# The cost-effectiveness of removing television advertising of high-fat and/or high-sugar food and beverages to Australian children

## Abstract

### Objective:

To model the health benefits and cost-effectiveness of banning television (TV) advertisements in Australia for energy-dense, nutrient-poor food and beverages during children's peak viewing times.

### The intervention

The intervention modelled was an extension of existing regulations within the Children's Television Standards to preclude advertising for EDNP foods, as well as for beverages and fast food outlets, during specified children's TV viewing hours and where a substantial proportion (15%) of children aged 5–14 years were in the viewing audience. This would effectively remove food and beverage advertising from TV during peak child viewing times in the morning for 1 to 2 h, and in the afternoon/evening for 5 h (up to 2100 hours).21

### Current practice

Current practice was specified according to the Children's Television Standards 16–23,22 which limit advertisements to 5 min every 30 min for 5 h per week of designated children's timeslots and prohibit advertisements for  h per week of designated preschool children's timeslots, as well as place some limitations on the content of advertisements. The content limits include accuracy of claims, sensitivity to cultural diversity and disabled groups, a quantum of Australian content, as well as not be frightening to children, nor display unsafe situations.

### Assessment of benefit

Benefits of intervention are calculated using a two-stage process. The first stage involves estimation of the health gain that could be attributed to the intervention measured in DALYs, an outcome measure combining premature years of life lost (YLL) and years lived with disability (YLD). The second stage involves a consideration of issues (filters) that either influence the degree of confidence that can be placed in the ICERs or broader issues that need to be taken into account in decision making with regard to resource allocation in health. The secondary filters considered qualitatively by the Working Group in this study include strength of evidence, equity, acceptability, feasibility, sustainability and side effects.

### Evidence of the impact of reduced advertising on food and beverage consumption

Randomized controlled trial evidence of the effect of reducing TV advertising of EDNP foods to children on BMI does not exist. Instead, we used the best available evidence, which included a combination of limited randomized controlled trial evidence of food consumption (behaviour change), together with modelled behavioural change to BMI change, using a mix of cross-sectional and longitudinal evidence.18, 23, 24 Parallel evidence of behavioural change measured with reductions in advertising of other products, such as toys, tobacco and alcohol, supported the conclusions reached in this analysis for EDNP foods.25, 26, 27, 28

The most relevant study to assess the effectiveness of a reduction in advertising was a randomized controlled trial that compared the impact of exposure to different food advertisements on food and beverage selections in 288 5- to 8-year old children on holiday camp for a period of 2 weeks in Quebec in 1982.23 The children were exposed to sweet commercials, no commercials, fruit commercials or nutritional public service announcements. Children who viewed sweet commercials chose (and ate) significantly more sweets over fruit as snacks compared with children in the other three groups (75 versus 67%, 64 and 65%, respectively). Children who viewed sweet commercials also consumed more sugar-sweetened KoolAid rather than orange juice, compared with children in the other three groups (75 versus 65%, 55 and 60%, respectively).

Gorn and Goldberg23 controlled for age and tested all combinations simultaneously. From the reported summary data, we assumed that the reported statistical significance of food choice difference was maintained for all age groups and used the percentage of time sweet food was consumed in the sweet-commercials group, compared with the average of the other three groups, as the measure of intervention effect. We assumed that the intervention effect applied to children up to the age of 14 years. We converted the percentage of time sweet food choices were made in the exposed and unexposed groups to relative risks of making sweet food choices (RR 0.87; 95% CI 0.74, 0.99). We also used the percentage of time KoolAid was consumed in the sweet-commercials group compared with the other three groups in the same manner, to determine the relative risk of making sweetened beverage choices (RR 0.80; 95% CI 0.68, 0.95).

