( 6.8)
1 Because only selected food groups are included here, intakes from subgroups may not sum to totals.
2For descriptions of the foods included in the food groups named in this article, see "Table Notes" in reference 12.
3
The proportion of total fluid milk intake that could be classified as whole, lowfat, or skim on the basis of information obtained from respondents was much
higher in 1994-95 (nearly 98 percent) than in 1977-78 (75 percent or less).
6 Family Economics and Nutrition Review
Table 4. Mean percentages of individuals using items from selected food groups and standard errors of the
percentage, men 20 years and over, 1 day, NFCS 1977-78, CSFII 1989-91, and CSFII 1994-95
1977-78 1989-91 1994-95
Food group1 Mean (SEP) Mean (SEP) Mean (SEP)
Percent
Total grain products 96.9 (0.22) 95.8 (0.47) 96.8 (0.37)
Yeast breads and rolls 82.5 ( .49) 71.3 ( .86) 69.0 (1.17)
Cereals and pasta 36.5 ( .62) 40.5 (1.27) 40.8 (1.23)
Ready-to-eat cereals 19.7 ( .51) 21.7 ( .98) 22.0 ( .87)
Mixtures mainly grain 16.9 ( .48) 28.7 (1.14) 32.9 (1.19)
Total vegetables 87.9 ( .42) 82.9 ( .79) 86.1 ( .71)
Dark-green vegetables 7.2 ( .34) 10.6 ( .82) 10.7 ( .74)
Deep-yellow vegetables 8.7 ( .37) 10.1 ( .69) 12.2 ( .75)
Tomatoes 25.8 ( .57) 32.6 (1.20) 41.7 (1.09)
Total fruits 51.0 ( .64) 47.8 (1.37) 49.9 ( .99)
Citrus juices 23.9 ( .56) 21.1 (1 .03) 19.8 ( 1.15)
Bananas 8.4 ( .37) 10.4 ( .71) 14.8 ( .72)
Noncitrus juices and nectars 4.4 ( .27) 4.9 ( .48) 5.0 ( .46)
Total milk and milk products 78.0 ( .53) 74.7 (1.17) 74.7 (1.31)
Total fluid milk2 61.6 ( .63) 54.8 (1 .32) 49.4 (1.30)
Whole milk 28.4 ( .58) 21.3 (1.20) 15.0 ( .87)
Lowfatmilk 12.1 ( .42) 25.4 (1.22) 24.5 (1 .26)
Skim milk 3.0 ( .22) 8.6 ( .79) 9.7 ( .68)
Milk desserts 15.5 ( .47) 13.7 ( .87) 18.3 ( .88)
Cheese 26.0 ( .57) 27.1 ( 1.10) 32.1 (1.10)
Total meat, poultry, fish 95.8 ( .26) 91.7 ( .51) 90.6 ( .69)
Beef 38.8 ( .63) 25.1 (1.19) 24.9 ( .99)
Pork 31.0 ( .59) 19.6 ( .88) 17.9 ( .83)
Mixtures mainly meat, poultry, fish 31.4 ( .60) 38.5 (1.08) 40.1 ( 1.07)
Eggs 39.1 ( .63) 24.1 ( .89) 21.2 ( .80)
Legumes 11.7 ( .41) 13.4 ( .82) 14.0 ( .76)
Nuts and seeds 8.3 ( .36) 8.6 ( .61) 8.7 ( .49)
Total fats and oils 63.5 ( .62) 64.0 (1.12) 58.0 (1.18)
Table fats 48.6 ( .64) 40.3 (1.18) 32.4 (1.08)
Salad dressings 27.1 ( .57) 31.9 (1.16) 31.3 (1.08)
Total sugars and sweets 57.4 ( .64) 49.7 (1.20) 52.2 (1.14)
Sugars 42.0 ( .64) 33.9 (1 .22) 33.4 ( .99)
Candy 4.1 ( .26) 7.1 ( .61) 12.5 ( .64)
Total alcoholic beverages 19.0 ( .51) 21.6 (1.15) 23.7 ( .96)
Beer and ale 12.8 ( .43) 15.2 ( .83) 17.9 ( .78)
Total nonalcoholic beverages 90.4 ( .38) 89.6 ( .88) 92.2 ( .52)
Coffee 67.7 ( .60) 59.1 (1.30) 56.2 ( 1.03)
Total fruit drinks and ades 8.1 ( .36) 11.5 ( .78) 14.7 ( .79)
Total carbonated soft drinks 31.3 ( .59) 47.0 (1.33) 54.4 (1.30)
Regular carbonated soft drinks 27.7 ( .59) 35.0 ( 1.16) 42.6 (1.53)
Low-calorie carbonated soft drinks 3.3 ( .23) 13.2 ( .71) 13.4 ( .73)
1 For descriptions of the foods included in the food groups named in this article, see "Table Notes" in reference 12.
2The proportion of total fluid milk intake that could be classified as whole, low fat, or skim on the basis of information obtained from respondents was much
higher in 1994-95 (nearly 98 percent) than in 1977-78 (75 percent or less).
1997 Vol. 10 No.4 7
Sample sizes for adults age 20 years and
over were the following: 10,035 women
and 7,027 men in theNFCS 1977-78;
6,229 women and 4,219 men in the
CSFII 1989-91; and 3,284 women and
3,352 men in the CSFII 1994-95. Overall
analytic Day-1 response rates were 56.9
percent in the NFCS 1977-78, 57.6 per­cent
in the CSFli 1989-91, and 80 per­cent
in the CSFII 1994-95.
Mean intakes are presented "per individ­ual,"
meaning they include intakes by
both consumers and nonconsumers of
the food group. To calculate "per user"
intakes, divide the mean intake by the
percentage of individuals using the food
expressed as a decimal. Since only
selected subgroups are presented,
subgroup intakes will not sum to the
food group total.
For this article, a "trend" was defined
very narrowly. For a given food group,
a trend is claimed to exist only when
mean intakes of the food (or percentages
of individuals using the food) either
rose or fell continuously from 1977-78
through 1989-91 to 1994-95. Further
analysis with more complex methods
may reveal additional trends.
Results and Discussion
Time Trends in Food Intakes
Grain products.-As shown in tables 1
to 5, adults 20 years and over in the
United States ate more grain products,1
especially grain mixtures (such as pizza
and tacos), in 1994-95 than in 1977-78.
1For descriptions of the foods included in the food
groups named in this article, see "Table Notes" in
reference 12.
8
Table 5. Trends in food intakes, 1977-78 to 1994-95: Women and men
20 years and over, 1 day, NFCS 1977-78 and CSFII 1994-95
Grain products
Mixtures
Fruit
Bananas
Fluid milk
Whole milk
Food
Meat, poultry, and fish mixtures
Beverages
Beer and ale
Fruit drinks and ades
Soft drinks
Percent change in amounts consumed
Women Men
+44 +43
+107 +129
+10 +21
+112 +90
-11 -17
-48 -47
+42 +49
+25 +47
+100 +79
+100 +146
+114 +162
Table 6. Food mixtures: Mean intakes and percentages of total weight
by ingredient, adults 20 years and over, 1 day, 1994
Intakes and ingredients
Mean intakes
Ingredient:
Grain products
Vegetables
Milk and milk products
Meat, poultry, and fish
Water
Other
Grain mixtures Meat mixtures
Grams
109 105
Percent
32 14
24 28
8 6
8 34
19 10
9 8
Family Economics and Nutrition Review
Table 7. Vegetable intakes: Intakes of vegetables coded separately,
intakes from grain mixtures and meat mixtures, and adjusted total
intakes, adults 20 years and over, 1 day, CSFD 1989-91 and CSFD 1994
Women Men
Intake source 1989-91 1994 1989-91 1994
Vegetables coded separately 187
From grain mixtures 16
From meat mixtures 22
Adjusted total vegetables 225
Vegetables.-N o clear trend in vegeta­ble
intakes can be seen in this series of
tables because of the widespread con­sumption
of vegetables as part of grain
mixtures and meat mixtures. This issue
is explored further in the section called
"Breaking Down Mixtures."
Fruits.-Overall, the average intake of
fruit rose slightly, but the percentages of
women and men eating fruit or drinking
fruit juice2 in a day remained about the
same. In 1994-95, only 55 percent of
women and 50 percent of men ate fruit
or drank fruit juice in a day. Meanwhile,
intakes of fruit drinks and ades3 doubled
for women and more than doubled for
men between 1977-78 and 1994-95,
partly due to increases in the percentages
of individuals consuming those drinks.
Percentages of individuals eating bananas
also doubled for women and nearly
doubled for men between 1977-78
and 1994-95.
2
Fruitjuices (includes citrus juices plus noncitrus
juices and nectars) are I 00 percent juice.
3
Fruit drinks and ades contain less than I 00 percent
juice.
1997 Vol. 10 No.4
Grams
190 222 242
22 25 30
25 35 33
237 282 305
Milk and milk products.-The per­centages
of adults drinking whole milk
declined progressively for both women
and men, while the percentages drink­ing
skim milk increased.
Meat, poultry, and fish.-In the meat,
poultry, and fish group, the percentages
of individuals eating separate cuts of beef
and pork decreased. Higher percentages
of women and men ate meat mixtures
(such as stew or a hamburger on a bun),
and they ate larger amounts.
Beverages.-Observed increases in
intakes of beer and ale between 1977-78
and 1994-95 among both women and
men may in part reflect a decrease in
underreporting. This possibility is
suggested because the trend seen here
is not paralleled by per capita consump­tion
trends for the same time period
(see 11 ). It has been noted that "sensitive
items" for which "public awareness is
heightened" are subject to bias in self­reporting
( 12). Perhaps the level of
sensitivity associated with reporting
beer intakes has decreased. Another
... since 1989-91,
amounts of soft
drinks consumed
by both women and
men have surpassed
their intakes of milk.
9
Figure 1. Ingredients from breakdown of grain mixtures and meat mixtures added to foods coded
separately, selected food groups, adults 20 years and over, 1 day
Beef
Yeast breads
Cheese
Pasta
Poultry
Tomatoes
1989
1
-91
994
1989 -91
994
1989 -91
1 994
1989 -91
994
1989 -91
1 994
1989 -91
994
0
~
D Not part of a mixture
]
• From a grain mixture
~ From a meat mixture
]
I
10 20 30 40 50 60 70
Grams
Source: U.S. Department of Agriculture, Agricultural Research Service, Beltsville Human Nutrition Research Center, Food Surveys Research
Group. USDA Continuing Survey of Food Intakes by Individuals 1989-91 and 1994.
reason for increased intakes of beer
and ale, as well as other foods, may be
improved quality of the data collected
using the multiple-pass approach (see
"Design and Methods" above).
Both the percentages of adults 20 years
and over drinking carbonated soft drinks
and mean intakes of soft drinks increased
progressively and substantially between
1977-78 and 1994-95. In fact, since
1989-91, amounts of soft drinks con­sumed
by both women and men have
surpassed their intakes of milk. In 1994-
95, a higher percentage of men drank
10
soft drinks than drank milk. In 1994-95,
men's mean intake of beer exceeded
their intake of fluid milk, although it
was lower than their intakes of coffee
and soft drinks.
Breaking Down Mixtures
When CSFII food intake tables such as
tables 1 to 5 are created, a food mixture
that was reported and coded as a single
item is classified under the food group
of its major ingredient. For this reason,
secondary ingredients may be included
under food groups other than the ones
they would appear in if each ingredient
were reported and coded separately.
For example, cheese pizza is a mixture
of dough, tomato sauce, and cheese.
Pizza's major ingredient is dough, so
pizza is classified under grain products
in the subgroup "grain mixtures." Thus,
the secondary ingredients in the pizza
(cheese and tomato sauce) are included
in the grain category rather than under
milk and milk products (cheese) and
vegetables (tomatoes).
Family Economics and Nutrition Review
Table 8. Nutrient intakes as percentages of 1989 Recommended
Dietary Allowances (RDA): Selected mean intakes per day and
standard errors of the mean, adults 20 years and over, 1 day, CSFII
1994-95
Percentage of RDA
Women Men
Nutrient
Protein
Vitamin A1
Vitantin B6
Vitantin B 12
Folate
Vitamin C
VitaminE
Calcium
Iron
Magnesium
Zinc
Food energy
Mean
127
118
94
212
122
147
88
77
100
83
76
78
(SEM)
( 1.2)
( 3.3)
( 1.0)
( 7.7)
( 1.8)
( 3.2)
( 1.9)
( 1.1)
( 1.0)
( .9)
.9)
.7)
Mean (SEM)
Percent
154 ( 2.2)
116 ( 4.7)
110 ( 1.8)
335 (25 .7)
150 ( 2.5)
186 ( 4.2)
100 ( 2.3)
107 ( 2.2)
185 ( 3.6)
94 ( 1.5)
94 ( 1.9)
91 ( 1.3)
1The unit of vitamin A intake used in this calculation was J.l.g RE.
