Changing Perceptions of the ChildhoodObesity Epidemic

W. Douglas Evans, PhD; Jeanette M. Renaud, PhD; Eric Finkelstein, PhDDouglas B. Kamerow, MD; Derek S. Brown, PhD i

Objectives: To examine changesin public attitudes about child-hood obesity and support for pre-vention. Methods: RTI surveyedUS households (N = 1047 and N =1139) about perceived severity,causes, and support for specificobesity interventions. Logisticregressions examined differencesin obesity attitudes and supportfor prevention. Results: Perceivedhealth threat of childhood obesityincreased between the 2 surveys.Support increased for interven-

tions such as regulation of res-taurant portions and fast foodadvertising. Logistic regressionsrevealed differences amongsociodemographic groups. Con-clusions: Public support for child-hood obesity prevention is increas-ing. Policy makers can use theseHndings to develop appropriateprevention strategies.

Key words: childhood obesity,prevention, att i tudes , publicpolicy

Am J Health Behav. 2006;30(2)167-176

The United States and nations aroundthe world are experiencing epidemic-level increases in overweight and

obesity. As reported elsewhere,'-^ obesityis a leading cause of preventable death inthe United States, and it costs over $93biUion in medical expenditures amongadults per year, or about 9% of the totalEunount. Poor diet and physical inactivity,the primary modifiable contributors toobesity, have led to a 33% obesity rate(defined as BMI >= 30) and 67% rate ofoverweight (BMI >=25} among US adults.^Although there has recently been debateabout the number of deaths attributable

W. Douglas Evans, Division Vice President,RTI, Washington, DC. Jeanette M. Renaud, Re-search Analyst, RTI, Atlanta, GA. Eric Finkelstein,Senior Health Economist, RTI, Research TrianglePark, NC. Douglas B. Kamerow, Chief Scientist,RTI, Washington, DC. Derek S. Brown, ResearchEconomist, RTI, Research Triangle Park, NC.

Address correspondence to Dr Evans, RTI, 1615M Street, NW, Suite 740, Washington, DC 20036.E-mail: devans@rti.org

to obesity,"* it remains a rapidly risingpublic health threat; and there is a press-ing need to conduct prevention research,educate the public, and increase preven-tion and control efforts.

The obesity epidemic is especiallyalarming among children and adolescents.Today, 15% of children and adolescentscombined are overweight.^ Increasingrates of child and adolescent overweightcan be traced to the 1980s. In the timeinterval between the second NationalHealth and Nutrition Examination Sur-vey (NHANES II; completed in 1980) andfourth survey (NHANES III; completed in2000), the prevalence of obesity increasedfrom an estimated 7% to 16% amongchildren aged 6 to 11 years and from 5% to16% among adolescents aged 12 to 19years.^ This trend suggests that a newgeneration of Americans will enter adult-hood already obese or at risk for obesity.They will already have or be at risk formultiple related health conditions, suchas diabetes and cardiovascular disease,unless actions are taken to reverse the

Am J Health Behav.^ 2006;30(2):167-176 167

Childhood Obesity Epidemic

This paper reports on results of thesecond in a series of nationally represen-tative surveys of US adult opinions andattitudes about obesity in general, andsupport for interventions to prevent child-hood obesity in particular. The objectivesof this study are to examine changes inpublic attitudes about obesity in the con-text of a changing social and policy envi-ronment around the issue. The US publicis increasingly awaje of the public healththreat from rising obesity rates.^ Publicattitudes toward the obesity epidemicamong adults and children are evolvingas the public becomes more aware ofobesity health risks and the need to com-bat them. There are a number of socialinfluences on public attitudes toward obe-sity that have the potential to changesocial norms and promote support for in-terventions.'^ These include major publicand private programs in schools and com-munities, policy and legislative initia-tives, advertising and promotion of healthyfoods and active lifestyles, and news me-dia coverage.

In response to the obesity epidemic,governmental and nongovernmental pub-lic health organizations are funding anumber of major prevention initiatives.The Centers for Disease Control and Pre-vention (CDC) have funded the Well-inte-grated Screening and Evaluation forWomen Across the Nation (WISEWOMAN)program, which is aimed at increasingknowledge and skills of women in the 40-64 age range to improve diet and physicalactivity to prevent and control chronicdisease.^ CDC has also funded obesityprevention programs in 28 states and hasplans to fund all 50. The Department ofHealth and Human Services (DHHS) haslaunched Steps to a Healthier US to ad-dress obesity, diabetes, asthma, and theirmodifiable risk factors-poor nutrition,physical inactivity, and tobacco use.^

Social marketing is increasingly aimedat promoting healthy lifestyles.'° In nutri-tion, the longstanding 5-A-Day for BetterHealth Campaign, funded by the NationalInstitutes of Health, aims to increaseawareness of the health benefits andreductions in health risks from consum-ing 5 or more-preferably 9 or more-serv-ings of fruits and vegetables per day."