For food intake, the intervention was thus estimated to reduce consumption of EDNP foods by 13% (to 87% of current practice). The impact of this reduction on energy intake and energy balance was assessed using previously analysed dietary data from the 1995 National Nutrition Survey in Australia.5, 29 Bell et al. classified foods and beverages as noncore or core according to the food groups defined in the Australian Guide to Healthy Eating.30 A classified listing of foods and beverages is detailed in Table 1. Bell et al. analysed them on the basis of their contribution to the total diet of Australian children with regard to energy (kJ per day) and weight (g per day) (Table 2). Noncore foods correspond to what is generally called EDNP foods based on individual food nutrient profile. However, it should be noted that by adopting Bell et al.'s classification of all breakfast cereals regardless of sugar content as core foods, and all fruit juices as core beverages, it is likely that the difference in energy density between EDNP and other foods was underestimated.29 The 13% reduction in the consumption of EDNP foods from the intervention trial was applied to the currently consumed average food weight of 265 g per day,5 yielding 35 g per day of EDNP foods swapped for equal weight of other foods, assuming no change in the total weight (g) of daily food consumed.20 As other foods generated a lower energy density of 6.5 kJ g−1 compared with 13.8 kJ g−1 for EDNP foods, the substitution of other foods for EDNP foods resulted in a reduced average energy intake of 259 kJ per day (35 g per day × (13.8–6.5) or 7.3kJ g−1), which was 3% of the total daily energy intake of 8220 kJ.5 We further assumed that no interactions occurred affecting beverages consumed or physical activity levels of Australian children. By reducing the EDNP to other food ratio in the diet, reductions in total energy intake occurred because of the different energy densities of EDNP and other foods (Table 2).

We then used the validated equation from Swinburn et al.18 to assess the impact of a change in energy balance on the body weight. Every 1% change in energy intake led to a 0.45% (95% CI 0.38, 0.51) change in weight. The 3% change in energy intake led to a 1.4% (0.45 × 3%) change in weight. The average weight of male and female children in Australia was reduced by 1.4%, yielding an average weight reduction of 0.54 kg per child. The weight reduction converted to a reduction of 0.26 BMI units (95% CI 0.07, 0.46) for the Australian population of 5–14-year olds. This BMI effect was assumed to occur within 12 months.

To model the impact of reduced sweetened beverages intake on BMI change, we used a prospective observational study.24 Ludwig et al.24 examined the association between change in consumption of sugar-sweetened beverages in 548 school children (mean age: 11.7 years, s.d.: 0.8) and difference in measures of obesity, over a period of 19 months between 1995 and 1997. They found that each additional serving of sugar-sweetened drink consumed (comprising 8–12 ounces) predicted a BMI increase after adjustment for confounders of 0.24kg m−2 (95% CI 0.10, 0.39; P=0.03). We assumed that, in the absence of any other behavioural change, the same size and opposite effect on BMI would occur from a reduction in sugar-sweetened beverage servings per day. We used the average consumption of grams of sweetened beverages per day from the National Nutrition Survey 1995 as current practice in 2001,5 and calculated the reduction in grams by applying the relative risks derived from the Gorn and Goldberg trial results ((1–0.80) × 496 g). The change in grams (94 g per day) was converted into servings per day, assuming that a serve was between 8 and 12 ounces,31 which is approximately 296 g (237–355 g). The reduced consumption of sweetened drinks observed in the trial converted to 0.33 fewer servings per day (296÷94 g) and generated a reduction of 0.08 BMI units (95% CI 0.03, 0.15) per child (0.33 serves × 0.24kg m−2). The BMI effect of reduced beverages was assumed to occur within 12 months.

In consideration of the likely continuation of most other nontelevised advertising in Australia, and even some compensatory increased advertising of other forms and at other televised times, we have adjusted the Gorn and Goldberg trial efficacy results downwards by an arbitrary factor of 40–60%. In this manner, we have adjusted downwards the reported trial efficacy of the potential intervention to a level of modelled effectiveness under more realistic Australian conditions.

The final effective reduction in BMI units due to the reduced EDNP food consumption modelled was 0.13 (95% uncertainty interval 0.03, 0.25) and 0.04 (95% uncertainty interval 0.01, 0.08) for reduced sweetened beverage consumption. The concept of uncertainty interval is detailed in the Supplemental Information.