This method of categorizing mixtures
gives a good picture of what types of
dishes people eat, but it can mask what
is happening with intakes of foods that
are widely used as ingredients in mixtures.
Because of the increasing popularity
of mixtures, the ability to look at foods
that are ingredients in mixtures can
make an important contribution to
understanding trends in food intakes.
For this reason, the grain mixtures and
meat mixtures food groups were broken
1997 Vol. 10 No. 4
down into their ingredients to provide
the information shown in tables 6 and 7
and figure 1. Predictably, the predonti­nant
ingredient (by weight) of grain
mixtures is grain products, and the pre­dominant
ingredient of meat mixtures
is meat, poultry, and fish (table 6).
Vegetables are the second largest ingre­dient
of both grain mixtures and meat
mixtures. Illustrated in figure 1 are the
intakes of several foods that are often
ingredients in mixtures; for each food,
the corresponding bar represents the
intake from food coded separately plus
the intakes from grain ntixtures and
meat mixtures. In both 1989-91 and
1994,4 intake of pasta was doubled
when mixtures were broken down,
and intakes of beef and tomatoes were
nearly doubled.
As mentioned in the preceding section,
no clear trend in vegetable intakes was
apparent from tables 1 to 4. However, as
shown in table 7, when vegetable intakes
are adjusted by adding the amount from
mixtures to the intake of vegetables
coded separately, small increases over
time can be seen in vegetable intakes
for both women and men.
Energy, Nutrients, and Other
Food Components
Energy.-In 1994-95, mean food energy
intakes were estimated to be 1,633 kcal
for women and 2,470 kcal for men (21 ).
This level of energy intake represents
78 percent of the RDA (8) for women
and 91 percent for men (table 8). In
apparent contradiction to this finding,
31 percent of women and 32 percent of
men in the CSFII 1994-95 were over­weight,
5 according to body mass index
estimates from self-reported height and
weight data (21 ). A similar percentage
of overweight men and a slightly higher
percentage of overweight women were
found in the National Health and Nutrition
Examination Survey (NHANES) 1988-
94 (9).
4Mixtures breakdown data from the CSFII 1995
were not available as this article was being written.
50verweight is defined as body mass index (kglm2
)
:<:.27 .3 for women and :<:.27 .8 for men (85th percen­tiles
from NHANES 11 for ages 20 to 29 years).
11
Figure 2. Food energy intake from fat and carbohydrate, adults 20 years and over, 1 day
Fat
1977-78
Women 1989-91
1994-95
1977-78
Men 1989-91
1994-95
Carbohydrate
Women
Men
1977-78
1989-91
1994-95
1977-78
1989-91
1994-95
0 10
D 1977-78
• 1989-91
• 1994-95
20 30 40 50 60
Percent of calories
Source: U.S. Department of Agriculture, Agricultural Research Service, Beltsville Human Nutrition Research Center, Food Surveys Research
Group. USDA Nationwide Food Consumption Survey 1977-78and USDA Continuing Survey of Food Intakes by Individuals 1989-91 and 1994-95.
Underreporting and sedentary lifestyles
are two factors that may help to explain
the disparity between energy intake
estimates and prevalence of overweight.
It is widely recognized that some indi­viduals
in nutrition studies underreport
the food they eat ( 1, 7, 12 ). Also, the
average energy allowances are designed
for a light-to-moderate level of physical
activity. The actual level of physical
activity of many individuals in the United
States is lower than light-to-moderate.
In the CSFII 1994-95, in answer to the
question, "How often do you exercise
vigorously enough to work up a sweat?",
44 percent of women and 28 percent
of men answered "rarely" or "never"
(21 ). In 1996, the Surgeon General
12
concluded that 60 percent of American
adults are not regularly active and 25
percent are not active at all (20), based
on data from 1985 through 1994.
Energy-providing nutrients.-For
both women and men, the proportion of
food energy intake provided by protein
was 16 to 17 percent of calories in all
three survey periods. In contrast, the
proportion of energy from fat decreased
between 1977-78 and 1994-95, and the
proportion of energy from carbohydrate
increased as discussed below. In 1994-95,
the proportion of energy from alcohol
was 1 percent of calories for women
and 3 percent of calories for men.
For women, the proportion of energy
from fat fell from 40 percent of calories
in 1977-78 to 32 percent in 1994-95,
and the proportion of energy from carbo­hydrate
climbed from 42 to 52 percent
of calories (fig. 2). Likewise for men,
the proportion of energy from fat dropped
from 42 to 34 percent of calories, and the
proportion of energy from carbohydrate
rose from 40 to 49 percent of calories.
The shift to lower fat intakes has been
paralleled by a decrease in the average
serum total cholesterol level of individuals
age 20 through 74 years in the United
States from 213 mg/dL in 1976-80 to
205 mg/dLin 1988-91 (6).
Family Economics and Nutrition Review
In 1994-95, average fat intakes as a
percentage of calories by both women
and men still exceeded the Dietary
Guidelines recommendation of 30 per­cent
of calories or less, and saturated
fat intakes ( 11 percent of calories for
both women and men) still exceeded
the recommendation of less than 10
percent of calories ( 19).
When CSFII 1994 food intakes were
compared with the Food Guide Pyramid
( 18), discretionary fat6 intake accounted
for 25 percent of calories for both
women and men 20 years and over (2).
Although the Pyramid does not mention
a specific percentage of calories from
discretionary fat, 25 percent is most
likely much too high. Individuals will
get up to half of the recommended 30
percent of calories from fat if they eat
the recommended number of servings
from each Pyramid food group, choose
the lowest fat choices within each food
group, and add no fat to foods in prepa­ration
or at the table ( 18). Any addi­tional
fat-that is, discretionary fat­should
be low enough to keep the total
fat at 30 percent or less of calories.
6oiscretionary fat was defined as fat added to
foods in preparation or at the table and as excess
fat above amounts people would consume if they
selected only the lowest fat foods within the major
food groups. Examples cited were the fat absorbed
by french-fried potatoes during preparation, the
fat from margarine spread on bread at the table,
and the fat in whole milk.
1997 Vol. 10 No. 4
In 1994, sugar intake was also higher
than the Food Guide Pyramid recom­mends.
Women consumed almost 15
teaspoons of added sugars 7 in a diet
providing I ,587 calories, and men con­sumed
21 teaspoons of added sugars at
an energy intake of 2,403 calories (2 ).
The Pyramid suggests that Americans
try to limit their added sugars to 6 tea­spoons
a day if they eat about 1 ,600
calories, 12 teaspoons at 2,200 calories,
or 18 teaspoons at 2,800 calories.
Vitamins and minerals.-As shown in
table 8, mean nutrient intakes8 by both
women and men age 20 years and over
met or exceeded the RDA for protein,
vitamins A, B12, and C, folate, and iron.
(However, women age 20 to 49 years
had mean iron intakes below the RDA.)
Mean intakes by men (but not women)
also met or exceeded the RDA for
vitamins B6 and E and calcium. Neither
women nor men had intakes that met
the RDA for magnesium or zinc. In
general, the nutrients that were below
the RDA in 1994-95 are the same
nutrients that were below the RDA
in 1977-78. (Vitamin E and zinc were
not examined in 1977-78.)
7The definition of added sugars included white
sugar, brown sugar, raw sugar, com syrup, honey,
molasses, and artificial sweeteners containing
carbohydrate that were eaten separate! y or used
as ingredients in processed or prepared foods such
as breads, cakes, soft drinks, jams, and ice cream;
it did not include sugars such as fructose and lactose
that occur naturally in foods such as fruit and milk.
Quantities were standardized on a carbohydrate
equivalent basis. One teaspoon of added sugars was
defined as the quantity of a sweetener that contains
the same amount of carbohydrate as I teaspoon
(4 grams) of table sugar (sucrose).
8Nutrient intake estimates in table 8 are based on
intakes from food and do not include intakes from
supplements.
Other food components.-Between
1989-91 and 1994-95,9 cholesterol
intakes decreased slightly among both
women and men, and dietary fiber
intakes increased somewhat ( 14, 21 ).
The 1994-95 mean cholesterol intake
by women (217 mg) met the recommenda­tion
to consume less than 300 mg per
day, but the average cholesterol intake
by men (337 mg) still exceeded the rec­ommendation
( 19). Mean dietary fiber
intakes in 1994-95 were 13.7 g for
women and 18.5 g for men, in contrast
with the 1989-91 estimates of 12.3 g
and 16.7 g, respectively. Many expert
groups recommend increasing fiber
intakes by increasing intakes of vegetables,
fruits, and whole grain products without
specifying a fiber goal in grams; other
groups have proposed increasing fiber
intakes to 20 to 35 g per day ( 4, 10 ).
9Cholesterol and fiber intake data are not available
from the NFCS 1977-78.
13
Summary and
Recommendations
Adults in the United States ate some­what
differently in 1994-95 than their
counterparts did two decades before,
with mixed results. Adults' low intakes
of fiber, magnesium, and zinc could be
improved by increasing intakes of whole
grains, fruits, dark green vegetables,
legumes, and lean meats and meat alter­nates
as recommended by the Food
Guide Pyramid ( 18). The proportion of
milk intake that was skim milk increased,
but milk intakes overall declined.
Women's low calcium intakes could
be boosted by increasing intakes of
skim and lowfat dairy products.
The proportion of energy from fat was
lower and that from carbohydrate higher
in 1994-95 than in 1977-78. This shift
was consistent with other national data
showing a decrease in the average blood
cholesterol level among adults. On the
other hand, more adults were overweight
in 1994-95 than before. Adults' diets
would benefit overall by decreasing
intakes of foods and beverages that are
high in fats and sugars but provide few
other nutrients. In addition, when choosing
among more nutrient-dense foods, adults
would do well to shift toward items
lower in fat and sugar.
Low activity levels probably contrib­uted
to the high percentage of adults
who were overweight in 1994-95. For
better health, adults should become
more active.
Acknowledgments
The authors wish to thank Katherine E.
Sykes for her review of the manuscript
and helpful comments.
14
References
1. Bingham, S.A. 1987. The dietary assessment of individuals: Methods, accuracy, new
techniques and recommendations. Nutrition Abstracts and Reviews 51( 10):105-142.
2. Cleveland, L.E., Cook, A.J., Wilson, J.W., Friday, J.E. , Ho, J.W., and Chahil, P.S. 1997,
March I 0. Pyramid Servings Data: Results from USDA's 1994 Continuing Survey ofF ood
Intakes by Individuals, [Online]. ARS Food Surveys Research Group. Available (under
"Releases"): <http://www.barc.usda.gov/bhnrc/foodsurvey/home.htm>[ visited 1997,
September 2].
3. DeMaio, T.M., Ciochetto, S., and Davis, W.L. 1993. Research on the Continuing Survey
of Food Intakes by Individuals. In: 1993 Proceedings of the Section on Survey Research
Methods, papers presented at American Statistical Association, San Francisco, August 8-12,
1993, and American Association for Public Opinion Research, St. Charles, IL, May 20-23,
1993, vol. II, pp. 1021-1026. American Statistical Association, Alexandria, VA.
4. Federation of American Societies for Experimental Biology, Life Sciences Research Office.
1987. Physiological Effects and Health Consequences of Dietary Fiber. Bethesda, MD.
5. Guenther, P.M., DeMaio, T.J ., Ingwersen, L.A., and Berlin, M. 1996. The multiple-pass
approach for the 24-hour recall in the Continuing Survey of Food Intakes by Individuals
(CSFII) 1994-96. The FASEB Journal10(3):AI98.
6. Johnson, C.L., Rifkind, B.M., Sempos, C.T., Carroll, M.D., Bachorik, P.S., Briefel, R.R.,
Gordon, D.J., Burt, Y.L. , Brown, C.D., Lippe!, K., and Cleeman, J.I. 1993. Declining serum
total cholesterol levels among US adults: The National Health and Nutrition Examination
Surveys. Journal of the American Medical Association 269:3002-3008.
7. Mertz, W., Tsui, J.C., Judd, J.T., Reiser, S., Hall frisch, J., Morris, E.R., Steele, P.D., and
Lashley, E. 1991. What are people really eating? The relation between energy intake derived
from estimated diet records and intake determined to maintain body weight. American Journal
of Clinical Nutrition 54:291-295.
8. National Academy of Sciences, National Research Council, Food and Nutrition Board.
1989. Recommended Dietary Allowances (lOth ed.). National Academy Press, Washington, DC.
9. National Center for Health Statistics and Centers for Disease Control and Prevention.
1997. Update: Prevalence of overweight among children, adolescents, and adults-United
States, 1988-1994. Morbidity and Mortality Weekly Review 46(9): 199-202.
