Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

An Australian childhood obesity summit: the role of data and evidence in 'public' policy making

An Australian childhood obesity summit: the role of data and evidence in 'public' policy making Background: Overweight and obesity in Australia has risen at an alarming rate over the last 20 years as in other industrialised countries around the world, yet the policy response, locally and globally, has been limited. Using a childhood obesity summit held in Australia in 2002 as a case study, this paper examines how evidence was used in setting the agenda, influencing the Summit debate and shaping the policy responses which emerged. The study used multiple methods of data collection including documentary analysis, key informant interviews, a focus group discussion and media analysis. The resulting data were content analysed to examine the types of evidence used in the Summit and how the state of the evidence base contributed to policy-making. Results: Empirical research evidence concerning the magnitude of the problem was widely reported and largely uncontested in the media and in the Summit debates. In contrast, the evidence base for action was mostly opinion and ideas as empirical data was lacking. Opinions and ideas were generally found to be an acceptable basis for agreeing policy action coupled with thorough evaluation. However, the analysis revealed that the evidence was fiercely contested around food advertising to children and action agreed was therefore limited. Conclusion: The Summit demonstrated that policy action will move forward in the absence of strong research evidence. Where powerful and competing groups contest possible policy options, however, the evidence base required for action needs to be substantial. As with tobacco control, obesity control efforts are likely to face ongoing challenges around the nature of the evidence and interventions proposed to tackle the problem. Overcoming the challenges in controlling obesity will be more likely if researchers and public health advocates enhance their understanding of the policy process, including the role different types of evidence can play in influencing public debate and policy decisions, the interests and tactics of the different stakeholders involved and the part that can be played by time-limited yet high profile events such as Summits. Page 1 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Within this environment characterised by public policy Background Any policy-making process is complex – it deals with inertia, the issue of childhood obesity, and the need for human and political dynamics, the use of resources, and effective interventions, was brought to the forefront of power [1]. The development and implementation of pol- one Australian state government's agenda through the icy in a democracy seeks to meet multiple objectives [2]: NSW Childhood Obesity Summit (hereafter referred to as addressing major health and social policy problems, using 'the Summit') in 2002. While obesity had already been public resources wisely, satisfying a range of stakeholders, identified as a problem, how to respond was unclear. avoiding conflict, and ensuring that political and eco- With little evidence available to guide Government nomic objectives are met. Research is only one influence responses to the issue, the state health department's (here- in the ongoing process of policy-making [3]. In setting the after referred to as NSW Health) articulated purpose of the agenda, formulating policy, and implementing and evalu- Summit was to i) create better understanding in the com- ating it, various forms of evidence are sought and utilised. munity; ii) inform Members of Parliament; iii) hear and While conventionally such evidence is conceived as being consider the views of families, parents and young people; derived from "scientific and objective" research, it is iv) examine existing approaches and consider new ideas increasingly clear that a much wider range of sources and in a bipartisan forum; v) consider evidence; vi) identify forms of evidence are influential [4]. There has been sig- ways to improve existing strategies and services; vii) build nificant debate in Australia about the interface between community consensus about future directions, and viii) evidence and policy-making [5], but little detailed analy- recommend a future course of action so that the best avail- sis of the way evidence shapes the process of policy-mak- able strategies, both long and short term, would be imple- ing. This paper examines the role of data and evidence in mented to overcome the childhood overweight and public policy-making in response to childhood obesity in obesity problem [13]. This paper examines the role of evi- an Australian state, New South Wales (NSW). dence and data in entrenching childhood obesity on the policy agenda, in shaping the Summit debate and inform- Overweight and obesity (O&O) in Australia, as in many ing the outcomes and the policies that were subsequently countries, has risen at an alarming rate over the last 20 adopted. years. Overweight is classified as a body mass index (BMI) of 25 and above, and obesity as 30 and above [6]. Obesity Methods in men in Australia rose from 9.3% in 1980 to 17.1% in Data collection 2000 and for women from 8.0% to 18.9% [7,8]. O&O in Data were collected from the transcripts of the Summit children and young people has also increased markedly. proceedings [14-16], media articles, the Summit Commu- From 1985 to 1995 the level of combined O&O in chil- niqué [17] which outlined the agreed resolutions, the Gov- dren more than doubled in all but the youngest age group ernment Action Plan [18] published after the Summit and of boys whilst the level of obesity tripled in all age groups the announcement by the NSW Health Minister in and for both sexes [9]. December 2002 [19]. 