Journal of Public Health | Vol. 32, No. 3, pp. 379 –386 | doi:10.1093/pubmed/fdp117 | Advance Access Publication 7 January 2010
Discretionary calorie intake a priority for obesity prevention:
results of rapid participatory approaches in low-income US
communities
Deborah A. Cohen1, Roland Sturm1, Marielena Lara1, Marylou Gilbert1, Scott Gee2
1
RAND Corporation, Santa Monica, 1776 Main St, Santa Monica, CA 90407, USA
Kaiser Permanente, 1950 Franklin Street, 13th floor, Northern California, Oakland, CA 94612, USA
Address correspondence to Deborah A. Cohen, E-mail: [email protected]
2
A B S T R AC T
Background Since resources are limited, selecting the most promising targets for obesity interventions is critical. We examined the relative
associations of physical activity, fruit and vegetable consumption and ‘junk food’ consumption with BMI and the prevalence of relevant policies
in school, work, food outlets and health-care settings.
Methods We conducted intercept surveys in three low-income, high-minority California communities to assess fruit, vegetable, candy, cookie,
salty snacks and sugar-sweetened beverage consumption and self-reported height, weight and physical activity. We also assessed relevant
policies in selected worksites, schools and health-care settings through key informant interviews.
Results Data were collected from 1826 respondents, 21 schools, 40 worksites, 14 health-care settings and 29 food outlets. The average
intake of salty snacks, candy, cookies and sugar-sweetened beverages was estimated at 2226 kJ (532 kcal) daily, 88% higher than the US
Department of Agriculture/Department of Health and Human Services guidelines recommend. Energy from these sources was more strongly
related to BMI than reported physical activity, fruit or vegetable consumption. Policies to promote healthy eating and physical activity were
limited in worksites. Fruits and vegetables were less salient than junk food in community food outlets.
Conclusion Targeting consumption of salty snacks, candy cookies and sugar-sweetened beverages appeared more promising than alternative
approaches.
Keywords community interventions, discretionary food, nutrition, obesity
Introduction
Many communities are eager to take steps to halt the
obesity epidemic. The causes of population weight gain are
multi-factorial and reversing the obesity epidemic calls for
changes in individual behavior, social norms, physical
environments, public policy and organizational practice.
There are countless potential interventions, for example: fix
sidewalks, make streets safer, increase physical activity
opportunities, add more sports clubs and after school programs, build walking trails, increase the availability of fruits
and vegetables, have merchants place fruits and vegetables
in the front of the store, and change the options in vending
machines. The scope can be overwhelming for communities.
Diluted efforts and messages may contribute to the limited
effect of recent community-level interventions.1 How could
community-based organizations prioritize among all possible
strategies?
Deborah A. Cohen, Senior Natural Scientist
Roland Sturm, Senior Economist
Marielena Lara, Natural Scientist
Marylou Gilbert, Project Manager
Scott Gee, Medical Director of Prevention and Health Information, KP Northern
California
# The Author 2010, Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved.
379
380
J O U RN A L O F P U B L I C H E A LTH
Rapid assessment methodologies were developed more
than two decades ago and have been widely implemented in
several fields.2 Their goal is to quickly identify targets for
interventions that have a high probability of being adopted
and sustained at the local level. Aside from identifying
promising targets, rapid assessments could also provide the
information necessary to strengthen the saliency and provide
political support for change. Rapid assessments typically
include community participation and are pragmatic, streamlined and applied.3
In order to inform community-led efforts in three communities in Northern California, we used rapid assessment
approaches to find out how healthy diet and physical activity
are supported or inhibited as well as the role of discretionary
foods, high in sugar and fat, but low in essential nutrients.
To prevent further increases in obesity our goal was to
identify the most salient targets for community interventions. Before we started the rapid assessment, several very
different targets were already under debate. One approach
was to increase fruit and vegetable consumption, a laudable
goal because Americans eat fewer servings of fruits and vegetables than recommended. Increasing fruit and vegetable
consumption has the potential to control obesity, but only if
these items were to replace the energy dense snacks and
high-calorie beverages that have been implicated in weight
gain.4,5 Another approach would more aggressively focus on
reducing the consumption of discretionary foods directly.
