Childhood Obesity:  An Economic Perspective

work may be reproduced in whole or in part for study or training purposes, subject to the

inclusion of an acknowledgment of the source. Reproduction for commercial use or sale

requires prior written permission from the Commonwealth. Requests and inquiries

concerning reproduction and rights should be addressed to the Commonwealth Copyright

Administration, Attorney-General’s Department, National Circuit, Canberra ACT 2600 or

posted at www.ag.gov.au/cca.

Media and Publications

Productivity Commission

Locked Bag 2 Collins Street East

Melbourne VIC 8003

Tel: (03) 9653 2244

Fax: (03) 9653 2303

Email: maps@pc.gov.au

Tel: (03) 9653 2100 or (02) 6240 3200

Productivity Commission Staff Working Paper, Melbourne.

The Productivity Commission is the Australian Government’s independent research

and advisory body on a range of economic, social and environmental issues affecting

the welfare of Australians. Its role, expressed most simply, is to help governments

make better policies, in the long term interest of the Australian community.

The Commission’s independence is underpinned by an Act of Parliament. Its

processes and outputs are open to public scrutiny and are driven by concern for the

wellbeing of the community as a whole.

Further information on the Productivity Commission can be obtained from the

Commission’s website (www.pc.gov.au) or by contacting Media and Publications on

(03) 9653 2244 or email: maps@pc.gov.au

Contents

1.2 Approach of this paper 3

1.3 Childhood overweight and obesity prevalence in Australia 4

2.1 Decision making 11

2.3 The costs of obesity 27

3.1 Framework for possible causes of overweight and obesity 35

3.2 Child characteristics and behaviours 38

3.3 Parenting styles and family characteristics 44

3.4 Community, demographic and societal characteristics 46

3.5 Summary and conclusion 51

4.1 Potential policy measures 56

4.2 Summary 66

5.1 The evidence base and obesity prevention 69

5.2 Implications for policy 85

5.3 Conclusion 89

1.1 The health consequences of childhood obesity 2

1.2 Measuring childhood obesity 5

2.1 Externalities, information failures and behavioural limitations 17

2.2 The efficiency cost of taxation 21

2.3 Access Economics’ estimates of the costs of obesity 32

4.2 Self-regulation: the Responsible Children’s Marketing Initiative 64

5.1 Evidence-based policy 70

5.2 Interventions for preventing obesity in children (review) 72

1.1 The approach of this study 4

1.2 Weight classification of Australian children aged 5–17 years 7

1.3 Weight classification of Australian children 9

1.4 Weight classification of Australian children aged 2–18 years 10

2.1 Decision-making framework 13

2.2 The costs of obesity: is childhood obesity a public problem? 28

3.1 Framework for factors associated with obesity and overweight 36

3.2 Potential factors in the rise of childhood overweight and obesity 53

4.1 What is the scope for policy solutions to reduce childhood obesity? 67

3.1 Summary of evidence presented on factors associated with

overweight and obesity 37

5.1 Be Active Eat Well 81

5.2 Characteristics of different policy interventions to prevent childhood

obesity 91

A.1 Active After-schools Communities (AASC) program 96

A.2 Be Active Eat Well 98

A.3 Burke et al. 1998 99

A.4 Coalfields Healthy Heartbeat School Project 100

A.5 Dwyer et al. 1983 101

A.6 Eat Smart Play Smart 102

A.7 Engaging adolescent girls in school sport 103

A.8 Fitness Improvement and Lifestyle Awareness (FILA) program 104

A.9 Fresh Kids 105

A.10 ‘Get Moving’ Campaign 106

A.11 Girls Stepping Out Program 107

A.12 ‘Go for 2&5®’ Campaign 108

A.13 Learning to Enjoy Activity with Friends (LEAF) 109

A.14 Moorefit 110

A.15 Move It Groove It 111

A.16 New South Wales Walk Safely to School Day 112

A.17 Nutrition Ready-to-Go @ Out of School Hours Care (NRG @

OOSH) Physical Activity Project 113

A.18 Program X 114

A.19 Q4: Live Outside the Box 2007 115

A.20 Romp and Chomp 116

A.21 Smart Choices – Healthy Food and Drink Supply Strategy for

Queensland Schools 117

A.22 Start Right-Eat Right Award Scheme 118

A.23 Stephanie Alexander Kitchen Garden Program 119

A.24 Switch–Play 120

A.24 (continued) 121

A.25 Tooty Fruity Vegie Project 122

A.25 (continued) 123

A.26 Vandongen et al. 1995 124

A.27 Wicked Vegies 125

B.1 ACT Early Childhood Active Play and Eating Well Project 128

B.2 ACT Health Promoting Schools Canteen Project 128

B.3 Activate NT - MBF Healthy Lifestyle Challenge 129

B.4 Active-Ate 129

B.5 Active School Curriculum Initiative 129

B.6 Around Australia in 40 Days Small Steps to Big Things Walking

Challenge 130

B.7 Be Active – Take Steps 130

B.8 Childhood Healthy Weight Project 130

B.9 Cool CAP 131

B.10 Crunch&Sip 131

B.11 CSIRO Wellbeing Plan for Children 131

B.12 Eat Right Grow Bright 132

B.13 Eat Well Be Active 132

B.14 Eat Well Be Active, Healthy Kids for Life – Badu Island 132

B.15 Eat Well Be Active, Healthy Kids for Life – Logan-Beaudesert 133

B.16 Eat Well Be Active, Healthy Kids for Life – Townsville 133

B.17 Family Food Patch 133

B.18 Filling the Gaps 134

B.19 Fit4fun 134

B.20 Fitness Improvement and Lifestyle Awareness (FILA) Program 134

B.21 Foodbank School Breakfast Program 135

B.22 Free Fruit Friday 135

B.23 Fresh Tastes @ School – NSW Health School Canteen Strategy 135

B.24 Fruit ‘n’ Veg Week 136

B.25 Fun ‘n’ Healthy in Moreland 136

B.26 Get Set 4 Life – Habits for Healthy Kids Guide 137

B.27 Get Up & Grow: Healthy Eating and Physical Activity Guidelines

for Early Childhood 137

B.28 Girls in sport 138

B.29 Go for 2&5 Campaign 138

B.30 Go4Fun 138

B.31 GoNT 139

B.32 Good for Kids, Good for Life 139

B.33 Growing Years Project 139

B.34 Health Promoting Communities: Being Active Eating Well 140

B.35 Healthy Beginnings Study 140

B.36 Healthy Dads Healthy Kids 141

B.37 Healthy eating and obesity prevention for preschoolers: A

randomised controlled trial 141

B.38 Healthy food and drink 141

B.39 Healthy Kids Check 142

B.40 It’s Your Move 142

B.41 Jack Brockhoff Child Health and Wellbeing Program 142

B.42 Jump Start 143

B.43 Just add fruit and veg 143

B.44 Kids – ‘Go for your life’ 143

B.45 Kids GP Campaign 144

B.46 Live Fit 144

B.47 Live Life Well @ School 144

B.48 Make Tracks to School 145

B.49 Many Rivers Diabetes Prevention Program 145

B.50 Mend 2-4 145

B.51 Move Well Eat Well 146

B.52 Munch and Move 146

B.53 NOURISH 146

B.54 Nourish-the-FACTS 147

B.55 Obesity Prevention and Lifestyle (OPAL) 147

B.56 Osborne Division of General Practice’s Obesity Program – Healthy

Families for Happy Futures 147

B.57 Parental Guidance Recommended Program 148

B.58 Physical Activity and Nutrition out of School Hours (PANOSH) 148

B.59 Physical Activity in Culturally and Linguistically Diverse

Communities 148

B.60 Physical Activity Leaders Program 149

B.61 Play5 149

B.62 Premier’s Active Families Challenge 149

B.63 Remote Indigenous Stores and Takeaways (RIST) Project 150

B.64 Right Bite 150

B.65 School’s Out – Open Playground Program 150

B.66 StarCAP & StarCAP2 151

B.67 Start Right Eat Right award scheme 151

B.68 Start Right Eat Right - Tasmania 151

B.69 Start Right Eat Right - Victoria 152

B.70 Stephanie Alexander Kitchen Garden National Program 152

B.71 Stephanie Alexander Kitchen Garden Program – Victoria 152

B.72 Streets Ahead 153

B.73 Talk about weight 153

B.74 The Melbourne Infant Feeding Activity and Nutrition Trial

(InFANT) Program 154

B.75 The Responsible Children’s Marketing Initiative 154

B.76 Time2bHealthy 154

B.77 Tooty Fruity Vegie in Preschools 155

B.78 Transferability of a Mainstream Childhood Obesity Prevention

Program to Aboriginal People 155

B.79 TravelSMART Schools 155

B.80 Tuckatalk 156

B.81 Unplug and Play 156

B.82 WA Healthy Schools Project 156

B.83 Walktober Walk-to-School 157

B.84 Walk safely to school day 157

B.85 Walking School Bus 158

B.86 Wollongong Sport Program 158

• The weight of Australian children has increased markedly in recent decades, to the point where around 8 per cent are defined as obese (based on Body Mass Index), and 17 per cent as overweight.

• While the prevalence of obesity may have levelled off since the mid 1990s, it is still widely considered to be too high.

• Childhood obesity has been linked to a raft of physical and psychosocial health problems, including type 2 diabetes and cardiovascular disease, as well as social stigmatisation and low self-esteem.

• Simply put, obesity results from an imbalance between energy consumed and expended. But the underlying causes are complex and difficult to disentangle.

– An economic perspective considers how individuals respond to changes in incentives, and how they make decisions involving tradeoffs between different consumption and exercise choices, including how they spend their time. 

• Governments need to consider a range of issues in addressing childhood obesity.

– Most of the costs of obesity are borne by the obese themselves and their families.

– Market incentives to provide information about the causes and prevention of obesity are weak, creating a role for government. But unlike alcohol and tobacco consumption, the externalities (spillovers on unrelated third parties) associated with obesity are probably minor.

– Behavioural limitations can influence how people use available information about preventing obesity — even when it is available — and their responses to incentives and tradeoffs. Children are particularly susceptible to these limitations and have difficulty taking into account the future consequences of their actions.

– Obesity prevalence varies across the socioeconomic profile of the community, such that there can be important distributional issues.

– The obese also consume a disproportionate share of medical services, which, equity considerations aside, adds to the costs of our public health system.

• There is only limited evidence of interventions designed to address childhood obesity achieving their goals.

– This could reflect the inherent complexities and the multiple causes of obesity.

– But it might also reflect poor policy design and evaluation deficiencies.

• Notwithstanding the lack of evidence of interventions reducing obesity, some studies suggest that they can positively influence children’s eating behaviours and levels of physical activity, which in turn might influence obesity over time.

• The complex nature of the problem suggests that policies need to be carefully designed to maximise cost-effectiveness, and trialled, with a focus on evidence gathering, information sharing, evaluation and consequent policy modification.

In Australia, as in many other countries, the community has become increasingly

concerned about the rising prevalence of childhood obesity (box 1). The raft of

health consequences for obese children now, and particularly when they are adults,

has provided impetus for increased interest in the role for government in obesity

prevention strategies.

The prevalence of childhood obesity in Australia began increasing in the 1970s, and

by 2007-08 around 8 per cent of children (5 to 17 year olds) were estimated to be

obese, and 17 per cent overweight (figure 1). Most of the growth in childhood

obesity occurred up until the mid 1990s, with recent research suggesting it might

have levelled off sometime since 1995. Despite government and community focus

on obesity prevention over recent decades — as far back as the 1980s the ‘Life. Be

in it’ campaign sought to influence levels of activity among Australians — the

medical literature suggests the current level of obesity among Australian children is

too high.

Box 1 What is obesity?

The concern with childhood obesity arises from its association with poor

psychological and social outcomes, as well as physical health problems in the short

and long term. For example, obesity in children is linked with reduced self-esteem

XIV CHILDHOOD OBESITY

Figure 1 Proportion of children classified as overweight or obese

a Ages 7–15 years. b Ages 5–17 years.

and depression, and obese children can suffer from social discrimination. The range

of physical health problems associated with childhood obesity includes type 2

diabetes, liver disease, impaired mobility, asthma, sleep apnoea, and risk factors for

cardiovascular disease. Although some of these health problems are increasingly

being seen in children, most health problems arise in later life. International and

Australian research indicates that overweight and obese children are at higher risk

than normal weight children of becoming overweight and obese adults.

The costs of obesity appear to be substantial but are borne mostly by the obese. For

example, Access Economics estimated the total cost of obesity in 2008 to be

$58 billion, comprising $50 billion in lost wellbeing and $8 billion in financial costs

(such as productivity costs, health system costs, carer costs and transfer costs)

(box 2). These estimates reflect the disease burden in 2008, which predominantly

relates to adults. But to the extent that today’s obese children become obese adults,

these costs can also be thought of as the potential future costs for today’s obese

children.

OVERVIEW XV

Box 2 Access Economics’ estimates of the costs of obesity, 2008

• health system costs (such as hospital and nursing home costs, GP and specialist

• productivity losses

• carer costs

• transfer costs (that is, the deadweight loss from the higher level of taxation)

• other indirect costs (such as aids, modifications and travel).

The obese themselves (and their parents) bear most of the costs of obesity, primarily

through the loss of wellbeing (due to disability or shorter life span), but also due to

them bearing some of the financial costs. Overall, Access Economics estimated that

obese people bear 90 per cent of the total costs and 30 per cent of financial costs

arising from their obesity.

XVI CHILDHOOD OBESITY

Estimating the costs of obesity relies on a range of assumptions — including the

proportion of diseases attributable to obesity and the effects of those diseases on

obese people’s quality of life. Some assumptions are more robust than others, with

estimates of the value of lost wellbeing being particularly sensitive to the value

placed on a human life.

Understanding the causes of childhood obesity is important for explaining the

changes in prevalence, and considering if it is a problem that requires, and is

amenable to, government intervention.

In simple terms, obesity results from an imbalance between energy consumed and

expended. But there is a complex web of factors that affect weight outcomes in

children, some of which might account for the increased prevalence of obesity

(figure 2). Significantly, not all factors that affect children’s weight outcomes will

be completely within their control, and decisions about eating and exercise are not

made exclusively with weight in mind. Although there is a genetic component to

obesity, the literature suggests this of itself is not likely to explain the recent rise in

its prevalence.

OVERVIEW XVII

Figure 2 Potential factors in the rise of child obesity

Some broad trends that might have influenced eating and exercise decisions —

increasing the prevalence of overweight and obesity since the 1970s — include real

declines in food prices, rising incomes, increasing costs of exercise, and the higher

‘time cost’ of preparing food at home coupled with the increased availability of and

access to pre-prepared and takeaway meals. Some authors suggest that these

influences may well have combined with genetic programming that encourages

humans to store fat in times of plenty to prepare for possible future famines (that

have so far not eventuated). Other factors that can affect weight outcomes include

behavioural factors such as dietary mix, family characteristics (such as nutritional

knowledge), and broader cultural and community practices, which can be related to

demographic and societal characteristics (such as socioeconomic status and

ethnicity). Changes in societal attitudes about body image may have also had an

influence.

In many cases the evidence of the links between these factors and childhood obesity

is ambiguous, confounding or non-existent. The following broad conclusions are

drawn:

• Australian children’s energy intake appears to have risen since the 1980s.

• Food consumption

• Soft drink

• Physical activity

− Sport

− Walking to school

• Sedentary activities

− Television

− Computer/video

• Parents and families

− Income

− Parents’ work habits

• Advertising

• Physical environment

− Urban sprawl

− Access to

− Fast food outlets

− Safety

• Knowledge

• School environment

• Peer behaviours and

XVIII CHILDHOOD OBESITY

– Australian studies have shown a link between soft drinks, the associated

increase in energy intake, and childhood obesity — although the size of the

effect is small.

• Incidental exercise among Australian children appears to have declined,

• Children spend more time watching television and using computers and video

• Australian children are exposed to a relatively high number of advertisements

For most adults and children, being overweight or obese is a consequence of a

number of decisions over many years regarding food consumption and exercise. An

economic perspective considers the decision making framework — by taking into

account how individuals (children and parents) respond to changes in incentives,

and how they make decisions involving tradeoffs between different consumption

and exercise choices, including how they spend their time.

In the case of eating decisions, individuals might consider (even if only intuitively)

the potential benefits with the potential costs of eating certain foods. For example,

the benefits include intake of life sustaining calories and nutrients, and other

benefits such as the (short-term) pleasure of taste or of sharing a meal, which is

influenced by family, community or societal characteristics. The costs of eating

include the financial cost, the time cost of purchasing and/or preparing food at

home, as well as the (longer term) future health effects of that consumption

decision. The time cost of purchasing and/or preparing food might be influenced by

family income and work arrangements, which in turn can influence children’s diet.

To the extent such decisions by individuals were able to take into account

appropriate information about the benefits and costs to themselves, as well as any

‘spillover’ effects imposed on the community, and were rationally made, the

resulting weight outcomes could be seen as ‘optimal’, providing little basis for

policy intervention. But there are several reasons why decisions may typically be

deficient in these respects.

OVERVIEW XIX

Choices about eating and exercise might be distorted if, for example, useful

information is not easily available (for example, about the nutritional or energy

content of foods). Outcomes may also be less than ideal from a community

perspective if an individual makes choices without taking into account costs

imposed on others (referred to as an externality or spillover effect).

But looking at the underlying causes of obesity and overweight there appear to be

few such ‘market failures’ relating to obesity in children (or adults). There might be

a role for government to ensure provision of certain information to help individuals

make decisions that increase their own wellbeing, but the abundance of

obesity-related information on nutrition and exercise already available suggests that

information gaps alone may not be ‘the’ problem. Further, there are few, if any,

externalities relating to obesity, unlike other health-related policy matters — such as

smoking or drink driving — where there are significant detrimental effects imposed

on other people (though see section on health system costs).

Another potential problem is that decisions may be ‘distorted’ because individuals

are unlikely to always fully account for the future health, financial and lifestyle

consequences of their actions, nor consistently value the associated costs and

benefits (box 3). Even where information is available, behavioural limitations may

prevent individuals from maximising their own wellbeing. For example, evidence

suggests that when eating in groups the amount people eat is influenced by how

much their peers eat rather than their own food intake needs. People also often make

decisions and act in a way that is against their own long-term best interests (such as

eating more food today but postponing the diet or exercise to counter that higher

consumption). This can be thought of as their ‘short-term self’ trumping their

‘long-term self’.

XX CHILDHOOD OBESITY

Box 3 Behavioural economics provides insights

The effects of behavioural limitations on children are likely to be greater than on

adults. Children might be more prone to peer group pressure, and have even more

difficulty in accounting for the future consequences of their actions. So children’s

vulnerability to behavioural limitations in decision making is twofold — they are

affected by the limitations to which their parents are subject, and by their own

limitations.