10. National Institutes of Health, National Cancer Institute. 1987. Diet, Nutrition, and Cancer
Prevention: A Guide to Food Choices. NIH Pub!. No. 87-2878.
11. Putnam, J.J. and Allshouse, J.E. 1993. Food Consumption, Prices, and Expenditures,
1970-92. U.S. Department of Agriculture, Economic Research Service, Statistical Bulletin
No. 867 (table 36).
Family Economics and Nutrition Review
12. Schoeller, D.A. 1990. How accurate is self-reported dietary energy intake? Nutrition
Reviews 48( 10):373-379.
13. Tippett, K.S. and Cypel, Y.S. (Eds.). 1997. Design and Operation: The Continuing
Survey of Food Intakes by Individuals and the Diet and Health Knowledge Survey 1994-96.
Continuing Survey of Food Intakes by Individuals 1994-96, Nationwide Food Surveys Rep.
No. 96-1. U.S. Department of Agriculture, Agricultural Research Service.
14. Tippett, K.S., Mickle, S.J., Goldman, J.D., Sykes, K.E., Cook, D.A., Sebastian, R.S.,
Wilson, J.W., and Smith, J. 1995. Food and Nutrient Intakes by Individuals in the United
States, 1 Day, 1989-91. Continuing Survey of Food Intakes by Individuals 1989-91, Nationwide
Food Surveys Rep. No. 91-2. U.S. Department of Agriculture, Agricultural Research Service.
15. U.S. Department of Agriculture, Human Nutrition Information Service. 1983. Food Intakes:
Individuals in 48 States, Year 1977-78. Nationwide Food Consumption Survey 1977-78, Rep.
1-1. National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161;
(703) 487-4650. Technical report: Accession no. PB91-103523INZ.
16. U.S. Department of Agriculture. 1995. CSFJJIDHKS 1994 Data Set and Documentation:
The 1994 Continuing Survey of Food Intakes by Individuals and the 1994 Diet and Health
Knowledge Survey. National Technical Information Service, 5285 Port Royal Road,
Springfield, VA 22161; (703) 487-4650. Computer tape: Accession no. PB96-500095.
CD-ROM: Accession no. PB96-501010.
17. U.S. Department of Agriculture. 1996. CSFJJ/DHKS 1995 Data Set and Documentation:
The 1995 Continuing Survey of Food Intakes by Individuals and the 1995 Diet and Health
Knowledge Survey. National Technical Information Service, 5285 Port Royal Road,
Springfield, VA 22161; (703) 487-4650. Computer tape: Accession no. PB97-500771.
CD-ROM: Accession no. PB97-500789.
18. U.S. Department of Agriculture, Center for Nutrition Policy and Promotion. 1996. The
Food Guide Pyramid. Home and Garden Bulletin No. 252, sl. rev.
19. U.S. Department of Agriculture and U.S. Department of Health and Human Services. 1995.
Nutrition and Your Health: Dietary Guidelines for Americans (4th ed.). Home and Garden
Bulletin No. 232.
20. U.S. Department of Health and Human Services, Centers for Disease Control and
Prevention. 1997. Physical Activity and Health: A Report of the Surgeon General. Atlanta, GA.
21. Wilson, J.W., Enns, C.W., Goldman, J.D., Tippett, K.S., Mickle, S.J., Cleveland, L.E.,
and Chahil, P.S. 1997, April18. Data Tables: Combined Results From USDA's 1994 and
1995 Continuing Survey of Food Intakes by Individuals and 1994 and 1995 Diet and Health
Knowledge Survey, [Online]. ARS Food Surveys Research Group. Available (under "Releases"):
<http://www.barc.usda.gov/bhnrc/foodsurvey/home.htm>[ visited 1997, September 2].
1997 Vol. 10 No.4 15
16
Trends in Marketing and
Usage of Fat-Modified Foods:
Implications for Dietary
Status and Nutrition
Promotion
Nancy E. Schwenk
Joanne F. Guthrie
Center for Nutrition Policy and Promotion
The tremendous increase in the number of fat-modified foods is one of the
most notable changes in the retail food market that has occurred in the past
few years. These products have the potential to change Americans' eating
habits and could have significant effects on dietary status. The purposes of
this article are to inform nutrition and consumer educators of this trend and
the factors that have contributed to it; to examine concerns that have been
raised in relation to this trend; and to discuss the potential implications for
nutrition promotion.
[!] n recent years, there has
been an enormous increase
in the number of fat-modified
foods in the marketplace.
This increase has been spurred by
consumer interest and further fueled
by passage of the Nutrition Labeling
and Education Act (NLEA). The rapid
influx of fat-modified food products
into the marketplace has raised concerns
regarding their use, including potential
physiological effects, price, and poten­tial
impact on overall diet quality.
This paper will define fat-modified
foods, discuss the factors that have led
to their increased presence in the market­place,
and review the evidence on the
concerns raised. It will identify future
research issues and discuss the
challenges for nutrition education and
promotion in helping consumers under­stand
the role of fat-modified foods in
their diets.
Family Economics and Nutrition Review
Definition of Fat-Modified Foods Table 1. Functions of fat
reduction ingredients
Some fat-modified foods can be created
merely by removing fat from a conven­tional
food product-skim milk is a good
example of this type of modification.
For many foods, however, fat provides
important sensory and functional qualities,
and simply removing it would not result
in an acceptable product. In those foods,
fat substitutes are used to mimic the
qualities provided by fat, such as the
rich, creamy "mouth feel" of ice cream
or the moist texture of a cake (table 1).
In this paper, the term "fat-modified
foods" will be used to refer to all food
products in which the fat content has
been reduced compared with a standard
product, either by simply removing fat
or by replacing fat wholly or partially
with some type of fat substitute.
In some products, fat can be replaced
or reduced using relatively simple
techniques. For example, water can be
added to some foods, like margarine, to
dilute their fat concentration. Sometimes
fat reduction can be accomplished by
substituting familiar ingredients for fat
in a product. Widely used examples
would be fruit purees like prune paste
or applesauce that are used in both
homemade and commercially baked
goods as substitutes for fat in a recipe.
In addition to these modifications,
food technologists have developed or
modified other compounds for use as
fat substitutes. These fat substitutes can
be based on protein, carbohydrate, or a
reduced-calorie or noncaloric version of
fats ( 17). Table 2 provides examples of
some of the major fat substitutes that
fall into these categories.
• Bulking agent
• Gelling
• Moisturizer
• Provide mouth feel
• Stabilizer
• Texturizer
• Thickener
Source:IntemationaiFood Information Council
Foundation, 1995, Uses and Nutritionallmpact
of Fat Reduction ingredients, November issue.
Table 2. Types of fat substitutes1
Category
Protein-based
Carbohydrate-based
Reduced and noncaloric fat-based
Examples
Simplesse
Dextrins and modified starches
Polydextrose
Cellulose
Gums and pectin
Oatrim and Z-trim
Caprenin
Salatrim
Mono- and diglycerides
Olestra
Under Development.
DDM
EPG
TATCA
1 Adapted from International Food Information Council, 1995, Uses and Nutritional Impact of Fat
Reduction Ingredients, November issue.
1997 Vol. 10 No.4 17
Protein-Based Fat Substitutes
Protein-based fat substitutes, such as
Simplesse1 (made by the NutraSweet
Company) and similar products, are
created by a process called micro­particulation.
Proteins are heated and
blended at high speed, producing tiny,
spherical particles that give foods a
creamy "mouth feel" similar to that
of fats ( 3 ). They can be used in such
products as reduced-fat ice cream
and dairy desserts, mayonnaise, salad
dressings, and cheese but cannot be
used to make reduced-fat fried foods
such as potato chips and french fries ( 3 ).
Protein provides 4 calories per gram and
fat provides 9 calories per gram; these
substitutes provide 1-4 calories per gram,
depending on the amount of hydration
of the product (20 ). In addition, they
may be usable in smaller amounts than
the fats they replace; for example, 1 gram
of Simplesse can be used to replace
3 grams of fat in ice cream ( 17).
Carbohydrate-Based Fat
Substitutes
There are a number of carbohydrate­based
fat substitutes, and they are used
in a wide range of food products such as
salad dressings, baked products, cheeses,
processed meats, sour cream, yogurt,
and puddings (20). Like protein-based
fat substitutes, they cannot be used to
fry foods ( 3 ).
Among the most common carbohydrate­based
fat substitutes are dextrins and
modified starches. These compounds
absorb water, forming gels that have
a texture and mouth feel similar to fat.
Gums such as guar gum are used as fat
replacers, particularly in foods such as
1Use of commercial or trade names does not imply
approval or constitute endorsement by the U.S.
Department of Agriculture (USDA).
18
salad dressings where they also act as
stabilizers. Pectin is another gel-forming
carbohydrate sometimes used as a fat
substitute.
Indigestible fiber cellulose can be ground
into rnicroparticles that can form gels
that are used as fat substitutes ( 3 ).
Two new fiber-based fat substitutes are
Oatrim and Z-trim, which were developed
by scientists at USDA. Oatrim is pro­duced
from oat fiber (3), whereas Z-trim
is made from a mixture of plant fibers,
including oat, corn, rice, wheat, and
soybean fiber ( 34 ).
Polydextrose, a glucose polymer used to
replace sugar or fat in foods, helps keep
foods such as lowfat cakes and desserts
moist. It also acts as a bulking agent
replacing volume lost by the removal
of fat and/or sugar ( 3 ).
Like proteins, carbohydrates provide 4
calories per gram. Carbohydrate-based
fat substitutes can supply 0 to 4 calories
per gram, depending on the specific
ingredient and how it is used (20). Dry
forms of dextrins can supply 4 calories
per gram; hydrated forms, 1-2 calories
per gram; and gums and fibrous carbo­hydrates
such as cellulose contribute
virtually no calories (20 ).
Reduced and Noncaloric
Fat-Based Substitutes
Reduced and noncaloric fat-based sub­stitutes
can be used in chocolate, dairy
products, and other foods. Some reduced­calorie
and noncaloric fat-based substi­tutes
can be used to fry foods ( 17).
The reduced-calorie fat-based substitutes
Caprenin (Procter and Gamble Company)
and Salatrim (Nabisco, Inc.) are only
partially digested and absorbed and,
therefore, contribute fewer calories per
gram than conventional fats-5 calories
per gram compared with 9 calories per
gram ( 17). Other fat-based substitutes
have been created using mono- and
diglycerides. Although mono- and
diglycerides provide the same amount
of calories as other fats, they can be
used in smaller amounts, thus reducing
total fat and calories ( 17).
Olestra, a fat substitute recently approved
by the Food and Drug Administration
(FDA) for use in snack products, is
not digested and absorbed at all and is,
therefore, noncaloric. Despite its recent
approval as a fat substitute, its use
remains controversial (see box, pp. 21
and 22). Several potential noncaloric
fat-based substitutes are currently under
development. Among them are dialkyl
dihexadecylmalonate (DDM), esterified
propoxylated glycerol (EPG), and trialk­oxytricarballate
(TATCA) ( 17).
Reasons for the Increase in
Fat-Modified Foods
Health Concerns
Diseases resulting from poor diet or
inadequate physical activity cause an
estimated 300,000 deaths a year (25).
High fat intakes, in particular, are associ­ated
with an increased risk of coronary
heart disease, stroke, certain types of
cancer, hypertension, diabetes, and other
diseases ( 17). Researchers studying
the effects of reduced dietary intake
of saturated fat on the incidence and
economic costs of coronary heart
disease estimated that a 3-percent drop
in saturated fat intake could prevent
about I 00,000 new cases of coronary
heart disease by the year 2005 and save
nearly $13 billion in medical costs and
lost earnings (in 1993 dollars) (28).
Family Economics and Nutrition Review
The 1995 Dietary Guidelines for
Americans, issued jointly by the U.S.
Departments of Agriculture and Health
and Human Services as the official
statement of Federal dietary guidance
policy, recommend, for healthy persons
age 2 and over? limiting saturated fat to
less than 10 percent of calories consumed
and total fat intake to 30 percent or less.
In addition, numerous health organiza­tions
such as the American Heart Asso­ciation
(22) and the American Cancer
Society ( 1) recommend a lowfat diet­no
more than 30 percent of total calories
from fat.
Since fat supplies energy and essential
fatty acids and aids absorption of fat­soluble
vitamins A, D, E, and K, some
fat is needed in the diet ( 36). However,
most Americans consume more dietary
fat than what is considered to be a healthy
amount. In 1995, Americans averaged
33 percent of their total caloric intake
from fat (35). Although this is a decline
from the 1977-78 figure of approxi­mately
41 percent of calories from fat
( 40 ), it is still above recommendations.