### Assessment of costs

In Australia, compliance with regulations, together with the cost of noncompliance, rests with individual commercial broadcasters. We assumed that broadcasters would comply with tightened regulations to minimize any cost associated with noncompliance and subsequent complaint handling. The incremental costs of stricter monitoring and the enforcement of tightened regulations were estimated to be quite minimal with two extra staff at the Australian Communications and Media Authority (salary plus on-costs), as a regulatory framework already existed.

Cost offsets were assessed as future health sector costs saved because of fewer occurrences of obesity-related conditions in the adult life of children exposed to the intervention. Cost offsets were deducted from the cost of intervention to determine the net cost of intervention.

## Results

Quantitative results are presented in Table 3. Removing TV advertising for EDNP food and beverages when children make up a significant proportion of the audience had a gross ICER of AUD$3.70 (95% uncertainty interval$2.40, $7.70) per DALY. When the present value of potential savings in future health-care costs was considered (AUD$300 million), the intervention was ‘dominant’, because it resulted in both a health gain and a cost offset compared with current practice. The variables that most strongly correlated with the ICER were the relative risks of a reduced consumption of EDNP food (rs=0.88), followed by the effectiveness under Australian conditions assumption (rs=−0.28) and the relative risk of reduced beverages consumption (rs=0.19). Neither of the scenarios evaluated in sensitivity testing altered the major conclusions of the analysis. The intervention remained dominant under both scenarios (Table 4). These results assumed full maintenance of BMI benefit through adulthood. Threshold analysis identified that BMI benefit could be almost completely eroded over time, and that this intervention would remain dominant because of its modelled very low cost. However, if lost advertising revenue or lost sales of EDNP foods were to reach approximately AUD$2 billion per year without any substitution of replacement revenue, the intervention would then return a gross ICER of AUD$50 000 per DALY.32, 33

A complete consideration of the qualitative second-stage filters for this intervention is shown in Table 5. The limited evidence of the effectiveness of this intervention and lack of acceptability to some stakeholders of removing TV advertising were the key concerns of the Working Group with this analysis. Public health experts and industry representatives have very divergent views. The former Australian federal government was opposed to the increased regulation of the advertising industry in this form, preferring the self-regulation model, and the current government has it under review. Without political support, the intervention is neither feasible nor sustainable, despite its economic credentials.

## Discussion

Restricting TV advertising of EDNP food and beverages seems to be extremely cost-effective in reducing unhealthy weight gain in children aged 5–14 years. Although the BMI change per child was small, the total health benefit was high because of the large number of children affected and the low cost. In fact, this intervention was the most cost-effective at a population level of 13 interventions analysed in the ACE-Obesity project.20 However, an appropriate evaluation should accompany any implementation of the intervention, given the concerns with regard to the limited evidence base.

Economic analysis raises important issues as to what constitutes ‘value-for-money’, and it is not uncommon to use a threshold ICER as a guide for decision making. In the ACE-Obesity project, we used a threshold of AUD$50 000 per DALY saved and assumed that anything below this threshold was essentially a ‘good buy’. The probability that the cost per DALY saved for this intervention would fall below this threshold was 100%, suggesting that this intervention is a ‘very good buy’. In fact, 100% of the ICERs were less than AUD$10 per DALY.

Whether the reduction in BMI would be maintained over the lifetime of the child is unknown and difficult to predict. At a technical level, the maintenance assumption has the potential to distort the health benefit and cost-effectiveness results, but as there is a lack of available data for children, on which to base quantitative modelling, we applied threshold analysis, which indicated that even if most of the reduction in BMI was lost over time, the intervention would still be cost neutral, and probably save money, from a societal perspective. None of the scenarios examined in the sensitivity analysis altered the cost-effectiveness conclusions reached. Furthermore, it is possible that if policies are maintained over time, there could be a synergistic and multiplicative effect between obesity interventions, such that we would expect reductions in overall obesity rates.20

In addition, even if potential lost sales of EDNP foods to children were considered as a cost to industry, we estimated that AUD$2 billion sales per year to children would need to be lost before this intervention would reach the AUD$50 000 cost per DALY threshold that ACE studies have used as an arbitrary cutoff for cost-effectiveness evaluation.34 We were unable to determine whether the potential lost sales would be replaced by diversification within the food industry into adult food products or into healthier foods for children, but both options would be available. These potentially lost sales have similarly been excluded from the base analysis, but their possible impact has been explored in threshold analysis.