OVERVIEW XXI

Another potential reason for governments to reduce obesity in children is to

decrease the health care costs borne by the rest of the community. Although most of

the costs associated with obesity are borne by the obese themselves, universal

access to healthcare and community rating mean the health care costs of obese

people are subsidised by other taxpayers and private health insurance members. Of

course, the obese are not the only group in the community that benefits from

universal health care and community rating, but that does not lessen the argument

for taking action to decrease these costs where it would be practical and cost

effective to do so.

Governments can employ a range of policy tools including price instruments (such

as taxes or subsidies), helping consumers be better informed (education and

information), and regulatory measures that influence consumer or producer choices.

Ideally, policies should directly target the source of the problem — for example,

information ‘failures’ and behavioural limitations might warrant ‘softer’ style

interventions, such as information provision and education. Given that obesity

occurs more in some groups of the population than others, such as those in lower

socioeconomic groups, there are also important population targeting considerations.

Moreover, the complex, multifaceted causes of obesity — which are yet to be fully

understood — suggest that effective policy solutions are likely to involve a mix of

tools acting on a range of levels (such as child, family and school, and energy in,

energy out). But these tools should be targeted carefully at the causes of obesity and

focus on improving individual decision-making over the longer term. Measures that

constrain behaviour indiscriminately are rarely effective, equitable, or improve

community wellbeing. Bans or taxes on particular energy-dense nutrient-poor foods,

for example, face design difficulties, affect all consumers regardless of their weight

status, and in the case of taxes, can have perverse budgetary and health effects

particularly for the neediest groups.

Importantly, intervention will clearly only be desirable when it delivers a better

outcome than not intervening. Any policy options being contemplated should be

rigorously assessed to ensure they deliver net community benefits over time.

XXII CHILDHOOD OBESITY

Interventions to address childhood obesity include both interventions targeted at

reducing childhood obesity and overweight (experimental trials or studies)

conducted by governments and others (mostly in school settings), along with

broader, community-wide government interventions (such as taxes on energy-dense

foods or television advertising restrictions).

International evidence on the effectiveness of interventions includes systematic

reviews of targeted interventions, along with studies on community-wide

government interventions. One of these is a review on behalf of the Cochrane

Collaboration, generally regarded as an authoritative voice in systematic reviews in

healthcare. Many interventions intended to prevent childhood obesity do not appear

to have been effective in preventing weight gain to any significant degree, but show

promise in improving lifestyle behaviours (healthy diet and physical activity levels).

Better information is needed to understand if this might translate into improved

weight outcomes in the future.

Evidence on community-wide interventions is also mixed (box 4).

OVERVIEW XXIII

Box 4 Some international evidence on community-wide interventions

Australian interventions addressing childhood obesity are primarily of a targeted

kind, focusing on providing information, increasing education and influencing

physical activity. Few interventions list reducing or preventing obesity in children

among their stated objectives, although many seek to influence physical activity or

dietary awareness or both. Given that some of these have measured body

composition (such as BMI or waist circumference), they provide some insights into

how well they work in terms of reducing or preventing obesity.

In general, the interventions studied have had mixed success in improving body

composition. But in some cases they were successful in promoting other desirable

outcomes, such as increasing the level of physical activity. The results from some

other interventions were less positive (box 5). Further, long-term follow up to assess

the sustainability of outcomes has not been undertaken for many Australian

interventions.

XXIV CHILDHOOD OBESITY

Box 5 Success of selected Australian interventions

Be Active Eat Well was one of the first community-based interventions in Australia with

school. Be Active Eat Well delivered positive (short-term) results for most of the body

Switch–Play focused on physical activity through two components — behaviour

Switch–Play had a significant effect on BMI for the children participating in a combined

Engaging Adolescent Girls in School Sport aimed to increase physical activity by

On balance, the evidence to date suggests that while many interventions to prevent

childhood obesity show promise in improving lifestyle behaviours, they have not

been effective in stabilising or reducing obesity prevalence to any significant

degree. That said, methodological issues may affect the reliability of conclusions

drawn from some of the research.

Possible reasons why many interventions do not appear to have been effective

include:

• the complexity of the problem and its inherent challenges

OVERVIEW XXV

• policy tools might have been poorly designed, for example not properly targeting

• institutional and other constraints mean otherwise well-designed interventions

• effective policy tools might have been designed, but methodological flaws in

• experiment first with low-cost programs that have low risks of collateral impacts

• for programs with potentially high costs, initially implement trials to gather

• roll out programs gradually, which can allow continual evidence gathering and

• evaluate programs rigorously

• share information to enable wider learning from successes and failures.

INTRODUCTION 1

Being overweight or obese as a child has implications for the child’s health now and

as an adult. It is a policy concern in Australia and for governments internationally.

Preventative health policy aims to manage risk factors such as obesity to decrease

the incidence and effect of subsequent health problems. However, such expenditure

needs to be justified in terms of effectiveness and value for public money. Programs

to prevent and reduce childhood obesity can be difficult to design and implement

successfully, particularly given the complexity and multitude of different

determinants of obesity.

This paper analyses the issue of childhood obesity within an economic policy

framework. It also reviews the evidence of trends in obesity in children and

provides an overview of recent and planned childhood obesity preventative health

programs. In this introductory chapter we discuss the policy background in

Australia and outline the approach taken by this paper. Following this the estimated

prevalence of obesity in Australian children is reported and discussed.

Childhood obesity is associated with a range of health problems, emerging in

childhood and later adult life (box 1.1). These include psychosocial problems such

as social discrimination and reduced self-esteem, and physical health problems such

as type 2 diabetes and risk factors associated with cardiovascular disease.

The prevalence of childhood obesity in Australia has been increasing since the

1970s, particularly in the decade from the mid 1980s to the mid 1990s (Norton et

al. 2006). While government and community focus on obesity increased from the

mid 1990s, broader preventative health programs focusing on physical activity and

nutrition date from the 1980s (for example, the iconic ‘Life. Be in it’ campaign and

the National Better Health Program (Lin and Robinson 2005)).

In 1997, the National Health and Medical Research Council released Acting on

Australia’s Weight: a Strategic Plan for the Prevention of Overweight and Obesity

(NHMRC 1997).

2 CHILDHOOD OBESITY

Box 1.1 The health consequences of childhood obesity

• Cardiovascular risk factors — include high blood pressure, high cholesterol and

• Type 2 diabetes — previously mostly occurring in adults, the diagnosis rate in

• Hepatic complications — include hepatic steatosis (fatty liver disease) and

• Other complications — include orthopaedic complications such as Blount disease

INTRODUCTION 3

Box 1.1 (continued)

Australian Health Ministers agreed that overweight and obesity was a sufficiently

significant public health problem to establish a taskforce in November 2002.

Healthy Weight 2008 — Australia’s Future followed, stating the initial focus of the

This paper covers five main areas that are important in understanding childhood

obesity in Australia and why it might require government intervention (figure 1.1).

The causes of overweight and obesity in children are explored, including individual

characteristics, such as genetics, that give a predisposition to weight gain, the

behaviours of individuals, such as their food consumption and exercise, and

environmental factors that influence these behaviours, such as advertising and the

physical environment.

Understanding the causes is important for explaining the changes in the prevalence

of childhood obesity and considering if childhood overweight and obesity is a

problem that requires government intervention. Some of the costs of overweight and

obesity are borne by obese individuals themselves (such as the health consequences)

and other costs are borne by the community (such as shared costs under a public

health system). Both the causes and the consequences of overweight and obesity

need to be considered when assessing the problem and deciding if government

action is warranted. If there is a case for intervention, the next step is to identify

which individual characteristics and behavioural and environmental causes are

4 CHILDHOOD OBESITY

amenable to policy influence. Then, consideration needs to be given as to which

potential policy options can be used to target them.

When making policy decisions it is important to consider the available evidence or

any evidence of the effectiveness of previous interventions, and if the benefits

warrant the costs of undertaking the policy option, including the cost of taxation or

diverting funds from other programs, and costs that might be imposed on

consumers.

Figure 1.1 The approach of this study

Assessing the prevalence of obesity in the community requires appropriate

measurement and collection of data. The Body Mass Index (BMI) is the most

commonly used method to measure overweight and obesity in large

• Costs to the individual

− Health

− Social

− Work and income

• Are the causes open to influence by government?

• What are the policy options available to government?

• Which policy actions are effective?

• What are the benefits?

• What are the costs?

• Costs to the community

− Shared health costs

− Shared other costs

• Individual characteristics

• Behaviour

• Environment

INTRODUCTION 5

population-level surveys. There are advantages and disadvantages to using BMI

(box 1.2), however, it is the measure most commonly used in surveys, mainly due to

its ease of measurement.

Box 1.2 Measuring childhood obesity

Sources: Cole et al. (2000); Cole et al. (2007).

• it is easier, and considered less invasive, to measure height and weight in a school

• it is less costly

6 CHILDHOOD OBESITY

Box 1.2 (continued)

• some studies have shown BMI is correlated with body fat in children (for example,

• it does not give any indication of the distribution of fat on a person, such as central

• it does not quantify body composition — people with high muscle mass, but not

• females with slender frames but significant excess fat may record misleadingly low

• it is not fully independent of height, especially for children, and is influenced by body

• the cut off points for children (Cole et al. 2000) were estimated using mostly

The data on overweight and obesity in Australian children are limited and ‘patchy’.

(For the purposes of this paper we consider children as anyone under 18 years of

age.) There are very few national surveys and they are often at least a decade apart.

Since the mid 1990s, many states have completed at least one survey. However,

data sets will often include different age groups and have different methodologies.

Olds et al. (2009) presents a comprehensive list of Australian measured (as distinct

from self-reported) data sources.

Two recent national surveys of childhood overweight and obesity reveal similar

results, except for obesity rates for male children. According to the latest National

Health Survey (NHS) for 2007-08 (ABS 2009a), 75 per cent of children were not

overweight or obese. The proportion of children classified as overweight was

17 per cent with a slightly higher prevalence of females (18 per cent) than males

(16 per cent). Overall, 8 per cent of children were classified as obese, with a

significantly higher prevalence for males (10 per cent) than females (6 per cent)

(figure 1.2).

The 2007 Australian National Children’s Nutrition and Physical Activity Survey

(National Children’s Survey) (CSIRO and University of South Australia 2008a)

reported similar results, with 72 per cent of children not overweight or obese. The

proportion of children classified as overweight was 17 per cent, with a slightly

INTRODUCTION 7

higher prevalence for females than males. Overall, 6 per cent were classified as

obese, with a slightly higher prevalence of females (6 per cent) than males (5 per

cent).

Figure 1.2 Weight classification of Australian children aged 5–17 years,

Source: ABS (2007-08 National Health Survey, Cat. no. 4364.0).

While the two surveys report very similar prevalence rates for overweight in

Australian children, the NHS reports a higher obesity prevalence (8 per cent),

compared to the National Children’s Survey (6 per cent). This result was driven by

a significant difference in the proportion of obese males, which was 10 per cent in

the NHS and 5 per cent in the National Children’s Survey, and in particular, the

proportion of 13–17 year old males classified as obese in the NHS at 13 per cent.

Differences in the age groups (2–16 years in the National Children’s Survey and

5–17 years in the NHS) and the survey response rates between the two surveys are

likely to account for part of the differences in results (ABS 2009c; CSIRO and

University of South Australia 2008b).

The prevalence of overweight and obesity is higher in the adult population than in

the child population, with 37 per cent of adults being overweight and 25 per cent

obese in 2007-08 (ABS 2009a).

8 CHILDHOOD OBESITY

In the twentieth century, the estimated prevalence of childhood overweight and

obesity was relatively constant until the 1970s when it began to increase, and

continued to grow for the rest of the century, according to BMI data (Norton et

al. 2006). Three national surveys conducted in 1985, 1995 and 2007-08 provide

some indication of trends in the prevalence of obesity in recent decades:

• The 1985 Australian Health and Fitness Survey included about 8500 children

• The 1995 National Nutrition Survey, conducted by the ABS, involved about

• The 2007-08 NHS (discussed earlier) included 5000 children aged 5–17 years.

These three national surveys indicate that overweight and obesity prevalence in

children has been increasing (although generally at a declining rate). However,

recently published research suggests it might have levelled off sometime between

the 1995 and 2007-08 national surveys. Olds et al. (2009) pooled together

41 different studies that included data on the BMI of Australian children conducted

between 1985 and 2008.

INTRODUCTION 9

Figure 1.3 Weight classification of Australian children, 1985 to 2007-08

a Ages 7–15 years. b Ages 5–17 years.

Sources: ABS (2007–08 National Health Survey, Cat. no. 4364.0); Magarey, Daniels and Boulton (2001).

Although the 2007-08 NHS was not included due to the timing of its publication, its

inclusion would not have changed the overall findings (Olds, T.S., University of

South Australia, pers. comm., 9 June 2009).

The authors found that since about 1996, the prevalence of children classified as

overweight or obese has stabilised, or only slightly increased. The prevalence of

obesity in males increased from 1 per cent in 1985 to 5 per cent in 1996 and

remained at 5 per cent in 2008. Similarly for females, it increased from 1 per cent in

1985 to 6 per cent in 1996 and was 6 per cent in 2008. For overweight and obesity

combined, the prevalence in males increased from 10 per cent in 1985 to 22 per cent

in 1996 to 24 per cent in 2008, and for females it increased from 12 per cent in 1985

to 24 per cent in 1996 to 25 per cent in 2008 (figure 1.4).

However, as highlighted later (chapter 3), childhood obesity prevalence is greater in

some subgroups than others and Olds et al. (2009) does not report on obesity trends

in different subgroups. Regardless of whether childhood obesity prevalence has

levelled off, current levels are still considered by many observers to be higher than

desirable (Olds et al. 2009).

10 CHILDHOOD OBESITY

Figure 1.4 Weight classification of Australian children aged 2–18 years,

Source: Olds et al. (2009).

For most adults, being overweight or obese is a consequence of a number of

decisions they have made over many years regarding food consumption and

exercise. If made voluntarily and with full information of the consequences these

are largely personal decisions, and there the matter might otherwise rest. But to

some extent the decisions some people make are sub optimal. They might have

made better choices if they had had the benefit of foresight or had been better

informed. Being obese might also have consequences for the rest of the community.

In this chapter we examine obesity in an economic framework to consider the case

for government intervention. Even if the basis for intervention in the case of adults

is relatively weak, there may be a stronger case for intervening in the affairs of

children. We first consider how people make decisions about eating and energy use,

before considering why these decisions may be ‘distorted’, for example because of

incomplete information or price signals or cognitive and behavioural limitations.

The costs of obesity are then considered.

While some people are more predisposed to becoming obese than others (chapter 3),

as a general rule people put on weight when the energy in the food they consume

exceeds the energy they expend. It is therefore necessary to understand the

influences on both sides of this obesity ledger: energy in and energy out.

Considering obesity in a cost–benefit framework can help to explain why a greater

proportion of the population have become more overweight or obese over time. At

its simplest this could occur where food costs decrease and income rises, where the

cost of expending energy through exercise increases, or where people’s preferences

change (such that they prefer more or different food and/or less exercise than they

previously might). Using an economic framework can also help identify market

failures or behavioural limitations that might constrain the potential for individuals

to maximise their own wellbeing over time (section 2.2).

People make a myriad of different decisions about the way they live their lives,

many of them having implications for their weight. Faced with multiple competing

12 CHILDHOOD OBESITY

activities and purchasing choices, rational consumers will make decisions to

allocate their scarce resources of time and money to maximise their wellbeing.

In the context of childhood obesity, decision making about eating and exercise is to

a large extent made by adults on behalf of children, though not entirely. Moreover,

decisions are probably not made to optimise weight per se — rather people make

decisions about what and when to eat and how much exercise they do, with weight

being a possible factor in these decisions (but weight is certainly an outcome of

such decisions).

Eating provides benefits to individuals including providing calories and nutrients

that keep them alive and healthy, providing pleasure through taste (palatability) and

being an important social activity (figure 2.1). In choosing what, when and where to

eat, people will compare these potential benefits against the potential costs of

eating, which can include the financial cost of purchasing food, the time cost of

purchasing/preparing food and future health effects (mortality, morbidity, quality of

life effects). For example, people might purchase energy-dense nutrient-poor

takeaway food, even in the knowledge that it may be less nutritious for them than a

home prepared meal, because it frees up time to engage in other pursuits (helping

the kids with their homework, walking the dog etc) that they value more highly.

Energy expended is often a function of other decisions and factors (such as the type

and location of work, location of home and transport options, and preferred leisure

activities) as well as specific decisions to exercise to improve fitness or control

weight. Energy expended therefore could be affected by a wide range of factors

including technological changes in the workplace, prices of transport and fuel, and

prices of housing. Exercise can deliver health benefits, enjoyment and social

benefits, but these can be weighed against the time and financial costs of exercise

and potential health (injury) risks (figure 2.1).

In making decisions about eating and undertaking exercise, tensions can occur

between the benefits and costs. Some people will effectively choose to be

overweight and knowingly incur additional health risks as the ‘price’ to be paid for

enjoying food/company or undertaking less exercise today. It follows that the

optimal prevalence of overweight and obesity in society will not be zero.

The decision-making framework in figure 2.1 is a simplified representation, and the

value (or perception of the value) of the benefits and costs of eating and exercise are

influenced by many factors. For example, while it is perhaps difficult to talk about

the price of ‘food’ in a generic sense in the Australian context, in some countries the

price of staples will act as a significant constraint on total food consumption. In

Figure 2.1 Decision-making framework

Eating Decision

14 CHILDHOOD OBESITY

Australia, price influences are more likely to involve relative prices between food

types (such as the relative prices of fast food and home cooked meals, for which a

substantial input is effort and time). People’s decisions will also be influenced by

their income (better quality foods and more ‘outsourcing’ to restaurants) and their

education, as well as access to information about nutrition, health benefits and risks.

Advertising might also influence or ‘distort’ individual preferences.

As noted earlier, decisions about children’s eating and exercise are to a large extent

made by adults on their behalf. Nonetheless, to the extent that these factors

influence parents’ assessment of the benefits and costs of eating and exercise in

making their own weight-related decisions, the same factors are also likely to

influence parents’ decisions on behalf of their children.

As relative prices, income or preferences change, so would the mix of activities that

maximises the wellbeing of that person. The following discussion canvasses some

possible broad trends that might have influenced eating and exercise decisions,

increasing the prevalence of overweight and obesity since the 1970s. Specific

factors that might cause, or at least are associated with, obesity — such as

socioeconomic status, education and advertising — are discussed further in

chapter 3.

A decline in the price of food will lead to an increase in consumption, other things

being the same. But the demand for food in total is ‘inelastic’ (that is, consumption

is not very responsive to price changes), so the effect of price decreases on total

consumption would be expected to be modest. Cutler, Glaeser and Shapiro (2003)

cite a price elasticity of demand of –0.6 (indicating that a 10 per cent decrease in the

price of food would increase consumption by 6 per cent), but suggest that if this

were adjusted for quality effects, the demand for caloric intake would be more

inelastic. Similarly, studies on the effectiveness of food taxes suggest demand is not

very responsive to price changes (section 5.1, chapter 5).