Recently, DHHS has also launched theSmall Steps public service announce-ment (PSA) campaign, which includesover 100 video advertisements aimed at

encouraging improved nutrition and in-creased physical activity.'^ The CDC'sVerb, It's What You Do, campaign aimedat "tweens" - youth aged 9 to 13 - seeks tobrand physical activity as popular and cooland to promote it over sedentary behav-iors such as television viewing.'^'^ Likeprevious initiatives in cardiovascular dis-ease and tobacco control,'^'"' these socialmarketing campaigns all seek to promotehealthy behaviors and lifestyle choices.

National and local news media havealso focused on obesity as a major publichealth issue. In June 2004, Time Maga-zine and ABC News cosponsored a sum-mit on obesity in Williamsburg, VA. An-chor Peter Jennings broadcast the na-tional nightly news from the conferenceand focused the program on the obesityepidemic. ̂ ^ Obesity-related stories havebecome common in health reporting innewspapers and on television and theInternet. Major publications such as Na-tional Geographic have put obesity ontheir covers and devoted extensive cover-age to the topic.'^

Americans also report that news cover-age of obesity has reached high levels.Evans et al̂ reported that over 65% of USadults reported seeing or hearing a newsstory on obesity at least once in the previ-ous week, and fewer than 11% had neverseen or heard a story on obesity.^ Asdiscussed below, these self-reported ratesof exposure to news coverage have con-tinued to increase, suggesting that theissue is highly visible and the public isincreasingly aware of it.

Given the increased amount of obe-sity-related programs, marketing, andnews coverage, we hypothesized that pub-lic perceptions of the severity of the obe-sity epidemic would increase and supportfor preventive interventions would alsorise. To better understand the changingsocial environment around this issue, weasked the following research questions:

• How have public perceptions of child-hood obesity changed over a 9-monthperiod of intensive news coverage andpublic debate?

• Has public support for specific child-hood obesity interventions changed dur-ing this 9-month period?

• What sociodemographic factors are as-sociated with support for childhood obe-sity interventions? Do these factorspersist over time?

168

Evans et al

METHODSAs reported elsewhere,^ in the fall of

2003, we developed an instrument de-signed to capture attitudes and opinionson the seriousness of childhood over-weight and obesity compared with otheryouth issues; support for specific inter-ventions; and potential barriers to sup-port, such as negative consequences andincreased taxes; and adult obesity.

The questions on support for specificinterventions were grouped into 3 do-mains: schools, media, and communi-ties. These domains were selected notonly because they represent areas inwhich promising interventions have al-ready been developed and implementedin some communities, but also becausethey are likely targets for future govern-ment and private funding. Within eachdomain, we created scales to rank pos-sible interventions by degree of intensityor restrictiveness. The scales were de-signed to be ordinal, based on intensity ofthe intervention.

In the summer of 2004, we revised theinstrument to include additional ques-tions on specific topics, such as advertis-ing and media; obesity-related healthcare, intentions, and behavior; andworksite and other adult prevention in-terventions. Some questions from thefirst instrument were deleted, such asthose that achieved extremely high lev-els of agreement (and were therefore sub-ject to ceiling effects for potential change).We retained all questions on specific child-hood obesity prevention interventions andmost of those on other obesity attitudesand opinions.

The Odom Survey Research Instituteat the University of North Carolina con-ducted the second data collection on be-half of Research Triangle Institute (RTI}.The study protocol received institutionalreview board approval in August 2004,and the survey was administered betweenSeptember 12 and October 31, 2004.

