Causes and Consequences of Food Choice Kiyah Duffey Department of Nutrition The University of North Carolina at Chapel Hill November 16, 2006.

Causes and Consequences of Food

Choice

Kiyah Duffey Department of Nutrition

The University of North Carolina at Chapel HillNovember 16, 2006

Outline

Nutritional Epidemiology at UNC Factors affecting food choice

Environmental factors Consequences of food choice: obesity

Background Epidemiology/ Current Trends

Specific Research Fast Food Consumption Beverage contribution to overall diet

Nutrition Epidemiology

Epidemiology is the study of the distribution of disease in a population. Population-level causes of disease

At UNC, focus is on the link between diet & physical activity and obesity; nutrition transition in developing countries US, China, Philippines, Russia, Brazil

Factors Influencing Food Choice

An Ecological Model of Diet, Physical Activity, and Obesity

Developed by Dr. Penny Gordon Larsen, UNC Chapel Hill Dept of Nutritionfor the NHLBI Workshop: Predictors of Obesity, Weight Gain, Diet, and Physical Activity; August 2004, Bethesda MD

Food Choice

Dietary Patterns & Nutrient Intake:Type, Amount,

Frequency of foods eaten

Biological & Demographic

Age, sex, SES, genetics, disease states

Psychological

Preferences, emotions, body image,

motivation, knowledge

Social/Cultural

Family factors, peer influencesocial norms, acculturation

Physical Environment

Access, urban design, transportation, advertising

Policies/Incentives

Cost & availability: Environmental regulation

Organizational

Programs & policies in schools, worksite,

community orgs

Factors Affecting Food Choice Physical Environment: Location

Social/Cultural: Visual Cues, portion size

Policies: Avertising

Location

Location of Fast Food Outlets to Chicago Schools: 35% w/ in 400m 80% w/ in 800m

3 - 4 times more FF w/in 1.5 km from schools than expected if distributed randomly

More highly clustered: outside downtown mid-upper income more commercialized

Austin et al. American Journal of Public Health 2005; 95( 9):1575-1581

Neighborhood Affects Access ARIC, nationwide data Examined neighborhood

influence on diet 8% of Blacks vs. 31%

Whites lived in census tract with ≥1 supermarket

Blacks: 1.31 and 2.18 times more likely to meet guidelines for F&V consumption if there was 1 or 2 supermarkets in their census tract

Moreland et al. Am J Public Health. 2002 Nov;92(11):1761-7

Black neighborhoods have Higher Density of Fast Food Outlets Orleans Parish, LA Modeled probability of

fast food outlets (FFO) given race and income

Black neighborhoods had more FFO/mi2 than White neighborhoods: 2.4 vs. 1.5

Black neighborhood explained 19% of the variance in distribution of FFO

Block et al. Am. J. Preventive Med. 2004:27(3):211-7.

Visual Cues

Are all containers created equal? People consumed more

beverage from short fat glasses than tall thin ones

4, 2L bowls vs. 2 4L bowls Self-serve Serving from 4L bowls

resulted in 53% more taken, 56% more calories consumed

≠

Brian Wansink, Cornell University: Nutrition Psychology. Unpublished data

≠

Would you like popcorn with that?

158 moviegoers

Popcorn buckets weighed before and after movie

2-week old popcorn tasted “stale”, “soggy”, “terrible”

Fresh

240 g 120 g

2-Weeks old

240 g 120 g

<< >

33%

>

41%Wansink & Kim. 2005. J Nutr Educ Behav: 37; 242-45

“Bottomless” Bowls Increase Lunch Calories

54 adults 20 min, free lunch Half given 18 oz, half

given “bottomless” bowls 20% more soup (113

kcals) eaten from bottomless bowls

No differences in estimated caloric intake or reported satiety

Brian Wansink. 2005. Obesity Research: 13(1); 93-100

Portion Size

Portion Sizes Increasing, Cheeseburger

Calorie difference: 257 calories

590 calories

20 Years Ago Today

333 calories

Increased Portion Sizes, Bagels

140 calories 3-inch diameter

Calorie Difference: 210 calories

350 calories 6-inch diameter

20 Years Ago Today

Effect of Portion Size on Energy Intake

Rolls et al. 2004, 2005 Altered portion sizes of

sandwiches, macaroni dishes & pre-packaged chips

Participants free to eat at will

Measured plate waste, and asked respondents about level of satiety

Pasta Entrée Size Affects Total Caloric Intake

0

500

1000

1500

2000

2500

100% 150%

Portion Size of Entree

Ene

rgy

Inta

ke (kJ

)