Even so, Lakdawalla and Phillipson (2002) claim that as much as 40 per cent of the

increase in weight of Americans over the period 1976 to 1994 can be attributed to

price decreases associated with agricultural innovation.

Furthermore, relative prices matter. An increase in the price of one food item can

lead to a switch from the relatively higher priced item to lower cost items. Research

in the United States indicates that there is an inverse relationship between energy

density and energy cost (price per unit of energy), so energy-dense foods may be

cheaper (Drewnowski and Specter 2004). If this is the case, consumers face a

tradeoff between buying healthy (and relatively more expensive) foods or buying

cheaper energy-dense alternatives and having additional income to buy more of

other goods or services.

Similarly, the time costs associated with food preparation will also influence

consumption decisions. Cutler, Glaeser and Shapiro (2003) note that ‘… reductions

in the time required to prepare food reduced the per calorie cost of food by 29 per

cent from 1965 to 1995’ (p. 112). Their research found that the lower time cost of

food preparation has led to more frequent consumption of a greater variety of food,

and to higher weight, and that this helps explain a ‘good share’ of the observed

increase in Body Mass Index (BMI) over the study period (Cutler, Glaeser and

Shapiro 2003, p. 110).

The effect of food prices on children’s consumption would be expected to be quite

muted, given that they are not usually the ones making household decisions

regarding food purchase and preparation. But Cawley (2007) cites evidence that

school children are sensitive to changes in the relative prices of high-fat and low-fat

foods. French et al. (2001) found similar results through changing prices of low-fat

snacks in secondary school vending machines.

The income elasticity of food is generally regarded to be quite low. As incomes rise

individuals reach a point of satiation, and begin to spend their income on other

goods and services. However, rising incomes give them access to a wider variety of

foods and enable them to consume more food away from home. Rising incomes

also increase the opportunity cost of time for preparing food or exercising if it

comes at the cost of work.

The effect of income on obesity changes with economic development — while in

less advanced countries there is a positive relationship between income and weight,

there is evidence that in rich countries there is a negative relationship between

obesity and both income and education (refer to chapter 3 for discussion on

socioeconomic status). Although technological advancement has led to lower food

prices, and may provide some explanation for the rise in obesity prevalence,

Philipson and Posner (1999) propose that the growth in obesity as a result of this

factor may be self limiting when it makes workers sufficiently well off (and

consequently increases the demand for ‘thinness’).

The income effects on children are not direct (as they are unlikely to earn their own

income). Rather such effects are likely to arise through their parents’ income, which

can also shape the family environment (chapter 3).

16 CHILDHOOD OBESITY

In addition to the effects of lower prices of food and increases in income, the cost of

exercise has increased. Due to production-related technological advancement,

workers are less engaged in paid work that entails expending significant physical

energy. This shift toward more sedentary jobs corresponds with an increase in the

cost of physical activity. Workers were once, in a sense, paid to exercise through

their active labour, but in more sedentary jobs they must now forgo their free time

to exercise. As well, in some cases, individuals choose to pay to engage in exercise

(such as through gym membership). In theory, the higher the wage, the higher is the

opportunity cost of time devoted to engaging in exercise, but also the greater the

capacity to pay for exercise. That said, people enjoy physical activity to varying

degrees, and the cost of forgoing more passive leisure in order to engage in physical

activity will also vary. There are many competing demands on children’s time, and

to undertake exercise can mean giving up other pursuits. Further, it can involve the

cost of a parent’s time where supervision or transport is required.

Just as obesity may be an unintended side effect of economic development, it may

also be an unintended consequence of policy action designed to address other

economic, social, or health goals. For example, Chou, Grossman and Saffer (2004)

found a positive relationship between the rising prevalence of obesity in adults (in

the United States) and cigarette prices. Other research indicates that an increase in

cigarette taxes increases female BMI (Rashad, Grossman and Chou 2006). For

children, providing a school bus might improve safety, but might reduce the

opportunities for children to exercise by walking or riding to school.

In making decisions about eating and exercise, people take into account a range of

information. This includes current and future market prices of goods and services,

the implicit value of their own time and efforts, as well as health and other

‘non-price’ impacts, such as social interaction and enjoyment.

If information and prices and costs are distorted, then these decisions may not be

optimal. For example, food producers might not have an incentive to provide

information about the nutritional or energy content of their foods. Taxes or other

interventions might distort food prices.

Furthermore, as discussed below, behavioural limitations can mean consumers

make decisions about food consumption and exercise that are not in their own best

interests (through an inability to process information or by not behaving in rational

ways) (see also box 2.1).

That said, the presence of spillovers, information gaps or behavioural limitations is

a necessary but not sufficient condition for government intervention. Intervention

should only occur if it leads to a better outcome than would occur in its absence,

after allowing for the costs of implementing the intervention (chapter 4).

Box 2.1 Externalities, information failures and behavioural

Externality (or spillover) effects occur when consumption or production of a product

Information failures arise where there is insufficient or inadequate information about

income and time) at their disposal. Behavioural economics suggests that people can

Information failures come in various forms. Information asymmetry — for example,

where a firm has information that consumers do not — can lead to consumers

purchasing goods that they might not want if they were in possession of full

information and were able to process that information. For example, a consumer

may unwittingly purchase and consume food containing ingredients that they would

rather avoid. Consumers, however, are not always less informed than firms — for

18 CHILDHOOD OBESITY

example, consumers purchasing insurance may know more about the risks they face

than the insurer.

In the case of obesity, if individuals are to make informed choices they need

information on the (eating and exercise) behaviour that can lead to obesity

(including the energy and nutritional content of food options), the health risks

associated with being overweight or obese, and the likelihood of these risks

occurring. The extent to which such information will be effective will be linked to

how they convert information to appropriate action, and the ease of engaging in that

action.

Children are usually neither well-informed, nor always free to make their own

choices, with parents often making decisions about eating and exercise for children.

For children who do make decisions (such as at school), or at least influence

parental decisions, the potential for information failure is very apparent, as they do

not get access to all the necessary information and generally have a reduced

capacity to evaluate information about their own health.

When parents make decisions on behalf of children, a special case of the

principal–agent problem applies, where the interests of the parent may not align

with those of the child’s. In this case, where interests conflict, the child has limited

ability to define and defend their own interests. In most cases though, parents are

expected to act in the best interests of their children. Therefore, any relevant

information failure relating to childhood obesity should be addressed by targeting

parents (or guardians) and its success would depend on parents acting in the best

interests of their children.

There may be a role for government provision and dissemination of information

about obesity and exercise, especially if it helps to address the costs to the

community of obesity (see below) and meets the cost–benefit test mentioned above.

Government may be seen as a more credible source of information than the private

sector. Alternatively, government might compel information provision by industry,

such as through nutritional labels on food products. Other government roles may

include social marketing campaigns about healthy eating patterns and desired

physical activity levels for children at school or in the broader community.

Regulations may be introduced to prevent misleading advertising of, say, food

products or health-related products such as diet products. Different policy

instruments are discussed further in chapter 4.

However, even when adequately informed, cognitive limitations may limit the

ability of consumers to act on that information to promote their own best interests

(see discussion in behavioural considerations section below).

Economics defines an externality as the side effects or spillovers of an activity that

are not taken into account in an individual’s or business’ decision and that affect

another party’s wellbeing. In these circumstances, the social and private marginal

costs differ. An individual or business only takes into account the private costs in

their decision and ignores the social costs (the private costs plus the cost it imposes

on another party). So, in the case of a negative externality, where social costs are

greater than private costs, too much of the good is produced or consumed, from a

community perspective (a positive externality would result in too little of the good

being produced or consumed).

For public policy purposes, it is helpful to distinguish between pecuniary and

technical externalities, and between (negative) externalities and costs imposed on

the community.

Pecuniary externalities are generally defined as externalities that are transmitted

through the price system (for more detail see Bohanon 1985). The typical example

is where a new, more efficient firm enters a competitive industry, driving down the

price received by all existing firms. The harm to the existing firms is more than

offset by a transfer of income to consumers.

Technical externalities on the other hand have a direct real effect on a third party,

and can result in market inefficiency, potentially warranting corrective action,

whereas pecuniary externalities do not have efficiency implications. In the case of

health issues such as smoking or alcohol consumption, technical externalities are

quite clear, such as the health or discomfort effects of passive smoking, or the

injuries or fatalities of others associated with drink driving.

By comparison, overweight and obese people cause few policy-relevant technical

externalities, that is, their weight does not materially affect the wellbeing of other

people. The loss of personal space that some people might experience sitting next to

an obese person on a plane (or on public transport, or at a concert) may be thought

of as a technical externality. Such externalities might be partly ‘internalised’

through affected passengers moving or adjusting their seating position. (Such

technical externalities would not normally be an issue with overweight children.)

The response by airlines or theatres to the growing size of their patrons may be to

increase the size of seating, reducing the overall number of seats within a given

venue, thus increasing the costs to normal weight people using those seats.

Alternatively, where a person’s weight is an important influence on the cost of

providing them with a service, businesses might choose to charge those people a

higher price, and/or provide them with a subtly different service. For example, some

20 CHILDHOOD OBESITY

airlines have experimented with higher charges for obese people, a recent example

being Air France/KLM offering obese people an option to buy a second seat at a

25 per cent discount, refundable if the plane is not fully booked. People who cannot

sit comfortably in a single seat and have not reserved an extra seat might not be

allowed to board if the flight is full (Air France 2010).

It is also important to distinguish between technical externalities that might be

policy relevant in their own right and costs imposed on the rest of the community

associated with deliberate policy decisions to provide a minimum level of public

health care for all.

On average, obese people consume more health care costs than normal weight

people. The obese bear some of these costs themselves, such as a portion of medical

costs and possibly lost wages (they may have more work loss days due to ill health),

and some of the costs will be borne by others. (Obese people also bear the

diminution of their own wellbeing.) Of the costs to others, some are borne through

collectively financed systems, such as medical care, sick leave, group life insurance,

and retirement pensions. Specifically, the greater healthcare needs of obese

individuals will increase health costs for all, either through increased taxes or

through increased healthcare premiums (where a policy of community rating

applies). These ‘external’ costs, which are sometimes characterised in the literature

as an externality (for example, see Bhattacharya and Sood 2005), are more in the

style of a pecuniary rather than a technical externality, and might simply be

recognised as part of the costs the community is prepared to bear to provide a

universal health care system. That is, they might be considered as a transfer from

one part of the community to the other. However, there is limited evidence on how

significant these obesity costs (for adults or children) might be in the Australian

context (although a US study estimated them to be of the order of $150 per capita

(Bhattacharya and Sood 2005)).

But some efficiency costs are also associated with public health care. The efficiency

costs arise from the deadweight losses of the additional tax revenue needed to fund

the public health system (box 2.2), and through distortions to the price of providing

health care services.

In addition, the fact that people do not face the full costs of the health services they

consume could be an efficiency issue if it reduces the incentives for people to

consider the full costs of obesity in making decisions about their food intake and

Box 2.2 The efficiency cost of taxation

exercise. Exposing obese individuals to the full costs of their health care might have

some marginal effect on obesity, however, on its own it is unlikely to significantly

reduce obesity prevalence in the community due to information gaps and

behavioural limitations. And there are several reasons that a policy response

designed to address this issue might not be practical. Taxing obese individuals, for

example, would unfairly single out their health-related risk behaviour for policy

treatment against other costly health-related risk behaviour (such as hang-gliding).

Additionally, any policy response to ensure obese individuals do face more or all of

the costs of the services they consume would negate the intent of publicly funded

health insurance programs designed to ensure financial contributions are

income-related rather than risk-related.

In summary, there are few substantial technical externalities arising from obesity in

adults or children (unlike alcohol or tobacco where effects on third parties can be

substantial). But the efficiency costs to the community from having to finance the

higher consumption of health care by the obese are an issue. As universal access to

health care and community rating have been key parts of the policy platform of

Australian governments for many years, this suggests that there may be important

dividends to the rest of the community from reducing obesity. However, the fiscal

effects of obesity need to be seen in terms of the total amount of community

resources that are consumed by the obese, not just health services. For example, the

costs of providing age pensions and aged care services may be reduced if obese

individuals have shorter life expectancies. The costs of obesity are addressed further

in a following section.

22 CHILDHOOD OBESITY

The preceding discussion is predicated on the assumption that individuals are

rational and act in their own best interests. In other words, individuals have the

cognitive skill and motivation to make decisions about how much and what they

consume (and how much they exercise) that maximises their own wellbeing, fully

accounting for the future health, financial and lifestyle consequences of their actions

(Cutler, Glaeser and Shapiro 2003). People who engage in behaviours that have

long-term negative effects (such as smoking or overeating) may do so in a fully

rational and forward looking way, by weighing up the benefits and the total

discounted costs of that activity, including monetary costs and health costs.

However, many individuals have difficulty in consistently making rational decisions

— that is, they do not consistently value the costs and benefits of their actions and

so do not choose options that maximise their wellbeing, subject to available

resources. At least some food consumption does not appear rational — for example,

people overeat, or eat the wrong foods, despite wanting to lose weight. To the extent

that people make less than ideal decisions about their own wellbeing, the wellbeing

of the community generally is less than it might otherwise be.

A large number of behavioural limitations have been described in the literature that

help to explain the way people actually behave, but it is difficult to build these into

a predictive model of human behaviour. That said, some behaviours are far more

predictable than others, and may be useful in helping policy makers develop better

approaches to managing obesity.

Behavioural limitations fall into one of three broad categories: bounded rationality;

bounded willpower; and bounded self interest. Of these, the last — being the notion

that we have some regard for the wellbeing of others when we make decisions about

our own wellbeing — is probably the least relevant for obesity policy, and is not

discussed further in this paper.

Bounded rationality refers to the fact that people have limited cognitive abilities and

are not always able to make decisions that maximise their wellbeing, even where

they may recognise these limitations and implement strategies to address them. That

people have difficulty computing the benefits and costs of all options open to them

about their diet and exercise, would come as no surprise to anybody used to

shopping, or sifting through the comments in the media on what is good for you and

what is not.

Individuals might not act in their own best interests even where adequate

information is available. The high ‘cost’ of gathering and processing the

information (transactions costs) and the difficulties in processing it, can lead us to

take short-cuts, either consciously or unconsciously. We can make decisions that,

instead of maximising our wellbeing, are merely satisficing, in the sense that they

produce a result that is ‘good enough’.

However, taking such short-cuts might be optimal, if these transactions costs

outweigh the additional potential benefits of a different decision. Many people use

rules of thumb (or heuristics) to assist decision making. For example, people might

consume breakfast cereal high in sugar content in the belief that cereals are

generally good for you and a key part of a balanced diet. If they use this rule of

thumb, and do not read labelling information on sugar and fat content, they may

make poor choices or fail to update their consumption decisions as new (healthier

cereal) alternatives come on the market, or new information becomes available.

Although such a person may increase their wellbeing — as the transactions costs of

reading and processing the nutrition information may be higher than the health

benefits from switching to a healthier cereal because of that information — a

‘better’ rule of thumb (say, eating cereal with a tick from the Heart Foundation)

might increase their wellbeing even more.

The context or environment in which we make decisions about diet and exercise can

have important effects on our ability to make sound decisions (see, for example,

Shiv and Fedorikhin 1999 or Ariely 2000). When people eat while undertaking

some other activity they spend less effort on monitoring consumption and tend to

eat more than they otherwise would (Bertrand and Schanzenbach 2009). They also

tend to focus less on future consequences (Shiv and Fedorikhin 1999). Eating while

watching television is an example.

The amount people eat is also influenced by the behaviour of those around them

including family, peers, or social groups. When eating in groups people tend to eat

as much as their peers (Birch and Fisher 2000). While this might lead the obese

within that group to reduce their consumption, the others generally increase their

consumption (Just 2006).

The amount people eat, or the choices of food they make, can also be influenced by

the choices available to them (for example, between healthy and energy-dense

nutrient-poor alternatives in a canteen), and the way food is presented. Even when

faced with essentially the same choices, people react differently to essentially the

same propositions, depending on how they are framed (Gibbs 1997).

24 CHILDHOOD OBESITY

The environmental influences on food consumption extend to the way food is

packaged and the size of servings. For example, people are subject to ‘unit bias’

(Geier, Rozin and Doros 2006), that is, the amount they consume is related to

portion size, or package size (see also, Rolls, Ello-Martin and Carlton Tohill 2004).

In commenting on Wansink (1996), Just (2006, p. 214) found that: ‘Doubling

portion sizes increases consumption anywhere from 18 to 25 per cent for

meal-related foods and up to 45 per cent for snack foods.’ Evidence also supports

the idea that people tend to under estimate their calorie consumption (Wansink and

Chandon 2006), and that this effect is exacerbated by portion size (Just 2006).

While context is important, so too are consumers’ starting points. Research

indicates that people sometimes demand much more to give up something than they

would be willing to pay to acquire that same thing. This is called the endowment

effect or loss aversion (Kahneman, Knetsch and Thaler 1991). Most people also

have a ‘status quo bias’, meaning a reluctance to change from established

behaviours (Samuelson and Zeckhauser 1988). These biases can have implications

for diet, consumers being willing to add good foods to their diets but reluctant to

give up bad foods, making it difficult to reduce overall energy intake.

Bounded willpower refers to the fact that people often make decisions and act in a

way that they know is against their own long-term best interests. To some degree

we all make decisions about tradeoffs between our current and future wellbeing (or

between our current selves and our future selves, as some analysts have

characterised the decision framework). Thus for example, we might forgo current

consumption now given the prospect of greater consumption at some time in the

future. This requires some consideration of the costs and benefits of the two options.

We might consider that the elevated risks of ill health that arise from, for example,

eating energy-dense nutrient-poor foods or smoking, is for some people more than

outweighed by the pleasure these pastimes give them in the current period.

The rate at which we trade off current and future wellbeing is what economists call

the rate of time preference. This rate can vary between people and for any one

person may vary over time. The higher the rate of time preference, the greater an

individual values the present over the future.

The standard discounted utility model (Samuelson 1937) assumes a constant rate of

time preference over time. This would mean for example, that if someone preferred

$100 today over $200 in one year’s time, they would also prefer $100 in ten years’

time to $200 in eleven years’ time. The choices are the same, just ten years apart.

Research indicates that people value the present over the future differently, which

can have implications for obesity. A high valuation of the present over the future (a

high rate of time preference) could, for example, lead to more food consumption

today at the expense of health in the future. A number of studies have considered

the links between the rate of time preference and obesity or risky behaviour. Zhang

and Rashad (2008) found some evidence of a positive association between time

preference and weight, particularly for males. A Netherlands study (Borghans and

Golsteyn 2006) considered the links between body mass and individuals’ ‘discount

rates’, and trends over time. The authors found that being ‘overweight’ might be

related to the way people discount future health benefits. However, they did not

consider it likely that the trend in increasing body mass in the population could be

attributed to an increase in the rate of time preference. They also found the link

between time discounting and BMI was stronger for women than for men.