A total of 1139 interviews were com-pleted with respondents in all 50 statesand the District of Columbia. The surveyachieved a relatively low response rate of28%. This rate is consistent with our firstsurvey, which achieved a response rate of30%, and recent results obtained for simi-lEir health-related surveys (eg, the Behav-ioral Risk Factor Surveillance System[BRFSS] survey).^

Response rates for telephone-based

surveys have generally been declining inrecent years due to factors such as cellphone use, prevalence of call screeningusing caller ID, Internet use, and relatedsocial and technological changes. Giventhat surveillance surveys are achievingconsistently lower response rates due tothese secular trends, there is an ongoingdebate about the effects of lower tele-phone-survey response rates on data qual-ity.̂ 'o

Recruitment and sampling methodswere identical to the first survey, as re-ported elsewhere.^ Survey respondentswere recruited using random-digit-dial-ing (RDD) methodology. The sample frameconsisted of the set of all telephone ex-changes that met residential telephoneexchange geographic criteria. The samplewas nationally representative accordingto American Association of Public Opin-ion Research standards.2' Advanced let-ters encouraging participation were sentto potential respondents identified for thesample. This procedure was not used inthe first survey.

As part of a methodological experiment,the introductory script for half of the house-holds (randomly assigned) stated that theinterviewer was calling on behalf of RTI;the script for the other half stated that theinterviewer was calling from the Univer-sity of North Carolina (the organizationthat conducted data collection). In allcases, a participant who asked for addi-tional information or clarification aboutthe study was informed of the roles of bothorganizations. The results of the method-ological experiment will be reported else-where.

Prior to analysis, Odum Institute andRTI analysts conducted extensive logicchecks, data cleaning, and validation toensure data quality. SAS version 8 forWindows software (SAS Institute, Cary,NC) was used for all data managementand analysis.

To explore whether support for each ofthe 18 childhood overweight and obesityinterventions varied systematically byspecific sociodemographic characteris-tics, we conducted a series of logisticregression analyses. The dependent vari-able for each regression was coded as 0 foroppose or strongly oppose and 1 for favor orstrongly favor. Independent variables in-cluded education (0 = less than high school,high school degree, or GED; and 1 = somecollege, 2-year college degree, 4-year col-

Am J Health Behav.™ 2006;30(2}:167-176 169

Childhood Obesity Epidemic

Table 1Sample Characteristics:Comparison

Second

DemographicGroup

Overall

Age GroupAges 1810 24Ages 25 to 34Ages 35 to 49Ages 50 to 64Ages 65+Refused/don't know

GenderFemaleMale

Race/EthnicityAsianAmerican IndianBlack/African AmericanHispanic/I^atinoWhite/CaucasianMultiracialOtherRefused/don't know

EducationLess than a high school

degreeHigh school graduate or

GEDSome co! lege or 2-yearcollege degree4-year college degreePostgraduate studyReftjsed/don't know

IncomeLess than $25,000$25,000 - $49,999$50,000 - $74,999$75,000 - $99,999$100,000-$124,999$125,000-$149,000$150,000 or moreRefused/don't know

Employment StatusEmployedNot employedRefused/don't know

Marital StatusMarriedSeparated or not

Currently marriedRefused/don't know

Has Overweight orObese Child

YesNo

of FirstSurveys

and

2004 PercentJan-Mar(n=1047)

100.00

7.1313.3731.9826.7318.122.67

59.0140.99

0.860.86

11.376.40

72.782.480.864.39

7.62

30.59

26.5319.7014.650.91

16.1426.4417.4311.093.761.884.55

18.71

67.2332.180.59

53.07

46.140,79

4.4933.62

Sept-Oct(n=1139)

100.00

5.0912.7327.7427.5721.07

5.80

65.2134.79

1.490.97

10.544.48

74.362.550.615.00

6.23

27.30

26.7818.0917.384.22

17.6525.5118.5310.715.362.114.04

16.09

60.6735.214.12

54.87

40.834.30

5.9728.36

Table 2Perceived Seriousness of

Childhood Obesity - Comparisonof 1st & 2nd Surveys

Percent RespondingThat It Is a VerySerious Problem

(n)2004 Percent

IssueJan-Mar(n=1047)

Sept-Oct{n=ll39)

Underage drinking

Underage smoking

Teen sex

Childhood obesity

Drug abuse amongadolescents

Youth violence

47.78(496)42.98(444)45.18(464)

41.54**(432)

55.38(571)50.97(528)

48.10(544)

43.10*(487)44.93(505)47.40(537)

59.25**(666)

53.51**(602)

Note.Significant differences in percentageresponding that it is a serious problembetween Jan-Mar 2004 and Sept-Oct 2004 areshown in bold. Significant differencesbetween obesity and other issues for Sept-Oct2004 are italicized.**P<O.OI* P< 0.05

lege degree, or postgraduate study), gen-der (0 = male and 1 = female), income (0 =$50,000 or more and 1 =less than $50,000),and children under the age of 18 living inthe household (0 none and 1 = one ormore). Table 4 reports odds ratios (ORs)and 95% confidence intervals (CIs) show-ing the relationship between each of thesevariables and support for specific inter-ventions.