Yet a third one is to encourage physical activity, which the
food industry supports,6 since if people burn more energy,
they need to eat more food to maintain weight. Hill estimates that increasing energy expenditure by 418 kJ
(100 kcal)/day would be sufficient to prevent obesity,7 as
long as food consumption does not also increase.
Understanding the role of discretionary foods and their
relationship to fruit and vegetable consumption, physical
activity, and dieting is important to developing effective
interventions. A new concept that was developed for the
latest U.S. Dietary Guidelines is ‘discretionary calories’.8 The
guidelines calculate the amount of energy available after
people have satisfied the guidelines to meet essential nutrients through the consumption of fruits, vegetables, grains,
meats or legumes and dairy product equivalents. For the
average person who needs a daily 8,370 kJ (2000 kcal) diet,
the US Department of Agriculture/Department of Health
and Human Services (USDA/DHHS) guidance recommends that energy from discretionary foods should not
exceed 1117 kJ (267 kcal) to achieve energy balance while
satisfying adequate nutrient intake.8 Discretionary foods
include cookies, candy, salty snacks and sodas that are high
in energy, but low in nutrients.
Our goal was to use multi-faceted rapid assessment
approaches to find out how a healthy diet and physical
activity are supported or inhibited as well as the role of discretionary foods in three communities in Northern
California.
Methods
The data come from a project to improve healthy eating and
active living in three low-income, high-minority communities. Three community-based organizations formed collaboratives with members from multiple organizations
concerned about obesity. In all three sites, staff were trained
to recruit subjects over 18 years and conducted anonymous
intercept interviews (stopping people on the streets) in targeted neighborhoods, in English or Spanish, lasting 5 –
10 min. In the spring of 2007, people were surveyed in
several venues, retail outlets, health-care clinics, worksites
and their homes. Locations were based upon where future
interventions to promote healthy eating and active living
would be implemented. Respondents had to either live,
work or receive medical care in the target area (total n ¼
1826; Site 1 n ¼ 547; Site 2 n ¼ 959, Site 3 n ¼ 320). All
methods were approved by the RAND Human Subjects
Protection Committee.
For fruit/vegetable consumption, the brief survey instrument included the same items as the Center for Disease
Control’s Behavioral Risk Factor Surveillance Study
(BRFSS).9 We asked whether the respondent is ‘eating fewer
calories to lose weight or to keep from gaining weight’,
another BRFSS question.10 The survey also asked about
consumption of four types of snack foods: salty snacks,
cookies, candy and sugar-sweetened beverages, all four of
which we abbreviate as SCCSs. One way to minimize the
bias is to have short recalls, so the dietary questions asked
about the past 24 h only.11,12
The 2005 US Dietary Guidelines provides a quantitative
target: ‘discretionary calories’8 addressing both energy
balance and the need for essential nutrients and depends
upon age, physical activity levels and diet quality. People
who exceed recommended discretionary calories will either
have an energy imbalance and/or be at risk for malnutrition.
Adjusting for age, gender, physical activity and BMI, we calculated each respondents’ recommended daily discretionary
calories.13 While some of the items contain ingredients that
are not automatically considered discretionary, for example
the nuts in a candy bar, we counted all the energy as an
indicator of diet quality.
A serving of salty snacks was defined as a ‘handful’
(about 28 gm), a serving of cookies as about 3 average
D I S C R E T I O N A RY CA LO R I E S A P R I O R I T Y FO R O B ES I T Y P R E V EN T ION
cookies (28 gm), a serving of candy as equivalent to a
medium-sized Snickers bar (about 42 gm) and a serving of a
sugared-sweetened beverage (including sodas) as a 355 ml
(12 oz) can. We imputed energy intake assuming that a
serving of salty snacks averages approximately 586 kJ
(140 kcal), a serving of cookies 586 kJ (140 kcal), a serving
of candy 837 kJ (200 kcal) and a 355 ml (12 oz.) can of
sugared-sweetened drinks 628 kJ (150 kcal) based upon the
USDA National Nutrient database and product nutrient
labels.14
For physical activity, people reported the frequency of
moderate and vigorous physical activity (MVPA) as well as
the minutes per session using these items based on the
International Physical Activity Questionnaire.15 Moderate
activity was defined as brisk walking, bicycling, vacuuming,
gardening, or anything else that causes some increase in
breathing or heart rate and vigorous activity was defined as
running, aerobics, heavy yard work or anything else that
causes large increases in breathing or heart rate. In contrast
to the diet items, the physical activity questions about a
‘usual’ week. However, validation studies suggest that these
kinds of questions can distinguish between broad categories
of sedentary and active individuals.16 By multiplying the
number of days per week and the minutes per day we estimated weekly minutes of MVPA. We categorized minutes of
MVPA into three categories, sedentary, if total weekly
minutes were less than 150, moderately active, if weekly
minutes were greater than 150 and less than 300 and active,
if they exceeded 300 min/week. Additional variables
included sex, age, race/ethnicity and we asked respondents
to self-report their height and weight, which were used to
calculate the body mass index (BMI). The surveys did not
contain items about individual income or education.
