National Academy of Science/Institute of Medicine Committee:

October 2005 – December, 2006

Trends: Children’s Obesity

Food Marketing Quiz

How much per year is spent on food & beverage marketing directed to children & youth?

a.$500 millionb.$1 billionc.$5 billiond.$10 billion

Food Marketing Quiz

At what age can children differentiate advertising from program content on television?

a. 2 yearsb. 4 yearsc. 6 yearsd. 8 years

At what age do children understand the persuasive intent of advertising?

a. 4 yearsb. 6 yearsc. 8 yearsd. 10 years

Growth in New Food Products Targeted to U.S. Children and

Youth 1994 to 2004

New products targeted to total market

New products targeted to children & youth

Source: Williams J. 2005b. Product Proliferation Analysis for New Food and Beverage Products Targeted to Children, 1994–2004. University of Texas at Austin Working Paper.

Policy History

Tobacco • Surgeon General’s Report -1964• Labels and Warnings – within weeks• TV & Radio Advertising – Banned by Law in 1970

1971 FTC (Sweets and Cavities)• Rules proposed to ban ads for children <8• Congress intervenes to hold back FTC• CARU formed 1974

1990s• Alarming rise in Obesity

Committee Timeline- Congress asks CDC to study Food

Marketing and Children’s Diets- CDC commissions National Academy

of Science/Institute of Medicine- IOM Committee Formed- Committee Works Tirelessly (for

FREE!)- Press Release- Book in Print- $2,000,000,000 lawsuit

- Policy? E.g., Law, Monitoring body, etc.??

2004 SpringSummerFall

2005 Winter Spring

SummerFall

2006 Winter Spring

SummerFall

2007 2008

Main Committee Charge

Comprehensive evidence based review of what is known about:

The causal influence of food and beverage marketing on the diets and diet-related health outcomes of children and youth

Committee Work

15 Committee members

2 - ½ time IOM staff members

4 (2 ½ day) Meetings

Subgroups:• Diet• Marketing• Ecological Model• Evidence Review• Public Policy

6 months – almost weekly conference calls

Systematic Evidence Review

• explicit criteria for study inclusion

• explicit criteria for study relevance, ratings, etc.

• replicable

Study Inclusion

1. Only Peer-reviewed, published research

2. English

3. Any country

4. Any date

5. Only original research, no review articles

6. Only research that reports a quantitative relationship between a variable involving marketing, and a variable involving either a pre-cursor to diet, diet, or diet-related health

Dimensions for Study Evaluation:• Sample size, year published, population studied• Cause variable, effect variable• Research Method• Statistically significant association between

cause variable and effect variable?• Strength of evidence for causation

(Causal Inference Validity)• Quality of Measures• Generalizability (Ecological Validity)

Causal Framework Used for the Systematic Evidence Review

Mediators/Precursors

Food & Beverage Preferences, Beliefs,

Purchase Requests

Diet

Short--Term Consumption

Usual Dietary Intake

Diet ---Related Health Outcomes

Adiposity, Other

Moderators

Age, SES, Gender,Race/Ethnicity

Marketing

Product, Place,

Price, Promotion

Cause Variable, Effect Variable• Cause variable, e.g.:

– TV ad exposure– Product Placement in Film– Print Ad exposure– Radio Ad exposure

• Effect variable, e.g.,:– Food Preferences– Food Purchase requests– Short-term consumption– Adiposity

Research Method

• CS (cross-sectional)• Exp (randomized trial)• Exp-N (natural experiment)• L-Pnl (longitudinal study – panel)• L-Trnd (longitudinal study – trend)

Statistical Significance

Cause _||_ Effect | Measured Covariates

E.g., in a regression

BMI = 21.7 + .52*TVhours/day - .64*Mother’s Education

P-value = .0074

Strength of Evidence for Causation

• Experimental studies

• Observational studies

Strength of Evidence for Causation:

Experimental studies

• Randomized Assignment of Treatment

• High Quality Measures

• If dropout - not associated with treatment

Strength of Evidence for Causation:

Experimental studies

Cause Effect

Cause Effect

Cause Effect

Confounder

I

I

I

Evidence: Causeset _||_ Effect

Causeset _||_ Effect

Causeset _||_ Effect

Causeset _||_ Effect

Strength of Evidence for Causation:

Observational studies

• Quality Measures

• All potential confounders measured (well) and appropriately controlled for statistically,

• Reverse causation can be eliminated, perhaps by time-order

Strength of Evidence for Causation:

Observational studies

Cause Effect

Cause Effect

Cause Effect

Unmeasured Confounder

Measured Confounder

Cause Effect

Measured Confounder

Measured Confounder

Evidence: Cause _||_ Effect | Measured Confounders

Cause _||_ Effect | Measured Confounders

Cause _||_ Effect | Measured Confounders

Cause _||_ Effect | Measured Confounders

Cause _||_ Effect | Measured Confounders

Strength of Evidence for Causation:

Observational studiesCause _||_ Effect | Measured Confounders

Good Evidence for causation if:

Cause Effect

Cause Effect

Cause Effect

Unmeasured Confounder

Measured Confounder

Cause Effect

Measured Confounder

Measured Confounder

• Effect not prior to cause

• All confounders measured well

Measurement

• Validity – measure directly and accurately measures what it is intended to measure

• Reliability – same measurement technique applied repeatedly, yields same outcome

• Precision – fineness vs. coarseness of measure

Ecological Validity

Generalizability:

Similarity of observational conditions to real-life.

E.g., - after-school lab study vs. survey

Results

1.Tabular Overview of Evidence

Base2. Marketing Precursors

3. Marketing Diet

4. Marketing Health

Example Results: Summary

Example Results: Summary

Results

1. Tabular Overview of Evidence Base

2.Marketing Precursors3. Marketing Diet

4. Marketing Health

Results: Pre-cursors

• Precursors:

o Preferences

o Food Purchase Requests

o Beliefs about Food

Results: TV Ads and Food Preferences

TV Ads and Purchase Requests

• (Young 2-5): Strong evidence for causation• (OC 6-11) Moderate evidence for causation• (Teens 12-18) Weak/Insufficient for causation

Results: Summary of Pre-cursors

Results

1. Tabular Overview of Evidence Base

2. Marketing Precursors

3.Marketing Diet4. Marketing Health

Results: Short-term Consumption

• (Young 2-5): Strong evidence for causation• (OC 6-11) Strong evidence for causation• (Teens 12-18) No evidence

Results: Usual Dietary Consumption

• (Young 2-5): Moderate evidence for causation• (OC 6-11) Weak evidence for causation• (Teens 12-18) Weak/Insufficient against causation

– (not including French, et al., study on vending machine price

Results

1. Tabular Overview of Evidence Base

2. Marketing Precursors

3. Marketing Diet

4.Marketing Health

Marketing Health

Results: Summary

• TV Ads Precursors (esp. for young children)

• TV Ads Short-term consumption (esp. for YC)

• TV Ads Usual Diet (???)

• Moderate evidence for YC

• Weak for OC

• Weak against for teens

• Marketing Health (???)

• TV Ads _||_ Obesity : strong evidence

• TV Ads Obesity: insufficient evidence

Results

• ONLY TV Ads studied – marketing much broader

• Only effects of ads for high-calorie, low nutrition foods studied

• Effects of ads for healthy foods – virtually unstudied!

Lawsuit

Center for Science in the Public Interest (CSPI)

vs. Viacom, Kellog

Charge: Unfair and deceptive advertising and marketing to children under 8

Request: $1 billion each

Committee on Food Marketing and the Diets of Children and

YouthJ. Michael McGinnis (Chair),

Institute of MedicineDaniel Anderson, University

of Massachusetts, AmherstJ. Howard Beales III, George

Washington UniversityDavid Britt, Sesame

Workshop (retired)Sandra Calvert, Georgetown

UniversityKeith Darcy, Ethics Officer

AssociationAimee Dorr, University of

California, Los AngelesLloyd Kolbe, University of

Indiana

Dale Kunkel, University of Arizona

Paul Kurnit, Kurnit Communications & KidShop

Robert Post, Yale Law SchoolRichard Scheines, Carnegie

Mellon University Frances Seligson,

Pennsylvania State University

Mary Story, University of Minnesota, Minneapolis

Ellen Wartella, University of California, Riverside

Jerome Williams, University of Texas, Austin

Co-study Directors: Jennifer Gootman and Vivica Kraak

For more information. . .

Read about the project and download fact sheets

at: www.iom.edu

The book is available at: www.nap.edu

Download the executive summary free . . .