'Factiva', a searchable archive of print media, was used to identify articles that referred to Despite rising obesity, the policy response has been lim- childhood obesity in the four main NSW statewide news- ited and hampered by a lack of evidence concerning effec- papers and one national newspaper (the Daily Telegraph, tive interventions. The World Health Organisation Sunday Telegraph, Sydney Morning Herald, The Sun Her- (WHO) has highlighted "globesity" and released the Glo- ald and the Australian) in the three months prior to the bal Strategy on Diet, Physical Activity and Health [10]. Ear- announcement of the Summit in July 2002 until the first lier, the United States (US) Surgeon General's Call to public response from government in December 2002. Action emphasised the need to create supportive environ- There were 127 articles retrieved from this search. ments which provide accessible and affordable healthy food choices and convenient opportunities for regular Seven semi-structured key informant interviews [20] and physical activity [11]. one focus group discussion (FGD)[20] with three health staff involved in the Summit's organization were also con- Australia was one of the first countries to produce an inte- ducted. The key informants included NSW health staff grated national strategy for the prevention of O&O. The and experts in human nutrition, physical activity, and National Health and Medical Research Council population health. The interviews and focus group discus- (NHMRC) report 'Acting on Australia's weight: a strategic sion used a guide to elicit opinions on the stimulus for, plan for the prevention of overweight and obesity' [12], was and organization of, the Summit and its outcomes. The released in 1997, but its recommendations, which focus group discussion was transcribed for analysis and included strategies such as promoting physical activity, the interviews were used as background material. dietary monitoring, and encouraging the development of school canteen policies remained largely unaddressed. Page 2 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Data analysis Results The transcripts of the Summit proceedings [14-16], media Three phases were discernible in the process of policy articles and other key documents were reviewed and con- making that occurred as part of the NSW Childhood tent analysed [20] to examine what type of evidence was Obesity Summit: 1) building and maintaining the used, by whom (eg. experts, industry, advocates) and for momentum 2) summit debate and 3) outcomes and pol- what purpose. Evidence that was valued or contested in icy formulation. the Summit debates and the media coverage received par- ticular attention. The type of evidence used was catego- 1) Building and maintaining the momentum rised into three types based on a model adapted from Obesity had been recognised as a longstanding and Bowen & Zwi [4] who outlined five types of evidence. The increasingly important public health problem. Ebbeling categorisation used in the current study were empirical et al (2002) pointed to publications decades earlier high- research (Type 1), such as randomised controlled trials, lighting the issue and the need for a policy response[24]. case control and cohort studies, time series analyses, Media interest in the issue of obesity in Australia was stim- observational studies, case reports and qualitative studies; ulated by available data highlighting "the doubling and tri- ideas and opinions (Type 2) which incorporated the two pling" of rates of obesity and concerns around the "second categories of 'knowledge and information' and 'ideas and fattest kids in the world" (FGD). Obesity was seen as "the interests' outlined by Bowen & Zwi, and included evi- new tobacco" – the public health issue which was being rec- dence such as the results of consultation processes, opin- ognised as demanding attention. Articles published in the ions and views of "experts", interest groups and peer review literature around this time [24,25] were trig- community members; and economic data (Type 3) which gers for media coverage and interviews with key inform- focused on economic evaluation, finance and resource ants and the focus group with NSW Health staff all implications. emphasised the importance of media coverage in bringing the issue to public and policy attention: Rigour Rigour was addressed through triangulation, clear exposi- "It [media coverage] was partly driven by data...the MJA tion of methods and reflexivity [21]. [Medical Journal of Australia] also carried some data on child- hood obesity and ... reinterpreting existing data sets. ...so that Triangulation is the use of different approaches, such as put it on the radar, that doesn't mean you've got [a] Summit interviews and document analysis, to answer the same happening yet... the data is essential – it is necessary, but not question which strengthens the rigour of a study and the sufficient." (FGD) interpretations made [22]. In the current study, inter- views, analysis of transcripts from