Other research shows that time preference, as measured through certain behaviours

such as education, smoking, use of nutritional information, and motivation for

accessing or acquiring nutritional knowledge, significantly affects the odds of

choosing a risky diet (Finke and Huston 2003). Individuals’ degree of impatience

may be related to factors such as income or education (Becker and Murphy 1988).

Education appears to be a key determinant of eating a healthy diet not just because

education helps you to understand the tradeoffs being made, but also because the

educated may have lower discount rates (as demonstrated by the decision of many

to defer earning a full-time income during studying for the prospect of earning

higher incomes later on).

The standard model of discounting our future selves against our current selves can

be a rational way to make decisions, provided the individual’s discount rate remains

consistent over time. But many people seem to apply different discount rates to

future benefits at different points in time, that is, their preferences are inconsistent

over time.

Typically people will apply a much higher discount rate to near decisions than they

do to those that occur in the future. This is called hyperbolic discounting and can be

illustrated by a variation of the example used above. If offered the choice between

$100 today and $200 in one year’s time an individual might choose the $100 today.

But when offered the choice between $100 in ten years’ time versus $200 in eleven

years’ time, that individual might choose the latter. This displays ‘time inconsistent

preferences’ because although the choices on offer are essentially the same, they are

valued differently when a time delay is involved. The inconsistency is highlighted

when, ten years later, that individual wants the $100 ‘today’, and does not want to

wait one year for the $200, as opted for years earlier.

26 CHILDHOOD OBESITY

To put this in an obesity context, it is easy to envisage many people deferring

decisions over diet because of the cost of forgoing the pleasure of over-eating today.

A rational consumer would weigh up the gratification of eating additional food

against the costs to their health in the future and also weigh up the ‘cost’ of going

on a diet or exercising against the benefits to their health, and reach an optimum

body weight. But hyperbolic consumers will consume more food today than is in

their long-term interests. This means their short-term self trumps the long-term self,

leading to procrastination and deferment of the diet or exercise that might have

countered their higher levels of consumption today.

Research indicates that obese dieters can exhibit behaviour consistent with time

inconsistent preferences (Scharff 2009), and leads to the conclusion that consumers

with hyperbolic time preferences need not only to be educated about the risks of

overeating but that they may need to ‘… overcome their temporally inconsistent

preferences through the use of commitment mechanisms’ (Scharff 2009, p. 19).

However, not all such consumers may need external assistance. Those who are

aware of their behaviour may design strategies to help address their tendency to

succumb to temptation despite their desire to do otherwise (such as buying limited

quantities of unhealthy foods during shopping trips, buying smaller quantities of

food per shopping trip, and using a peer group, such as Weight Watchers, to help

ensure that commitments are adhered to).

Other research indicates that time preferences are affected by emotional state —

positive mood is known to sometimes have effects on cognition and behaviour, and

may increase willingness to delay gratification (Gibbs 1997).

Adults can exhibit bounded rationality and bounded willpower, which can lead to

obesity. But generally speaking these are largely the concern of the individual. On

the other hand, parents (or guardians) are often required to make decisions for

children, including weight-related decisions. These decisions will normally include

what the child eats and to some extent how that child spends their time. As children

grow older, they may influence the decisions parents make on their behalf (for

example, Turner, Kelly and McKenna 2006), and increasingly make more decisions

for themselves. They may also be increasingly influenced by factors outside the

family.

Children display many of the same biases as adults, but more strongly. In the

absence of supervision, children are more likely to indulge in behaviours that can

have consequences for their weight and future health. For example, children may be

more prone to peer group pressure than their parents, and have even more difficulty

in accounting for the future consequences of their actions.

The effect on children of cognitive limitations and distortions to decision making is

therefore twofold. First, the individual choices made by parents or guardians will be

imposed on their children, so any distortions to which parents are subject —

through for example, cognitive limitations or advertising — will affect their

children. Second, the children themselves will be subject to their own distortions

from inside and outside the family.

While the case for hard interventions on the basis of behavioural biases in adults is

relatively weak, it is much stronger for young children. But the role of parents and

guardians makes for a difficult policy environment. While policy might be directed

at minders, and the environment within which children live (for example, by

regulating eating and exercise options while at school), it needs also to consider the

potential for their behaviour to be influenced by external stimuli when they are

largely left to their own devices (for example, while watching television). Policy

options for addressing behavioural biases are discussed further in chapter 4.

As the preceding discussion has noted, some costs of obesity can be policy relevant.

In this section we explore this issue further by identifying the different types of

costs and considering the available evidence. Again, unfortunately, what little

evidence there is relates to obesity generally, and not to obesity among children

per se.

The costs of obesity will vary according to numerous factors including obesity

prevalence, the health costs associated with treatment, and the severity of obesity

related health consequences.

The economic costs of obesity can be viewed in various ways, including in terms of

direct costs, intangible costs and indirect costs, or whether the costs are borne by the

obese themselves or by others in the community (figure 2.2).

Direct costs are the costs to the community from diagnosis and treatment

(WHO 2000). Direct costs relating to obesity and its consequences can include

medical services, hospital-related costs, and personal health care costs (such as

medication). Intangible costs are the effects on health and quality of life for the

individual and others including family members. The obese tend to suffer more

disability (morbidity) than others and die earlier, thus resulting in less wellbeing in

aggregate.

28 CHILDHOOD OBESITY

Indirect costs are the loss of wellbeing and economic benefits to other members of

the community through less goods and services being produced (WHO 2000).

Figure 2.2 The costs of obesity: is childhood obesity a public

Indirect costs can relate to lost output as a result of reduced productivity (due to

illness or early death). In the case of childhood obesity, indirect costs might include

potential productivity losses incurred by parents (if they were required to spend

more time as a carer) but not by the child itself (as generally they are not working).

Nonetheless, the specific costs included in each category vary between studies.

Personal costs are those borne by the obese themselves, which in the case of obese

children are also borne by their parents. Personal costs include those portions of

medical services, hospital-related costs, and personal health care costs incurred by

the individual, as well as lost wages. They also include more intangible costs

associated with loss of wellbeing from outcomes such as not being able to engage in

some physical activities they enjoy and social reactions to their weight.

• Physical health problems

• Psychological and social problems

• Increased probability of being

• Increased risk of type 2 diabetes

• Increased risk of cardiovascular disease

• Increased risk of certain cancers

• Increased risk of pulmonary, orthopaedic

• Higher costs of production

• Shared cost due to public health system

• Other ‘shared costs’

– where price discrimination for health

– in workplaces (such as collectively

• Health effects on wellbeing

• Health effects on household budget

• Health effects on employment and

Costs imposed on others (such as employers, government, and the rest of society)

include the costs incurred by the rest of the community as a result of collectively

financed programs such as medical costs, insurance and pensions (see earlier

discussion under externalities). Costs to employers might include lost output or

search and hiring costs to replace lost labour, and workers’ compensation

(depending on the degree to which employers can pass on costs through lower

wages). Costs also include the cost of raising taxation revenue. However, not all

costs are financial costs.

While there are substantial costs imposed on others from these pecuniary

externalities, it is argued in section 2.2 that there are few pure technical externalities

associated with obesity, unlike those related to some other health risk factors such

as smoking (passive smoking) and alcohol consumption (violence or road fatalities).

Estimating the costs of obesity is inherently difficult, as estimates rely on a range of

assumptions, some of which may be more robust than others. Data limitations can

also constrain the construction of estimates. These deficiencies should be

considered when interpreting estimates of the costs of obesity.

There have been several attempts to estimate the costs of obesity in Australia. The

Australian Institute of Health and Welfare (AIHW) and the Centre for Health

Program Evaluation (CHPE) estimated the total costs of obesity were $736 million

in 1989-90 (cited in NHMRC 1997; DoHA 2009f). This estimate included direct

costs of $464 million and indirect costs of $272 million (although not all

obesity-related conditions were included in the analysis). DoHA (2009f) also cites a

later, unpublished, study by Crowley in 1995-96, in which the costs of obesity were

estimated at between $0.68–1.24 billion. Colagiuri et al. (2010) estimated the

annual total direct costs due to overweight and obesity (above the costs for

normal-weight individuals) for Australians aged 30 years or older (based on BMI or

waist circumference thresholds) was $10.7 billion in 2005. Direct costs included

health costs such as ambulatory services, hospitalisation, prescription medication

and other costs including transport to hospitals, supported accommodation, home

service and day centres, and purchase of special foods. Cost estimates were based

on survey responses on health services use and expenditure, including medication

use.

A recent report (Access Economics 2008) commissioned by Diabetes Australia

estimates the total cost of obesity in 2008 as $58.2 billion, which even after

accounting for the time difference is much higher than that estimated by

30 CHILDHOOD OBESITY

AIHW/CHPE. However, part of the difference can be accounted for by the broader

classification of costs adopted by Access Economics, including the loss of

wellbeing by the obese. Access Economics’ total cost estimate comprised

$8.3 billion of financial costs and $49.9 billion in lost wellbeing.

The financial costs of obesity (adults and children) were estimated to include:

• direct health system costs ($2.0 billion)

• productivity costs ($3.6 billion)

• carer costs ($1.9 billion)

• transfer costs (that is, the deadweight loss from the higher level of taxation)

• other indirect costs (for example, for aids, modifications and travel)

• changes in the method of valuing a statistical life (for estimating lost wellbeing)

• an increase in the assumed proportion of key diseases attributable to obesity

cent in the 2008 estimates. The updated attributable fractions are based on

revised work from the AIHW (Begg et al. 2007).

Population growth, higher prevalence and cost increases also contributed to the

higher estimated costs in the later Access Economics study.

Measurement issues aside, the Access Economics report highlights that the great

majority of the costs of obesity are borne by the obese themselves. This occurs

primarily through the loss of wellbeing (due to disability or shorter life span) but it

also results from them sharing some of the financial costs. Overall, the obese bear

90 per cent of the total costs, and 30 per cent of financial costs (box 2.3).

32 CHILDHOOD OBESITY

Box 2.3 Access Economics’ estimates of the costs of obesity

Source: Access Economics (2008).

While the preceding discussion suggests that the costs of obesity can be very high,

there is little evidence on the costs of obesity among children. A US study suggests

that the medical expenses of obese children are not greatly different from other

children. Johnson, McInnes and Shinogle (2006) estimate that paediatric obesity

costs in the United States was $127 million annually. However, they note that

overweight children are likely to become overweight adults and that the economic

burden of adult obesity is large.

The previous chapter examined obesity in an economic framework and the decision

making about eating and energy use that could lead to an individual becoming

overweight or obese. A range of individual, family-level and community-level

factors will influence these decisions about eating and energy use by influencing the

benefits and costs of each activity that are weighed up when making such decisions.

This chapter looks at some of the evidence on these possible causes of childhood

overweight and obesity. Understanding these possible causes is important for

understanding what causes are amenable to policy influence.

As mentioned earlier, obesity is caused by an imbalance between energy consumed

and energy expended, with excess energy stored as fat, and it has numerous health

consequences (box 1.1, chapter 1). The increase in childhood obesity in Australia

since the 1970s could be a result of an increase in energy intake, a decrease in

energy expenditure or, more likely, a combination of the two. It is likely that some

of the factors discussed in this chapter have changed over time, changing the value

(or perception of value) of the benefits and costs of eating and exercising, leading to

the increase in childhood obesity.

We base our discussion of some of the possible causes of, or factors associated

with, childhood overweight and obesity on the framework presented in Davison and

Birch (2001) (figure 3.1). The framework sets out three main categories:

• ‘Child characteristics and behaviours’, which includes genetics (child

• ‘Parenting styles and family characteristics’, which can affect a child’s

36 CHILDHOOD OBESITY

• ‘Community, demographic and societal characteristics’, which can influence

Figure 3.1 Framework for factors associated with obesity and overweighta

a Child behaviours (in upper case lettering) are associated with the development of overweight and obesity.

Source: Davison and Birch (2001).

The factors in the outer layers affect those in the inner layers, culminating in the

child’s behaviour. For example, SES might influence the types of food available in

the home, which can influence dietary intake of children. Or crime rates and

neighbourhood safety might directly affect the preparedness of parents to allow

their children to partake in discretionary physical activity outdoors.

Much of the data and research presented in this section have limitations. Many

studies use self-reported (rather than measured) data to measure some factors (such

as dietary intake and physical activity) and rely on proxy measures, as many factors

are difficult to measure precisely. Much of the research presented in this section is

also cross-sectional in nature. While such research may suggest a correlation

between the factor and obesity, the scope to make causal inferences is limited.

Familial susceptibility ACTIVITY

Table 3.1 summarises the overall findings of the Australian and international

evidence in relation to each possible cause.

Table 3.1 Summary of evidence presented on factors associated with

Genetics None included 0 Association 14a,b

Dietary intake Ambiguous 2 Ambiguous Unstateda,b

Soft drinks Positive association 2 Positive association Unstateda

Physical activity Ambiguous 4 Inverse association 56a,b

Ambiguous 3 Positive association 30a,b

Ambiguous 1 Ambiguous 66a,b

Advertising None included 0 Ambiguous Unstateda,b

Ambiguous 7 Ambiguous 35a,b

Ethnicity Associationc 3 None included 0

Ambiguous 6 Ambiguous Unstateda,b

a Included in sourced international reviews. b Includes Australian studies, some of which are discussed

separately. c Some ethnicities may predispose an individual to being relatively heavier and others relatively

38 CHILDHOOD OBESITY

A strong genetic basis exists for the development of obesity. Numerous genes have

been linked with a predisposition to excess fat. At least six very rare mutations of

single-genes causing severe early-onset obesity have been identified. In addition,

there are also a number of rare syndromes that cause obesity, among other

conditions, such as Prader–Willi syndrome and Bardet–Biedl syndrome (Baur and

O’Connor 2004). In addition, an international review of twin and adoption studies

found that genetics had a strong effect on Body Mass Index (BMI) variation at all

ages, and the effect was stronger than that of environmental influences

(Silventoinen et al. 2010).

However, biological factors alone, including genetic composition, are unlikely to

account for the rise in obesity that has occurred since the 1970s, as it has occurred

too quickly to be explained in evolutionary terms (Crawford 2002; Philipson and

Posner 2008). It is more likely that the rise is due to changes in the social and

physical environment (Baur and O’Connor 2004).

A child’s birth weight appears to be associated with childhood weight outcomes. An

international review of the literature concluded that a consistent and positive

relationship exists between birth weight and BMI in childhood (Parsons et al. 1999).

Most of the studies included in this review used BMI as their indicator of body

weight, but as discussed earlier (chapter 1), BMI does not quantify body

composition and captures both fat and lean mass.

More recent studies have used measures of body weight that distinguish between fat

and lean mass and have found that birth weight was positively associated with lean

mass but not fat mass (Labayen et al. 2008; Wells et al. 2005). Other studies found

an inverse association between birth weight and central fat (Dolan, Sorkin and

Hoffman 2007; Garnett et al. 2001; Labayen et al. 2008; Oken and Gillman 2003).

Central fat is associated with risks of cardiovascular disease and may carry more

risk than total body fat (Oken and Gillman 2003). Other factors, such as SES, can

influence both birth weight and later body weight (Parsons et al. 1999).

Dietary intake contributes directly to energy consumed. Dietary intake in children

may have changed over time, possibly contributing to the rise in childhood obesity

in Australia. A number of factors may have influenced the financial and time cost of

food consumption, leading to a change over time. First, agricultural and food

processing innovation may have lead to reductions in both the financial cost of

food, and the time cost for preparing food. Second, rising incomes increase the

opportunity cost of the time spent preparing food. Third, increasing working hours

also increases the time cost of preparing food (chapter 2).

The quantity of food consumed is not the only important consideration. The energy

density of food is also important as different macronutrients (such as fat, protein

and carbohydrates) contribute different amounts to energy intake. Also, fat, in

particular, is stored more readily as fat in the body than other macronutrients

(Davison and Birch 2001). Different macronutrients have different satiety effects

that will promote or suppress additional dietary intake:

• energy density influences the palatability of food, which will influence

• different macronutrients have different thermic effects, which will influence

• energy storage in the body will be influenced by food composition and the

40 CHILDHOOD OBESITY

longitudinal study, found that fat intake was directly related, and carbohydrate

intake inversely related, to subscapular (bottom point of shoulder blade) skinfolds.

However, they were not related to BMI or triceps skinfolds. The authors concluded

that macronutrient intake when young did not predict body fatness when older.

Another Australian study, Sanigorski, Bell and Swinburn (2007), found significant

positive relationships between daily servings of fruit juice/drinks and soft drinks

and the probability of being overweight/obese. Surprisingly, children who

consumed the highest amount of fruit and vegetables were also more likely to be

overweight/obese than children who had consumed no fruit and vegetables the

previous day. This result could be due to a number of factors, including

overweight/obese children eating a higher overall volume of food, overweight/obese

children positively changing their diet in response to their weight, or reporting bias

being stronger in parents of overweight/obese children. There was no significant

relationships between the proportion of overweight and obese and daily

consumption of fast foods and packaged snacks.

An international review (Newby 2007) found that, overall, there is no consistent

association between childhood obesity and most dietary factors. The evidence on

the relationship between total energy intake and obesity was the most inconsistent,

but there was some evidence to support positive relationships between fat intake,

and consumption of sugar-sweetened drinks and obesity. However, several

methodological weaknesses in the studies covered by the review could at least

partly explain the inconsistent findings, including interaction effects with other

factors not taken into accounted, underreporting of dietary intake, genetic

influences, different growth stages and generalisability of studies.

Soft drink consumption is likely to influence obesity. Evidence suggests that people

do not compensate for the increase in energy consumed by drinking soft drink, and

that soft drinks may provide insufficient satiety signals when compared with solid

food (DiMeglio and Mattes 2000). In addition, soft drink consumption can stimulate

appetite, as consuming high glycaemic carbohydrates can cause glucose levels to

fall (Wolff and Dansinger 2008). Also, when processing soft drink the body may

use less energy than when processing other food (lower thermogenesis) (Olsen and

Heitmann 2009).

The most recent national Australian data (1995) on soft drink consumption are

presented by Gill, Rangan and Webb (2006). They found that about half of all

teenagers and 36 per cent of 2–3 year olds had consumed soft drink in the past

24 hours. More recent data for New South Wales for 2004 showed that almost

60 per cent of males and almost 40 per cent of females in years 6, 8 and 10 drank

more than 250ml of soft drink daily (Booth et al. 2006). In addition, between

7–12 per cent of males and a smaller proportion of females drank more than 1 litre

of soft drink daily (Booth et al. 2006). It appears that male children consume more

soft drink than female children (Abbott et al. 2007; Booth et al. 2006; Gill, Rangan

and Webb 2006), and soft drink consumption increases with age (Abbott et al. 2007;

Gill, Rangan and Webb 2006).