Consumer Reaction
The belief that food and nutrition have
a strong impact on personal health is
growing among consumers. The impor­tance
of nutrition as a determinant in
food-buying decisions is at an all-time
high (78 percent). In 1996, 97 percent
of food shoppers claimed to have
changed their eating habits to reflect a
healthier diet, up from 92 percent the
previous year ( 30 ). The leading changes
in dietary behavior reported in 1996
were an increase in fruits and vegetables
and a decrease in fats and oils (table 3).
2F at t.n takes by children ages 2 to 5 should be
gradually decreased to no more than 30 percent of
calories from fat (36).
1997 Vol. 10 No.4
Table 3. Leading changes in dietary behavior, 1992-961
Dietary change
More fruits/vegetables
Less fats/oils
Less meat/red meat
Less sugar
Less snack foods/junk foods
Less salt/sodium
Eating more chicken/turkey/
white meat
Less fried foods
Eating more fish
Less dairy/butter/cheese/
whole milk
More fiber
More starch/rice/potato/pasta
Less cholesterol
More whole grain
More fresh foods
1992
60
28
31
12
12
8
14
7
10
7
8
5
8
2
3
1993
62
26
30
15
12
8
14
6
10
4
8
5
6
3
3
Year
1994
Percent
63
32
31
13
11
7
13
6
8
6
7
7
3
5
1995
63
34
34
15
12
6
16
6
8
6
5
8
3
6
3
1996
77
42
32
20
18
13
12
9
8
7
7
6
6
4
3
1 Shoppers were asked, "What, if anything, are you eating more of to ensure that your diet is healthy?
What, if anything, are you eating less of to ensure that your diet is healthy?" (Multiple responses accepted.)
Source: Food Marketing Institute, 1996, Trends in the United States: Consumer Attitudes and the
Supermarket, 1996.
The most common motivation for making
these changes was weight control,
rather than concern about disease ( 15 ).
Among food shoppers with nutritional
concerns (an overwhelming 94 percent
in 1996), fat is likely to be the main
concern ( 14 ). When asked their specific
concerns about the nutritional content
of food, these shoppers indicated that
their top concerns were: Fat (60 percent),
salt/sodium (28 percent), cholesterol
(26 percent), sugar (12 percent), calories
( 12 percent), and vitamin/mineral
content (12 percent) (multiple answers
were accepted) (table 4, p. 22). Concern
about fat peaked in 1995 at 65 percent ( 14).
It can be difficult for a person to adopt
a lowfat diet. It may require changing
basic food selection and preparation
patterns. Often, many culturally based
or favorite foods must be limited ( 3 ). In
addition, many consumers (nearly half
of food shoppers) are confused and
unsure about how to modify fat intake.
Substituting a fat-modified food for its
regular counterpart may seem a simpler
way to reduce fat than eating different
types of foods.
19
20
In 1995, Americans
averaged 33 percent
of their total caloric
intake from fat.
Olestra
In the past year, foods containing olestra, a noncaloric fat substitute manufactured
by Procter & Gamble, have begun to appear on American supermarket shelves.
Olestra, chemically referred to as sucrose polyester, is formed by joining sucrose
with fatty acids obtained from vegetable oils ( 17). Olestra is not broken down
by digestive enzymes. It passes through the gastrointestinal tract without being
digested and absorbed and therefore provides no calories.
Olestra tastes like fat and, unlike protein- and carbohydrate-based fat substitutes,
it can be used in frying to make reduced-fat potato chips, french fries, and other
fried foods. It can also be used as a fat substitute in baked goods, spreads, salad
dressings, candy, ice cream, and other foods ( 17).
Most fat substitutes currently used by food manufacturers are classified by the
FDA as GRAS (generally recognized as safe). Olestra, as a novel ingredient, had
to be approved by the FDA as a food additive. Procter & Gamble first petitioned
FDA for approval of olestra as a food additive in 1987 ( 18). Between 1987 and
1995, Procter & Gamble submitted to FDA results of more than 150 studies on
olestra's physiological effects (8).
Two major concerns about the safety of olestra have been raised: The first is olestra' s
effects on the absorption of some nutrients; the second is its gastrointestinal
effects. Olestra has been found to impair the absorption of the fat-soluble vitamins
A, D, E, and K. It also interferes with the absorption of carotenoids. Beta caro­tene
is a precursor of vitamin A; some research suggests that other carotenoids
may also have important health effects (21 ). In addition, it has been suggested
that olestra may interfere with the absorption of some potentially beneficial
phytochemicals (21 ). 1
Olestra's gastrointestinal effects occur because it passes through the digestive
tract unabsorbed. As originally formulated, consumption of olestra could result
in anal leakage. Reformulation appears to have alleviated this problem, at least
partially, but olestra consumption still may result in loose stools and gastro­intestinal
distress ( 18 ).
The data presented by Procter & Gamble were reviewed by FDA's Center for
Food Safety and Nutrition (CFSAN), Center for Veterinary Medicine, and
Food Advisory Committee (8). Most members of the Food Advisory Committee
accepted olestra's safety, but a minority advised rejection of the petition (32).
There also was vocal public opposition to approval of olestra; opposing groups
included the Center for Science in the Public Interest (CSPI), the American Public
1Phytochemicals are defined as non-nutrient, physiologically active components of plants (21 ).
Investigation of their potential health properties is an emerging area of research.
Family Economics and Nutrition Review
Health Association, the American Academy of Ophthalmology, and Consumers
Union (32). Nevertheless, in January 1996, FDA approved use of olestra as an
additive in savory snacks, for example, potato chips and crackers. Olestra can be
used to replace up to 100 percent of the conventional fat or oil used in one of
these products, but for olestra to be used in other products, a new petition would
have to be filed by Procter & Gamble and approved by FDA ( 1 8).
As a condition of approval, FDA required olestra-containing foods to be fortified
with vitamins A, D, E, and K. Carotenoid fortification was not required; previously,
the FDA Food Advisory Committee had considered carotenoid depletion by
olestra, with most members concluding that they were reasonably certain that
no harm would result from the effects of olestra on carotenoids ( 10). In addition,
FDA requires that olestra-containing products bear the label statement:
"This product contains olestra. Olestra may cause abdominal cramping
and loose stools. Olestra inhibits the absorption of some vitamins and
other nutrients. Vitamins A, D, E, and K have been added."
Finally, FDA required Procter & Gamble to conduct studies monitoring olestra's
consumption and its long-term effects. FDA announced its plans to review the
data within 30 months of olestra's approval (June, 1998) ( 10).
Procter & Gamble is now marketing "olean," its brand name for olestra, as an
ingredient in snack products. Frito-Lay has become the first company to market
a snack-potato chips-made with olestra ( 11 ). Procter & Gamble has also
announced plans to market its own brand of olestra-containing potato chips ( 1 8).
Despite FDA approval of olestra, consumers have continued to protest the product.
In particular, the Center for Science in the Public Interest has campaigned
actively to inform the public of what it perceives to be the negative effects of
olestra ( 12). With olestra-containing products still available only in limited markets,
the degree of consumer acceptance they will enjoy remains to be seen. However,
during the first 4 weeks of test marketing in Indiana, 4 million servings of chips
were sold and no complaints from consumers were received ( 13 ). The popularity
of snacks containing olestra (as measured by sales), along with the results of
FDA's postmarket review of olestra's effects, could have a critical effect on
future development and marketing of olestra-containing products. If sales are
good and no negative effects are found, it may encourage the manufacturer to
seek approval for olestra's inclusion in a broader range of products. If either the ­FDA
review or consumer reaction is negative, olestra may fade from the scene.
1997 Vol. 10 No.4
Growth of Fat-Modified Foods
·The Public Health Service's 1990
publication Healthy People 2000:
National Health Promotion and Disease
Prevention Objectives made a request
of the food industry-a doubling of
available lower fat foods by the year
2000 (38). With consumer demand
spurring industry on, this objective was
accomplished quickly, as more than
1,000 fat-modified foods were intro­duced,
on average, each year since
1990 (20). By comparison, only 38 fat­modified
products were introduced in
1981 (26 ). At that time, there was less
consumer demand for such products,
and advanced technologies to produce
reduced-fat foods were not available.
Types of Foods
Processed foods provide the greatest
opportunity for fat modification because
they can be reformulated to control their
fat content. Even though the fat content
of foods such as milk, meat, and eggs
occurs naturally, it can be decreased
through breeding or feeding practices
(26). For example, according to the
National Cattlemen's Association and
the Beef Industry Council, "the beef
industry has succeeded in reducing the
amount of fat in its products. Beef today
has 27 percent less trimmable fat" than
a decade ago (27). Leaner meats are the
result of improved breeding and meat­trimming
practices ( 16).
Frazao and Allshouse ( 16) used super­market
scanner data for 1989-93 to study
the size and growth of the nutritionally
improved foods market compared with
their traditional counterparts. The 37
food categories used in their study
accounted for 71 percent of volume
sales captured by the scanner data in
1993. These nutritionally improved
foods represented a steadily increasing
proportion of retail food sales. Food
21
categories with nutritionally improved
versions that had the largest percent
growth in sales volume between 1989
and 1993 were dairy puddings, spaghetti
sauce, cookies, whipping creams, popcorn,
cream cheese, sour cream, and salad/
cooking oils. The largest contributors to
volume sales in 1993 were beverages­carbonated
beverages, milk, fruit
juices/drinks/ades, and beer ( 16).
Fat-modified products make up a very
large share of total products available
among several product categories,
including yogurt, refrigerated milk,
and cottage cheese. Nonfat yogurt is
so popular that whole milk yogurt
accounts for only a small percentage
of the yogurt market (9 ).
Snacks
Fat-modified snack foods were eaten
by 49 percent of consumers in 1996
compared with 39 percent in 1993 (6).
According to the Snack Food Association,
fat-modified snack foods are the "fastest­growing
products in the supermarkets."
Between 1994 and 1995, fat-modified
pretzels gained 339 percent in sales,
whereas fat-modified varieties of tortilla
chips gained 67 percent; potato chips,
48 percent; and cookies, 38 percent.
Low fat dips and salsas also gained in
sales in 1995 ( 11 ).
The Influence of
Federal Legislation
The Nutrition Labeling and Education
Act (NLEA), which was signed in
November of 1990 and went into effect
in August of 1994, amends the Federal
Food, Drug, and Cosmetic Act (33).
Under the NLEA, virtually all processed
foods for sale must carry nutrition label­ing
that provides the amount of total fat
and saturated fat per serving (as well as
other nutrients). Also, the number of
22
Table 4. Leading concerns about nutritional content of food, 1992-961
Year
Concern 1992 1993 1994 1995 1996
Percent
Fat content, low fat 50 54 59 65 60
Salt/sodium content, less salt 21 26 18 20 28
Cholesterol levels 30 23 21 18 26
Sugar content/less sugar 13 18 14 15 12
Calories/low calorie 9 15 7 13 12
Vitamin/mineral content 8 10 6 8 12
Preservatives 11 8 10 11 8
Chemical additives 9 6 8 10 7
Food/nutritional value 5 10 4 8 6
Freshness/purity /no spoilage 5 3 5 7 5
Desire to be healthy/eat what's
good for us 2 5 4 3 5
Ingredients/contents 5 5 2 6 4
As natural as possible/not overly
processed 2 2 4 5 3
Chemicals 4 2 3 4 3
Getting a balanced diet 4 4 2 2 3
1 Shoppers who were very or somewhat concerned about the nutritional content of their food were asked,
"What is it about the nutritional content of what you eat that concerns you most?" (Multiple answers
accepted.)
Source: Food Marketing Institute, 1996, Trends in the United States: Consumer Attitudes and the
Supermarket, 1996.
calories per serving derived from fat
and any other source must be included
( 3 3 ). As a result, consumers are able to
compare the fat content of different
types of foods and different brands of
the same food. In addition, the NLEA
restricts the types of claims that manu­facturers
can make on food products. The
number of food products bearing fat­related
health claims increased until 1993,
when the number decreased dramatically.
Although consumer demand for fat­modified
products remained strong, the
decrease in fat-related health claims can
be attributed to the NLEA. By 1993,
food manufacturers were labeling their
products in anticipation of the
August 1994 implementation date ( 31 ).
FDA researchers found that between
the spring of 1994 and the fall of 1995
(before and after NLEA implementation)
there was a significant increase in the
percentage of adults who often used
food labels to obtain information on a
food's nutrient content or for health
claims applied to that food (7). It appears
that consumers are increasingly using
food labels to make food choices based
on fat content and other nutrients. In
addition, consumers are more confident
about the accuracy of label information
(31 ).