the summit debate and "The doubling and tripling was the most used [news] grab eve- related documents and media coverage were used to rywhere, in every article, and it is still used." (FGD) answer questions posed in relation to the role of evidence in the NSW Childhood Obesity Summit. Why was NSW Health interested? The issue was shown to be important to the public. It provided the opportunity to It is important to consider the ways in which researchers divert attention away from other health issues which are and authors' past and present experiences may have considered solely the responsibility of government, for shaped the way data was collected and interpreted – often example, health care service provision. NSW Health also referred to as reflexivity [22]. All the authors of this paper wanted to show leadership in an area where there was are involved, at some level, in public health advocacy and arguably Federal Government inaction. In New South support a range of initiatives to address public health Wales there was a clear perception that "prior to the Summit problems, including childhood obesity. The paper arose there was a national leadership vacuum" around childhood from a desire by the authors to better understand and obesity (FGD). An earlier government summit on illicit reflect upon the role of evidence and its use by the differ- drugs [26], had mobilised massive public attention and ent stakeholders in the Summit debate and how the resources and it was hoped by NSW Health that a child- debates around evidence were seen to influence the reso- hood obesity summit would draw in funds and resources lutions agreed. The involvement of all authors in the anal- to address this public health problem. A summit was seen ysis and interpretation of the data presented in this paper, as providing scope to debate interventions in an area data triangulation, clear exposition of methods, conduct where there was no scientific or political clarity at the of a focus group with some of the key actors involved, and time: reflection on alternate ways of viewing the data were all important in enhancing the rigour of the study and the "there was interest, we were asked to do things, write things, credibility of the interpretations made [22,23]. pull things together... there were lots of false starts...we had things in train that were going to take another 5 or 10 years Page 3 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Table 1: Number of media articles and evidence type used Number of articles Month Total Articles Type 1 Type 2 Type 3 April 5550 May 7630 June 9952 July 15 740 August 10 6 3 0 September 40 15 16 2 October 11 450 November 11 8 8 2 19 12 12 3 December Announcement of Summit Month in which Summit was held Month in which government initial policy response was announced and they said they wanted a solution today... a summit was sug- tions [16]. Once the Summit was underway, Type 2 data gested as a way forward."(FGD) were more widely reported and ideas from experts, com- munity members and key stakeholders concerning the Table 1 shows the number of media articles by month way forward, were presented in the media. In putting for- between April 2002 and December 2002. Within each ward their views, these stakeholders called on common month the percentage of articles that used Type 1, Type 2 sense understandings, research studies or pointed to a lack or Type 3 evidence are identified. All the articles drew on of conclusive evidence to support inaction. more than one type of evidence. Peak months of coverage were July when the Summit was announced (n = 15), Sep- Food advertising to children was a case in point. Prior to tember when the Summit was held (n = 40), and Decem- the Summit, debates about evidence in the media focused ber when the Health Minister announced the preliminary on taxing 'high fat foods' and banning food advertising to government response (n = 19). In the months prior to the children. The soft drink industry spoke about the lack of announcement of the Summit, childhood obesity was good evidence for the effectiveness of such initiatives and covered 1–2 times per week in the newspapers studied. the negative economic impact of a "fat tax". Physical activ- ity and the role of parents as an influence on obesity were In the lead up to the Summit, most of the articles cited evi- highlighted by the advertising and food industries as dence of at least one type concentrating on Type 1 evi- being the major influences on childhood obesity. Results dence focussed on the magnitude of the problem, backed from Sweden which were stated by the food and advertis- up by expert opinion (Type 2). In the month before the ing industry as showing obesity rising despite an advertis- first announcement by government in December eco- ing ban were used to demonstrate that "there is no evidence nomic data (Type 3), always referring to the cost of obesity that advertising makes children eat more fatty foods" (The Aus- to the health care system, were also reported. tralian Newspaper, 1 July 2002)[28]. It became clear from the media coverage during the Summit that a ban on food Prior to the Summit and throughout the study period, advertising was the critical concern for industry who were Type 1 evidence was widely reported and largely uncon- calling a 'clear link' between harmful childhood behav- tested, quoting authoritative sources such as the Lancet iour and commercials, with editorials suggesting that [24] and the Medical Journal of Australia [25] concerning instead "parents are the dominant influence on food choices" the magnitude of the problem. Media representations (Daily Telegraph, 12 September 2002) [29]. drew on such data to present 'sound bites' to stimulate debate. The most commonly reported statistics were that 2) Summit debate either one in four, or one in 5 children in Australia was The Summit was held in September 2002 at Parliament overweight or obese and that overweight and obesity had House in NSW. An across-government organising com- doubled between 1985 and 1995. These data from mittee oversaw delegate selection and sought to ensure Magarey et al (2001) [9] were also contained in the back- balanced representation including: i) children and young ground document prepared for the Summit [27] and people; ii) families, parents and community perspectives; included in the factual preamble to the Summit resolu- iii) experts; iv) relevant peak bodies; v) special population Page 4 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 perspectives, such as the socially disadvantaged, people data were provided. Data from the US concerning changes from culturally and linguistically diverse communities, in levels of physical activity were presented: "most people in Aboriginal and Torres Strait Islanders, rural and remote my generation walked to school, today less than a third of chil- communities, and people with disabilities. dren in the United States walk to school" (US expert, Day 1, pg 12) and Australian data on sedentary activity: "97% of The Summit provided an opportunity for delegates to our adolescents watch television ... between 60 and 80% play present their case for action during plenary sessions. Dur- computer or video games." (Australian expert, Day 1, pg 16). ing the Summit, nine working groups (WGs) were con- Anecdotal observations about changing societal behav- vened: i) Early Childhood, ii) Family and Community, iii) iours and environments were widely cited and seen as School Education, iv) Health, v) Sport, Recreation & Fit- important factors to address despite the lack of reliable ness, vi) Local Government, vii) Commercial Food Indus- trend data and research evidence: try, viii) Media, and ix) Transport and Planning. "we do not yet have evidence that any single one of these factors The WGs were requested to put forward 10–15 resolutions is driving the epidemic" (US expert, Day 1, pg 13) for the Communiqué to be presented to government. The importance of evidence for the resolutions was made clear "we know very little in any, firm solid way about the factors that by the NSW Premier on the first day of the summit when influence young people to be active or sedentary – all we have he referred to the NSW Drug Summit [26], which had to work with over the next three days are some recently been held in May 1999. informed guesses and some far less well-informed speculations" (Australian expert, Day 1, pg 12) "The Drug Summit emphasised looking at evidence, basing pol- icies on evidence ... I would like that to be your guide too." The views of young people, were an integral part of the (NSW Premier, Day 1 pg 33) Summit process and provided an emotive appeal to take action. Young people's stories were shown on video and The case for action to tackle childhood obesity was uncon- they addressed the Summit. However, there appeared to tested from the outset. In opening the Summit the NSW be little attempt to draw these views together, articulate Minister for Health referred to strong empirical evidence common threads or examine whether and how such views of the magnitude of childhood obesity: related to other empirical data and expert opinion. A young person opening the Summit stated that "It is genu- "In Australia the level of combined overweight and obesity inely important that our voice be heard"(young person, Day 1, between 1985 and 1995 has more than doubled... Today in pg 2). The FGD participants saw young people's stories as New South Wales one in five children aged between seven and being powerful in stimulating action: 15 are classified as being either overweight or obese." (NSW Minister for Health, Day 1, pg 5) " we had to have lots of consultation processes that included the voice of the child... engaging the children... it was the most Experts from government departments, academic institu- powerful thing." (FGD) tions and the health service put forward similar statistics that highlighted the magnitude of obesity. Many outlined A young person at the Summit, however, expressed frus- the consequences of overweight and obesity for type 2 dia- tration about the focus on evidence: betes in particular. The use of simple statistical concepts such as "doubling in rates" and "one in five of our children" "Unfortunately we have been bombarded with statistics. They were commonplace. Economic evidence highlighted the have been repeated over and over again ... we are almost scared cost of the "obesity epidemic" to society – "it costs us a com- to put up a decent suggestion." (young person, Day 2, pg 30) munity $830 million a year" (Minister for Health, Day 1, pg 7) and individuals – "in one year the personal cost to individ- In contrast to the research evidence supporting the magni- uals who are obese is $19 billion" (expert, Day 1 pg 16). Such tude of the problem and the influencing factors, evidence data were uncritically and widely accepted during the supporting calls for action