Between 1969 and 1999, soft drink consumption by Australian adults and children

more than doubled from an average of 47 litres per person per year to 113 litres per

person per year (ABS 2000). A number of factors could explain this increase. First,

increased availability of soft drinks, such as more vending machines, making it a

relatively more convenient purchase. Second, a reduction in the relative price of soft

drinks. Although Australian price data are not available, in the United States

relative soft drink prices have decreased and consumption has increased over the

past 20 years — soft drink consumption of 6–11 year olds roughly doubled between

1977–78 and 1998, and between 1982–1984 and 2000 the price of soft drinks

increased by only 26 per cent, much lower than the overall consumer price index

(80 per cent) and the price of fresh fruits and vegetables (158 per cent)

(Sturm 2005).

A study conducted in regional Victoria (Sanigorski, Bell and Swinburn 2007) found

that 4–12 year olds who consumed three or more servings of soft drink ‘yesterday’

were significantly more likely to be overweight/obese than those who consumed

0–2 servings. Another Australian study (Tam et al. 2006) examined the relationship

between soft drink/cordial consumption in mid-childhood (average age 7.7 years)

and BMI in early adolescence (average age 13.0 years). The results suggest that

increases in soft drink/cordial consumption may have contributed to the

development of adolescent obesity.

A large international meta-analysis (Vartanian, Schwartz and Brownell 2007) found

a significant correlation between soft drink consumption and energy consumed.

However, the average size of the effect was small for children. Evidence for an

association between soft drink consumption and body weight was mixed, and was

influenced by how body weight was measured. (The authors found that studies that

received funding from the food industry on average reported smaller results.)

Physical activity affects children’s weight status through increasing energy

expenditure. But it also links directly to children’s health outcomes — for example,

42 CHILDHOOD OBESITY

low physical activity in children may be associated with a higher risk of developing

cardiovascular disease (Ruiz and Ortega 2009). In adults, physical inactivity and

obesity have similar health consequences (Blair and Church 2004). Further,

physically active, obese individuals have lower morbidity and mortality rates than

those who are normal weight and sedentary (Blair and Brodney 1999). Research

also indicates that moderate exercise can improve mental wellbeing (Fox 1999).

Deciding to undertake physical activity involves allocating scarce time and possibly

money. For children, undertaking physical activity could be at the expense of

studying, relaxing or other activities. How these activities are valued may have

changed over time, potentially increasing the costs of undertaking physical activity,

leading to reduced energy expended and increased obesity. In addition, concerns

about safety and changes to the physical environment (section 3.4) may have also

increased the costs of physical activity.

The National Physical Activity Guidelines recommend that children and adolescents

aged 5–18 years undertake at least 60 minutes of moderate to vigorous physical

activity every day (DoHA 2004a, 2004b). The recent National Children’s Survey

found that physical activity guidelines were met by the majority of 9–16 year olds

by most methods of measurement. Boys were more likely to meet the guidelines

than girls, and older children were less likely to meet the guidelines than younger

children (CSIRO and University of South Australia 2008a).

Many claim that physical activity in Australian children has decreased over time

(for example, Waters and Baur 2003). However, the evidence and data surrounding

physical activity in Australian children over time are sparse and patchy and overall

do not point to clear trends. A national study found that organised and informal

sport participation rose slightly from 59 per cent in 2000 to 62 per cent in 2006

(ABS 2006a). Further, Okely et al. (2008) found that both the prevalence and

minutes per week spent in self-reported moderate to vigorous physical activity in

NSW school children increased between 1985 and 2004 (by 12–20 percentage

points and 135–175 minutes per week, respectively). However, a NSW study found

that in 5–14 year olds the proportion that walked to school dropped substantially

and the proportion that were driven to school rose substantially between 1971 and

2003 (van der Ploeg et al. 2008). In addition, some smaller longitudinal studies have

observed declines. In a three year study, Cleland, Crawford et al. (2008) found that

average moderate to vigorous physical activity was significantly lower three years

later. Ball et al. (2009) also observed a decrease in physical activity over a three

year period.

Australian studies have examined the relationship between physical activity and

childhood obesity. Abbott and Davies (2004) found that physical activity levels

were significantly inversely associated with BMI and body fat. Another study, Ball

et al. (2001), found that physical activity level was inversely associated with body

fat in boys but not girls. However, Spinks et al. (2007) did not find a significant

relationship. A longitudinal study that examined the relationship between the level

of compulsory physical activity at school and overweight in childhood and 20 years

later, also found no significant relationship (Cleland, Dwyer et al. 2008).

However, an international review found that 12 of 14 longitudinal studies and 25 of

42 cross-sectional studies identified a significant inverse association between

physical activity and child and adolescent body weight (Trost 2005).

Sedentary behaviour includes activities that do not involve physical activity, and

could include activities such as watching TV and playing computer/video games

(also referred to as small screen recreation (SSR)), or other activities such as

reading or studying.

SSR can influence weight outcomes in a number of ways. First, SSR may substitute

for more physically demanding activities, reducing energy expended. Second,

higher exposure to advertising may result in children choosing more energy-dense

nutrient-poor foods (discussed later in this chapter). Finally, more energy-dense

nutrient-poor foods may be consumed while engaging in SSR than might have been

otherwise (eating as a secondary activity).

Sedentary behaviour may have increased over time due to technological change

increasing the variety of SSR activities available, and possibly reducing the relative

cost of undertaking SSR. Changes in SSR need to be considered in the

decision-making framework discussed in chapter 2. Presumably children are to

some extent making choices about the amount of time they spend on SSR after

considering the benefits they obtain from it and the tradeoffs they are making.

A recent national survey (CSIRO and University of South Australia 2008a) found

that the mean number of minutes 9–16 year olds spent on SSR was nearly double

the national guidelines (DoHA 2004a, 2004b). Using the most generous criteria for

meeting the guidelines, about a third of children met the guidelines on an average

day. The majority of SSR time was watching TV and the mean number of minutes

was lower for girls than boys. TV viewing peaked at age 12–14 years. The

relationship with age appeared to differ with gender and the type of SSR.

Marshall, Gorely and Biddle (2005) reviewed international trends in TV viewing in

youth between 1949 and 2004 and found that, while TV ownership increased

44 CHILDHOOD OBESITY

dramatically, TV viewing by youth who had access to a TV remained relatively

constant.

Various Australian studies have attempted to estimate the relationship between SSR

and weight. Wake, Hesketh and Waters (2003) found a significant cross sectional

relationship between watching TV and BMI, but not for playing computer/video

games and BMI. However, only 1 per cent of total BMI variance could be explained

by watching TV. Bogaert et al. (2003) found no significant relationship between

hours of watching TV and weight change in a 12-month study. Hesketh et al. (2007)

found in a longitudinal study a significant relationship between SSR and BMI with

an extra hour of SSR increasing the odds of being overweight three years later by

3 per cent. The causal effects appeared to go in both directions: higher BMI leading

to higher SSR, and higher SSR leading to higher BMI.

Similar results have been found internationally. Marshall et al. (2004), in a

meta-analysis, found a significant, but very small, positive relationship between

different types of SSR and body fatness. They also found that TV viewing may

displace more physically demanding pursuits.

Parents shape the family and home environment for their children. Parenting styles

and family characteristics can influence children’s dietary intake and activity levels.

Parental body weight is often described as a predictor of obesity in children, both in

their childhood and adulthood. This is not only due to the shared genetic

characteristics between parent and child, which may cause a predisposition to being

obese, but also shared attitudes and behaviours of parents with their children.

Australian literature indicates there is a relationship between parental BMI and their

children’s BMI in childhood and early adulthood. Magarey et al. (2003) found that

mother’s and father’s BMI were shown to be significantly but weakly correlated

with children’s BMI in some age groups. Another Australian study found that

children with overweight or obese fathers or mothers had consistently higher BMIs,

and BMI in 18 year olds was significantly predicted by father’s and mother’s BMI

(Burke, Beilin and Dunbar 2001). A study of preschoolers found that children with

an overweight mother were nearly twice as likely, and children with an obese

mother nearly three times as likely, to be in a heavier BMI category when compared

with a child with a non-overweight mother (Wake, Hardy et al. 2007). Other

international studies (Lake, Power and Cole 1997; Lee et al. 2006; Whitaker et al.

1997) have also found a relationship.

Parenting style and behaviour are also thought to influence their child’s weight

status. An Australian study using data on 4–5 year olds found no relationship

between mothers’ parenting behaviours and style and the odds of their child being

in a heavier BMI category, but they did for fathers. For example, children with more

controlling fathers had lower odds of being in a higher BMI category. A one point

increase in paternal control reduced the odds of the child being in a heavier BMI

category by 26 per cent (where control is measured as the frequency with which

fathers reported parenting behaviours that set and enforced clear expectations and

limits for their children’s behaviour, on a Likert scale from 1 (never/almost never)

to 5 (all the time)) (Wake, Nicholson et al. 2007).

In adults, education has often been linked to health outcomes, including adult

obesity. Higher levels of education might provide greater access to health-related

information and improved ability to handle such information, clearer perception of

the risks associated with lifestyle choices, and improved self-control and

consistency of preferences over time (Sassi et al. 2009). In Australia, an additional

year of education was found to be associated with a lower chance of being obese in

adulthood. There was no significant difference in the relationship between people of

different ethnicity (Sassi et al. 2009). To the extent that higher levels of education

provide parents with greater access to information and clearer perceptions of risks

of obesity, it might influence their child’s weight status as well. Education is also

used as an indicator of SES status (discussed below).

An international review found causal evidence that parenting affects children’s

eating patterns. However, there was insufficient evidence to determine whether or

not parenting affected childhood weight status via eating patterns (Ventura and

Birch 2008).

Other literature on the influence of parents on childhood obesity has looked at the

influence of maternal employment on overweight children. One study in the United

States found that children were more likely to be overweight the more hours their

mother worked per week. This study was conducted among higher SES families,

despite these children being least likely to have weight problems (Anderson,

Butcher and Levine 2003). The study found that if a mother in the top income

quartile worked an extra 10 hours per week, the child was between 1.4 and

3.8 percentage points more likely to be overweight. A recent study in Australia of

children at ages 4–5 years and 6–7 years found the children of mothers who worked

part-time watched less TV and were less likely to be overweight than children of

mothers who were not employed or who worked full-time (Brown et al. 2010).

An Australian study of 10–12 year olds (using data over three years) found family

physical activity and sedentary environments influenced weight status, but results

46 CHILDHOOD OBESITY

differed by gender. Among boys, sedentary equipment in the home was associated

with greater increases in BMI-z score. (A BMI-z score indicates the relativity of a

particular BMI to the mean for that age and gender.) Among girls, sibling physical

activity and physical activity items in the home were associated with greater

decreases in BMI z-score (Timperio, Salmon et al. 2008).

Another Australian study of primary school students examined whether aspects of

the family food environment were associated with weight status. Few significant

associations were found. Nonetheless, among older children, more frequent dinner

consumption while watching TV was associated with a higher BMI z-score and less

frequent breakfast consumption and more frequent fast food consumption at home

was associated with higher likelihood of being overweight (Macfarlane et al. 2009).

Advertising on TV and other media is often cited as a potential cause of child

obesity. Some evidence suggests that Australian children have in the past been

exposed to more advertisements directed at them than in many other countries

(Dibb 1996). Advertising is thought to influence childhood obesity through

influencing children’s preference for energy-dense nutrient-poor foods. A recent

analysis of Australian TV advertising found that food advertisements made up 31

per cent of all advertisements. Of these, 81 per cent were for ‘unhealthy/non core

foods’. Overall, advertisements for ‘unhealthy/non core foods’ accounted for 25 per

cent of advertisements between 7am and 9pm on weekdays. The most concentrated

timeslot for ‘unhealthy/non core foods’ advertisements was the early morning

timeslot on Saturday (Chapman, Nicholas and Supramaniam 2006).

The development of children’s knowledge and understanding of advertising is

related to their cognitive development and social maturation. As they age, children

develop the ability to differentiate advertising and non-commercial content, and

understand and interpret the persuasive intent of advertising (John 1999; Kunkel

et al. 2004).

Although the ability to distinguish between TV shows and advertising starts to

develop in the pre-school years (ages 3–4 years), this does not necessarily reflect an

understanding of the intent and bias behind advertising, which does not usually

emerge until ages 7–8 years. However, even if a child understands the intent and

bias of advertising, some research suggests that this knowledge has little effect on a

child’s desire for advertised products (John 1999; Kunkel et al. 2004).

What is the effect of TV food advertisements on knowledge, preferences,

behaviours and obesity? According to Brand (2007), there appears to be evidence of

a correlation between advertising and knowledge and preferences. Also, as noted

earlier, a number of studies have found a link between TV viewing and obesity, but

the size is usually modest and causation has not been established. In addition, it is

difficult to isolate the effect of advertising from other factors that affect the TV

viewing and obesity relationship, such as the sedentary nature of TV viewing.

Carter (2006) has reached similar conclusions.

Socioeconomic status (SES) is linked with obesity. SES is a measure of an

individual or group’s relative social and economic standing and ideally takes into

account income, education and occupation. In Australia, there is a higher prevalence

of obesity in low SES groups.

For example, if low energy-dense food were relatively more expensive than less

healthy energy-dense food, it may be that low SES groups could not literally afford

to be thin. In addition, as discussed in chapter 2, jobs in developed countries have

become increasingly sedentary and, as a result, more people are now having to give

up alternative pursuits to exercise.

Another explanation might be that being obese in adulthood results in less job

opportunities if it affects an individual’s ability to do their job, or if there is wage or

job discrimination, suggesting a causal relationship from adult obesity to SES.

However, there is evidence to suggest that, in Australia, the obese do not suffer a

wage penalty (Kortt and Leigh 2009).

Research suggests a link between parent SES and the SES of their child in

adulthood, and also a link between parent obesity and child obesity (Sobal and

Stunkard 1989). If social mobility is limited, the relationship between SES and

obesity may persist.

The recent National Health Survey found that children living in the area of greatest

relative disadvantage had higher overweight prevalence, and more than twice the

prevalence of obesity, than the least disadvantaged children (ABS 2009b). Many

other Australian studies also find this relationship (Booth et al. 2006; O’Dea 2008;

Salmon, Timperio et al. 2005; Wake, Hardy et al. 2007), although not all (Garnett

48 CHILDHOOD OBESITY

et al. 2005). Some find a link for one gender but not the other (for example, Booth

et al. 2001).

An international review (Sobal and Stunkard 1989) found that in developed

societies, such as Australia, the relationship between SES and obesity for both male

and female children was weak and inconsistent. However, different measures of

SES (for example, income or occupation) make comparing the results across studies

difficult. A more recent review (Ball and Crawford 2005) studied the relationship

between SES and weight change in adults and found that when occupational-based

indices (for example, low versus high skilled occupations) were used as an indicator

of SES, there was evidence of an inverse relationship for both men and women. The

results were more inconsistent when education or income were used as an indicator

of SES.

Ethnicity is also often associated with obesity. The influence on obesity could be

through cultural factors. For example, in many western cultures thinness is seen as

the ideal and overweight is seen as undesirable, but in many other cultures this is

not the case. The influence may also be genetic.

A recent national level survey found significant variation in obesity prevalence for

different ethnicities. Pacific Islander and Middle Eastern/Arabic adolescents were

most likely to be obese when compared with other adolescents. Anglo/Caucasian

and Asian children of all ages were the least likely to be obese (O’Dea 2008). In

addition, Wake, Hardy et al. (2007) found that Indigenous 4–5 year olds were

1.5 times more likely of being in a higher weight category than non-Indigenous

children. The difference between Indigenous and non-Indigenous weight outcomes

appears to be even more pronounced in adulthood (ABS 2006b). Other Australian

studies have also found a relationship between ethnicity and obesity (for example,

Booth et al. 2006).

These studies all used BMI as their measure of obesity. Evidence suggests that the

relationship between percentage body fat and BMI in adults differs among people of

different ethnicities, suggesting the general cut-off points for BMI might over or

underestimate obesity for people of different ethnicities (Deurenberg, Yap and van

Staveren 1998). If this is also true for children it may influence the interpretation of

the results of these studies.

O’Dea (2008) also examined weight perceptions and found that there appeared to be

a trend towards females of an ethnic background with higher prevalence of obesity

being less likely to perceive themselves as ‘too fat’ than other groups. This could

support the idea that overweight is seen as desirable in some cultures.

The physical environment is another potential factor affecting childhood obesity

through its influence on levels of physical activity. For example, it is argued that

areas with parks and bike paths may facilitate physical activity among residents,

whereas areas designed for motorised transport or with few facilities for physical

activity, would not. In addition, the urban environment may influence eating habits,

if the density and proximity of different food stores in an area affect relative costs

and hence dietary choices. The physical environment has changed over time, and

this has likely influenced changes in decision making related to dietary intake and

physical activity, potentially leading to increased obesity.

While there is a growing body of research in this area, there are some study

limitations. A lack of consensus exists on how to measure many environmental

variables and the size of the area that influences an individual (Dunton et al. 2009).

In addition, people will often spend time in multiple geographic areas, making it

difficult to identify the environmental factors that influence an individual. For

example, children will spend a lot of their time at home and at school. There is also

a wide array of potential physical environmental factors that can be studied. Often

the choice of variables is based on data available rather than any theoretical

underpinning (Ball, Timperio and Crawford 2006). In addition, due to the design of

many of the studies, the direction of the relationship cannot be determined. For

example, it may be that, rather than living near a park resulting in people being

more active, more active people choose to live near parks.

Australian studies have examined the prevalence of obesity in rural and urban areas.

Booth et al. (2001) included data from three surveys and found that, while the

prevalence of obese boys was greater in urban areas in two 1997 state surveys, there

was no difference for girls. In addition, no relationship was found in the National

Nutrition Survey data. However, in a different study, no relationship was found for

males, but for females BMI and skinfolds had increased by a significantly greater

amount in urban areas relative to their rural counterparts over a five-year period

(Dollman and Pilgrim 2005). Several other studies (Dollman, Norton and Tucker

2002; Cleland et al. 2010) found no significant relationship.

Some Australian studies have examined the relationship between specific aspects of

the physical environment and childhood obesity. Timperio et al. (2005) examined

the link between perception of the local environment and childhood overweight and

50 CHILDHOOD OBESITY

obesity. They found that a parental view that there was heavy traffic in their area

and concern for road safety was associated with higher overweight and obesity in

older children; however, there was no relationship in younger children. Crawford

et al. (2008) examined the association between childhood obesity and

neighbourhood fast food outlets. This provided little support for the notion that

exposure to fast food outlets in the local neighbourhood increased obesity. They

found older children who had at least one fast food outlet within 2km of their home

actually had significantly lower BMI scores. There was no relationship for younger

children.

A recent international literature review found the association between the physical

environment and childhood obesity was mixed and differed with factors such as

age, gender, population density and SES (it also found that the association differed

according to whether reports were made by the child or parent). For adolescents,

obesity appeared to be related to urban sprawl, neighbourhood pattern and access to

equipment and facilities. There was not sufficient empirical evidence to determine if

a relationship existed between obesity and most of the environmental variables

considered (Dunton et al. 2009).