The multiple-country ecological analysis by Lobstein and Dibb19 supports these results. The relationship they described was the association between the national mean number of advertisements for sweet and fatty foods during children's TV and the national prevalence of overweight and obesity. Reducing the number of EDNP food advertisements from 10 per hour (about the rate in Australia) to 5 per hour (about the rate in Germany) could, assuming causality, reduce the prevalence of overweight and obesity by 0.05 percentage points.

Supportive evidence is also gained from the evaluation of other products (toys in children, tobacco products and alcohol in adults) for which reduced exposure to advertising has been evaluated. Robinson26, 35 found fewer demands for toys (OR=0.29), Saffer found reduced demand for tobacco products (RR=0.93),27 and for alcohol (RR=0.75).28 We assumed that adjusted odds ratios were a good estimate of adjusted relative risks. Although acknowledging the fact that some of these products were evaluated in adults and are addictive (in the case of tobacco), the magnitude of the effects of reduced advertising for these products is parallel evidence supporting the reduced demand for EDNP foods that we model (RR=0.87).

We have also assumed that not only was the cost of other food that was substituted for the EDNP food removed from the diet of equal dollar value, but the food preparation time component was also equivalent. If this were not the case, the intervention would become less acceptable to parents. Any cost differential would particularly affect lower socioeconomic families. A study conducted in the United States found that, in terms of cost per kJ, high-fat, high-sugar products were cheaper, but in terms of cost per gram, fruit and vegetables were the cheaper alternative.36 Our substitutions were estimated using weight in grams; therefore, replacing fruit bars, for example, with an equal weight of fresh fruit, or soft drinks with an equal weight of water, is likely to be cost-saving for consumers.

The first and third most important parameters influencing the results were derived from the Gorn and Goldberg trial, conducted many years ago in Quebec, on children out of their home environment, at camp. The trial was conducted in 1982 and the level of exposure to advertisements during the trial (5 min within 30 min of TV) was less than under Australian conditions in 2001.37 The impact of greater restrictions on televised advertising in Australia may thus be more or less than that found in the Gorn and Goldberg trial. To account for any possible overestimation of impact and to present a conservative approach, the reported efficacy of the trial was halved in the base model to reflect the different conditions in Australia. For these reasons, the intervention could be trialled and should be formally evaluated in an Australian setting to address the issues raised and better inform the policy debate.

Other limitations of the Gorn and Goldberg trial used as evidence include the inability to control for age as in the original study, and we have assumed that the results apply equally to all ages as the age range involved in the trial was quite narrow (5–8 years). The nature of advertising has also altered in the intervening years. The cumulative effects of other promotional activity apart from TV advertising cannot be ignored.10 The relative importance of family values and habits must also be taken into consideration.38

The only evaluated population-wide action to decrease TV advertisements to children is the Quebec experience. In 1980, Quebec passed a regulation to restrict all commercial advertisements for sweets or foods in addition to toys directed at children 13 years. An evaluation of the changes in regulation found a reduced number of high-sugar breakfast cereals in households exposed to less advertising.25 However, the study results were limited as they were unable to control for important confounding factors of family size and cereal box size. The recent UK strategy restricting advertising to children was informed by reviews of international regulation (acknowledged as culturally specific) and evaluations. The Office of Communication (Ofcom) recognized both the ethical and realistic difficulties of conducting formal trial-based evaluations to inform policy,7 which explains the absence of other useful international evaluations of the effect of restrictive regulations.