The community organizations also conducted assessments of healthy eating and active living policies and practices in 21 schools, 40 worksites and 14 health-care settings.
Members of the community collaborative interviewed
knowledgeable staff at each institution using an interview
guideline prepared by a working group with members from
all three sites. Two of the three communities observed a
total of 29 food outlets and documented the presence,
location and price of fruits, vegetables and selected items
high in energy from discretionary foods.
Data analysis
The main analytic approaches are descriptive statistics, stratified by subpopulation. We also used multivariate linear
regression of caloric intake from SCCSs to isolate the
unique contribution of particular socio-demographic
381
characteristics, as well as whether attempts at dieting or
higher consumption of fruits/vegetables are associated with
consuming less energy from SCCSs. Our second set of
multivariate analyses examines the contribution of health
behaviors and individual characteristics to reported BMI.
Only surveys with complete responses were included in this
analysis. Because sites are quite different in terms of resident
composition, we included site indicator variables in the
regressions.
In order to get a general picture of how worksites,
schools and health-care settings are addressing the obesity
epidemic, we aggregated the responses across the three communities and calculated the percentage of sites with and
without selected health eating, active living policies.
Results
Sample characteristics
The three communities targeted served between 39 800 and
58 000 individuals; between 12.3 and 19.3% of households
were classified as having incomes below the federal poverty
level. Altogether 1826 individuals from the three communities provided responses to the rapid intercept surveys with
complete information. Of this group 41.4% were Hispanic,
32.1% were African-American, 16.1% were White and 9.1%
were Asian. Their average age was 39.4 years and 59.9% of
respondents were female (Table 1). Compared with the local
population as measured in the US census, our sample had a
higher proportion of African-Americans and Hispanic
respondents. However, the intervention planned is intended
to reach minorities more than non-minority groups.
Healthy eating and active living behaviors
The average reported consumption of fruits and vegetables
is 4.2 or 16% less than 5-a-day; 38.7% of respondents
reported eating at least five servings of fruits and vegetables
in the past day. Across the three sites, 54.4% reported
obtaining at least 150 min/week of moderate to vigorous
physical activity and 46% reported they were eating less to
control their weight. The average servings consumed in the
last 24 h was 0.85 servings of salty snacks, 0.83 servings of
cookies, 0.64 servings of candy and 1.2 servings (12 oz
equivalents) of sugar-sweetened beverages. This converts to
an average 1 day intake of 2226 kJ (532 kcal) (95% CI:
2088, 2360), a level of consumption that is on average 1.88
times (95% CI: 1.74, 2.01) higher than the Dietary
Guidelines recommend from discretionary sources (Table 1).