Another review found no clear link between urban environment and childhood

obesity. However, it did reveal that characteristics of the built environment have an

effect on other factors, that in turn are believed to influence overweight and obesity

— such as active recreation and ‘active transport’ (such as walking or cycling)

(Sallis and Glanz 2006).

While the literature examining the relationship between the physical environment

and obesity is quite small, the literature looking at the association between the

physical environment and obesity-related factors, such as physical activity and

healthy eating, is more expansive. Some Australian studies have looked at these

relationships. Timperio et al. (2009) examined the relationship between availability

of fast food close to home and on the route to school and children’s intake. They

found that only availability of fast food within 800m of home was associated with

intake, but it was an inverse relationship. Each additional outlet within 800m was

associated with lower odds (3 per cent) of consuming take away or fast foods at

least once weekly. This result is consistent with Crawford et al. (2008) (discussed

above).

Timperio, Giles-Corti et al. (2008) examined the associations between public open

spaces and children’s physical activity and found that overall, there were no clear

relationships between many aspects of the closest public open space to where

children live and their physical activity. Hume et al. (2009) conducted a

longitudinal examination of environmental predictors of children walking or cycling

to school and found that most of the individual preference, social and physical

environmental factors had no influence on active commuting over time. However,

children whose parents were satisfied with the number of pedestrian crossings were

more likely to increase their ‘active transport’, while this was more likely to

decrease if their parents perceived there to be insufficient traffic lights.

Davison and Lawson (2006) conducted a review of the relationship between the

physical environment and children’s physical activity and found that physical

activity appeared to be related to recreational infrastructure, transport infrastructure,

and local conditions. However, the findings were more mixed for recreational

infrastructure and local conditions compared to transport infrastructure.

Much of the evidence of the relationships between different possible factors and

childhood obesity appears ambiguous. While it is easy to make a theoretical

argument for how and why each of these possible factors can affect body weight,

the evidence for many is at best limited, often mixed, and sometimes counter

intuitive. Methodological and data limitations are also issues with many studies of

childhood obesity.

The following broad conclusions are drawn from the literature.

• Australian children’s energy intake appears to have risen since the 1980s, but in

• The relationship between physical activity and obesity may be stronger than for

• Parents with high BMIs tend to have children with high BMIs, but this does not

52 CHILDHOOD OBESITY

family characteristics. Eating patterns, such as eating in front of the TV, the

working status of mothers, and the degree of parental control may all be

important factors.

• Australian children are exposed to a relatively high number of advertisements

• There is a higher prevalence of obesity in Australian children from low

• The physical environment is an area receiving increasing attention in the

Figure 3.2 Potential factors in the rise of childhood overweight and

• Food consumption

• Soft drink

• Physical activity

− Sport participation

− Walking to school

• Sedentary activities

− TV

− Computer/video

• Parents and families

− Income

− Parents’ work habits

• Advertising

• Physical environment

− Urban sprawl

− Access to

− Fast food outlets

− Safety

• Knowledge

• School environment

• Peer behaviours and

In chapter 2 the possible rationales for government intervention to address

childhood obesity were considered, including information gaps, pricing distortions

and behavioural limitations. In this chapter the different policy instruments and

factors that should be taken into account when considering alternative areas for

intervention are considered.

This can be difficult because obesity occurs in a complex economic and social

environment, which makes it hard to isolate and evaluate the effects of

interventions. Multiple causes of obesity make the task more difficult because of the

potential for interactive effects of various policies and for unintended consequences.

As for any policy intervention, clear specification of policy objectives is essential

both to promote appropriate policy design and to facilitate assessment of its

effectiveness (involving comparison with other policies). The primary goal should

be to increase aggregate wellbeing of the community, taking into account all of the

benefits and costs (financial and other), with the best policy option being the one

that generates the highest net benefit. The benefits of childhood obesity

interventions include the enhanced wellbeing of those obese children whose weight

is reduced. The costs include the resources consumed in developing, delivering,

monitoring and evaluating the intervention and the efficiency costs of raising

revenue (discussed in box 2.2 in chapter 2), and the costs the intervention imposes

on others.

Interventions that have the best likelihood of achieving this overall goal will need to

be effective, feasible, and have limited unintended consequences. Policy

effectiveness will be enhanced by addressing policy objectives as directly as

possible, and flexible policies will be more effective under changing circumstances.

Feasible policies will be consistent with current legislation and other policy settings,

and within the constitutional powers of the jurisdiction involved. For example, State

Governments do not have the constitutional powers to tax energy-dense

nutrient-poor foods. The risk of unintended consequences will be minimised by

matching the instrument to the objective as clearly as possible. This can involve

carefully targeting the group that give rise to the social cost, and avoiding those

consumers where the policy imposes undue costs, or targeting the foods that give

rise to the costs. For example, in the United Kingdom, a regulatory ban on fat,

56 CHILDHOOD OBESITY

sugar, and salt in fact unexpectedly reduced the availability of cheese and yoghurt

— foods that are valued for their nutritional content.

Because people might value incremental gains or losses differently, the effects of

policy measures on different groups might need to be considered. For example,

taxes on foods can be regressive (they affect lower income groups more

significantly) (Chouinard et al. 2007). But matters of equity and fairness are

difficult to assess as they are inherently subjective and perceptions about them vary.

A variety of policy mechanisms can be aimed at directly or indirectly influencing

childhood obesity. These mechanisms may include price instruments (such as taxes

or subsidies) that seek to change the incentive structure for decision making,

measures that seek to make consumers better informed (education and information),

or regulatory measures that influence consumer or producer choices.

Relevant obesity-related policy instruments, although not mutually exclusive, can

be categorised as follows:

• taxes and subsidies to affect relative prices

• information

• regulation to constrain production or consumption, or enforce particular

Box 4.1 Assessing the costs of policy interventions

• set-up and implementation costs

• administration costs

• enforcement costs.

• administrative costs (‘paper burden’) — such as the cost of reporting to government

• fees and charges levied by government (which may offset some of the costs to

• changes required in production, transportation and marketing procedures (including

• loss of sales revenue — this could be offset by higher sales for other businesses

• costs associated with doing business in multiple states — where there are

• restricted access to markets.

• loss of choice

• compliance and participation costs

• risk associated with the intervention being changed or withdrawn

• higher prices of goods and services (including where the costs to business are

• distributional effects

• prejudice and stigma

• reduced consumer vigilance (‘lulling’ effect)

• information overload

• adoption of unhealthy lifestyle habits (for example, extreme dieting)

• poorly targeted taxes may adversely affect demand for goods that might not have

58 CHILDHOOD OBESITY

Weight-related consumption decisions, made by parents on behalf of children or by

children themselves, may be influenced by prices (see section 2.1 in chapter 2).

Taxing or subsidising certain foods (and drinks) is one possible way to attempt to

reduce obesity. For example, Taiwan is currently planning to introduce a tax on

‘junk food’. The Bureau of Health Promotion is drafting a bill that would see a

special tax on foods containing high quantities of sugar and fat, and on alcoholic

drinks (The Age 2009).

Taxes can shift preferences away from consumption of certain goods or services,

and subsidies can direct preferences towards certain goods or services. For example,

a tax may be introduced on energy-dense nutrient-poor foods or a subsidy provided

on fresh fruit and vegetables, or in the case of physical activity, a subsidy to

increase children’s participation in sport.

Taxes and subsidies on specific goods or services can be a useful tool for addressing

significant externalities (negative or positive). From an economic efficiency

perspective, an ideal tax aimed to correct for market failure would be set equal to

the difference between private and social costs. In the case of smoking, for example,

the ideal tax would be set just equal to the marginal externalities associated with

smoking, so that the smoker would face the full social costs of smoking. As a result,

the subsequent consumption level chosen by the individual would be the socially

optimal level of smoking. However, as discussed in chapter 2, there are limited

externalities associated with obesity, which can usually be dealt with in low cost

ways (such as through passengers moving or adjusting their seating position to

address loss of personal space), suggesting taxes and subsidies are not justified on

externality grounds.

To be practically implemented, criteria for determining which goods or services to

tax or subsidise would be required. The ease or difficulty of such as task, as well as

the effectiveness of the resulting policy, depends on how well defined the links are

between the target of the tax and obesity.

Food is a necessity good, and its consumption only results in obesity in some

situations. In the case of healthy children with well balanced diets, most, if not all,

foods might be ‘healthy’, but for obese children, food high in fat content may be

unhealthy no matter what nutrients it contains. The efficiency losses of reducing

food consumption by low-risk individuals could outweigh the efficiency gains of

reducing food consumption of obese individuals. Also, obesity may arise from

consuming large amounts of food, rather than consuming particular types of food

deemed unhealthy and therefore taxable.

Particular foods could be taxed (for example, doughnuts) or, instead, foods with

particular characteristics (for example, foods high in sugar). It might be relatively

simple to tax well recognised categories of food that play little useful role in

nutrition rather than the underlying characteristics of those foods, but this might

have unintended consequences (Jacobson and Brownell 2000). For example, taxing

an underlying characteristic such as high fat could tax nuts, which have nutritional

benefits as well.

For a food tax to improve community wellbeing, policy makers would need to

measure the amount of obesity-related social costs attributed specifically to certain

foods, compared to other obesity-related consumption or behaviours (such as

sedentary lifestyle). Further, any tax considerations would be further complicated if

the relationship of costs to consumption is nonlinear. Strnad (2005) observed that

for food consumption, at extreme levels (low and high), an additional unit of

consumption would likely generate very little increase in disease risk. On the other

hand, individuals with medium-level food consumption likely have a higher

marginal increase in disease risk for each additional unit of food consumption.

In addition, taxing an energy-dense nutrient-poor food could be a cumbersome way

of trying to curb consumption of energy content (such as fat) and may become

redundant if changing tastes or changing technologies lead to use of different

ingredients. A targeted tax, such as a per kilojoule or per unit of fat tax, may be

more appropriate, as obesity and its consequences are more closely correlated with

the food itself rather than the cost of food (Freebairn 2010).

The effectiveness of taxes or subsidies will depend on the purchaser’s (parent or

child) responsiveness to changes in price (consumer price elasticity of demand), and

if there are substitute goods available (see section 5.1 in chapter 5). In some cases

demand for certain goods will not be responsive to changes in price, meaning that

taxes or subsidies will have little effect. A good example (Access Economics 2006)

is tap water. In Australia tap water is an essentially free and healthy drink, but its

near zero price does not necessarily induce people to drink it rather than more

expensive and less healthy alternatives, such as soft drink. Several studies indicate

consumers have limited responsiveness to food taxes as a general rule, but with

some exceptions (see section 5.1 in chapter 5).

Taxes on energy-dense nutrient-poor food affect consumer demand in two ways —

demand for all goods typically falls as purchasing power falls (income effect), and

demand for non-taxed food items increases as consumers substitute other foods

60 CHILDHOOD OBESITY

(now relatively cheaper) for the newly taxed energy-dense nutrient-poor food

(substitution effect). Taxes on foods tend to disproportionately burden the poor

(they are regressive) (Chouinard et al. 2007), as food costs make up a higher

proportion of household expenditure and consequently the income effect for them is

likely to be greater than for others. Taxes on foods may also impose costs on

low-risk individuals that moderately consume those foods.

The distribution of costs and benefits resulting from, say, a food tax may also

influence the acceptability, and therefore feasibility, of some policies.

Internationally, many food taxes have been withdrawn after short periods of time

due to industry pressure and popular concern (Caraher and Cowburn 2005).

Effectively functioning markets require information so that people and

organisations can make informed decisions. In some cases, for example, parents

may not be aware of the health risks associated with the food their children eat.

Alternatively, parents may lack the time or cognitive ability to comprehend that

information, or consider such an investment in understanding is not warranted by

the potential rewards of more informed choice.

Food producers generally have greater access to relevant food product information,

and greater certainty about its content than consumers. Further, consumer and

producer interests in information may not align. Policies that aim either to increase

information, improve the quality of information available, or improve access to

information can help people to make informed decisions, according to their own

preferences and in their own best interests.

Governments can assist in addressing information deficiencies either through

directly providing information (such as pamphlets with nutritional information) or

by encouraging (or requiring through regulation) the provision of better information

(such as the fat content of food) by businesses or others.

Depending on how it is framed, providing information may have little effect on

relative prices and hence be less distortionary than other measures (such as bans on

certain products, which reduce choice for everyone). At the same time, it has

negligible impacts on those consumers who already make better-informed choices

(or for whom such products do not involve health effects).

In the case of labelling, in addition to the administrative costs imposed on

government, labelling can impose costs on firms. Costs to firms of food ingredient

labelling would include testing food content as well as the labelling itself (which are

likely to be passed on to consumers).

Labelling of food can be simple or detailed and can help disseminate information

about products. Labelling by businesses may be introduced either through voluntary

means or through regulation (mandatory labelling).

Information provision or education components are common among many obesity

prevention interventions targeted at children.

One example of an intervention that includes information provision is

‘Get Up & Grow’. This Australian Government funded strategy includes

non-mandatory healthy eating and physical activity guidelines for early childhood

settings, such as centre-based care, family day care and preschools (DoHA 2009c).

The guidelines, aimed at directors/coordinators, staff/carers, and families cover

healthy eating (including guidelines on breastfeeding, composition of diet and

eating breakfast) and physical activity (including how much active play and

sedentary behaviour that children under five years of age should undertake). They

encourage childcare and preschool providers to enforce their own guidelines on

healthy eating and physical activity, and encourage providers to help parents to

make decisions in relation to healthy eating and physical activity when the child is

at home.

Other possible interventions include information and guidelines to control

children’s eating environment. A body of evidence suggests that the eating and food

environment can have important effects on our food consumption decisions

(chapter 2). Guidelines might encourage schools to have groups of children sit

together to eat their lunch, and adjust the size and content of the eating groups to

regulate how much children eat (Just 2006).

Other examples of information provision relating to preventing obesity include:

• pamphlets with nutritional information provided to parents on contributory

• social marketing campaigns that provide fact sheets and communicate the need

• education for parents on preparing healthy food

• voluntary labelling of particular characteristics of foods (for example,

• obesity education in schools.

62 CHILDHOOD OBESITY

Information can be delivered in various ways (such as through information labels on

products or through mass-media advertising) and the best format depends on a

number of factors — do consumers generally take notice of labels or advertising

campaigns? Do they understand the information, and will they act on it?

Consumers face vast amounts of information, which can create a ‘poverty of

attention’. Also, providing more information to consumers can reduce the perceived

need for precaution and lead to unintended consequences. New generations or new

waves of consumers may not be ‘reached’ unless information strategies are

implemented on a repeated basis.

Achieving behavioural change through providing information is likely to work best

where it targets behaviour resulting mainly from ignorance, and where consumers

are motivated to change once they have that information. Numerous social

marketing campaigns have aimed to raise awareness of obesity, and improve

lifestyle behaviours associated with obesity (refer to appendixes A and B for

Australian policies aimed at addressing obesity in children). Despite this, a high

proportion of Australian adults are overweight or obese, a high proportion of

Australian children are overweight or obese, and half of the overweight Australians

think they are healthy (Zurich Financial Services Australia and Heart Foundation

2008). But it is difficult to say how effective these campaigns have been —

overweight and obesity prevalence might have been higher in their absence.

The insights from behavioural economics suggest that people may have cognitive

limitations (chapter 2) that can limit the effectiveness of information (including

labels), or that some ways of providing information may be more effective than

others. Information may have limited effect for many reasons — for example,

people have limited capacity to process information, even if they have access to it.

As a result, policy measures, such as nutrition labelling requirements, can benefit

from knowledge about consumers’ abilities to access and compute information. On

the other hand, information provision can play an important role in countering

distortions to consumer preferences. Australian evidence suggests that a third of

consumers refer to labelling information on products when purchasing them for the

first time, and referring to labelling was positively related to health consciousness

(FSANZ 2008). Some evidence on the effectiveness of labelling is presented in

chapter 5 (section 5.1).

Some research indicates that knowledge about health reduces the likelihood of an

individual being obese (Nayga 2000), but there is mixed evidence relating to the

effectiveness of health information on food consumption and diet. Downs,

Loewenstein and Wisdom (2009) found there is little evidence that information

alone affects diet. Higher education appears to be linked to the use of nutrition

labels (Kim, Nayga and Capps 2000 in Finke and Huston 2003). The motive to

acquire specific nutrition and health knowledge seems to be linked to an

individual’s rate of time preference, illustrated by behaviours such as exercise,

smoking and education (Huston, Finke and Bhargava 2002 in Finke and Huston

2003).

Regulation encompasses a broad spectrum of interventions, ranging from

self-regulation (such as industry codes of conduct where industry is solely

responsible for enforcement), quasi-regulation (such as government-endorsed

industry codes of conduct), co-regulation (where government provides legislative

backing to industry arrangements), to explicit government regulation of activities

(black letter law). Regulation might also require information provision by

businesses or other third parties (discussed above).

Regulation may be more or less prescriptive. More prescriptive regulation can

provide more certainty about outcomes but it is less flexible to accommodate

different or changing circumstances than principles-based or performance-based

regulation, which can enable businesses and individuals to choose the most

effective way of complying. Principles-based or performance-based regulation often

has greater overall benefits.

Examples of regulation relating to preventing obesity include (from those that are

designed to inform choice to more ‘heavy-handed’ approaches that limit choice):

• mandatory labelling of foods (including nutrition information labels or warning

• self-regulation to limit children’s exposure to unhealthy food and drink

• banning television advertising of certain foods to children

• guidelines (where there is pressure from government to comply) to ban certain

• standards prohibiting certain types of food and drinks being sold in schools

• nutritional standards limiting the content of specific ingredients in foods

• nutritional standards controlling the types of foods that can be sold.

64 CHILDHOOD OBESITY

Box 4.2 Self-regulation: the Responsible Children’s Marketing

The Responsible Children’s Marketing Initiative is a self-regulation initiative of the food

A particular example of regulatory action is ‘Fresh Tastes @ School’, a NSW

school canteen strategy. The strategy aims to encourage students to eat healthier

foods, and is a mandatory set of rules regarding what NSW Government school

canteens can sell. It has been in operation since 2005 (Healthy Kids Association

?2010b; NSW Department of Health and NSW Department of Education and

Training 2006). It uses the traffic light colours of red, amber and green to categorise

different canteen foods. Green foods are ‘fill the menu’ foods, which should make

up the majority of the canteen menu and be promoted to students as the best choice,

amber foods are ‘select carefully’ foods, which should only be offered for sale on

certain days of the week, and red foods are ‘occasional’ foods, which cannot be sold

on more than two occasions in a school term.

Another example is legislation recently introduced in Japan, in response to concerns

about rising healthcare costs, setting a maximum waist circumference —

85 centimetres for men and 90 centimetres for women — for all employees over

40 years old. Japan’s obesity rate is under 5 per cent, although levels of obesity

have risen over the past few decades and total healthcare costs are expected to

account for 11.5 per cent of gross domestic product by 2020. Employees undergo

check-ups once a year and employees found to exceed the maximum waist

measurement are required to undergo counselling. Companies are required to reduce

their number of overweight employees by 10 per cent by 2012 and 25 per cent by

2015. If targets are not met, companies may have to increase their payments into a

health fund for the elderly (Nakamura 2009).