There is also likely to be concern with regard to the generalizability of international findings to local settings.7 Behavioural interventions may work differently in other countries because of cultural and educational factors, as well as because of current practice in obesity prevention. The impact on total BMI units and DALYs will depend on population size and structure, existing BMI distributions and rates of disease within countries.20 Costs are likely to vary depending on the degree of current regulation of advertising in place. For these reasons, implementation of the intervention should be accompanied by appropriate evaluation.

Some commentators have expressed concern that a decrease in TV station revenue due to a reduction in food advertising to children would have a negative impact on the quality of children's programming.39 However, the Quebec experience was examined for impact of advertising on children's programme variety and quantity, and the impact was found to be far less than expected.25

The method used in this study to convert the change in energy balance of food to a change in weight and BMI in children was novel and recently published.18 The approach taken by Swinburn et al. for estimating the impact of changes in energy intake on weight gain differs from the approaches published by other researchers such as Wang et al. and is thus not directly comparable.40 Swinburn et al.'s approach can be used to estimate the expected change in population weight when a population moves from one level of energy intake to another, allowing sufficient time for the population weight to arrive at a new settling point (equilibrium). In contrast, the approach used by Wang et al. (and Hill41) can be used to estimate changes in weight resulting from a daily gap between energy intake and energy expenditure—with the daily energy gap sustained over several years. We believe that Swinburn et al.'s approach is better suited to the type of intervention examined in this study and it has been validated against data from three longitudinal studies (mean duration: 3.4 years), showing that the equations predicted weight gain to within 0.5%.

This notwithstanding, it is possible to make some rough comparisons between the Swinburn et al. method and the Wang et al. method. Wang et al. estimated that a 131 kcal per day energy intake excess over energy expenditure (energy gap) would explain the increase in weight in 2- to 7-year olds of 0.43 kg per year over a 10-year period. Using Swinburn et al.'s method, an increased energy intake of 131 kcal per day over 2 years would result in an approximate increase in weight in 2- to 7-year olds of about 0.81 kg. The 2-year period between settling points is chosen because the half-life of the time taken for a population to arrive at a new settling point for weight is approximately 1 year,42 indicating that the 0.81 kg change in weight estimated using Swinburn et al.'s equation can be compared with a 0.86 kg change in weight using Wang et al.'s methods (0.43 kg per year for 2 years).

We chose to be conservative in our analysis, using the Ludwig method to assess the impact of reduced sweetened beverages, because it generated lower BMI change results than did the Swinburn et al. method.

Relying on the classification of sugary breakfast cereals and all fruit juices regardless of sugar content compared with other foods and beverages, respectively, has introduced the potential to understate the difference in energy density and importance of EDNP/other foods in the diet of Australian children. The most recently analysed national survey of Australian children's food consumption was conducted some years ago in 1995, but a more recent (2007) national survey indicated that total energy intake has not increased greatly since then. However, if the relative contribution of EDNP foods has increased in 2007, the impact of the intervention would be greater and the cost-effectiveness ratio would be improved. For these reasons, this analysis should also be regarded as conservative.

## Conclusion

Restricting televised advertisements targeting children could potentially be one of the most cost-effective population-based obesity prevention interventions available to governments today. Despite its economic credentials from a public health perspective, this initiative involved genuine concerns for some important stakeholders, which need to be addressed in a constructive manner. The second-stage filter analysis identified that the key decision points for implementing this intervention were twofold. There was an identified need to accompany implementation with an appropriate evaluation plan under the current Australian conditions, as the best available evidence was limited. Second, there was the need to address conflicting ideological positions with regard to the responsibility of individuals, industry and the government towards children in the public interest. This paper adds economic evidence to the growing international public support for government intervention to protect children by regulation.

## Conflict of interest

The authors declare no conflict of interest.

## Disclaimer

The opinions, analysis and conclusions in this article are those of the authors and are not necessarily endorsed by the Victorian Government Department of Human Services, the Victorian Minister for Health or by the Government of Victoria.

## References

1. 1

Cole T, Bellizzi M, Flegal K, Dietz W . Establishing a standard definition for child overweight and obesity worldwide: international survey. Br Med J 2000; 320: 1240–1245.