Baked Pasta Accompaniments Side Dishes Desserts Beverages

600

400

200

En

erg

y Inta

ke

(kcal)

600

400

En

erg

y Inta

ke

(kcal)

Diliberti et al. 2004. Obesity Research: 12(3);562-568

*

*

Sandwich Size Affects Energy Intake

Females (n=37)

Males (n=38)

Rolls et al. J Amer Diet Assoc. 2004; 104:367-372

0

200

400

600

800

1000

1200

6 8 10 12 6 8 10 12

Sandwich Size (inches)

En

erg

y In

take (

kcals

)

+ 159 kcals + 355 kcals

Advertising and Other Influences

Advertising Children (<12y) viewed ~4,900 food & restaurant

commercials/year http://www.msnbc.msn.com/id/15095189/

Snickers budget alone 5 x greater than 5-a-day

$11.6 billion $9.5 million

Other Influences Changing food supply

~3900kcal/person available* Significant increases in added sugars

Pricing Availability at home

Carrots slices vs. Doritos Parental Influences

Overweight children are more likely to have overweight parents

*Data are based on ERS estimates of per capita quantities of food available for consumption, imputed consumption data, and on estimates from USDA's. Source: USDA/Center for Nutrition Policy and Promotion, March 3, 2006. http://www.ers.usda.gov/Data/FoodConsumption/NutrientAvailIndex.htm

Consequences of Food Choice: Overweight & Obesity

Senate bill bans obesity lawsuits By Marguerite Higgins

THE WASHINGTON TIMES

'Freshman 15' really 5 or 7, but the gains don't stopCNN , POSTED: 11:37 a.m. EDT, October 23, 2006,

Defining Overweight & Obesity Excess of body fat for a given height and

weight

In adults, defined using BMI (kg/m2)

Example 6’0”, 160 lbs, BMI=21.7; 200 lbs, BMI=27.1 5,4”, 145 lbs ,BMI=24.9; 170 lbs, BMI=29.2

Defining Adult Overweight, Obesity

BMI (kg/m2)BMI Weight Status

<18.5 Underweight

18.5-24.9 Normal weight

25.0-29.9 Overweight

≥ 30.0

30.0-34.9 Obese

35.0-39.9 Severe obesity

≥ 40.0 Morbid obesity

Adapted: National Heart, Lung, and Blood Institute Guidelines. Obes Res.1998:6(suppl 2):51S-209S.

BMI Correlates with % Body Fat

Evidence of a Problem

About 127 million US adults are overweight

60 million obese, 9 million severely obese (equivalent to BMI ≥40)

48

15

56

22

66

32

0

10

20

30

40

50

60

70

Overweight Obese

197619882004

Ogden et al. JAMA 2006: 1549-1555

No Data <10% 10%–14%

(*BMI ≥30, or ~ 30 lbs overweight for 5’ 4” person)

Obesity Trends* Among U.S. AdultsBRFSS, 1985

http://www.cdc.gov/nccdphp/dnpa/obesity/trend/maps/

Obesity Trends* Among U.S. AdultsBRFSS, 2005

(*BMI ≥30, or ~ 30 lbs overweight for 5’ 4” person)

No Data <10% 10%–14% 15%–19% 20%–24% 25%–29% ≥30%

Health Consequences of Obesity

Hypertension Dyslipidemia (for example, high total cholesterol

or high levels of triglycerides) Type 2 diabetes Coronary heart disease Stroke Gallbladder disease Osteoarthritis Sleep apnea and respiratory problems Some cancers (endometrial, breast, and colon)

Study 1: Fast Food vs. Restaurant Food Consumption

Effects on Energy Intake

Duffey, Gordon-Larsen et al., AJCN. In press

Background

Away-From-Home (AFH) eating provides >30% total daily energy intake among adolescents/adults

Evidence suggests that: AFH food typically higher in total calories, saturated fat,

refined carbohydrates and cholesterol

Higher frequency of AFH consumption may be associated with BMI and weight change

Little is known about nutritional differences between restaurant and fast food as they are usually studied together

Study Questions

Do fast foods and restaurant foods have differential effects on weight gain?

What happens to these effects if the two sources of away from home eating are combined into a single measure?