Regulation that reduces obesity will deliver benefits to society, but the net benefit

will depend on the costs. The costs associated with regulations — those incurred by

government, business and consumers — will depend on their specific nature and

their complexity. Regulations will likely involve higher costs if they change the

behaviour of some groups that are not the target of regulation, such as physically

active children forced to do more sport at the expense of academic study.

Costs to government and business may be more where regulation does not reflect

accepted commercial practices. Costs to consumers will depend on how much their

choices and preferences are restricted — for example, their costs would be higher

under a strategy that restricts the sale of energy-dense nutrient-poor foods than a

non-mandatory strategy that encourages reduced consumption of such goods.

Were regulation introduced to restrict sales of certain foods, costs to business would

include loss of profit from lower sales, although there may be offsetting gains from

the sale of healthier alternatives.

Some arrangements will be more flexible than others. A strategy that allows, say,

individual childcare and preschool providers to enforce their own guidelines would

be more flexible than one prescribed by government, and would allow different

providers to tailor strategies specific to their particular target group (such as

children from different cultural backgrounds), and respond to changes that may

impact its effectiveness.

Regulation can also have unintended consequences. Mandated changes that seek to

reduce individual risk can have adverse consequences if they result in reduced

consumer vigilance. For example, Viscusi (1984) found that the ‘lulling’ effect of

child-resistant safety caps on aspirin and other drugs negated the otherwise

desirable effect of reduced child poisonings, and may have lead to additional

poisonings. Currie and Hotz (2004) explored the relationship between childcare

regulation and accidental injuries. The authors found that where regulation was

introduced, the risk of fatal and non-fatal injuries were significantly reduced for

children in childcare, but that the increased expense discouraged some families

66 CHILDHOOD OBESITY

from using childcare and increased the risk for those children. Obesity relevant

examples include the banning of certain foods in school canteens, resulting in

falling profit, or possibly the development of black markets or absenteeism (which

might be detrimental to the school and community in other ways).

The effectiveness of regulation aimed at children’s eating habits is discussed in

chapter 5 (section 5.1).

This chapter considered different policy instruments and the factors that should be

taken into account when considering those alternatives, including their effectiveness

(figure 4.1).

The considerable uncertainty about the causes of obesity suggests that hard

interventions, such as taxes or subsidies on specific goods and services, would be

difficult to justify. Further, the practical challenges of designing taxes on specific

goods and services limit the likelihood of them being effective in addressing obesity

(and may lead to perverse outcomes). Softer interventions, targeted at addressing

information failure and education, appear to be on stronger ground. The complex

nature of obesity suggests that multi-pronged strategies addressing multiple risk

factors may be more effective than other strategies that focus on a single risk factor.

In the next chapter the evidence base on the effectiveness of past and present

interventions to address childhood obesity is considered. Interventions include both

targeted interventions conducted by government or others (mostly in school

settings), along with broader, community-wide interventions (such as taxes on

energy-dense nutrient-poor foods or television advertising bans). In the next chapter

the following question is considered: how might policy makers best address the

challenge of childhood obesity in the future?

Figure 4.1 What is the scope for policy solutions to reduce childhood

• Information

• Externalities

• Urban environment

• School environment

• Parental labour force

• Time

• Persuasive

• People do not

• Information provision

• Insurance that can

• Energy-dense

• Exercise programs (for

• Advertising

• Commitment

• Can it be implemented?

• Is there evidence the option is effective?

• Will the option impose unnecessary costs on those not overweight or obese?

• Is the option equitable?

• Will the option increase net social benefits to the community?

This chapter examines the evidence on the effectiveness of interventions designed

to address childhood obesity (both overseas and in Australia). Interventions

encompass both measures that target particular groups — whether conducted by

government or others — along with broader measures that are rolled out at a

community-wide level, such as social marketing campaigns or national television

advertising bans (some of these broader initiatives are not just targeted at children),

usually delivered by governments. Some implications for future interventions are

presented.

This section considers the effectiveness of obesity-related interventions, by

examining international and Australian evidence. International evidence includes

systematic reviews of trials that tested various interventions, along with research on

other policy measures (such as advertising bans). Australian evidence includes the

outcomes of various interventions, including trials and other studies and programs.

This section also examines methodological issues that may affect the reliability of

conclusions drawn from the research. In assessing evidence associated with obesity

interventions, it is first necessary to consider the concept of evidence and what

constitutes good evidence.

Central to evidence-based policy is the use of rigorous and tested evidence

(box 5.1). Good evidence that assesses the effect of policy is required to ensure that

interventions are effective and provide net benefits to the community. Some

evidence is more relevant than others but can be difficult and/or costly to obtain.

70 CHILDHOOD OBESITY

Box 5.1 Evidence-based policy

• Evidence-based policy transparently uses rigorous and tested evidence in the

• Features of good policy design and evaluation methodologies include:

– carefully defining the policy problem and establishing clear objectives

– developing a range of policy options drawing on a coherent framework of theory

– using evidence to test those options, in a cost–benefit framework where possible

– explicitly addressing the counterfactual, that is, what would have happened in the

– examining direct and indirect effects on the economy and the community.

• The choice of methodology depends on the task and the type of evidence available.

• Evidence-based policy requires more than good policy formulation methodologies

– improving transparency

– building in and financing evaluation from policy commencement

– using sequential roll-out, pilots and randomised trials where appropriate

– establishing channels to disseminate evaluations and share results across

– strengthening links between evidence and the decision-making process.

Source: PC (2010).

Some study designs are more robust than others. This has given rise to the concept

of an evidence hierarchy to rank evidence based on different research methods (for

example, see Leigh 2009). Systematic reviews of randomised controlled trials

(RCTs) are placed at the apex of the hierarchy. Such reviews use systematic and

explicit methods to identify, select, and critically appraise research in a

comprehensive and unbiased way. Lower down the hierarchy are before–after

(pre−post) studies. In some cases, ‘higher’ level evidence will be difficult to obtain,

and findings from less robust studies may need to be used as a means of shaping

policy.

Various systematic reviews provide an understanding of the success of

obesity-related interventions in other countries (and sometimes in Australia), and a

range of these focus on interventions that target children. Further research provides

an assessment of the likely effectiveness of taxes, advertising bans and mandatory

caloric labelling in fast food restaurants. Interventions include those delivered by

government and other institutions.

Summerbell et al. (2005) undertook a systematic review (update) of a large number

of interventions for preventing obesity in children on behalf of the Cochrane

Collaboration (which is generally regarded as an authoritative voice in systematic

reviews in healthcare). The study included 22 interventions from 1990 to February

2005, of which 12 were short-term and 10 were long-term. The interventions

reviewed included those designed to act on diet and nutrition, exercise and physical

activity, and/or lifestyle and social support. Two interventions focused on dietary

education (both long-term studies), 6 on physical activity (2 long-term and 4

short-term studies), and 14 on a combination of dietary education and physical

activity (6 long-term and 8 short-term studies) (box 5.2). Although many

interventions were not effective in preventing weight gain, they were more often

effective at improving lifestyle behaviours.

Of the two long-term dietary related interventions, one focused on fruit and

vegetable intake, and the other on reducing carbonated drink consumption. Neither

intervention resulted in statistically significantly anthropometry (body

measurement) differences between the intervention and control groups, although

there was a reduction in self-reported soft drink consumption in the intervention

group in one study.

Of the long-term physical activity related interventions, one focused on physical

activity for kindergarten children, and the other on a physical education program

with a self-management component. For the first intervention, at the short-term

follow-up a reduction in obesity prevalence among the intervention groups almost

reached statistical significance. In the case of the second intervention, small

differences in Body Mass Index (BMI) for girls and boys were observed relative to

the control group, although the statistical significance was not estimated.

72 CHILDHOOD OBESITY

Box 5.2 Interventions for preventing obesity in children (review)

• increasing fruit and vegetable intake in US children (with at least one obese parent)

• reducing carbonated drink consumption in UK children (randomised controlled trial)

• a regimen of exercise for Thai kindergarten children (randomised controlled trial)

• a physical education program with a self-management component for school

• a multi-component program addressing body fat in American Indian school children

• a physical activity and nutritional program addressing obesity and promoting

• a physical activity, nutritional and sedentary behaviour program to reduce Body

• a nutritional education and ‘active breaks’ program for children in Germany (KOPS,

• a multi-disciplinary program targeting UK school children (APPLES, randomised

• a 20 week physical activity and nutrition program for children in England, involving

Box 5.2 (continued)

• a 12 week physical activity program in the United States, involving dance

• a physical activity program for girls in the United States, with BMI at or above the

• a physical activity program for fourth grade school children in the United States

• a six month education and monitoring program to reduce screen-based activity

• a program for African-American girls from middle-income families, involving day

• a program for African-American girls from low-income families, some involving

• a program for African-American girls from low-income families, involving dance and

• a program for African-American girls from low-income families, involving after school

• a program for children in day care centres in the United States (randomised

• a 16 week home visiting intervention for young children (average age 21 months)

• a Chilean program for children in grades 1 to 8 (Kain et al. 2004)

• a 12 week program for inner-city girls with African-American backgrounds, focusing

Source: Summerbell et al. (2005).

Of the short-term studies that focus on physical activity, two reported significant

improvements to BMI, although Summerbell et al. (2005) noted the methodology

appears weak in one of these. The other successful intervention used education,

monitoring and reporting to reduce television and other screen-based recreation. It

74 CHILDHOOD OBESITY

reported a statistically significant reduction in BMI and all other measures of body

fat. Because there was no assessment beyond six months, it is not known if the

outcomes were sustained. Of the two other interventions, one used education,

physical activity (such as kick-boxing, self-defence and water aerobics) and social

support sessions to increase girls physical activity levels and improve nutrition, and

the other promoted play behaviour, used teacher-directed activities and encouraged

self-directed activities in girls and boys. Where neither reported significant

reductions in BMI, positive changes in behaviours were reported (for example, in

the first study, girls were significantly more active in some of the target groups

compared to the girls in the control group).

Of the six long-term combined dietary and physical activity interventions,

Summerbell et al. (2005) noted one as a ‘good quality’, school-based,

‘multi-component’, ‘multi-centre’ RCT, that included family involvement.

Nonetheless, this intervention showed no significant differences in BMI (or other

related measures) at the end of the three year intervention. Other similar

interventions include one aimed at reducing energy, fat and sodium of school meals

and increasing physical activity; one ‘high quality’ RCT among ethnically diverse

school children aimed at reducing sedentary behaviour; and a multidisciplinary

RCT that included modification of school meals. These reported respectively: no

impact on obesity on follow-up; obesity reduced among intervention girls but not

boys; and no differences in change in BMI between the intervention and control

groups. However, in a number of these studies there were positive behavioural

changes. For example, the multi-component study found reduced school lunch

calorie intake from fat, and another reported improved fruit and vegetable intake

(and an associated smaller daily increase in total energy intake) among girls.

Of the short-term combined studies, four are pilot studies and although all reported

positive trends in anthropometry, the differences were not significant. Nonetheless,

one study reported that on follow-up girls were consuming fewer sweetened drinks,

and in another, several significant improvements in dietary practices were observed.

The remaining four studies in this group also had some limited success for

anthropometry measures. More promisingly, in one study, behaviours such as

television viewing were significantly improved (the number of children watching

more than two hours per day was lower in the intervention group), and in others,

energy intake was reduced.

Other reviews of childhood obesity-related interventions have come to broadly

similar conclusions.

• Micucci, Thomas and Vohra (2002) found that interventions were more effective

• Thomas et al. (2004) noted that although some interventions found statistically

• Ammerman et al. (2002) found that interventions appeared to be more successful

• Livingstone, McCaffrey and Rennie (2006), in their review of methodological

• In a review of healthy eating interventions, French (2005) found that overall,

The reviews above do not cover the full spectrum of obesity prevention policy tools

(outlined in chapter 4). The reviews mostly tend to focus on interventions involving

information provision, education, physical activity and nutritional programs. In this

section we review the available evidence on the effectiveness of some

community-wide interventions.

Several studies indicate consumers have limited responsiveness to food taxes as a

general rule, but with some exceptions. For example, Kuchler, Tegene and Harris

(2005) estimated that demand for snack foods was unresponsive to price change

(inelastic), after taking into account quality variation. Some more recent research

indicates that relative price changes (between ‘unhealthful’ and ‘healthful’ foods)

could only explain about 1 per cent of the growth in BMI and the incidence of being

overweight or obese (Gelbach, Klick and Stratmann 2009). In addition, US research

using individual-level data and data on state-level soft drink taxes found no

76 CHILDHOOD OBESITY

significant association between soft drink taxes and BMI (Powell et al. 2009; Sturm

et al. 2010). In contrast, Mytton et al. (2007) found that a carefully targeted ‘fat tax’

could produce modest changes in food consumption (although this research did not

extend to its effect on body weight). Furthermore, the effects of price instruments

have been shown to be stronger for lower socioeconomic status groups than in other

groups in the population (Smed, Jensen and Denver 2007). There is also some

evidence that children respond to relative prices of low and high-fat foods (French

et al. 1997).

Taxing particular foods can affect the consumption of other foods, with

unpredictable health effects (Mytton et al. 2007). A US study of soft drink taxation

found that the tax reduced soft drink consumption, but that the consumption of other

high energy-dense drinks (such as milk and fruit juice) increased, resulting in no

statistically significant reduction in overall energy intake (Fletcher, Frisvold and

Tefft 2009, cited in Freebairn 2010).

A recent review (Powell and Chaloupka 2009) of empirical evidence on food and

restaurant prices and weight outcomes found that when statistically significant

associations were found between food and restaurant prices and weight outcomes,

the effects were generally small in magnitude. In some cases, however, they were

larger for low socioeconomic status populations and for those at risk of overweight

or obesity. The authors concluded that small taxes (or subsidies) are unlikely to

produce significant changes in obesity prevalence, although non-trivial pricing

interventions might.

The effects of changes in food prices on demand for snacks have been tested in

some school-based interventions. One study (CHIPS in French 2005) found that

demand for lower fat snacks was responsive to price change, and a related pilot

study found the demand for fresh fruit was also price responsive.

A recent seven-month study, exploring the effects of financial incentives (totalling

US$750 over the course of a year) on weight loss, suggests they might be effective,

particularly if implemented over a long period (Volpp 2009). At 16 weeks,

individuals in the incentive groups had lost significantly more weight than those in

the control group. Although the incentive groups regained weight after the study’s

completion, they remained at a lower weight than when the study began. The

incentive plans were based on small, frequent rewards, which provided immediate

and tangible feedback that made it easier for individuals to do in the short term what

is in their long-term best interest. The results indicate that interventions that

incorporate insights from behavioural economics (such as time inconsistent

preferences, discussed in chapter 2) can be effective.

As outlined in chapter 3, Australian children are exposed to a relatively high

number of advertisements for energy-dense nutrient-poor foods. This has led to

many calling for a ban on advertising of these foods. Yet, while research shows that

television viewing and childhood obesity are related, the direction of causation and

the magnitude of the contribution of food advertising to obesity is uncertain. In

addition, the link between television viewing and childhood obesity is very small

(Carter 2006). While research shows correlations between advertising and

children’s preferences, there is no strong evidence of a causal relationship between

advertising and children’s food preferences and weight outcomes. It is also difficult

to isolate the effect of advertising from other factors that affect the television

viewing and obesity relationship, such as the sedentary nature of television viewing

(Brand 2007).

If, as the evidence suggests, the link between television viewing and childhood

obesity is tenuous, or at most small in magnitude, it is unlikely that banning the

advertising of energy-dense food would significantly address childhood obesity

prevalence (Carter 2006).

Where restrictions on television advertising of energy-dense nutrient-poor foods

aimed at children have been implemented (Sweden, Norway, the Canadian province

of Quebec and the United Kingdom), the evidence is inconclusive. In some places

restrictions may have been undermined by a number of factors. For example,

restrictions in Sweden and Quebec do not apply to advertising on foreign TV

channels, and yet both source much of their TV from outside their jurisdiction. In

addition, Sweden’s regulations banned advertisements designed to ‘attract the

attention of children’, and Quebec’s bans applied to advertisements ‘directed at

children’. As a result, the intent of the advertising ban could be side-stepped by

aiming advertising at people other than children. A lack of national obesity data in

Sweden has also limited assessment of the effectiveness of the restrictions

(Handsley et al. 2007; National Preventative Health Taskforce 2009).

Two studies have examined the effectiveness of mandatory posting of calories on

menus in chain restaurants in New York, introduced in 2008. Using Starbucks data,

Bollinger et al. (2010) found a decrease in calories per transaction, and a greater

decrease for higher calorie consumption individuals. However, the decrease was

largely driven by food purchases (rather than drinks), and there were no data on

whether consumers increased calorie consumption elsewhere. The estimated effect

was small in magnitude — the reduction in calorie consumption would reduce body

78 CHILDHOOD OBESITY

weight by 1 per cent, and this ‘back-of-the-envelope calculation suggests that

average reductions resulting from calorie posting in chain restaurants will not by

themselves have a major impact on obesity’ (p. 24). Another study of the same

mandatory posting of calories policy (Elbel et al. 2009) found no significant change

in calories purchased among adults using data from fast-food restaurants in

low-income, minority New York communities.

Most reviews of childhood obesity prevention strategies are international and as a

rule they contain few Australian interventions. As a systematic review of all

Australian interventions is outside the scope of this study, this section draws on

some specific Australian interventions to provide a general indication of their

effectiveness.

Appendixes A and B list interventions in Australia to address childhood obesity.

Appendix A includes interventions for which evaluation reports have been

published and appendix B provides a stocktake of the remainder, based on publicly

available information. They include research studies and programs that include

children. They specifically include prevention-related interventions and exclude

those designed to treat obesity. Due to the inconsistency with which organisations

publicly release program information and the vast number of interventions that may

indirectly affect obesity, it is unlikely to be a definitive list. Nonetheless, the

appendixes contain more than 100 interventions. They mostly date from the mid

1990s, although one earlier cluster-randomised trial report is included (Dwyer et al.

1983). It is one of few Australian studies included in international reviews.

Most of the interventions include an education component (for example, nutritional

education), and some more directly target behaviour by requiring participation in

physical activity. Some interventions attempt to influence consumption of certain

foods, for example by influencing school canteen menus or promotions. The

funding sources for these interventions are varied and include Australian, state,

territory and local governments. They include interventions with multiple

components, and a small number of interventions focus on building community

capacity.

Although few Australian interventions list obesity prevention among their stated

objectives, they relate to obesity because they focus on at least one part of the

‘energy equation’, that is:

• ‘energy in’

• ‘energy out’

• both ‘energy in’ and ‘energy out’.

• body composition

• level of fitness

• behaviours (such as physical activity levels or diet)

• cholesterol or blood pressure

• knowledge or attitudes.