2. 2

Margarey A, Daniels L, Boulton T . Prevalence of overweight and obesity in Australian children and adolescents; reassessment of 1985 and 1995 data against new standard international definitions. Med J Aust 2001; 174: 561–564.

3. 3

Olds T . Prevalence of obesity and overweight in Australian children 9-16 years. 2007, Adelaide, personal communication.

4. 4

Australian Government Department of Health and Ageing, Australian Food and Grocery Council and Australian Government Department of Agriculture Fisheries and Forestry. 2007 Australian National Children's Nutrition and Physical Activity Survey. 2008.

5. 5

Australian Bureau of Statistics. National Nutrition Survey. Nutrient intakes and physical measurements. Australia. 4805.0. 1995. ABS: Canberra.

6. 6

Institute of Medicine. Preventing childhood obesity: health in the balance, in Fact Sheet September 2004. 2005.

7. 7

Livingstone S In: Department of Media and Communications (ed) A Commentary on the Research Evidence Regarding the Effects of Food Promotion on Children Prepared for the Research Department of the Office of Communication (OFCOM) 2004. London School of Economics and Political Science: London.

8. 8

Wake M, Hesketh K, Waters E . Television, computer use and body mass index in Australian primary school children. J Paediatr Child Health 2003; 39: 130–134.

9. 9

Neville L, Thomas M, Bauman A . Food advertising on Australian television: the extent of children's exposure. Health Promot Int 2005; 20: 105–112.

10. 10

Hastings G, Stead M, Mc Dermott L . Review of research on the effects of food promotion to children 2003. Centre for Social Marketing, University of Strathclyde: Glasgow.

11. 11

Institute of Medicine. Food Marketing to Children and Youth. Threat or Opportunity? 2006. National Academy of Sciences: Washington.

12. 12

Kotz K, Story M . Food advertisements during children's Saturday morning television programming: are they consistent with dietary recommendations? J Am Diet Assoc 1994; 94: 1296–1300.

13. 13

Oates C, Blades M, Gunter B . Children and television advertising: When do they understand persuasive intent? J Consumer Behav 2001; 1: 238–245.

14. 14

United Nations System Standing Committee on Nutrition. Statement by the Working Groups on Nutrition throughout the Life Cycle and Nutrition, Ethics and Human Rights, Rome. 26 February–5 March 2007. 2007. [cited 6 June 2007]. Available from: http://www.unsystem.org/scn/Publications/AnnualMeeting/SCN34/34_humanrights.htm.

15. 15

World Health Organisation (WHO). WHO European Charter on counteracting obesity. WHO European Ministerial Conference on Counteracting Obesity, Istanbul, 15–17 November 2006; Denmark. 2006. [cited 6 June 2007]. Available from: http://www.euro.who.int/Documnet/E89567.pdf.

16. 16

World Health Organisation (WHO). Prevention and control of non-communicable diseases: implementation of the global strategy. Sixtieth World Health Assembly, Resolution WHA60.23, Geneva, 23 May. 2007. [cited 6 June 2007]. Available from: http://www.who.int/gb/ebwha/pdf_files/WHA60/A60_R23-en.pdf.

17. 17

Ofcom. Ofcom publishes final Statement on the television advertising of food and drink products to children. [cited 21 September 2007]. Available from: http://www.ofcom.org.uk/media/mofaq/bdc/foodadsfaq/.

18. 18

Swinburn B, Jolley D, Kremer P, Salbe A, Ravussin E . Estimating the effects of energy imbalance on changes in body weight in children. Am J Clin Nutr 2006; 83: 859–863.

19. 19

Lobstein T, Dibb S . Evidence of a possible link between obesogenic food advertising and child overweight. Obesity Rev 2005; 6: 203–208.

20. 20

Haby MM, Vos T, Carter R, Moodie M, Markwick A, Magnus A et al. A new approach to assessing the health benefit from obesity interventions in children and adolescents: the assessing cost-effectiveness in obesity project. Int J Obes 2006; 30: 1463–1475.