Family Economics and Nutrition Review
Figure 1. Percent of U.S. adults purchasing food products
labeled low fat, 1995
Percent
Regularly
41 ~ ~ ::: : I
Sometimes
39
Hardly ever
11
Never
9 1
Source: Derby, B.M. and Levy, A.S., 1996, Consumer and Market Impacts of the Nutrition
Labeling and Education Act, Presentation made at the Marketing and Public Policy Conference,
Rosslyn, VA, May 17.
Purchase Behavior/Consumption
According to the Calorie Control
Council, 88 percent of American adults
(90 percent of women and 87 percent of
men) consumed fat-modified foods and
beverages in 1996 (6). Only 20 percent
of adults never or hardly ever purchased
lowfat food products (fig. 1) (7).
As of 1996, 81 percent of shoppers had
sought out and bought products labeled
"low-fat" ( 14). Fat is the main reason
why shoppers selected the foods they
did. Information about fat content on
the nutrition label prompted 72 percent
of shoppers to start buying a new food,
compared with those who were moti­vated
by information about sodium
(13 percent), calories (9 percent),
cholesterol (6 percent), and preserv­atives
and additives (4 percent) ( 1 5).
In 1996, consumers who used any fat­modified
products on a regular basis
were likely to use at least five: Milk,
salad dressing/sauce/mayonnaise, cheese/
dairy products, margarine, and chips/
1997 Vol. 10 No.4
snack foods (6). Three-quarters of these
consumers used these products at least
several times a week, and over half used
them daily.
Concerns Regarding
Consumption of
Fat-Modified Foods
Increased consumption of fat-modified
foods may seem to be a positive trend in
American eating habits, since it could
help achieve the Dietary Guidelines
recommendation to reduce fat and satu­rated
fat intake. Nevertheless, numerous
concerns have been raised about these
products-including those of safety,
price, and diet quality.
Safety Concerns
The safety of food additives must always
be considered, but fat substitutes merit
particular study because, unlike most
additives, they could potentially be
consumed in very large quantities (37).
The FDA is responsible for the safety of
food additives. Under the Federal Food,
Drug, and Cosmetic Act, a substance
is exempt from the definition of food
additive and from premarket approval
requirements if it is generally recognized
as safe (GRAS) by qualified experts.
Generally, substances that are identical
or similar to traditional food ingredients
are regulated under the GRAS exemp­tion,
whereas those that are truly novel
are regulated as food additives.
Many of the products used as fat replacers,
such as starches, gums, and cellulose,
are also in common use as emulsifiers,
gelling agents, stabilizers, and bulking
agents and under those uses have long
been approved for use by FDA. Thus,
it is assumed that their use poses no
health concerns. FDA, however, may
re-examine approved substances should
any concerns surface.
The microparticulated protein product
Simplesse was the first ingredient
announced by FDA specifically for
use as a fat substitute. It is considered
to be a GRAS product. Currently, there
are two non-GRAS ingredients that
have been approved as food additives
for use as fat substitutes: Polydextrose
and olestra. Although polydextrose was
approved, it does have laxative effects
when consumed in large amounts, and
foods containing large amounts of poly­dextrose
must carry a notification label.
Olestra's possible side effects have
raised considerable controversy (see
box, pp. 20 and 21), but its use as a food
additive was approved by FDA on the
recommendation of the FDA Food
Advisory Committee. Currently, its use
is limited to one food category-savory
snacks-and FDA plans to review the
impact of olestra consumption no later
than June 1998 (30 months after
approval) ( 10 ).
23
24
... more than 1,000
fat-modified foods
were introduced,
on average, each
year since 1990.
Petitions have been filed with FDA
regarding two other new fat substitutes,
Caprenin and Salatrim (2,3 ). These
petitions affirm to FDA that Caprenin
and Salatrim are GRAS. As a part of the
process of filing a GRAS Affirmation
Petition, FDA requires that any safety
issues concerning the ingredient be
addressed. The agency does not file
petitions that do not address all criteria
for GRAS status. Food manufacturers
often begin using these products once
the GRAS Affirmation Petition is filed
with the FDA, because they believe it
is unlikely that an ingredient's GRAS
status will be challenged once the peti­tion
is filed. Salatrim, for example, is
currently being marketed under the
brand name Benefat™ (2).
Other Physiological Effects
Some fat substitutes may have positive
health properties. For example, Oatrim
contains soluble fiber, which has been
found to lower cholesterol ( 3 ). Z-trim
also has been cited as a potentially
valuable fiber source (34). Whether
either of these substitutes will be con­sumed
in sufficient quantities to have
a significant impact on Americans'
fiber intakes remains to be seen.
Price
Fat-modified food products may offer
increased options for those following a
lowfat diet, but only if the consumer can
afford them. Results of a recent study
by Frazao and Allshouse indicate that
"nutritionally improved" food items
(a category that included fat-modified,
sodium-modified, and other foods that
were modified nutritionally) generally
cost more per pound in 1993 than their
regular versions (for 30 of 37 food
categories studied) ( 16).
This price differential may have affected
consumer perception of the cost of
healthful diets. A 1996 survey of shoppers
found that 58 percent agreed with the
statement "it costs more to eat healthful
foods," up from 43 percent in 1992
( 15 ). It is, of course, possible to eat a
healthful diet by making appropriate
choices from regular food items. But if
consumers perceive that it is necessary
to purchase fat-modified food products
in order to follow a lowfat diet, the
higher prices of many of these items
may be a barrier to dietary change.
Impact on Diet Quality
Fat-modified foods' potential role in
decreasing fat intake has generated con­siderable
interest. Since many, though
not all, of these foods are also reduced
in calories, it has also been suggested
that their use might help Americans
achieve or maintain a healthy weight.
In addition, some research suggests
that a lower fat diet may be beneficial
for weight loss because fat compared
with carbohydrate is more efficiently
converted to adipose tissue ( 17).
Detractors have argued that there is no
proof that use of these products leads to
an overall diet that is lower in fat and/or
calories. They further contend that if
consumers rely on fat-modified foods
rather than changing basic eating patterns,
this may lead to consumption of diets
that are less healthful in terms of intake
of vitamins, minerals, and other dietary
essentials.
Effects of Fat-Modified Food Use
on Overall Fat and Caloric Intake
Researchers have attempted to model
the potential impact of regular use of fat­modified
food products on diet quality.
Lyle et al. (24) estimated that fat intake
could be reduced to 30 percent of calories
and average caloric intake decreased by
Family Economics and Nutrition Review
approximately 800 calories per week if
fat-free products in several food categories
(cheeses, sour cream, frozen desserts,
commercial sweet baked goods) were
substituted for their regular versions.
In real life, however, consumers may
act differently than the model suggests.
Consumers may eat reduced-fat foods in
larger amounts or otherwise compensate
for the reduced-fat (and sometimes calorie)
content of some foods by eating more
of other foods. To confirm whether
consumption of reduced-fat foods is
associated with an overall diet reduced
in either fat or calories, studies of.actual
eating behavior are needed.
Experimental Data
Studies in which subjects were surrepti­tiously
fed reduced-fat, reduced-calorie
versions of familiar foods have generally
shown that healthy, normal-weight
subjects compensated for the reduced
calories in the lowfat foods by eating
more of other foods. They did not seek
out high fat foods to compensate for the
reduction in fat calories. Instead, they
ate more of their usual mix of foods,
resulting in a net decrease in fat intake
but not in caloric intake ( 17). One study,
conducted with overweight individuals,
showed incomplete caloric compensa­tion
when subjects were surreptitiously
fed reduced-fat, reduced-calorie foods,
resulting in a net reduction in both fat
and calorie consumption ( 19 ).
A special concern is the potential impact
of consumption of fat-modified food
products on children's energy intake.
If a child's caloric intake were reduced
by excess consumption of reduced-fat
products, this might compromise growth.
1997 Vol. 10 No.4
Even if these products are not marketed
towards children, it could reasonably
be expected that if they are widely
consumed there will be at least some
children who will eat them regularly.
Few studies of the impact of consump­tion
of fat-modified food products on
children's diets have been done. One
study, conducted by Birch et al. (5),
indicates that preschool-age children,
like normal-weight adults, compensate
for the reduction in calories but not for
the reduction in fat when fed reduced­fat,
reduced-calorie foods. More studies
with children of differing ages would be
helpful in fully assessing any potential
impacts of these products on childhood
diets (4).
Potential Impacts of Consumers'
Beliefs About Fat-Modified Foods
on Diet
The experimental studies described
above were conducted with individuals
who did not know they were consuming
foods that were reduced in fat and calories.
Therefore, their changes in eating behavior
reflect the influence of physiological
factors only. In normal circumstances,
individuals know they are consuming
fat-modified food products, and this
awareness may influence their behavior.
For example, individuals may react by
eating more of the fat-modified food
product or by eating more of other
foods. Such a reaction could negate the
reduction in fat or calories associated
with consumption of reduced-fat foods
or even lead to a diet higher in fat or
calories if the individual miscalculates.
Advertising and promotional strategies
may encourage consumers to believe
that eating more of fat-modified food
products is acceptable. According to
one national survey conducted in 1996,
about one-third of shoppers believed
that it was alright to eat larger amounts
of lowfat or fat-free products than the
regular variety. This percentage, though
large, was down from 42 percent the
previous year, indicating that consumers
may be learning that fat-free is not
calorie-free ( 15 ).
Evidence that consumption of fat­modified
food products may lead to
overconsumption of other foods was
provided by a study by Shide and Rolls
(29 ), who found that when women
knew they had eaten a lowfat food for
breakfast, they ate more fat and calories
at later meals than they would normally.
However, this was a short-term study,
examining food consumption behavior
only over a 1-day period in a laboratory
setting. Whether this effect would
persist over time and/or would occur
in a real-life situation would need to be
established by other studies, including
studies conducted with a free-living
population.
Food Consumption Survey Data
The best information to date on the
impact of consumption of fat-modified
foods on the average consumer's eating
habits comes from analysis conducted by
vanderReitetal. using data from USDA's
Nationwide Food Consumption Survey
1987-88 and Continuing Survey of
Food Intakes by Individuals 1989-91
and 1994 (39). These surveys collected
dietary data from national samples of
individuals including data on the use
of fat -modified food products and its
impact on the overall diet.
25
The researchers found no evidence of a
tendency to overconsume reduced-fat
foods. These foods were usually eaten
in the same or smaller amounts as their
regular versions. Use of reduced-fat
foods was associated with a decrease in
total fat intake. Nonusers of reduced-fat
foods averaged 35 percent of calories
from fat; those who used one or two
reduced-fat foods averaged 32 percent
calories from fat, and those consuming
three or more reduced-fat foods had
average fat intakes of 26 percent of
total calories (fig. 2).
Saturated fat intake showed a corre­sponding
decline, with nonusers of
reduced-fat foods consuming 12 percent
of calories from saturated fat, compared
with an intake of 8 percent of calories
from saturated fat among those using
three or more reduced-fat products ( 39 ).
These differences in fat and saturated
fat intakes may reflect not only the use
of reduced-fat food products but also a
tendency to choose traditional foods
that are naturally lower in fat and/or to
prepare foods with less fat. Neverthe­less,
these findings do not support the
assertion that consumers of fat-modified
food products may compensate by
increasing fat intake from other foods.
These decreases in fat and saturated fat
intake were consistently found among
population subgroups defined by age
and sex or by racial/ethnic identification.
The associations between use of reduced­fat
foods and total caloric intake were
less consistent. Among the total popula­tion,
users of reduced-fat food products
had total caloric intakes that were slightly
higher than those of nonusers: 1,805
kcal/day for nonusers; 1,860 kcal/day
for users of one to two reduced-fat food
products; and 1,837 kcal/day for those
using three or more reduced-fat food
26
Figure 2. Fat intake by use of reduced-fat food products, 1987-941
Percent
calories
40
30
20
10
0
35 32 26
No reduced·
fat foods
1·2 reduced­fat
foods
3 or more
reduced-fat foods
1AII food and nutrient consumption data are derived from the following surveys conducted by
USDA: 1987-88 Nationwide Food Consumption Survey, Individual Intake Component; 1989-90
CSFII; 1990·91 CSFII; 1991-92 CSFII; and 1994 CSFII.
Source: van der Reitet a/., 1996, Impact of the use of reduced fat foods on nutrient adequacy.
Speech presented at the Conference on Nutritional Implications of Macronutrient Substitutes,
New York Academy of Sciences, Arlington, VA, October 29.
products (fig. 3). But within population
subgroups, the associations were not
consistent. For example, teenage males
who used reduced-fat food products had
higher caloric intakes than those who
did not, but for female teenagers, the
reverse was true (39). It is possible that
different subgroups are using reduced­fat
foods for different purposes-some
to restrict both fat and calories and some
to reduce fat only. It also is possible
that some are simply more successful
in using reduced-fat foods as part of an
overall calorie reduction strategy. More
examination of the motives for using
fat-modified foods and related effects
on caloric intake among specific popula­tion
subgroups is needed.