This average discrepancy is more than 5 times larger than
the discrepancy in fruit/vegetable consumption (88% too
382
J O U RN A L O F P U B L I C H E A LTH
Table 1 Descriptive statistics by study site
Site 1, n ¼ 547
Site 2, n ¼ 959
Site 3, n ¼ 320
All sites, n ¼ 1826
Neighborhood SESa
Population living in target area
57 957
39 838
45 843
Median household income
$43 554
$37 213
$52 926
Percent households with income ,$15 000
16.9
19.3
12.3
Mean age (adults 18– 65 only) (SD)
44.0 (15.9)
37.0 (16.5)
38.4(11.6)
Percent Female
58.3
56.6
70.0
59.9
Percent African-American
56.4
30.6
1.2
32.1
Sociodemographics
39.4 (15.7)
Percent Hispanic
37.6
33.3
64.9
41.4
Percent non-Hispanic White Only
5.3
18.3
27.5
16.1
Percent Asian
4.0
14.9
4/7
9.1
Healthy eating, active living behaviors
Average number servings of fruit (SD)
2.2 (1.8)
2.0 (1.6)
2.4 (1.6)
2.1 (1.7)
Average number servings vegetables (SD)
2.4 (1.8)
1.8 (1.4)
2.3 (1.6)
2.1 (1.6)
Average minutes of weekly physical activity (SD)
428 (597)
261 (420)
287 (364)
315 (475)
Percent 150þ minutes of weekly physical activity
62.3
48.2
59.2
54.4
Percent 5þ servings of fruit/vegetables
38.6
36.0
45.3
38.7
Percent eating less/dieting
48
41
56
46
Energy from salty snacks, candy, cookies and
2544 (3448) [608]
2238 (2866) [535]
1632 (1929) [390]
2226 (2933) [532]
2.12 (3.43)
1.89 (2.69)
1.39 (1.62)
1.88 (2.78)
55.3
54.3
48
54
BMI (SD)
28.1 (6.3)
26.4 (5.7)
27.6 (5.8)
27.1 (6.0)
Percent obese
30.0
17.5
27.8
23.0
sugar-sweetened beverages, kJ (SD) [kcal]
Mean ratio of energy consumed to recommended
energy from discretionary sources (SD)
Percent exceeding recommended energy from
discretionary foods
Related health outcomes
Note: (standard deviations) for continuous variables are represented in parentheses. n is the number of respondents with complete data on food items.
a
Source: ESRI 2006, not individual level data.
high versus 16% too low). Moreover, the average SCCS consumption among the 45% of dieters who exceeded the discretionary calories recommendations was 3059 kJ (731 kcal).
Figure 1 breaks down SCCS energy into the measured
components stratified by site and by key socio-demographic
variables. Site 1 had the highest consumption and the main
components were candy and sugar-sweetened beverages,
consumed even more than in other sites. Women consumed
fewer sugar-sweetened beverages than men (P , 0.01).
Figure 1 also shows the large differences in SCCS consumption, depending on whether or not individuals reported
eating less to lose/maintain weight (P , 0.001 for each
component and the total). The total energy in the ‘eat less’
group averaged 1619 kJ (387 kcal) (95 percent CI: 1456,
1778) or 1.4 times the recommended energy from discretionary foods (95% CI: 1.3, 1.6). Respondents who did not
try to eat less consumed an average of 2745 (656 kcal) of
energy from SCCSs (95% CI: 2536, 2950), or 2.3 times the
recommended amount (95% CI: 2.1, 2.5) with 61% exceeding recommended discretionary calorie intake (95% CI: 57,
64).
Table 2 attempts to disentangle the effects of individual
characteristics. There are two dependent variables, total
energy from SCCS (Model 1) and reported BMI (Model 2).
Model 1 indicates that trying to eat less, eating more fruit
and more exercise were associated with reduced SCCS consumption. Yet there is little substitution of fruit for energy
from SCCS: an additional serving of fruit reduces energy
from SCCSs by 130 kJ (31 kcal), which is less than such a
serving would add (e.g. a medium apple is 310 kJ (72 kcal).
Physical activity was associated with a small reduction in
SCCSs—only 17 kJ (4 kcal).
D I S C R E T I O N A RY CA LO R I E S A P R I O R I T Y FO R O B ES I T Y P R E V EN T ION
383
Fig. 1 Daily energy from snack and sugar-sweetened beverages by sociodemographic group, obesity status and related health behaviors.
In terms of socio-demographics, age was associated with
lower energy consumed from SCCS, but had no effect on
excess SCCS energy, while the opposite was true for
females, in that they consumed more excess SCCS energy
than men, likely due to their smaller size and lower energy
needs. We found no significant difference in consumption
of SCCS energy between Hispanics and non-Hispanic
Whites, but African-American and Asian differed significantly from Hispanics.