21. 21

CFAC. Children's Health or Corporate Wealth? The Case for Banning Television Food Advertising to Children 2003. Coalition on Food Advertising to Children: Adelaide.

22. 22

Australian Broadcasting Authority. Children's Television Standards & Australian Content Standards, 1999.

23. 23

Gorn G, Goldberg M . Behavioural evidence of the effects of televised food messages to children. J Consumer Res 1982; 9: 200–205.

24. 24

Ludwig DS, Peterson KE, Gortmaker SL . Relation between consumption of sugar-sweetened drinks and childhood obesity: a prospective, observational analysis. Lancet 2001; 357: 505–508.

25. 25

Goldberg M . A quasi experiment assessing the effectiveness of TV advertising directed to children. J Market Res 1990; 27: 445–454.

26. 26

Robinson T, Nichols Saphir M, Kraemer H, Varady A, Farsih Haydel K . Effects of reducing television viewing on children's requests for toys: a randomized controlled trial. J Dev Behav Pediatr 2001; 22: 179–184.

27. 27

Saffer H, Chaloupka F . The effect of tobacco advertising bans on tobacco consumption. J Health Econ 2000; 19: 1117–1137.

28. 28

Saffer H . Alcohol advertising bans and alcohol abuse: an international perspective. J Health Econ 1991; 10: 65–79.

29. 29

Bell A, Kremer P, Magarey A, Swinburn B . Contribution of ‘noncore’ foods and beverages to the energy intake and weight status of Australian children. Eur J Clin Nutr 2005; 59: 639–645.

30. 30

Australian Government Department of Health and Ageing. The Australian Guide to Healthy Eating, 1998. [cited 26 May 2009]. Available from: http://www.health.gov.au/internet/main/publishing.nsf/Content/health-pubhlth-strateg-ffood-guide-index.htm.

31. 31

United States Department of Agriculture. Measure Equivalents. 15 May 2001 [cited 8 February 2005; Nutrient Data Laboratory]. Available from: http://www.nal.usda.gov/fnic/foodcomp/Bulletins/measurement_equivalents.html.

32. 32

Australian Bureau of Statistics. Retail Industry Commodity Sales. Australia 8624.0. 2000 [cited 10 November 2005]. Available from: http://www.abs.gov.au/AUSSTATS/abs@.nsf/ProductsbyReleaseDate/69AE6720398498C1CA257339900121F8A?OpenDocument.

33. 33

Australian Bureau of Statistics. Retail Trade Australia 8501.0. 2005 [cited 10 November 2005]. Available from: http://www.abs.gov.au/AUSSTATS/abs@.nsf/ProductsbyReleaseDate/06CC0034C5B3DC95CA2570C8800744BCA?OpenDocument.

34. 34

Haby MM, Carter R, Mihalopoulos C, Magnus A, Andrews G, Vos T . Assessing cost-effectiveness (ACE)—mental health: introduction to the study and methods. Aust NZ J Psychiatry 2004; 38: 569–578.

35. 35

Robinson TN . Reducing children's television viewing to prevent obesity. A randomised controlled trial. J Am Med Assoc 1999; 282: 1561–1567.

36. 36

Drewnowski A, Darmon M . The economics of obesity: dietary energy density and energy cost. Am J Clin Nutr 2005; 82 (Suppl): 265S–273S.

37. 37

Zuppa JA, Morton H, Metha KP . Television food advertising: counterproductive to children's health? A content analysis using the Australian Guide to Healthy Eating. Nutr Diet 2003; 60: 78–84.

38. 38

Bolton R . Modeling the impact of television food advertising on children's diets. In: Leigh J, Martin Jr C (eds) Current Issues and Research in Advertising. Division of Research, Graduate School of Business Administration, University of Michigan: Ann Arbor, MI, 1983.

39. 39

Caron A . Children, advertising and television. Choices in a new media environment. In: Frith S, Biggins B (eds) Children and Advertising: A Fair Game?. New College Institute for Values Research. University of New South Wales: Australia, 1994.