These data suggest that for many indi­viduals,
use of fat-modified foods may
be an effective strategy for decreasing
the fat content of their diets. Research
by Krista! et al. (23) indicates that it
may also be a strategy that consumers
find relatively easy to adopt and main­tain,
compared with other strategies that
require more basic changes in eating
habits, such as limiting the use of fats
as spreads or flavorings.
Family Economics and Nutrition Review
Figure 3. Calorie intake by use of reduced-fat food products
1987-941 '
Calories/day
2,000
1,500 f..
1,000
500 ~
0
1805
No reduced­fat
foods
1860
I
1-2 reduced­fat
foods
1837
3 or more
reduced-fat foods
1AII food and nutrient consumption data are derived from the following surveys conducted by
USDA: 1987-88 Nationwide Food Consumption Survey, Individual Intake Component; 1989-90
CSFII; 1990-91 CSFII; 1991-92 CSFII; and 1994 CSFII.
Source: van der Reit et a/., 1996, Impact of the use of reduced fat foods on nutrient adequacy.
Speech presented at the Conference on Nutritional Implications of Macronutrient Substitutes,
New York Academy of Sciences, Arlington, VA, October 29.
Effects of Fat-Modified Food Use
on Overall Diet Quality
Some nutritionists are concerned that
use of fat-modified food products as
a primary strategy for decreasing fat
intake may result in diets of lower quality
than if people lowered dietary fat by
making changes in their basic eating
patterns (21,32). Traditionallowfat
diets may be valuable not only because
of their decreased fat content, but also
because an increased intake of fruits,
vegetables, and grains usually accompa­nies
the decrease in fat. Diets rich in
fruits, vegetables, and grains may
1997 Vol. 10 No.4
contain more vitamins, minerals, fiber,
and potentially beneficial phytochemicals
than lowfat diets that consist mainly of
reduced-fat versions of cookies, chips,
and other products (21 ).
Little research has been done on the
overall quality of the diets of people
consuming fat-modified food products.
Van der Reit et al. examined the associa­tion
of consumption of reduced-fat food
products with intakes of two vitamins­A
and E; and two minerals-iron and
zinc (39). The researchers found that
vitamin A and iron intakes were higher
Little research has
been done on the
overall quality of
the diets of people
consuming fat­modified
food
products.
27
among users of reduced-fat food products
in the total population and among sub­populations
identified by age, sex, and
ethnicity. In the total population, users and
nonusers of reduced-fat food products had
similar intakes of vitamin E. Within popula­tion
subgroups, there were differences in
vitamin E intake between users and non­users
that seemed to parallel differences
in caloric intake.
Similarly, zinc intakes of users of
reduced-fat food products were higher
in the total population, but within popu­lation
subgroups, there were differences
that corresponded to differences in
caloric intake ( 39 ). These results do not
demonstrate any negative effects of
inclusion of reduced-fat food products
in the diet. However, examination of
any effects on intakes of other dietary
components, particularly those such
as carotenoids and dietary fiber for
which plant-based foods are sources,
is indicated.
Conclusions and
Recommendations
Fat-modified foods have rapidly become
a significant part of the American food
supply, but their value as a means of
improving the diets of Americans remains
to be established. More research is needed
to verify the nutritional benefits claimed
for these products and/or identify any
negative effects they may have.
Several research studies have been
consistent in finding that consumption
of fat-modified food products may be
helpful for reducing fat intake. Their use
does not seem to reduce calorie intake
among most population segments; there­fore,
their value for weight control is
doubtful. There is no evidence that
regular use of fat-modified food products
results in reduced vitamin or mineral
28
intake or other decreases in diet quality.
The research in this area is so limited,
however, that it is premature to draw
conclusions. Much more research
examining the impact of the use of
fat-modified food products on diet
quality needs to be done.
The growing popularity of fat-modified
food products offers challenges and
opportunities for those engaged in
nutrition education and promotion.
Educational and prcmotional strategies
are needed to help consumers realize
that consumption of fat-modified foods
will not automatically result in either a
lowfat diet or weight reduction (2 1 ).
There is evidence that consumers are
able to assimilate these messages; for
example, consumer beliefs that one can
eat more of foods if they are lowfat have
declined ( 15 ). Where economics is a
concern, it is important that consumers
realize that a lowfat diet can also be
achieved using appropriately chosen
conventional foods, since many fat­modified
foods are more expensive
than their regular counterparts.
Finally, those engaged in nutrition
education and promotion must strive
to develop messages that convince
consumers that eating no more than
30 percent of calories from fat, while
important, is not the only characteristic
of a healthy diet. These messages should
also motivate consumers to consume
a diet with recommended amounts of
fruits, vegetables, and grains. Ideally,
it will be possible to capitalize on the
consumer interest roused by fat-modified
food products and use that interest to
change overall eating habits that con­form
to the recommendations of the
Food Guide Pyramid and the Dietary
Guidelines for Americans.
Family Economics and Nutrition Review
References
1. The American Cancer Society 1996 Dietary Guidelines Advisory Committee.
1996. Guidelines on Diet, Nutrition, and Cancer Prevention: Reducing the Risk of
Cancer with Healthy Food Choices and Physical Activity. American Cancer Society.
2. American Dietetic Association. 1996. Salatrim means less fat...fewercalories ...
more taste! The American Dietetic Association Fact Sheet for Dietetics Professionals.
National Center for Nutrition and Dietetics, Chicago.
3. American Dietetic Association. 1991. ADA reports: Position of The American
Dietetic Association: Fat replacements. Journal of the American Dietetic Association
91(10): 1285-1288.
4. Birch, L.L. and Fisher, J.O. 1997. Food intake regulation in children. Annals of
the New York Academy of Sciences 819:194-220.
5. Birch, L.L., Johnson, S.L., Jones, M.B., and Peters, J.C. 1993. Effects of a
nonenergy fat substitute on children's energy and macronutrient intake. American
Journal of Clinical Nutrition 58:326-333.
6. Calorie Control Council. 1996. Calorie Control Commentary 18(1):1-5.
7. Derby, B.M. and Levy, A.S. 1996. Consumer and Market Impacts of the Nutrition
Labeling and Education Act. Presentation made at the Marketing and Public Policy
Conference, Rosslyn, VA, May 17.
8. FDA Backgrounder. 1995. Olestra and other fat substitutes. BG95-18, November
28 issue.
9. Fensholt, C. 1996, February. Reiterating the 30% rule. Supermarket Business, pp.
37-39.
10. Food and Drug Administration. 1996. FDA Approves Fat Substitute, Olestra.
Press Release. January 24 issue.
11. Food Labeling and Nutrition News. 1996. Snack sales drop. Vol.4 No.41 , p.12.
12. Food Labeling and Nutrition News. 1996. CSPI campaigns in 'Diarrhea Capitals' to
inform public of alleged olestra dangers. Vol.4 No.37, p.12.
13. Food Labeling and Nutrition News. 1997. CSPI takes anti-olestra campaign to
Indianapolis. Vol.5 No.26, pp.7-8.
14. Food Marketing Institute. 1996. Trends in the United States: Consumer Attitudes
and the Supermarket, 1996.
1997 Vol.JONo. 4 29
30
15. Food Marketing Institute and Prevention Magazine. 1996. Shopping for Health
1996: Americans Look for Answers About the Foods They Eat.
16. Frazao, E. and Allshouse, J.E. 1996. Size and Growth of the Nutritionally Improved
Foods Market. U.S. Department of Agriculture, Economic Research Service.
Agriculture Information Bulletin No. 723.
17. Gersh off, S.N. 1995. Nutrition evaluation of dietary fat substitutes. Nutrition
Reviews 53(1 1):305-313.
18. Giese, J. 1996, March. Olestra: Properties, regulatory concerns, and applications.
Food Technology, pp. 130-131.
19. Glueck, C.J., Hastings, M.M., Allen, C., Hogg, E., Baehler, L., Gartside, P.S.,
Phillips, D., Jones, M., Hollfmbach, E.J., Braun, B., and Anastasia, J.V. 1982.
Sucrose polyester and covert caloric dilution. American Journal of Clinical Nutrition
35:1352-1359.
20./nternational Food Information Council Foundation. 1995, November. Uses
and nutritional impact of fat reduction ingredients.
21 . Kantor, M.A. 1996, July-August. Olestra: Questions still unanswered. Journal
of Nutrition Education, pp. 191-192.
22. Krauss, R.M., Deckelbaum, R.J., Ernst, N., Fisher, E., Howard, B.V., Knopp,
R.H., Kotchen, T., Lichtenstein, A.H., McGill, H.C., Pearson, T.A., Prewitt, E.,
Stone, N.J., Van Horn, L., and Weinberg, R. 1996. Dietary Guidelines for healthy
American adults: A statement for health professionals from the Nutrition Committee,
American Heart Association. Circulation 94:1795-1800.
23. Krista!, A.R., White, E., Shattuck, A.L., Curry, S., Anderson, G.L., Fowler, A.,
and Urban, N. 1992. Long-term maintenance of a low-fat diet: Durability of fat­related
dietary habits in the Women's Health Trial. Journal of the American Dietetic
Association 92:553-559.
24. Lyle, B.J., McMahon, K.E., and Kreutler, P.A. 1992. Assessing the potential
dietary impact of replacing dietary fat with other macronutrients. Journal of Nutrition
122:211-216.
25. McGinnis, J.M. and Foege, W.H. 1993. Actual causes of death in the United
States. Journal of the American Medical Association 270( 18):2207-2212.
26. Morrison, R.M. 1994. Fat substitutes in foods: Growing demand and potential
markets. In C. C. Akoh and B.G. Swanson (Eds.) Carbohydrate Polyesters as Fat
Substitutes (pp. 237-249). Marcel Dekker, Inc., New York.
27. National Cattlemen's Association and the Beeflndustry Council. 1996. Press
Release: Statement Regarding the 1995 Dietary Guidelines for Americans. January 2.
Family Economics and Nutrition Review
28. Oster, G. and Thompson, D. 1996. Estimated effects of reducing dietary saturated
fat intake on the incidence and costs of coronary heart disease in the United States.
Journal of the American Dietetic Association 96(2): 127-131.
29. Shide, D.J. and Rolls, B.J. 1995. Information about the fat content of preloads
influences energy intake in healthy women. Journal of the American Dietetic
Association 95(9):993-998.
30. Sloan, A. E. 1996, July. America's appetite '96: The top I 0 trends to watch and
work on. Food Technology, pp. 55-71.
31. Stillings, B.R. 1994. Trends in foods. Nutrition Today 29(5):6-13.
32. Thomas, P.R. 1996, July-August. Olestra: Another technological fix for the food
supply. Journal of Nutrition Education, pp. 193-194.
33. U.S. Congress. 1990. Public Law 101-535: Nutrition Labeling and Education
Act of 1990.21 USC 301.
34. U.S. Department of Agriculture, Agricultural Research Service. 1996. USDA
Develops Tasty No-Cal, High-Fiber Fat Substitute. USDA press release, August 26,
PR-1396.
35. U.S. Department of Agriculture, Agricultural Research Service, Food Survey
Research Group. 1996. Data Tables: Results from USDA's 1995 Continuing Survey
of Food Intakes by Individuals and 1995 Diet and Health Knowledge Survey. On
CSFIIIDHKS 1995 CD-ROM.
36. U.S. Department of Agriculture and U.S. Department of Health and Human
Services. 1995. Nutrition and Your Health: Dietary Guidelines for Americans. (4th
ed.) Home and Garden Bulletin No. 232.
37. U.S. Department of Health and Human Services, Food and Drug Administration.
1995, November. Olestra and other fat substitutes. [On-line]. Available:
http://www .fda.gov ... nders/olestra.html
38. U.S. Department of Health and Human Services, Public Health Service. 1991 .
Healthy People 2000: National Health Promotion and Disease Prevention Objectives.
DHHS Publication No. (PHS) 91-50212.
39. van der Riet, B.E., Heimbach, J.T., and Egan, S.K. 1996. Impact of the use of
reduced fat foods on nutrient adequacy. Speech presented at the Conference on
Nutritional Implications of Macronutrient Substitutes, New York Academy of
Sciences, Arlington, VA, October 29.
40. Welsh, S. and Guthrie, J.F. 1991. Changing American diets. In A. Bendich and
C.E. Butterworth, Jr. (Eds.) Micronutrients in Health and in Disease Prevention
(pp. 381-408). Marcel Dekker, Inc., New York.