Model 2 shows the multivariate analysis of BMI and the
coefficients represent units of BMI in kg/m2. Individuals
who try to lose/maintain weight by eating less have the
highest BMI, equivalent to 4.4 kg (9.6 lbs) more for a
person 1.65 m (50 500 ) tall. Energy from SCCSs is the most
significant predictor of BMI. Consumption of the average
number of SCCSs per day was associated with a BMI
increase of 0.5 kg/m2 or 1.4 kg for a 1.65 m person. Fruit
consumption alone was also associated with reductions in
BMI (1.1 kg less for a 1.65 m) person, but vegetables were
not. In contrast to the results in Table 2, physical activity,
gender or African-American race were not associated with
BMI. Once the SCCS variables were included, these variables provided no additional explanatory power for BMI. In
a sensitivity analysis, we used obesity status in a logistic
regression model and the qualitative results were unaffected.
Table 3 shows findings from the community assessments
of worksites, schools, health-care settings and local food
outlets. With the exception of the near elimination of beverage and snack vending machines available to students in
schools, most settings had a limited number of practices
that would reduce SCCS consumption. Half of the worksites
and more than half of the health-care settings sold
sugar-sweetened beverages and snacks in vending machines.
The prominence of discretionary foods high in energy was
further noted by more than half the schools selling candy
and cookies for fundraisers and nearly half of the worksites
and health-care settings serving cookies, cakes and pastries
at meetings. Furthermore, candy, chips and pastries dominated the local food outlets, and although one-third did not
sell fresh fruits and vegetables, nearly all residents had
access to fruits and vegetables within a 1 mile radius of
their homes.
Physical activity programs were below state guidelines in
the majority of the schools surveyed (only 38% met guidelines) and most worksites and health-care settings did not
offer any structured physical activity programs during the
384
J O U RN A L O F P U B L I C H E A LTH
Table 2 Models predicting energy from salty snacks, cookies, candy
Table 3 Selected healthy eating, active living policies by community
and SCCS and BMI
sectors
Variable, n ¼ 1575
Model 1: SCCS
Model 2: BMI
(kJ)
(kg/m2)
Energy from SCCS
—
2.42 (.56)***
(per 10 000 kJ)
Number of servings of fruit
2130 (49)**
20.41 (.11)***
47 (51)
0.11 (0.12)
Number of servings of
Policies
% of sites that
have policy
Schools (n ¼ 21)
Have formal, written wellness policy
100
Beverage or snack vending machines
5
(for students)
Sells cookies/candy as fundraisers
57
Eat less
2888 (141)***
1.6 (.32)***
PE time meets state guidelines
38
Hours of weekly physical
217.0 (9)*
0.013 (.019)
Organized PA during lunch break
67
School grounds open after school hours
67
vegetables
activity
Age
228 (5)***
0.003 (0.01)
Worksites (n ¼ 40)
2310 (141)*
20.47 (0.31)
Have formal, written wellness policy
White-Non Hispanic
117 (195)
20.75 (0.44)
Beverage or snack vending machines
50
African-American
580 (172)**
20.16 (0.38)
Cookies, cakes, pastries or sweetened
45
Female
Asian
2 1217 (235)***
2 2.2 (0.52)***
25
beverages served at meetings
Site 2
2148 (169)
2 1.6 (0.38)***
Structured PA available
Site 3
2 561 (235)**
20.31 (0.57)
Has a fitness center
23
Offers discounts, incentives for using mass
28
Note: standard errors are represented in parentheses. Bold refers to
significance at P , 0.01. Reference categories are not trying to eat less,
40
transit or active commuting
Health-care settings (n ¼ 14)
male, Hispanic and Site 1. n ¼ 1575, observations with complete data
Measure and track BMI routinely
only. Models use continuous measures for physical activity and fruit/
Have weight management programs
64
vegetable servings.
Have beverage or snack vending machines
57
Cookies, cakes, pastries or sweetened
50
***P , 0.001.
**P , 0.01.
*P , 0.05.
workday. A few sites did have fitness centers and only a
small percentage of worksites and health-care settings
offered incentives to employees to use active or mass transit
to get to work. The majority of health-care settings appeared
to be providing services related to obesity, but only a minority had policies that would help employees to avoid SCCSs
or to engage in physical activity at work.
Discussion
Main finding of this study
The main finding of this study was that the consumption of
large amounts of SCCS was more closely tied to BMI than
fruit and vegetable consumption or physical activity and,
therefore, could justifiably be deemed the most important
target in these communities’ obesity prevention campaigns.