RESULTSSample CharacteristicsTable 1 presents sociodemographic

characteristics of respondents and showsthat the 2 samples are comparable inmost respects. Nearly 60% of respon-dents in the first survey, and 65% in thesecond survey, were female. Approxi-

170

Evans et al

Figure 1Perceived Seriousness of Childhood Obesity Relative

to Other Youth Issues

hat

HU)e

•5GOQ.

Res

e•uw41

E

2o

flu

3

o

Sen

>

65

60

55

50

45

40

35

• • Adolescent drug abuse

O - — -O Youth violence

fl - - - fl Underage drinking

Q H Teen sex

4 # Cbildhood obesity

> E> Underage smoking

Jan-Mar 2004 Sept-Oct 2004Survey Period

mately 62% of respondents in the firstsurvey and 66% of respondents in thesecond survey had attended at least somecollege. Approximately 57% of respondentsin both surveys reported incomes of$50,000 or more. Nearly 40% of respon-dents in both the first survey and thesecond survey reported having at leastone child at home. The distributions ofage and race/ethnicity were very similarbetween the 2 surveys.

It is important to note that 18.7% ofrespondents in the first survey, and 15.1%of respondents in the second survey ei-ther refused to report or did not knowtheir household income. Cross-tabula-tion analyses were conducted to examinepotential associations between demo-graphic character is t ics (ie, race/ethnicity, gender, and education) andthose who reported their income versusthose who did not report their income. Nosignificant associations were found. Thus,the subsample of respondents who did notreport their income was no different demo-graphically from the larger sample of re-spondents who did report their income.

Perceptions of Childhood ObesityIn the first survey, over 41% of Ameri-

cans perceived childhood overweight andobesity to be a very serious problem, simi-

lar to tobacco use (42%) but not as seriousas drug abuse (55%, P<0.05). In the sec-ond survey, over 47% of Americans per-ceived childhood overweight and obesityto be a very serious problem. This repre-sents a statistically significant increase(P<.01) compared to the first survey. Noother youth health threat increased sig-nificantly between the 2 surveys. In thesecond survey, respondents consideredchildhood obesity to be more serious thanunderage smoking, statistically equal tounderage drinking and teen sex, but notas serious as drug abuse and youth vio-lence (Table 2). Difference-in-differenceregressions further supported these find-ings. Controlling for the overall change inthe other youth issues, the increase inthe perceived seriousness of childhoodobesity remained significant, p = -.06, t =-2.95, P<.01 (Figure 1). Comparing thechange in the perceived seriousness ofchildhood obesity to the changes in per-ceived seriousness of each of the otheryouth issues independently revealed asignificant increase for childhood obesityin comparison to teen sex, p = .06, t = 2.02,P<.05. All other comparisons were notsignificant.

Childhood Obesity InterventionsPublic support for specific childhood

Am J Health Bebav.™ 2006;30(2):167-176 171

Childhood Obesity Epidemic

Table 3Support for Childhood Obesity

Prevention InterventionStrategies - Comparison of

1st 8B 2nd Surveys

Intervention

School Vending MachinesIncreasing promotion andmarketing of healthy foodsand drinks in school vendingmachines

Percent (%) Favor(N)

Jan-Mar Sept-Oct(n=1047) (n=1139)

85.39(894)

Increasing cost of less healthyfoods and drinks in school 45.32vending machines (474)

Allowing only the sale ofhealthy foods and drinks inschool vending machines

Removing all vendingmachines from schools

School CafeteriasIncreasing the cost of lesshealthy foods and drinks inschool cafeterias

Restricting the availabilityof less healthy foods anddrinks in school cafeterias

Allowing only the sale ofhealthy foods and drinks inschool cafeterias

School CurriculumRequiring more physicaleducation classes in school

Requiring schools to teachstudents healthy eating andexercise habits

Evaluating Children'sWeight in SchoolsRecording students' weighton a regular basis

70.89(738)

35.92(374)

44.00(455)

74.51(769)

67.44(696)

82.30(846)

93.86(963)

49.51(508)

(continued next

86.57(986)

45.22(515)

72.17(822)

35.29(402)

42.84(488)

73.75(840)