1997 Vol. /0 No.4 31
32
Supplement Use May Not
Be Associated With
Better Food Intake in
All Population Groups
David L. Pelletier
Anne Kendall
Division of Nutritional Sciences
Cornell University
This study tests the generalization from earlier studies that food intake of
supplement users is more healthful than that of nonusers and finds that
the association varies across major population groups. Data on adults
from the 1989-91 Continuing Survey of Food Intakes by Individuals (CSFII)
were analyzed by multiple regression to test for (1) confounding by socio­demographic
factors; and (2) whether the strength and direction of the
relationship among supplement use and healthfulness of diet is similar
across population groups defined by ethnicity, age, sex, and nutrition
knowledge and attitudes. The results indicate that sociodemographic factors
do not account for the positive association between supplement use and
nutrient intake from food. However, the presence of statistically significant
interactions between supplement use and ethnicity, age, sex, and nutrition
knowledge, attitudes, and beliefs reveals that the strength and direction of
this association varies across major population groups. This is in contrast
to the generalization that emerged from 13 previous studies. These findings
deserve confirmation in larger national samples and have implications for
nutrition education and regulatory policies concerning nutrient supplements.
or many years, the profes­sional
medical and nutrition
communities in the United
States have taken the posi­tion
that the nutritional requirements for
most nutrients in most people can and
should be met by consuming an appro­priately
balanced diet. A corollary is
that dietary improvement should be the
first line of intervention to prevent or
correct nutritional imbalances when
they do occur, and that the use of nutrient
supplements should not be part of a
broad, population-based strategy for
health promotion and disease prevention.
This view is evident in a joint statement
from the American Dietetics Association,
the American Institute of Nutrition, the
American Society for Clinical Nutrition,
and the National Council Against
Health Fraud (2).
Family Economics and Nutrition Review
In stark contrast to this professional
opinion, the use of nutrient supplements
is quite common in the U.S. population,
and the prevalence increased from
1977-78 to 1986 (1 5 ). The 1987 Health
Interview Survey indicates that 51 per­cent
of the adult population used one or
more supplements in the past year and
23 percent did so daily (24 ). Whites,
women, and older adults are more likely
to consume supplements regularly than
are Blacks, Hispanics, men, and younger
adults.
The National Research Council report
on diet and health noted that physicians
and pharmacists have a strong influence
on patient decisions to use supplements
in addition to the effects of marketing,
the media, and other sources of informa­tion
( 15 ). The influence of industry
marketing may increase in the future
as a result of the Dietary Supplement
Health and Education Act enacted in
October 1994, which requires a much
more permissive regulatory policy
than that sought by the Food and Drug
Administration ( 17). The proposal to
permit the use of food stamps for pur­chasing
supplements (H.R. 236, 1995),
although defeated, could have further
increased supplement use among low­income
households.
One area of disagreement about the
appropriate role of supplements and food­based
approaches involves the relation­ship
between so-called antioxidant
nutrients and chronic disease. Although
chronic disease morbidity and mortality
show strong and consistent associations
with the consumption of fruits and
vegetables, they have inconsistent
associations with antioxidant intake
from diet and/or supplements ( 1, 7,15,16).
1997 Vol.10No. 4
This has led to the suggestion that the
preventive effects of fruits and vegetables
may relate to other constituents of food
and that serum antioxidant levels may
be simply markers of a diet that is
habitually high in these foods ( 12). If
valid, this hypothesis would add further
support for a food-based approach and
call into question the advisability of
using a supplement-based approach.
These findings raise the question of
the extent to which the public is using
supplements as a substitute for healthful
diets or as a complement to such diets.
Most studies suggest that the nutrient
intake from food is either similar to or
higher among supplement users as com­pared
with nonusers (4,6,10,11,22,25).
These same studies have established
that supplement use is associated with a
variety of socioeconomic and attitudinal
characteristics that may confound, modify,
and/or mediate that relationship, but
most have not taken these factors into
account in the analysis.
The purpose of the present study is:
(1) to examine the association between
supplement use and dietary intake after
accounting for socioeconomic factors;
(2) to examine the consistency of this
finding across various sociodemographic
groups; and (3) to examine the consis­tency
of these findings across groups
that vary in nutrition knowledge, attitudes,
and beliefs.
Methods
This study uses data from the 1989-91
Continuing Survey of Food Intakes of
Individuals (CSFII). The CSFII is a
multistaged, stratified probability sample,
representative of the 48 contiguous
States and is intended to be representative
of the U.S. population in these States.
The present study is limited to adults
(20 years or older) of both sexes with
complete sociodemographic and dietary
data, of which 8,865 were included in
the survey. A sample of7,361 individuals
is used in the sociodemographic analysis,
representing those with nonmissing and
nonextreme values (i.e., within three
standard deviations of the mean) for the
dietary and socioeconomic variables. A
further subsample, consisting of 2,895
females, is used in analyses involving
nutrition knowledge, attitudes, and
beliefs (KAB) variables. This is because
the KAB module was administered to
only the main meal planner/preparer
in the CSFII, 83 percent of whom are
female in this sample.
The dietary data used here are based on
a single 24-hour recall for each subject.
Total nutrient intake from food and
beverages for seven key nutrients is
examined-total fat and saturated fat,
dietary fiber, vitamin A (retinol equiva­lents),
vitamin C, calcium, and iron.
All nutrients are expressed in relation
to energy intake on the day of the recall
to form nutrient density indices that are
independent of energy intake and, thus,
more indicative of diet quality. These
indices were examined for skewness
prior to analysis and transformed as
recommended elsewhere ( 14). Total fat
was found to be normal in its natural
scale, a square root transformation was
applied to saturated fat, and a natural
log transformation was applied to the
other five nutrients.
33
Each of the seven nutrients is examined
separately, and they are also combined
into an overall dietary score. The dietary
score is calculated for each individual as
the average Z-score of the seven trans­formed
nutrients, using the sample of
7,361 individuals to estimate the internal
means and standard deviations. The sign
of the Z-score was reversed for total and
saturated fat prior to summing across all
nutrients, so that the average Z-score
would have a consistent interpretation
(i.e., positive or higher Z-score values
indicate relatively nutrient dense or
healthful diets). This dietary score
is considered important in this context
because it more closely reflects the
overall quality of the diet (with respect
to these seven nutrients) than any one
nutrient by itself.
Socioeconomic and demographic
(hereafter termed "sociodemographic")
variables include age, sex, ethnicity,
household income, education, and
employment status. The categories
defined for each of these variables are
shown in table 1 along with the sample
sizes and percentage of individuals report­ing
supplement use in each category.
Three ethnic groups are defined:
non-Hispanic Whites ("Whites"),
non-Hispanic Blacks ("Blacks"),
and anyone reporting Hispanic origin
("Hispanic").
Ten indices of nutrition knowledge,
attitudes, and beliefs were created from
the KAB module and used in preliminary
analyses. Four of these were chosen for
further analysis, based on the consistency
of their statistical interaction with sup­plement
use and nutrient intake. These
four relate to: (1) knowledge of diet
and heart disease relationships (binary
responses to six questions); (2) agree­ment
with the statement that, "There are
34
so many recommendations, it is hard to
know what to believe" (six-point Likert
scale); (3) agreement with the statement
that, "Things I eat are healthy, so there
is no need to change" (six-point Likert
scale); and (4) responses indicating an
attitude that many dietary recommenda­tions
are not important (six-point Likert
scale for each of 12 questions). The
total score for each of these four items
was converted to a binary variable for
analysis, in each case using the median
score for the total sample as the cutoff
point.
Users of supplements are defined based
on the question, "How often, if at all,
do you take a vitamin or mineral supple­ment?"
Users are defined as those
reporting the use of any type of supple­ment
"every day" or "every so often,"
and nonusers are defined as those
reporting "not at all."
The relationships among dietary intake,
supplement use, sociodemographic vari­ables,
and KAB variables are examined
using multiple regression analysis.
Three sets of analyses were conducted:
(1) main effects models were used to
test whether the positive association
between supplement use and dietary
intake can be accounted for by socio­demographic
variables. Each nutrient
(and the dietary score) is used as a
dependent variable in its own model,
and the statistical effect of supplement
use is observed before and after adjusting
for the sociodemographic variables as a
set (i.e., income, education, occupation,
age, sex, and ethnicity); (2) interaction
models were used to test whether the
strength or direction of the association
is uniform across age, sex, and ethnic
groups while controlling for income,
education, and occupation. This is done
by testing the significance of an entire
block of interactions between supplement
use and age, sex, and ethnicity after
controlling for the above-mentioned
variables; (3) in similar fashion, interac­tion
analysis is used to test whether the
strength or direction of the association
varies according to individual nutrition
knowledge, attitudes, and beliefs. In
this case, a set of interactions between
supplement use, KAB variables, and
ethnicity is tested. This third set of
analyses was conducted separately for
younger (20-49 years) and older (50 years
and older) females. This was done in
order to simplify the interpretation of
interaction terms.
The statistical methods described above
are designed to permit a valid test of
the null hypothesis that the strength or
direction of the association between
supplement and nutrient intake from
food is uniform across groups defined
by sociodemographic or KAB variables.
In this study, such a test is obtained by
comparing the proportion of variance
explained by the full model (containing
all interaction terms involving the
"supplement use" variable) versus the
reduced model (containing no such
interaction terms). Since the array of
model coefficients is difficult to inter­pret
in the presence of higher order
interaction terms, graphs are used to
gauge the direction and magnitude
of the differences in nutrient intake
implied by these models.
Preliminary analyses were conducted
using SAS ( 18) and proportionate
sample weights. These analyses were
confirmed using SUDAAN to account
for the effects of the complex sample
design on variance estimates, and all
results presented here are derived from
SUDAAN ( 19).
Family Economics and Nutrition Review
Results
Table 1, p. 36 is presented for descriptive
purposes, in order to understand the
variation in supplement use across the
various sociodemographic and KAB
variables used in later regressions.
Supplements are reportedly used by
roughly 33 to 43 percent of U.S. adults
in this sample. Usage is higher among
Whites than non-Whites and among
females compared with males. Usage
is higher among older adults. Usage is
also higher among those with incomes
greater than 170 percent of poverty,
among those with 12 or more years of
education, and among the employed.
Usage is positively associated with
greater knowledge about diet-heart
relationships, with a belief that one's
diet does not need to change, and with
the belief that nutrition is unimportant.
In agreement with previous studies,
table 2, p. 37, reveals that supplement
users have slightly lower mean fat and
saturated fat densities and higher mean
densities of all other nutrients (fiber,
vitamins A and C, calcium, and iron).
When averaged across all seven nutrients,
as seen in the mean diet score, the average
diets of users are superior to those of
nonusers. As shown, controlling for
sociodemographic variables diminishes
the differences in mean density between
users and nonusers for most nutrients and
for the diet score, but all of the differences
remain statistically significant (or
nearly so in the case of total fat, with
p=0.054).
The test of uniformity of these results
across major population groups is con­tained
in table 3, p. 38. This test takes
the form ofF-statistics for the R2 im­provement
associated with the total
block of 11 interactions, as shown in
the bottom row. The table also shows
1997 Vol. 10 No.4
the regression coefficients for all vari­ables
in the models (which are used in
the next section to calculate predicted
means). The F-statistic reveals that the
strength and/or direction of the associa­tion
between supplement use and nutri­ent
density is not uniform across all
population groups. The interaction
reaches statistically significant levels
for fat {p<O.OOI), saturated fat (p<O.OOl),
fiber (p<0.001), vitamin A (p<0.001),
vitamin C (p<O.OOI), iron (p<O.Ol), and
the diet score (p<O.OOI ). The only equa­tion
in which the block of interactions is
not significant is that for calcium.
Having determined that the association
between supplement use and nutrient
density does vary significantly across
major population groups, a series of
predicted means was estimated from the
regression equations in order to assist
interpretation. Inspection of these pre­dicted
means reveals that supplement
use is indeed associated with more
healthful nutrient profiles in some groups.
However, these predicted means also
reveal that supplement use is associated
with little or no differences in nutrient
profiles in other population groups. The
example of middle-aged females and
males is illustrated in figure 1, p. 39.
When the dietary means for supplement
users are expressed as a percentage of
that for nonusers (to standardize the
comparison across nutrients), figure 1
shows that female Hispanic users have
14 percent lower total fat density, 22
percent lower saturated fat density, and
higher densities for all other nutrients
(ranging from 19 to 86 percent higher)
than their nonusing counterparts. Thus,
the mean diet score is .38 Z-scores
higher for users compared with nonusers.
However, this is in contrast to the pattern
in several other groups. For instance,
the figure shows that female White and
... the association
between supplement
use and nutrient
density does vary
significantly across
major population
groups ...