80 CHILDHOOD OBESITY

Only about one-third of the prevention interventions measured body composition

outcomes, such as BMI or percentage of body fat. About one-half measured

physical activity levels, and about one-third measured physical fitness. About

one-third also measured dietary habits (behaviours). Three interventions measured

cholesterol or blood pressure (factors that are linked to obesity and risk factors for

long-term health outcomes associated with obesity, such as cardiovascular disease).

About one-third measured knowledge or attitudes, such as beliefs regarding fruit

and vegetable consumption. Fewer than one-third measured sedentary behaviour.

Very few interventions considered the potential for unintended consequences.

Few interventions included, or plan to include, a long-term follow-up evaluation to

assess the sustainability of the effects of the intervention. Only one intervention

included a (planned) economic evaluation.

The Australian interventions have been grouped according to whether body

composition was measured or not.

Australian interventions that measured body composition results show mixed results

and there has been limited long-term follow-up to assess sustainability of short-term

outcomes. In some cases, however, these interventions have recorded success for

other measured outcomes, such as level of physical activity:

• The Be Active Eat Well multi-strategy intervention (table 5.1) has received

Table 5.1 Be Active Eat Well

• N (Intervention baseline): 4 preschools, 6 primary schools,

• N (Intervention post-intervention): 4 preschools, 6 primary schools,

• N (Comparison baseline): 4 preschools, 12 primary schools,

• N (Comparison post-intervention): 4 preschools, 12 primary schools,

Resultsa The intervention group had statistically significant lower increases in

a Results adjusted for baseline variable, age and height at follow-up, gender, duration between measurements

Sources: Bell et al. (2008); Sanigorski et al. (2008).

• Promising results were released recently for Romp and Chomp (table A.20), an

82 CHILDHOOD OBESITY

2 year-olds and 3.4 percentage points in 3–4 year-olds. While these results

appear positive, there appears to be no plan for a long-term follow-up to assess

their sustainability.

• The body composition changes for Switch–Play (table A.24), which focused on

• The FILA program (table A.8), which focused on physical activity and diet,

• The Burke et al. (1998) (table A.3) health promotion intervention in Western

• The Vandongen et al. (1995) (table A.26) fitness and nutrition intervention had

• A much earlier physical activity related intervention (Dwyer et al. 1983)

The Australian interventions that did not measure body composition outcomes

varied in terms of objectives. Some focused on enhanced physical activity levels,

increased enjoyment or awareness of the benefits of physical activity, improved

health knowledge and fitness, and improved dietary habits (for example, increased

fruit/vegetable/water consumption) or awareness. The interventions varied in terms

of outcome measures. Some included objectively recorded measures (such as actual

physical activity), self-reported behaviours (such as subjects’ own report of dietary

habits), actual level of physical fitness, and self-perception (such as body image or

knowledge). The interventions varied in terms of target population. Two national

social marketing campaigns (‘Get Moving’ and the ‘Go for 2&5’ campaign) were

included. Some aimed at primary school children and some targeted specific groups

(such as children in economically disadvantaged areas, adolescents girls, or students

from mostly non-English speaking backgrounds). They also varied in terms of study

design, and include quasi-experimental designs, pre and post, one pilot randomised

controlled trial, and a cluster-randomised controlled trial.

Some interventions showed more positive results for some measured outcomes than

others. Selected examples are presented below:

• Engaging Adolescent Girls in School Sport (a pilot study), succeeded in

• Learning to Enjoy Activity with Friends (LEAF) promoted physical activity, and

• Tooty Fruity Vegie Project, a long-term intervention, aimed at increasing

There is little evidence of the effectiveness of other community-wide interventions

in Australia, partly because other than general information and awareness programs,

few have been attempted. For example, specific food taxes and mandatory energy

content on restaurant menus have not been implemented in Australia (although

Victoria recently announced it will introduce mandatory calorie content on menus

for certain fast-food outlets in 2012 (VicHealth 2010)). The GST, introduced in July

2000, and its GST-free status of ‘basic’ food and drinks including fresh fruit and

84 CHILDHOOD OBESITY

vegetables, could provide important price sensitivity evidence but there appear to be

no research studies on this (this may be because there is no reliable pre and

post-data). A range of initiatives are in place to reduce children’s exposure to

advertising of energy-dense nutrient-poor foods and drinks (Australian Government

2010), including the Responsible Children’s Marketing Initiative (box 4.2,

chapter 4). However, there has been no evaluation on its effectiveness.

A range of issues and shortcomings affect the robustness of reported results of the

interventions to prevent obesity. Some international systematic reviews found

methodological weaknesses in all studies under review. Some study design and

evaluation issues included:

• In undertaking obesity prevention interventions, the reliability of results can be

• Differences in ethnicities and socioeconomic position can also limit the

• Few studies report on sub-groups (Ammerman et al. 2002 and Thomas et al.

• Many studies are ‘under-powered’ (Summerbell et al. 2005). Power is the

• Many of the studies included in the Summerbell et al. (2005) review have ‘unit

analysis when allocation is by school can result in overly narrow confidence

intervals. If clustering is not taken into account, it may lead to false conclusions.

• Involvement in a trial alone can induce behaviour change. In many cases the

• ‘Confounding bias’ is also a problem as some studies do not control adequately

The apparent shortcomings of many policy interventions and the limitations of the

assessment of their outcomes, to a large extent reflect the inherent complexity and

the multiple causes of obesity itself. This complexity poses serious challenges for

policy design and can confound attempts to measure the contribution of obesity

prevention and reduction strategies. The mixed results of past and current

interventions suggest that policy making in this area could benefit from improved

data collection and new approaches to their implementation and design.

Policy makers in many instances face insufficient access to, or simply inadequate,

data and information. Policy relevant data are important at every stage of the policy

cycle, from identifying the policy problem, through assessing policy options, to

ex-post evaluation of interventions. Population level data and intervention specific

data need to be collected and presented consistently to enable robust comparisons.

However, data can be costly to obtain, and interventions require appropriate funding

allocation between the intervention itself and evidence gathering. Too, though, the

information gathering and evaluation process can form part of the actions affecting

outcomes, resulting in misleadingly optimistic results.

• Consistent data are lacking to fully understand childhood obesity prevalence in

that included measured child height and weight data.1

86 CHILDHOOD OBESITY

• Consistent data are not always collected at the intervention level (see

• Also, long-term follow-up evidence is necessary to understand the sustainability

(Wong 2009, p. 7227).2 Also, the National Health and Hospitals Reform

reviews found different results for girls and boys), and interventions might be better

tailored to the different groups (Micucci, Thomas and Vohra 2002). This also

suggests that community-wide interventions may be more effective on some

sub-groups than others.

In order to be effective, future interventions may need to be of greater length and

intensity to successfully modify weight (Summerbell et al. 2005).

Some of these factors may be influenced by practical limitations, such as the

particular nature of research project funding arrangements. For example, funding

may not allow for interventions over longer periods, or longer term follow-up. The

evaluation component of projects is important — the costs of not evaluating

interventions can be greater if in the long run ineffective interventions continue or

effective ones are ended — but costs can be substantial. For example, the support

and evaluation component was at least 50 per cent of Be Active Eat Well project

funds (Bell et al. 2008). Evaluation expenditure should be assessed in a cost–benefit

framework, taking into account any potential wasted expenditure if the intervention

proves ineffective.

Building supportive environments conducive to behavioural change appear to be

important. Most interventions focus on short-term effects and behaviour change

rather than body composition change. Interventions imposed on a community are

less likely to influence behaviours than those that actively involve that community

in their design and implementation.

Greater experimentation may be desirable. Australia’s federal system provides a

useful model for experiment, enabling comparison and contrasting of alternative

policies across state borders (such as restrictions on television advertising of

energy-dense nutrient-poor foods aimed at children).

Policy design should include consideration of potential unintended consequences.

For example, targeting a group with a disproportionately high level of obesity may

make sense, but there is a risk of stigmatisation and marginalisation. For example,

Dobbins et al. (2009) note this in the context of school-based physical activity

programs. Care may need to be taken to ensure obesity-related interventions do not

contribute to body image issues or eating disorders (O’Dea 2005). Studies included

in systematic reviews appear to not identify underweight children (typically these

are grouped with normal weight children), so it is not known if the interventions

have unintended effects on these children. Van Wijnen et al. (2009) found that very

few childhood obesity interventions investigate the effects of their programs on the

psychosocial wellbeing of children and adolescents. There may be value in

embedding obesity prevention interventions into existing broader programs that

already target groups that are high-risk for obesity.

88 CHILDHOOD OBESITY

Although many past interventions to address childhood obesity have struggled to

demonstrate effectiveness, there are some early signs that those that are community

based, multifactorial, and long-term in nature may prove more effective. The

authors of the Cochrane review (Summerbell et al. 2005) stated that the most

promising interventions were underway and are yet to report findings, and that more

recent studies are conducting trials with more attention to participant involvement

and more comprehensive evaluations.

Several recent community-based interventions aimed at preventing obesity in

children include these aspects in their design. In Australia, one is the Be Active Eat

Well program (mentioned earlier, table 5.1). The intervention was designed and

implemented by parents and local organisations (such as schools and community

agencies). The strategy includes nutrition strategies, physical activity strategies and

screen time strategies to promote healthy eating and physical activity. This

long-term intervention ran for several years. As outlined above, the early results

were promising, showing significant results for most of the body composition

measures in the intervention group.

EPODE (Ensemble prévenons l’obésité des enfants, which translates to ‘together,

let’s prevent obesity in children’) is a community-based intervention launched in

France in 2004. In each town the intervention is led by a committee, and

suggestions are received for different community initiatives, activities and diets.

Initiatives may include organising games at school playtime, walk to school groups,

and learning about vegetables in the classroom. Children aged between 5 and

12 years have their BMI calculated annually, with overweight children, or children

at risk of being overweight, being encouraged to see a doctor (Westley 2007).

Half of the towns showed a statistically significant decrease in overweight and

obesity combined between 2005 and 2007. Although the results look very

encouraging, there is no clear and reliable control group against which they can be

assessed (Borys 2008; Westley 2007).

EPODE is in place in a large number of towns in France and has spawned similar

interventions in Belgium, Greece and Spain (European Public Health Alliance

2008). Mexico is also launching a new campaign to address obesity, also based on

the French model (The Independent 2010). Obesity Prevention and Lifestyle (see

table B.55), a South Australian intervention, launched in March 2009, was based on

EPODE (Hill and Lomax-Smith 2009).

The available evidence suggests that many interventions to prevent childhood

obesity have not been effective in preventing weight gain to any significant degree,

although they may have had greater success improving lifestyle behaviours

generating other benefits (such as a healthier diet and increased physical activity).

This is encouraging, and despite not improving children’s weight, may improve

their health and wellbeing. However, some of these interventions are very

expensive, making it unlikely that their limited benefits outweigh the costs (table

5.2).

On balance, governments appear yet to find a way to effectively intervene to reduce

obesity prevalence among children. This may be because, despite some

understanding of the factors driving obesity, an effective response to such a

complex problem is inherently challenging, or policy tools have been poorly

designed (for example, interventions have not properly targeted the causes of

obesity). Alternatively, effective policies may not be identified as such because of a

lack of robust evaluation. In some cases, potentially effective policies may be being

thwarted by institutions or other constraints.

This suggests the need for a measured, incremental approach to intervention in this

area, with a focus on program evaluation to weed out ineffective and inefficient

programs as well as to identify those that deliver net benefits. While data collection

and evaluation can be expensive, pilot programs and trials can help to contain both

program and evaluation costs. Successful programs can then be implemented more

widely. Evaluation expenditure should be assessed in a cost–benefit framework,

taking into account any potential wasted expenditure if the intervention proves

ineffective (and that the evaluation itself may influence outcomes).

Where evidence on benefits remains limited and uncertain, a number of strategies

can help ensure that interventions generate net community benefits:

• First, experiment with low-cost programs that have low risks of collateral

• Second, for programs with potentially high costs, implement trials that allow

• Finally, evaluate to facilitate sharing and thus quicker adoption of successful

90 CHILDHOOD OBESITY

In summary, given the lack of firm evidence that childhood obesity prevention

measures have been effective, policies should be designed based on

cost-effectiveness, and implemented gradually, with a focus on evidence gathering,

information sharing, evaluation and consequent policy modification. A greater

commitment to evaluation, with a focus on methodological rigour, may produce

more robust and conclusive evidence.

Table 5.2 Characteristics of different policy interventions to prevent childhood obesity

Table 5.2 (continued)

schoolsa

Table 5.2 (continued)

a Although non-mandatory guidelines are voluntary, to the extent that schools adopt them they become bans on types of food and drinks sold in schools.

A number of Australian interventions address childhood obesity directly or

indirectly by targeting diet and/or physical activity. This appendix includes

Australian childhood obesity related interventions for which we have located

evaluation reports. Most of the Australian interventions identified date from the

mid 1990s. The studies and programs included here focus on prevention, rather than

the treatment or management of childhood obesity. (Interventions for which no

published evaluations were identified are listed in appendix B.)

The interventions in this appendix result from desk-based research and are based on

publicly available information. Due to the inconsistency with which bodies publicly

release program information, this is unlikely to be a definitive list.

The information is as complete as possible — information provided for some

interventions is more detailed than others. In the case of funding details, funding

organisations and funding figures may not reflect the complete picture. Where

specific information was not able to be located, it is left blank.

96 CHILDHOOD OBESITY

Table A.1 Active After-schools Communities (AASC) program

• Computer assisted telephone interviewing: Each cohort had pre- and

• Web-based surveys with stakeholders: Each year (2005, 2006, 2007)

• N (Cohort 1, 2005–2006, intervention group): 664 parents of

• N (Cohort 1, comparison group): 750 parents of non-participating

• N (Cohort 2, 2006–2007, intervention group): 635 parents of

• N (Cohort 2, comparison group): 695 parents of non-participating

• N (Cohort 3, 2007–2008, intervention group): 936 parents of

• N (Cohort 3, comparison group): 936 parents of non-participating

• N (2005): 542 school and out-of-school hours care staff,

• N (2006): 1158 school and out-of-school hours care staff,

• N (2007): 1789 school and out-of-school hours care staff,

Table A.1 (continued)

• levels of and attitudes towards physical activity

• children, their parents, and organisations’ thoughts on AASC

• satisfaction with the AASC program

Resultsa Children participating in the AASC program almost doubled their

a Details given here are of the interim evaluation. The final evaluation was due mid-2009.

Source: Australian Sports Commission (2008); DoHA (2009a, 2009b).

98 CHILDHOOD OBESITY

Table A.2 Be Active Eat Well

• N (Intervention baseline): 4 preschools, 6 primary schools,

• N (Intervention post-intervention): 4 preschools, 6 primary schools,

• N (Comparison baseline): 4 preschools, 12 primary schools,

• N (Comparison post-intervention): 4 preschools, 12 primary schools,

Resultsa The intervention group had statistically significant lower increases in

a Results adjusted for baseline variable, age and height at follow-up, gender, duration between measurements

Source: Bell et al. (2008); Sanigorski et al. (2008).

Table A.3 Burke et al. 1998

Source: Burke et al. (1998).

100 CHILDHOOD OBESITY

Table A.4 Coalfields Healthy Heartbeat School Project

Source: Plotnikoff, Williams, and Fein (1999).

Table A.5 Dwyer et al. 1983

• First phase: Cluster-randomised controlled trial, randomised at class

• Second phase: Pre- (pre-data from first phase) and post-data (no

• Number of intervention groups: 2 (Skill group and Fitness group)

• Number of control groups: 1

• First phase: At end of 14-week intervention

• Second phase: 2 years after completion of the first phase

• First phase (1978): 14 weeks

• Second phase: 2 years, 1978–1980

• Both skill and fitness groups gained in endurance fitness and the

• The 1980 students had superior endurance fitness when compared

Source: Dwyer et al. (1983).

102 CHILDHOOD OBESITY

Table A.6 Eat Smart Play Smart

Source: National Heart Foundation of Australia (2006, 2009a).

Table A.7 Engaging adolescent girls in school sport

Resultsa The intervention group showed greater improvement in body image and

a This is a pilot RCT and as such is not sufficiently powered to know if between group differences were

Source: Dudley et al. (2009).

104 CHILDHOOD OBESITY

Table A.8 Fitness Improvement and Lifestyle Awareness (FILA)

Resultsa The intervention group had a smaller increase in BMI, greater reduction

a This is a pilot RCT and as such is not sufficiently powered to know if between group differences were

Source: Peralta, Jones, and Okely (2009).

Table A.9 Fresh Kids

• 2003–05: $60 000

• 2005–07: $60 000

Source: AIHW (2006b); Laurence, Peterkin and Burns (2007).

106 CHILDHOOD OBESITY

Table A.10 ‘Get Moving’ Campaign

Outcomes measureda Outcomes measured included:

• recall of campaign

• whether action was taken as a result of the campaign

• motivation as a result of the campaign

• awareness of guidelines on physical activity and electronic media use

• Significant increase in those who could recall the Get Moving

• About three quarters had seen the television commercial at follow-up

• 37 per cent of parents suggested that being exposed to the campaign

• Around 80 per cent of those who took action said the campaign was

• Over 95 per cent of children report seeing at least 1 element of the

• No change in organised sport as a result of the campaign

• Resulted in decrease in weekend sedentary time

• Over 80 per cent of children and teenagers said the campaign

• No obvious changes in attitude to physical activity, although attitude

• Increased awareness of the recommended levels of activity and

a All of the outcomes measured were self-reported.

Source: Elliot and Walker (2007a).

Table A.11 Girls Stepping Out Program

Source: Schofield, Mummery and Schofield (2005).

108 CHILDHOOD OBESITY

Table A.12 ‘Go for 2&5®’ Campaign

Description The ‘Go for 2&5®’ Campaign was a national social marketing campaign

Outcomes measureda Fruit and vegetable consumption

Go for 2&5® Campaign awareness

• Proportion who ate fruit did not change between surveys

• Between surveys there was a significant increase in the proportion

• Attitudes and beliefs about fruit consumption did not change between

• No improvement in the amount of parents that had attempted to

• High level of awareness of the 2&5® Campaign

• A significant amount of parents said they had taken action as a result

• No significant change in fruit and vegetable consumption

• Knowledge about recommended fruit and vegetable intake increased

• High level of awareness of the 2&5® Campaign and the main

• Large proportion said they had taken action as a result of the

a All of the outcomes measured were self-reported.

Source: Elliot and Walker (2007b).

Table A.13 Learning to Enjoy Activity with Friends (LEAF)

Source: Lubans and Morgan (2008).

110 CHILDHOOD OBESITY

Table A.14 Moorefit

Source: Cass and Price (2003).

Table A.15 Move It Groove It

Source: Barnett et al. (2009); van Beurden et al. (2003).

112 CHILDHOOD OBESITY

Table A.16 New South Wales Walk Safely to School Day

Source: Merom et al. (2005).

Table A.17 Nutrition Ready-to-Go @ Out of School Hours Care (NRG

• Proportion of OOSH care services with a physical activity component

• Amount of moderate-to-vigorous physical activity