Study Population & Methods

CARDIA: longitudinal data

Study years 0 (1985-86), 7 (1992-93) and 10 (1985-86); adults aged 18-30

Defined patterns of fast food and restaurant food intake

Modeled the association of eating away from home with weight change and total caloric intake

Higher 7-year Fast Food Intake is Associated with

Greater 7-year BMI Gain in White Females*

* Model 1: adjusted prevalence stratified multinomial logistic regression models controlled for: age, income, education, eating pattern (maintain low, decreased, increased, maintain high fast food) and study center.

** p<0.05, compared to referent category (maintained BMI).

0

10

20

30

40

50

60

70

80

Pe

rce

nt

of

wh

ite w

om

en

n=112 n=363 n=557

Fast Food IntakeFrom Year 0 to 7

11

33

56

10

33

57

9

23**

11

34

55

68**

Lost > 1 BMI unit Maintained BMI Gained > 1 BMI unit

Maintained Low

Decreased

Increased

Maintained High

Higher Fast Food (versus Restaurant) Intake is Associated with Higher BMI Gains over a 7-year Period in White

Females*

8

24

13

37

8

33

59

68**

50**

0

10

20

30

40

50

60

70

80

90

Lost > 1 BMI unit Maintained BMI Gained > 1 BMI unit

Pe

rce

nt

of

wh

ite w

om

en Higher Fast Food

Higher Restaurant

Higher Both

n=112 n=363 n=557

* Model 2: adjusted prevalence stratified multinomial logistic regression models controlled for: age, income, education, eating pattern at time 7 (high v. low intake of restaurant, fast food or combination) and study center.

** p<0.05, compared to the referent category (maintained BMI).

Conclusions

Increase in fast food consumption associated with greater likelihood of increase in BMI

Higher restaurant (versus higher fast food) consumption is associated with greater likelihood of maintenance of BMI

Combining AFH to a single measure masks the independent associations of these two food sources with long-term energy intake and BMI

Study 2: Dietary & Beverage Patterns

Predicting Water vs. Soda Consumption

Duffey & Popkin. Journal Nutr. 2006: 2901-07.

Adults are Consuming More Calorically-Sweetened Beverages Increased consumption of calorically-sweetened beverages over the past two decades

Average adult obtains 21% of calories from beverages: providing an additional 150-300 calories/day

Among consumers, there is greater consumption of calorically-sweetened beverages, such as soda, than for caloric beverages with nutrients, such as low-fat milk

52

22.325

26

12 12 11

21

28

0

10

20

30

40

50

60

Water Coffee Tea Diet Low-Fat Milk Fruit Juice VegetableJuice

Fruit Drinks Soda

Beverages

Among Consumers*, Mean Fluid Ounce Intake Greater for

Calorically-Sweetened Beverages

Mea

n C

on

sum

pti

on

(fl

. o

z.)

UnsweetenedUnsweetened

Caloric w/ Caloric w/ NutrientsNutrients

Calorically-Calorically-SweetenedSweetened

* Adults 18+ years from NHANES 1999-2002 survey

DietDiet

52

22.325

26

12 12 11

21

28

0

10

20

30

40

50

60

Water Coffee Tea Diet Low-Fat Milk Fruit Juice VegetableJuice Fruit Drinks Soda

Beverages

Study Questions

Do certain beverages tend to be consumed together (are there patterns of beverage intake)?

If so, how do food patterns associate with these beverage patterns?

Study Population & Methods

National Health and Nutrition Examination Survey 1999-2002, adults ≥19 years

First of two non-consecutive 24 hour recalls

UNC-CH Food Grouping System & cluster analysis Finds patterns in data and generates groups of like

individuals

Modeled effect of food cluster on beverage cluster

Final Food Clusters*

Normal55%

Cereal & LF Meat10%

Vegetables5%

Fruit & LF Dairy5%

Snacks & HF Foods14%

Fast Food11%

*Adults, 19+

Final Beverage Clusters*

Water & Tea14%

Coffee, Water & Tea19%

Diet14%

Coffee & Soda22%

Nutrients & Soda

14%

Soda17%

*Adults, 19+

Fast Food Membership Linked with Decreased Probability* of

Consuming Water-Containing Beverage Patterns

*Predictions from mlogit results controlling for age, race, gender, income, education & overweight status