35
Table 1. Supplement use stratified by sociodemographic and knowledge, attitudes, and beliefs variables,
U.S. population, 1989-91 CSFII
Both sexes Men Women
(N = 7,361) (N = 3,079) (N = 4,282)
Variable Percent1 N Percent N Percent N
Ethnicity
White (R)2 43.3 5,878 33.6 2,548 52.2 3,330
Black 34.7 865 29.0 275 39.1 590
Hispanic 32.5 618 22.2 256 43.0 362
Age (years)
20-49 (R) 38.1 4,329 28.8 1,891 47.3 2,438
50 and older 47.8 3,032 39.2 1,188 54.5 1,844
Income (Percent of poverty)
<100(R) 29.6 1,807 21.1 590 39.1 1,217
100- 170 31.5 1,696 18.0 681 41.2 1,015
>170 44.7 3,858 35.2 1,808 54.2 2,050
Education
Less than high school (R) 30.8 2,176 21.7 865 38.4 1,311
High school or some college 43.1 4,015 33.2 1,624 51.2 2,391
College degree or more 46.2 1,170 36.9 590 57.2 580
Occupation
Unemployed (R) 39.3 3,911 31.1 2,038 49.0 1,873
Employed 45.7 3,450 35.3 1,041 51.4 2,409
Both sexes Men Women
(N = 3,489)
Heart health knowledge3
(N = 594) (N = 2,895)
<Median (R) 42.9 1,309 25.6 241 47.5 1,068
>Median 50.2 2,180 37.8 353 53.1 1,827
Too many recommendations
Weak/no agreement (R) 50.4 1,293 33.9 207 54.2 1,086
Strong agreement 46.3 2,196 33.7 387 49.4 1,809
No need to change
Weak/no agreement (R) 43.2 1,412 30.1 230 46.3 1,182
Strong agreement 51.2 2,077 36.5 364 55.2 1,713
Nutrition is unimportant
Weak/no agreement (R) 44.2 1,624 33.4 348 47.7 1,276
Strong agreement 51.3 1,865 34.4 246 54.2 1,619
1The percentage using supplements reflects weighting; the sample size refers to the actual number of people with complete data as used in this analysis.
2(R) indicates the reference group in the regression analysis.
3Knowledge of diet-heart disease relationships.
36 Family Economics and Nutrition Review
Table 2. Nutrient intakes of supplement users and nonusers and coefficients for supplement use in multiple
regression models
Supplement coefficient
Mean Unadjusted Adjusted1
Users Nonusers
Nutrient (N = 2,844) (N = 4,517) B SE (B) Probability B SE (B) Probability
Fat(% kcals) 33.6 34.5 -.724 .216 .0008 -.423 .220 .054
Saturated fat (% kcals)2 9.2 11.4 - .070 .014 .0001 - .051 .014" .0004
Fiber (g/1 ,000 kcald 7.5 7.0 .082 .012 .0001 .057 .012 .0001
Vitamin A (RE/1 ,000 kcals)3 455 372 .196 .020 .0001 .141 .020 .0001
Vitamin C (mg/1 ,000 kcals)3 41.3 34.8 .169 .022 .0001 .117 .022 .0001
Calcium (mg/1,000 kcals)3 376 354 .066 .012 .0001 .042 .012 .0007
Iron (mg/ 1,000 kcals)3 7.2 7.0 .043 .010 .0001 .029 .010 .0036
Diet score (Z-score) +.08 -.06 .143 .014 .0001 .098 .014 .0001
1 Models include income, education, employment status, age, sex, and ethnicity in addition to a dummy variable indicating supplement use.
2Square root transformation applied in regressions; geometric means are shown to facilitate interpretation.
3Naturallog transformation applied in regressions; geometric means are shown to facilitate interpretation.
Black supplement users have nutrient
densities that are more similar to those
of their nonuser counterparts. Thus, while
the mean diet score for users is 13 per­cent
higher for Whites, it is only 1 per­cent
higher for Blacks.
In a second comparison, the figure shows
that the more healthful iliet ofHisparuc
females is not rephcated among Hispanic
males (the mean iliet score is .02 Z-scores
higher for users compared with nonusers).
A third example of subpopulation differ­ences
is that supplement use among
Black females has little or no associa­tion
with nutrient density (mean diet
score=.01 Z-scores higher than nonusers).
However, among Black males, supple­ment
use is associated with 10 to 12 per­cent
lower fat and saturated fat densities
and 9 to 54 percent higher densities for
the other nutrients, such that the mean
1997 Vol. 10 No.4
diet score is .37 Z-scores higher among
users compared with nonusers.
The magnitude of these differences is
not estimated with precision in these
analyses because of small sample sizes.
When combined with the significant F
statistics from table 3, however, they are
sufficient to indicate that significant
variation does exist in the strength and/
or direction of the association between
supplement use and nutrient density
across distinct population groups.
Building upon evidence from an earlier
study (25), it was hypothesized that the
above findings may reflect differences
in the underlying motivations for using
supplements and associated knowledge,
attitudes, and beliefs concerning food,
diet, and health. Table 4, p. 40, presents
the F-values for the block of interactions
between supplement use, ethnicity, and
the four KAB variables (each KAB vari­able
in its own model), after controlling
for sociodemographic variables. This
table contains the results for females
only because the KAB module was
administered to the main meal planners,
most of whom are female. Separate
analyses were conducted for younger
(20-49 years) and older (50 years and
older) females.
The F-statistics in table 4 reveal that the
association between supplement use and
many nutrient densities is significantly
dependent upon measures of diet-heart
knowledge, confusion about dietary
recommendations, beliefs about the need
to change diet, and attitudes about the
importance of food and diet to health.
These interactions are especially strong
and consistent for the two KAB variables
pertaining to " too many recommenda­tions"
and "no need to change my diet"
37
Table 3. Regression coefficients for main effects and interaction terms involving supplement use with ethnicity,
age, and sex (N = 7,361)
Nutrient
Saturated Diet
Variable Fat fat Fiber Vitamin A VitaminC Calcium Iron score
Main effects1
Intercept 35.442 3.521 1.841 5.823 3.331 5.972 1.914 -.135
Supplement users -.675 -.077 .113 .117 .137 .011 .0176 .111
Black -.368 -.108 -.071 -.200 .018 -.244 -.059 -.107
Hispanic -.519 -.087 .007 -.069 .067 -.151 -.044 -.027
Age50 -1.837 - .169 .199 .230 .339 .008 .079 .246
Males .933 .053 .010 -.057 - .091 -.058 -.017 -.069
Two-way interactions
Supp x Black 2.190 .139 - .234 -.001 .109 .044 -.028 -.117
Supp x Hispanic -4.050 - .3 13 .101 .087 .383 .059 -.016 .250
Supp x Age50 .151 .032 -.058 .101 -.011 .031 .004 -.002
Supp x Males -.329 .028 -.082 -.218 -.148 .028 -.006 -.079
Three-way interactions
Supp x Black x Age50 .353 .024 .303 .474 -.103 .012 .169 .203
Supp x His x Age50 8.835 .695 - .267 -.058 -.556 .201 -.100 -.455
Supp x Black x Male -4.945 -.309 .309 .440 .293 .028 .245 .451
Supp x His x Male 7.800 .333 -.425 .296 -.415 .093 .105 -.277
Supp x Age50 x Male 1.756 .070 .017 .233 -.066 -.051 -.008 -.026
Four-way interactions
Supp x Black x Age50 x Male 4.434 .266 -.384 -1.271 -.353 .050 -.450 -.653
Supp x His x Age50 x Male -10.664 -.635 .498 -.452 .468 -.290 - .015 .373
F Statistic for interactions2 5.569 5.667 6.160 6.562 3.969 1.243 2.439 5.166
Probability <0.001 <0.001 <0.001 <0.001 <0.001 NS <0.01 <0.001
1 Models also include income, education, and employment status.
2F values~ 1.95 are significant at p < 0.05.
Abbreviations: Supp =supplement users; His= Hispanic; Age50 =50 years and older.
38 Family Economics and Nutrition Review
Figure 1. Percent difference in mean nutrient intake between supplement users and nonusers 1
by ethnicity
Females 20-49 years, by ethnicity
Percent
100
B6
80 77
60
3B
-22
-40
Fat Saturated fat Fiber Vitamin A Vitamin C Calcium Iron Z-score x 100
Nutrient
Males 20-49 years, by ethnicity
Percent
100
80
60 54
40
20
0
-20 -12
-25
-40
Fat Saturated fat Fiber Vitamin A Vitamin C Calcium Iron Z-score x 100
Nutrient
1Adjusted for income, education, and employment.
1997 VoL 10 No_ 4
D White
• Black
D Hispanic
39
Table 4. F-Values for interaction tenns involving supplement use with ethnicity and knowledge, attitudes, and
beliefs variables1
Nutrient
Saturated Diet
Variable Fat fat Fiber Vitamin A Vitamin C Calcium Iron score
Females 20 - 49 years
Heart health knowledge2 6.13*** 4.14** 4.17** 2.06 0.30 5.75*** 2.21 2.18
Too many recomrnendations3 4.92*** 3.63** 3.23* 3.67** 3.80** 8.58*** 2.19 5.34**
No need to change4 8.50*** 4.79*** 1.95 2.96* 2.36* 7.95*** 1.66 6.87**
Attitudes about nutritions 4.16** 4.15** 1.24 3.63** 5.24*** 6.55*** 0.88 3.00*
Females 50 years and older
Heart health knowledge2 0.69 0.65 1.69 3.73** 1.73 0.89 3.13* 1.80
Too many recomrnendations3 1.84 0.88 2.70* 4.37** 1.00 0.74 3.34** 2.56*
No need to change4 2.57* 1.74 2.34 3.98** 1.48 1.98 3.29* 2.53*
Attitudes about nutritions 0.62 1.08 3.02* 4.69** 1.37 1.19 4.70** 2.45*
1 Models also include income, education, employment status, ethnicity, age, and supplement use. The models also include the two-way interactions between
supplement use and age, ethnicity, and the KAB variable, as well as the three-way interaction of supp x ethnicity x KAB. The F-values shown here correspond
to the R2 improvement for the entire block of interactions.
2Knowledge of diet-heart disease relationships.
3Degree of agreement with "There are too many dietary recommendations and it is hard to know what to believe."
4Degree of agreement with "Things I eat are healthy so there is no need to change."
5Degree to which respondent feels that each of 12 nutrition recommendations is important(salt, fat, saturated fat, cholesterol, five-a-day, sugar, alcohol, fiber,
starch, variety, desirable weight, 6 grain servings per day).
* p < O.OS; ** p <0.01; *** p < 0.001
and are statistically significant (p<0.05)
in 44 percent of the comparisons for
older women (14 out of 32 comparisons)
and 72 percent of comparisons for younger
women (23 out of32 comparisons).
These results provide strong evidence
that the strength and/or direction of the
association between supplement use and
nutrient densities in food is dependent
upon an individual's knowledge, attitudes,
and beliefs about diet and health.
To illustrate the implications of these
interaction models, figure 2 shows
the predicted means pertaining to the
"no need to change" model for younger
females from the three ethnic groups.
This case was chosen for illustration
because it has the strongest interaction
40
in table 4 (F=6.87 for the mean diet
score). The figure shows that among
women 20-49 years old who feel that
their diet needs to change, supplement
use is associated with nutrient
profiles that are similar to or more
healthful than those for nonusers. The
most healthful profiles are seen among
Hispanic supplement users, who have
30 percent lower fat and saturated fat
intakes and substantially higher vitamin
A (93 percent) and calcium (81 percent)
intakes compared with nonusers, with a
mean diet score that is .68 Z-scores
higher than nonusers. Similar healthful
patterns are seen among Black and White
women, but the differences between
users and nonusers are smaller than
those seen among Hispanic women.
Among young women who feel that their
diet does not need to change Oower panel),
the healthful nutrient profile for Hispanics
is replicated (though less marked), but
the nutrient profile for Blacks is reversed.
Specifically, in this comparison, Black
supplement users have 9 and 12 percent
higher fat and saturated fat densities and
5 to 26 percent lower densities for fiber,
vitamins A and C, calcium, and iron,
compared with nonusers. The result is
a mean diet score that is .24 Z-scores
lower among users than nonusers.
Among Black women who feel that
their diet does need to change (upper
panel), the mean diet score of supple­ment
users is .33 Z-scores higher than
that of nonusers.
Family Economics and Nutrition Review
Figure 2. Percent difference in mean nutrient intake between supplement users and nonusers 1
by no need to change model
Females 20-49 years, need diet change
Percent
100 93
80
60
40
20
0
-20
-4