70.85(807)

84.20(959)

93.85(1069)

52.33(596)

column)

Table 3 (continued)

Intervention

Sending parents a healthreport card of their childrenweight on a regular basis

Providing students whoare obese with weight-lossand exercise programs inschool

Promotion of Fast Foodand Other Less Healthy

Percent (%) Favor(N)

Jan-Mar(n=IO47)

s 57.11(586)

72.90***(748)

Food Marketed to ChildrenIncreasing the tax on fastfood and less healthy foodsmarketed to children

Restricting the amount offast food and less healthyfood advertisements during

39.10(400)

75.27children's television programs (770)

Prohibiting the advertisingand promotion of fast foodand less healthy foodsmarketed to children

Restaurant PoliciesPlacing health ratings onrestaurant menus to indicatehealthy choices

Requiring restaurantmenus to list nutritioninformation about foodproducts, similar to theNutrition Facts panel onfood packaging

Requiring standardizedfood portions in restaurants

Note.Significant differences in

47.89**(490)

82.04(859)

67.34(705)

42.31**(443)

percentagebetween Jan-Mar 2004 and Sept-Octshown in bold.

***P< 0.001**P<0.01* P< 0.05

Sept-Oct(n=n39)

58.30(664)

79.81***(909)

36.79(419)

77.17(879)

52.77**(601)

85.60(975)

75.17(787)

48.90**(557)

favor2004 are

172

Evans et aJ

interventions is presented in Table 3.The Percentage Favor column includesthose who either strongly support or sup-port the intervention. We observed verysimilar levels of support between the 2surveys, with 3 exceptions. First, supportfor providing students who are obese withweight-loss and exercise programs inschool increased significantly (P<.05) fromnearly 73% to nearly 80%. Second, sup-port for prohibiting the advertising andpromotion of fast food and less healthyfoods marketed to children increased sig-nificantly (P< .001) from nearly 48% tonearly 53%, thus representing majoritysupport. Third, support for requiring stan-dardized food portions in restaurants in-creased significantly (P<.01) from approxi-mately 42% to 49%.

Interventions based on increasing thecost of unhealthy foods were not wellsupported and often opposed in both sur-veys. However, the significant increasesin support for advertising prohibitionsand standardized food portions representincreased support for regulation. Regula-tory strategies continued to have mixedsupport in the second survey, but theconsistent pattern of weak support oropposition observed in the first surveyclearly changed. The previously observedpattern of opposition to cost and regula-tory strategies has been discussed indetail elsewhere.^

Logistic Regression AnalysesFor 5 of the 15 specific interventions in

the second survey, logistic regressionsrevealed that support was statisticallygreater among more educated respon-dents. These results are similar to thefindings for the first survey, in which 8 of18 specific interventions revealed greatersupport among more educated partici-pants. None of the regressions for thesecond survey revealed significantly lowersupport among the more educated. Re-sults of the first survey, however, didreveal that more educated respondentswere significantly less likely to supportweight evaluation in schools. In thesecond survey, more educated respon-dents had higher observed levels of sup-port for at least some interventions inevery area except weight evaluation, al-though none of these differrences weresignificant.

Similar to the first survey, women weremore likely to support school vending.

school cafeteria, and marketing inter-ventions than were men, but were lesslikely to support weight evaluation. For 6of the 15 regressions, the odds ratios werepositive and statistically significant forwomen, ranging from 1.29 (prohibit lesshealthy food ads during children's televi-sion programs) to 1.79 (allow only sale ofhealthy foods in vending machines). Inthe first survey, there were no cases inwhich women were significantly less likelyto support an intervention than men were.In the second survey, however, womenwere significantly less likely than men tosupport recording students' weight on aregular basis and sending parents a healthreport card of their children's weight on aregular basis.

Having a higher income was associ-ated with higher support for 5 of the 15interventions and with lower support for 3of the 15. Similar to the first survey, therewas no consistent pattern within the in-come analysis by category of interven-tion. However, findings for income weredirectly correlated with those for educa-tion: higher income respondents hadhigher levels of support for the same 5intervention strategies supported by moreeducated respondents. This is also con-sistent with the first survey.

Finally, respondents with children athome were less supportive of all forms ofweight evaluation in schools than werethose without children at home, similarto the first survey. The only interventionin the second survey supported more bythose with children compared to thosewithout was increased promotion ofhealthy foods in vending machines.