10.311.9

9.6

26.1

17.8

24.2

15 15.6

11.3

22.5

15.417.3

0

5

10

15

20

25

30

Water & Tea Coffee, Water

& Tea

Diet Coffee &

Soda

Nutrients &

Soda

Soda

Beverage Cluster

Per

cen

t o

f S

amp

le

Fast Food Member Fast Food Non-Member

Vegetable Group Membership Linked with

Decreased Probability* of Consuming Soda

20.9 20.8

8.5

24

17.9

10.9

14.1

17.8

11.2

22.8

15.7

18.4

0

5

10

15

20

25

30

Water &

Tea

Coffee,

Water &

Tea

Diet Coffee &

Soda

Nutrients &

Soda

Soda

Beverage Clusters

Per

cen

t o

f S

amp

le

Vegetable Member Vegetable Non-Member

*Predictions from mlogit results controlling for age, race, gender, income, education & overweight status

In Conclusion Caloric and non-caloric beverages tend to be

consumed independently

Healthier diet patterns associated with healthier beverage patterns

Substituting non-caloric beverages for caloric ones can reduce total energy intake

What’s Next? How does an individual’s

neighborhood impact their food choices, dietary intake, physical activity patterns and weight gain?

Do changes in the environment lead to changes in these outcomes?

Derived Measures Distance Matrices, Network Calculations,Connectivity, Community Classifications

GIS Database

Respondent LocationsPA Resources

Contextual Data

Land Use

Ancillary Data(Roads, Administrative

Boundaries, etc.)

GPS Data

Obesity & The EnvironmntThe University of North Carolina at Chapel Hill

Respondents

Block Group Boundary

Residential Neighborhoods

Obesity & The EnvironmentThe University of North Carolina at Chapel Hill

Respondents

Individual Buffer

Community Study Area

Sampled Block Group

Create Buffer Zones, 5 Mile from Residential Location

Obesity & The EnvironmentThe University of North Carolina at Chapel Hill

Respondents

Individual Buffer

Community Study Area

Sampled Block Group

Block Group Boundary

Food Sources

Measures Food Sources in 5 Mile Buffers

Intended Analysis

Examine average distance to and count of fast food and restaurant places within buffers around residential locations

Determine if proximity → times/week consumption

Determine if consumption → weight gain

8

For consideration… Continued need for unified, simple message

Conventional vs. Organic Whole Foods & Michael Pollan http://gristmill.grist.org/story/2006/6/29/143121/559?source=weekly

Local vs. long-distance foods

Nutrition Labeling at point of purchase; restaurant menus, menu boards

Selected References Books: Food Politics by Marion Nestle The Hungry Gene by Ellen Ruppel Shell Food Fight by Kelly Brownell

Websites http://www.consumersunion.org/pub/core_health_care/002657.html http://www.cdc.gov/nccdphp/dnpa/obesity/trend/maps/ The Food Trust: http://www.thefoodtrust.org/php/programs/farmers.market.program.php Food Policy Institute: http://www.preventioninstitute.org/npp.html Pro Restaurant/Food producers: http://www.consumerfreedom.com/ American Obesity Association: http://www.obesity.org/ North American Association for the Study of Obesity (NAASO): http://www.naaso.org/ Center for Science in the Public Interest: http://www.cspinet.org/ Whole Foods vs. Pollan blog: http://gristmill.grist.org/story/2006/6/29/143121/559?source=weekly Tufts University, Professor of Food Policy- blog :http://www.usfoodpolicy.blogspot.com/ Rudd Center for Food Policy, Yale: http://www.yaleruddcenter.org/home.aspx

Articles: Young & Nestle. Am J Pub Health 2002: 246-47 Gordon-Larsen et al. Obes Res. 2003;11(1):121-129. Albright et al. Health Educ Q. 1990;17(2):157-167. Weinsier et al. Am J Med. Aug 1998;105(2):145-150. Fiske & Cullen. J Am Diet Assoc. 2004;104(1):90-93. Diez Roux et al. N Engl J Med. 2001;345(2):99-106. Jeffery & Utter. Obes Res. 2003;11:12S-22S. Nielsen S et al. Obes Res. 2002;10(5):370-378. French S et al. Annu Rev Public Health. 2001;22:309-335. Rolls et al. Appetite. 2004;42(1):63-69. Block et al. Am J Prev Med. 2004;27(3):211-217. Gordon-Larsen et al. Pediatrics. Feb 2006;117(2):417-424.

My Contact Information:

kduffey@unc.edu

Vote with your Fork