The high amounts of SCCSs consumed was unexpected,
considering that consumption of these types of food is typically underreported (although we tried to minimize this bias
with 24-h recall) and because we did not assess many other
71
beverages served at meetings
Structured PA available
29
Has a fitness center
29
Offers discounts, incentives for using mass
7
transit or active commuting
Community retail food outlets (n ¼ 29)
Any fruits available
Any vegetables available
Any foods high in sugar and fat (candy, chips,
67
70
100
pastries)
discretionary foods, such as cakes and pastries, ethnic
sweets such as churros and pan dulce, and frozen treats. For
more than one-fourth of the population the level of SCCS
consumption in a given day is the energy equivalent of a full
lunch or dinner that consists solely of candy, soda, cookies
and chips.
The finding that fruits are associated with a lower BMI
suggest that the substitution of fruits for higher calorie
snacks is occurring, although the amount is relatively small,
accounting for only 130 kJ (30 kcal)/serving of fruit, which
is fewer calories than a typical fruit contains. In the USA,
fruits are typically consumed as desserts or snacks
in-between meals.
D I S C R E T I O N A RY CA LO R I E S A P R I O R I T Y FO R O B ES I T Y P R E V EN T ION
In the majority of community settings observed, foods
high in discretionary calories were promoted in a routine
way, either in cafeterias, in business meetings or consistently
made available in vending machines. Schools were the only
settings in which there were explicit policies to discourage
consumption of cookies, soda, candy and chips.
What is already known on this topic
According to USDA/DHHS guidelines, to lose weight
people should forego all discretionary calories, and only
consume the minimum amount of food with the necessary
nutrients to maintain their health. The recommendation that
everyone increase physical activity expenditure by an average
of 418 kJ (100 kcal) a day (about a 20 min walk) will not
compensate for some of the large intakes of energy from discretionary foods observed. A person weighing 68 kg
(150 lbs.) who consumes 2400 kJ (500 kcal) excess a day
would need to walk briskly for about 5 times the recommended amount to burn that amount of energy, or for
1 34 h. Physical activity is important to health, although in this
population more physical activity was not significantly associated with reductions in BMI, probably because people who
exercise also ate more, which has been found elsewhere.17
385
reports are typically biased in a direction that would make
our estimates conservative, since most people underestimate
the snacks they consume.22 Yet, the overall finding of the
excessive consumption of non-nutritious foods such as
SCCSs is consistent with other dietary analyses,23 and trends
indicating increases in the sales of sugar and oils over the
past two decades.24
The collaboratives identified a host of interventions to
promote eating more nutritious foods and increasing physical activity, and each intervention could probably be argued
is reasonable and necessary. Nevertheless, unless the excessive consumption of SCCS is curtailed, other interventions
are likely to have a limited impact on obesity control.
Acknowledgements
The authors would like to thank the Kaiser Permanente
project staff, especially Jodi Ravel, and the three collaboratives who helped develop and field the questionnaires.
We greatly appreciate their extraordinary efforts to promote
healthy eating and active living in their local communities.
Funding
What this study adds
While physical activity and the consumption of fruits and
vegetable were not optimal in any of the communities, identifying factors that are likely to play a more significant role in
obesity is helpful in determining where efforts will likely yield
greater impact. When facing other demands or when under
stress, people tend to select high sugar/fat snack foods automatically and preferentially.18 People need help to resist
sugar- and fat-laden snack foods and beverages for which
humans have inborn preferences and which interfere with
appetite regulation.19 – 21 Although improvements in fruit and
vegetable consumption and physical activity are necessary for
optimal health, consumption of SCCS emerged from this
rapid assessment as a much more promising target for
obesity control.
Limitations of this study
While a limitation of the survey is that it may not be representative of the total community, those responding to the
survey do represent the people whom the community collaboratives want to reach with healthy living campaigns.
Another limitation is that the modeling we did is based
upon self-reports of consumption, physical activity, height
and weight that are likely imprecise, and therefore may not
accurately reflect the true relationships. However, these
The study was conducted as part of an initiative sponsored
by Kaiser Permanente, Northern California.
References
1 Yancey AK, Kumanyika SK, Ponce NA et al. Population-based
interventions engaging communities of color in healthy eating and
active living: a review. Prev Chronic Dis 2004;1(1):A09.