DISCUSSIONThis study confirms that there is strong

and growing public support for interven-tions aimed at reducing overweight andobesity among children and adolescents.Public support evolved in the 9 monthsbetween the first and second RTI surveys.Perceived severity of the childhood obe-sity health threat increased. At the sametime, the public increasingly supportsregulatory strategies to prevent child-hood obesity.

Overall, the US public appears increas-ingly concerned about childhood obesityand wants action by governmental andother public health organizations to com-bat it. These data are likely to be of greatinterest to policy makers considering spe-

Am J Health Behav.™ 2006;30(2):167-176 173

Childhood Obesity Epidemic

Table 4Support for Childhood-Obesity Prevention Interventions

(Odds Ratios and Confidence Intervals From Logistic RegressionAnalyses) - Second Survey

Intervention

School Vending MachinesIncrease promotion of

healthy foodsIncrease cost of less

healthy foodsAllow only the sale

of healthy foodsRemove all vending

machines from schools

School CafeteriasIncrease cost of less

healthy foodsRestrict availability of

less healthy foodAllow only tiie sale of

healthy foods

School CurriculumRequire more physical

educationRequire teaching of

healthy eating & exercise

Weight Evaluation inSchoolsRecording weight on

regular hasisSend parents a health

report card of children'sweight on regular basis

Provide obese students withweight-loss & exerciseprograms

Marketing of LessHealthy FoodsIncrease tax on less healthy

foods marketed to kidsRestrict less healthy food ads

during kid's TV programsProhibit less healthy food

ads marketed to kids

Education(No collegeV!i at least

some college)

0.42*

0.62*

0,90

0.82

082

0.44*

0.90

0.85

0 4 4 '

0.86

111

1.21

1.16

0,60*

1.17

(0.29, 0.62)

(0.48, 0.80)

(0.68, 1.20)

(0.63, 1.07)

(0.63, 1.06)

(0.34, 0.59)

(0.68, 1.19)

(0.58, 1.23)

(0,26, 0.77)

(0.66, 1.10)

(0.86, 1.43)

(0.87, 1.70)

(0-90, 1.50)

(0.45, 0.81)

(0.91. 1.51)

1

Demographics

Gender(Female vs

1.64*

1.18

1.79*

1,16

0,95

1.42*

1.58*

0.98

0.78

0.72*

0.70*

0.90

1.29

1.60*

1.29*

Male)

(1.12, 2.40)

(0,92, 1.52)

(1.36, 237)

(0,90, 1.51)

(0-74. 1.23)

(1.07, 1.88)

(1.21. 207)

(0.67, 1.43)

(0.43, 1.42)

(0.56, 0.93)

(0.54, 0.91)

(0.64, 1,25)

(0.99, 1.67)

(1 19, 2 16)

(1 01. 1 66)

;Income(<$50.000

vs$50,000+)

0.23*

0.6!*

l.Ol

1.24

0-94

0.53*

1.03

0.77

0.40*

1.19

1-44*

1,44*

1.37*

0-71*

L27

(0.14, 0.38)

(0.47, 0.79)

(0.75, 1-35)

(0.95, 1.61)

(0.73, 1.22)

(0.39, 0-72)

(0.78, 1.37)

(0 .51 . 1.14)

(0.20, 0.78)

(0,92, 1.54)

(1.11, 1.88)

(1.02 2.03)

(1.05. 1.78)

(0.52, 0.98)

(0.98, 1.64)

Childrenunder 18(None vs

One or more)

0.58*

0.82

0.84

1,11

0-89

0.75

1-18

0,90

1.08

1.44*

1,64*

1.43*

0.93

0.91

0.99

(0.38, 0.89)

(0.64, 1.06)

(0.63, 1-12)

(0.86, 1.44)

(0.69, 1,14)

(0.55, 1.01)

(0.90. 1.55)

(0.61, 1-32)

(0.61, 1.92)

(1.12, 1,85)

(127. 2.11)

(1-03. 1.96)

(0,72, 1.20)

(0.67. 1-23)

(0.77. 1.27)

Note.Intervention items were coded 0 = oppose and i = support. Demographic characteristics were coded asfollows: Education: 0 = some college, 2-year college degree, 4-year college degree, or postgraduatestudy and 1 = less than high sehool, high school degree, or GED; Gender: 0 = Male and 1 = Female;Income: 0 = S50,000+ and I = less than S50,000; and Children under 18: 0 = none and I = one or more.Thus, the odds ratios ean be interpreted as the characteristic coded as 1 (the denominator) as comparedto the characteristic coded as 0 (the numerator).