40. 40

Wang C, Gortmaker SL, Sobol AM, Kuntz KM . Estimating the energy gap among US children: a counterfactual approach. Pediatrics 2000; 118: e1721–e1733 doi: 10.1542/peds.2006-0682.

41. 41

Hill JO . Can a small-changes approach help address the obesity epidemic? A report of the Joint Task Force of the American Society for Nutrition, Institute of Food Technologists, and International Food Information Council. Am J Clin Nutr 2009; 89: 477–484.

42. 42

Hall K . Time between settling points, 2008. Melbourne, personal communication.

43. 43

Gold MR, Siegel JE, Russell LB, Weinstein MC . Cost-effectiveness in Health and Medicine. Oxford University Press: New York, 1996.

## Acknowledgements

The Assessing Cost-Effectiveness in Obesity (ACE-Obesity) project was funded by the Victorian Government Department of Human Services, Australia. We thank members of the ACE-Obesity working group for their input into the project: Michael Ackland (Deputy Chair), Bill Bellew, John Catford, Elizabeth Devlin, Helen Egan, Bonnie Field, Tim Gill, John Goss, Robert Hall (Chair), Brian Harrison, Kellie-Ann Jolly, Mark Lawrence, Amanda Lee, Tony Mc Bride, Karen McIntyre, Jan Norton, Anna Peeters, Theo Vos, Melissa Wake and Rowland Watson. We also thank Leah Galvin, who worked on this topic as part of her postgraduate studies, and Professor Tim Olds, who assisted with information requests from the latest national children's nutrition and physical activity survey.

## Author information

Authors

### Corresponding author

Correspondence to A Magnus.

Supplementary Information accompanies the paper on International Journal of Obesity website (http://www.nature.com/ijo)

## Rights and permissions

Reprints and Permissions

Magnus, A., Haby, M., Carter, R. et al. The cost-effectiveness of removing television advertising of high-fat and/or high-sugar food and beverages to Australian children. Int J Obes 33, 1094–1102 (2009). https://doi.org/10.1038/ijo.2009.156

• Revised:

• Accepted:

• Published:

• Issue Date:

### Keywords

• children
• modelling health benefit
• cost-effectiveness

• ### Impact of interventions to reduce overnutrition on healthcare costs related to obesity and type 2 diabetes: a systematic review

• Michaela Olm
• , Renée G Stark
• , Nathanael Beck
• , Christina Röger
•  & Reiner Leidl

Nutrition Reviews (2020)

• ### A macro-level assessment of introducing children food advertising restrictions on children’s unhealthy food cognitions and behaviors

• May O. Lwin
• , Andrew Z. H. Yee
• , Jerrald Lau
• , Janelle S. Ng
• , Jocelin Y. Lam
• , Ysa Marie Cayabyab
• , Shelly Malik
•  & K. Vijaya

• ### Nutritional quality, child-oriented marketing and health/nutrition claims on sweet biscuit, breakfast cereal and dairy-based dessert packs in Argentina

• Lorena Allemandi
• , Luciana Castronuovo
• , Maria Victoria Tiscornia
• , Patricia Gutkowski
• , Julieta Gijena
•  & Celeste Nessier

• ### Ranking Preventive Interventions from Different Policy Domains: What Are the Most Cost-Effective Ways to Improve Public Health?

• Nina van der Vliet
• , Anita W.M. Suijkerbuijk
• , Adriana T. de Blaeij
• , G. Ardine de Wit
• , Paul F. van Gils
• , Brigit A.M. Staatsen
• , Rob Maas
•  & Johan J. Polder

International Journal of Environmental Research and Public Health (2020)

• ### Space-time analysis of unhealthy food advertising: New Zealand children’s exposure and health policy options

• Wei Liu
• , Michelle Barr
• , Amber L Pearson
• , Tim Chambers
• , Karin A Pfeiffer
• , Moira Smith
•  & Louise Signal

Health Promotion International (2020)