2 Fitch C, Stimson GV, Rhodes T et al. Rapid assessment: an international review of diffusion, practice and outcomes in the substance
use field. Soc Sci Med 2004;59(9):1819– 30.
3 Beebe J. Basic concepts and techniques of rapid appraisal. Hum
Organization 1995;54:42– 51.
4 Malik VS, Schulze MB, Hu FB. Intake of sugar-sweetened beverages and weight gain: a systematic review. Am J Clin Nutr
2006;84(2):274– 88.
5 Popkin BM, Nielsen SJ. The sweetening of the world’s diet. Obesity
Res 2003;11(11):1325 – 32.
6 SparksCo, GaleGroup. U.S. Food companies stress ‘science-based’
dietary guidelines—brief article. Food & Drink Weekly. 2004. http://
findarticles.com/p/articles/mi_m0EUY/is_5_10/ai_113299042.
7 Hill JO, Wyatt HR, Reed GW et al. Obesity and the environment:
where do we go from here? Science 2003;299(5608):853 – 5.
8 USDA. Part D: Science Base Section 3: Discretionary Calories. 2005. http://
www.health.gov/DIETARYGUIDELINES/dga2005/report/HTML/
D3_DiscCalories.htm.
386
J O U RN A L O F P U B L I C H E A LTH
9 CDC. Behavioral Risk Factor Surveillance System—Trends Data. 2004.
http://apps.nccd.cdc.gov/brfss/Trends/TrendData.asp.
10 CDC. Behavioral Risk Factor Surveillance System. Summary Data
Quality Report. 2006. http://ftp.cdc.gov/pub/Data/Brfss/2006
SummaryDataQualityReport.pdf.
11 Subar AF, Thompson FE, Kipnis V et al. Comparative validation of
the Block, Willett, and National Cancer Institute food frequency
questionnaires: the Eating at America’s Table Study. Am J Epidemiol
2001;154(12):1089 – 99.
12 Kipnis V, Subar AF, Midthune D et al. Structure of dietary measurement error: results of the OPEN biomarker study. Am J Epidemiol
2003;158(1):14– 21 (discussion 22– 16).
13 USDHHS. Dietary Guidelines for Americans, 2005. Stock Number
001-000-04719-1. Washington, DC: USDHHS, 2005.
14 USDA. USDA National Nutrient Database for Standard Reference. 2008.
http://www.nal.usda.gov/fnic/foodcomp/search/.
15 Craig CL, Marshall AL, Sjöström M et al. International physical
activity questionnaire: 12-country reliability and validity. Med Sci
Sports Exer 2003;35(8):1381 – 95.
16 Yore MM, Ham SA, Ainsworth BE et al. Reliability and validity of
the instrument used in BRFSS to assess physical activity. Med Sci
Sports Exerc 2007;39(8):1267 – 74.
17 Dutton GR, Napolitano MA, Whiteley JA et al. Is physical activity a
gateway behavior for diet? Findings from a physical activity trial.
Prev Med 2008;46(3):216– 21.
18 Shiv B, Fedorikhin A. Heart and mind in conflict: the interplay of
affect and cognition in consumer decision making. J Consum Res
1999;26(3):278– 92.
19 Drewnowski A. Why do we like fat? J Am Diet Assoc
1997;97(7):S58– S62.
20 Erlanson-Albertsson C. How palatable food disrupts appetite regulation. Basic Clin Pharmacol Toxicol 2005;97(2):61 – 73.
21 Keskitalo K, Knaapila A, Kallela M et al. Sweet taste preferences
are partly genetically determined: identification of a trait locus on
chromosome 16. Am J Clin Nutr 2007;86(1):55 – 63.
22 Subar AF, Kipnis V, Troiano RP et al. Using intake biomarkers
to evaluate the extent of dietary misreporting in a large sample
of adults: the OPEN study. Am J Epidemiol 2003;158(1):
1– 13.
23 Bachman JL, Reedy J, Subar AF et al. Sources of food group intakes
among the US population, 2001 – 2002. J Am Diet Assoc
2008;108(5):804 –14.
24 USDA. USDA Food Availability Data System. Economic Research Service
Datasets. 2007. http://www.ers.usda.gov/Data/FoodConsumption/.