'̂ Odds ratio significantly different from I (P<O.OS).

174

Evans et al

cific interventions.In terms of regulation, compared to the

first survey, significantly higher percent-ages of respondents in the second surveysupported "prohibiting the advertising andpromotion of fast food and less healthyfoods marketed to children" and "requir-ing standardized food portions in restau-rants." These changes suggest that in-creased public attention to obesity-whichcould be fostered by factors such as in-creased news coverage of the topic, pro-motion and marketing of specialized dietprograms (eg, Atkins), government warn-ings about rising obesity rates, and de-bate about potential solutions-may be rais-ing public concern and therefore vialling-ness to attempt to solve the problemthrough government regulation. Theseobserved trends deserve further investi-gation as they have potential to influencethe policy debate on obesity prevention.

The consistency in levels of support forand opposition to specific childhood-obe-sity interventions suggests that policymakers should take the public's supportfor obesity prevention seriously. However,the observed increases in support for pro-hibiting advertising of less healthy foodsand regulating portion sizes in restau-rants have additional policy implications.The increase in support for these strate-gies, both of which would require signifi-cant governmental regulation, suggeststhat increasing public attention to child-hood obesity may be heightening the de-sire for strong and decisive action and anincreased public willingness to limit per-sonal choice (for children in the form oftelevision choice and for everyone in re-gard to restaurant food choice).

At the same time, support for tax-basedinterventions to increase the price of lesshealthy foods (eg, fat tax) remained con-sistently low. So although increased pub-lic attention to childhood obesity mayaffect the public's intervention priorities,strategies that involve taxation remainlargely unsupported.

Policy ImplicationsThe public clearly wants to reduce un-

healthy, and increase healthy, food con-sumption among children and adoles-cents. They support reducing the avail-ability of low nutrient density/high calo-rie (ie, junk or fast food) independentlyand in relation to healthier foods. Themost consistently supported venue in

which to improve food consumption isschool. However, they are wary about ac-complishing these goals through pricecontrols or taxation. Interventions thathave been implemented on a local level,such as restricting vending-machine ac-cess or decreasing the number of un-healthy food choices available in schools,continue to enjoy broad public support andshould be further considered for imple-mentation.

Other more invasive approaches, suchas prohibiting advertising and promotionto adolescents, enjoy growing support. Al-though support for intensive regulatorystrategies in general is mixed, it appearsto be increasing. Based on findings from aperiod of heightened visibility for child-hood obesity, the public appears increas-ingly supportive of strong measures tocombat it.

Support for weight-loss and exerciseprograms in schools, which would mostlikely be accomplished through physicaleducation classes and intramural activi-ties, increased between our 2 surveys. Atthe same time, as noted above, previousresearch has showed that the public isnot willing to sacrifice basic education.Given this finding, school policy makersmay need to explore alternative strate-gies, such as decreasing elective time orcombining health and physical educationclasses, to find time for increased physi-cal Eind health education without decreas-ing time for standard subjects.

Policy makers should also recognizethe importance of nutrition and fitness inchildren's educational progress andachievement.^^'^^ Education and healthare not only compatible, but mutuallyreinforcing and essential to child andadolescent development.

This survey examined public supportfor specific interventions aimed at reduc-ing childhood overweight and obesity andchanges in support during a time ofheightened intervention activity and pub-lic attention. It showed that public con-cern and support for strong, in some caseshighly invasive, action is grovnng. It pro-vided information concerning which in-terventions might be feasible from a pub-lic policy perspective and showed thatthere are both consistent patterns of sup-port and some clear trends in which in-terventions are most supported. This in-formation vrill be valuable for policy mak-ers as they set their programmatic and

Am J Health Behav.^ 2006;30(2):167-176 175

Childhood Obesity Epidemic

funding agendas in the near future.However, as noted in several recent

studies, relatively little scientific infor-mation exists on which of these interven-tions are effective in actual school, com-munity, and media settings.^'' As calledfor in the recent Institute of Medicine(lOM) report Preventing Childhood Obe-sity: Health in the Balance, more re-search on interventions and the basicpsychosocial processes underlying themand their policy implications is needed.One implication is the need for policy mak-ers, researchers, and practitioners to col-laborate in developing shared and inte-grated policy, programmatic, and researchagendas to fund, develop, test, implement,and evaluate evidence-based interventionsto prevent childhood obesity.

