Claims by Professor Boyd Swinburn in Crikey that the Productivity Commission got it wrong on childhood obesity should be taken with a pinch of (artery-hardening) salt.

The PC report Childhood Obesity: An Economic Perspective, argues that there is little evidence for the cost-effectiveness of banning junk food advertising to children. Professor Swinburn disagrees, citing research to support his view, including the ACE-Prevention Studies as well as research from the World Health Organisation (WHO) and the US Institute of Medicine.

It is correct that both the WHO and ACE reports support Professor Swinburn’s position on junk food advertising. This is not surprising given that he played a key role in their production, acting as an advisor to WHO, providing background papers for the report, and sitting on both the Working Group and Obesity Research Team for the ACE report.

The Report from the Institute of Medicine did not conclude that junk food ad bans were cost effective but simply that there was good evidence that food advertising affects the beliefs about food and food choices of young children.

The evidence cited in the ACE-Prevention and WHO reports is based on research conducted by Professor Swinburn (among others) into the cost-effectiveness of junk food advertising bans. This research cites as its main source of evidence, a single randomised controlled trial on the impact of junk food advertising on children’s food choices.

Anyone bothering to look more closely at this study (and few journalists have the time or inclination to do this) would find that it was undertaken over 30 years ago in Canada. The study involved 288 children aged 5-8 from low income backgrounds attending a Summer Camp for two weeks in Quebec in the late 1970s (the paper was published in 1982). During the camp, the children were divided into four groups, with one group exposed to junk food advertising. The food choices of all children were monitored during the two week period of the camp and it was found that the children who had not seen the food advertising made healthier food choices than those who had.

Almost all current research making claims about the cost-effectiveness of junk food advertising bans relies, at some point, on the findings of this research. For example, Professor Swinburn’s model bases assumptions about how an advertising ban in Australia today would affect the food choices of children, over the course of their lifetime, on this study. The model discounts the effects of this intervention in Australia by a seemingly random 50% to account for the impact of the Internet and associated marketing changes on children today.

It hardly needs to be pointed out that a single, short-term, study, undertaken a generation ago, on a small group of children on the other side of the world, from a different social and cultural background, in a pre-Internet media environment, is hardly a robust evidence base for major public policy changes in Australia today. Changing any of the assumptions about the relevance of this study to contemporary Australian society (for example, altering projections of how changes in the food choices of children over a two week period would be sustained over their lifetime) would deliver significantly different results.

Despite the quasi-scientific gloss of academic studies in this area and their misleadingly precise findings (a junk food ad ban has a “gross incremental cost-effectiveness ratio of AUD$ 3.70 per Disability Adjusted Life Year”), their conclusions are only as robust as their assumptions. When put under the microscope, it is clear that these assumptions are based as much on conjecture, option and ideology as they are on science.

The PC was right to conclude that the cost-effectiveness of junk food ad bans has not been established. If the Preventive Health Agency is genuinely committed to evidence-based policy it will take advice from the PC and allocate resources into interventions which have a better chance of success in reducing childhood obesity.