AcknowledgmentsThis research was supported by an

internal grant from Research TriangleInstitute International. •

REFERENCESl.Mokdad A. Marks E. Stroup E, et al. Actual

causes of death in the United States, 2000.JAMA. 2004;291:1238-1245.

2.FinkeIstein E, Fiebelkom I, Wang G. Nationalmedical spending attributable to overweightand obesity: how much, and who's paying?Health Aff 2003;22(3):219-226.

3.U.S Department of Health and Human Ser-vices, Centers for Disease Control and Pre-vention, National Center for Health Statistics.Health, United States, 2003. Washington,DC. Available from the US Government Print-ing Office. (Cited as CDC 2003.).

4.McKay D. Obesity: CDC study inflated relateddeaths by tens of thousands'. Wall StreetJournal. 11/23/2004. Available at: http://www.wsj.com. Accessed on April 20, 2005.

5.0gden CL, Flegal KM, Carroll MD. et al.Prevalence and trends in overweight amongUS children and adolescents, 1999-2000.JAMA. 2002;288:1728-1732.

6.Evans WD, Finkelstein E, Kamerow D, et al.Public perceptions of childhood obesity. Am JPrevent Med. 2005;28(l):26-32.

7.Sallis JF, Prochaska JJ, Taylor WC, et al.Correlates of physical activity in a nationalsample of girls and boys in grades 4 through12. Health Psychology. 1999;18(4):410-415.

8.Centers for Disease Control and Prevention.WISEWOMAN program. Available at: http://www.cdc.gov/wisewoman/. Accessed onJune 12, 2005.

9.U.S. Department of Health and Human Ser-vices. Steps to a Healthier US program. Avail-

able at: hUp://www.healthienjs.gov/steps/.Accessed on June 12, 2005.

10-Maibach E. Designing Health Messages: Ap-proaches from Communication Theory andPublic Health Practice. Sage Publications:Thousand Oaks, CA., 1995:1-320.

11.National Cancer Institute. 5 A Day for BetterHealth Program Evaluation Report. Availableat: http://www.cancercontrol.cancer.gov/5ad_exec.html. Accessed on June 12, 2005.

12.U.S. Department of Health and Human Ser-vices. Small Step Program. Available at: http://www.smallstep.gov/. Accessed on June 12,2005.

13.Huhman M, Heitzler C, Wong F. The VERB{tin)campaign logic model: a tool for planning andevaluation. Prev Chronic Dis. Available at:http://www.cdc.gov/pcd/issues/2004/jul/04_0033.htm. Accessed on June 12, 2005.

14.Centers for Disease Control and Prevention.Verb: It's what you do campaign. Available at:http://www.cdc.gov/youthcampaign/. Ac-cessed on June 12, 2005.

15.Robinson TN. Television viewing and child-hood obesity. Pediatric C/inics of North America.2001;48(4):1017-1025.

le.Farquhar JW, Fortmann SP, Flora JA, et al.Effects of communitywide education on car-diovascular disease risk factors. The StanfordFive-City Project. JAMA. l990;264{3):359-365.

17.Evans D, Wasserman J, Bertolotti E, et al.Branding behavior: the strategy behind thetruthsm Campaign. Social Marketing Quarterly.2002:3:17-29.

18.The Time/ABC News Summit on Obesity.Available at: http:/ /www. time, com/time/2004/obesity/. Accessed on April 20, 2005.

19.Newman C. Why Are We So Fat? NationalGeographic. Available at: http:/ /magma.nationalgeographic. com/ngm/0408/feature3/. Accessed on April 20, 2005.

2O.Biener L, Flynn CA, Gilpin EA, et al. Impactof declining survey response rates on esti-mates of smoking prevalence. Prevent Med.2004;27(3):254-257.

21.American Association of Public Opinion Re-search. Guidelines and Tools for Document-ing Survey Results. Available at: http://aapor.org/AAPOR_counciI_guidelines.pdf.Accessed on June 12, 2005.

22.U.S. Department of Health and Human Ser-vices. The Surgeon General's Call to Action toPrevent and Decrease Overweight and Obe-sity. 2001:1-39.

23.Sallis JF, McKenzie TL, Kolody B, et al.Effects of health-related physical educationon academic achievement: Project SPARK.Research Quarterly for Exercise and Sport.1999:70:127-134.

24.Institute of Medicine. Preventing ChildhoodObesity: Health in the Balance. Institute ofMedicine of the National Academies; Wash-ington, DC. 2004:1-461.

176