were mostly opinion and ideas Summit. with some reference to overseas efforts. Nonetheless, much was made of the need for evidence-based strategies, On the opening day, experts, parents, community groups with a US expert claiming three strategies that were "defen- and industry talked anecdotally about societal changes sible, but not conclusive" (US expert, Day 1, pg 13): breast- over decades and their impact on physical activity and feeding, limiting television viewing and the promotion of food consumption. Statistics and studies were referred to physical activity. A Cochrane systematic review [30] cov- in support of these observations, such as an increased reli- ering 1985–2001 and encompassing 14,000 studies was ance on carbonated sugared drinks, although no actual reported (researcher, Day 1, pg 37). It found 11 studies of Page 5 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 a high enough quality to examine the effectiveness of the the Summit was to: "Frame evidence-based solutions within a intervention and it found only small or no effects with community-based 'reality check' perspective."(Day 3, pg1) those interventions that were most effective focussed on reducing sedentary behaviour. A few delegates questioned Evidence of the magnitude of overweight and obesity was the need for evidence from primary prevention trials, included with little debate. Statements about the influenc- pointing to broader experiences that tell us "what works". ing factors were carefully worded to reflect agreement on They highlighted other successful public health cam- importance and available evidence: paigns such as in tobacco control as evidence for the suc- cess of a range of strategies, including advertising controls "Although physical activity trend data is lacking, it is apparent and taxes: that children and adolescents are less physically active" (Day 2, pg 72) "we do need evidence, we do need to work at what has been shown to be the most effective, but that should not inhibit us "An increase in television viewing is associated with an increase from acting now. There have been a number of successful public in obesity in children. An increase in sedentary behaviour is health programs that have been introduced without definitive associated with an increase in obesity in children. Experts have evidence." (expert, Day 1, pg 39) advised that television viewing needs to be one of the targets for obesity control efforts" (Day 2, pg 78) The FGD and comments by Summit delegates highlighted the need for action coupled with thorough evaluation: Exposure to advertising messages was included in the fac- tual preamble referring to the range of potential influ- "There needs to be a recognition of the sense of urgency...that ences on food selection behaviours. The resolutions about policy won't wait for the data." (FGD) food advertising to children generated the most debate concerning the evidence-base for such interventions (see "This is about promising interventions, we have to just go with Figure 1) and the relationship between food choice and promising interventions, make sure they do no harm and just television viewing. This debate illustrates the use of differ- evaluate the heck out of them, and then maybe in ten years ent types of evidence by industry representatives as one time, if they weren't the best things to do, well at least we did means of opposing calls for a ban on food advertising to something" (FGD) children. "we need periodic surveys to tell us how we are doing with A resolution to ban food advertising to children was not respect to implementation of strategies ... we need causal mod- agreed. In its place agreement was reached to have an els, that is, longitudinal studies which allow us to link risk fac- independent review by the Federal government of the reg- tors like change in the food supply with changes in the ulatory arrangements for food advertising "in recognition prevalence of obesity." (US expert, Day 1, pg 13) that food advertising is one of the contributing factors to the prevalence of eating habits that may promote obesity"(Day 3, "we would like to see a regular – maybe five yearly – national pg 9) in addition to a review of a voluntary code to be nutrition, physical activity and health survey." (industry, Day undertaken by industry. Attempts by the Food Industry to 2, pg 2) have this statement deleted from the Communiqué were not successful. A systematic review of the impact of food "Certainly, we need to take action, but at the same time we advertising on diet, physical activity and childhood obes- need to be doing research. We cannot continue to act in an evi- ity was also recommended. dence vacuum." (expert, Day 2, pg 20) All other resolutions passed with minimal debate, includ- The most contentious issue centred on the role of food ing those addressing physical activity, school education, advertising to children (see Figure 1). The intensity of the transport and planning. Most of the resolutions agreed at debate between food industry representatives and the the Summit and taken up in the subsequent Government advocates of a ban on food advertising to children clearly Action Plan [18] were focused on physical activity and illustrates the way different types of evidence are drawn nutrition education. Mandatory guidelines for school upon to articulate a particular position or undermine that canteens also passed as a resolution despite some opposi- of opposing perspectives. tion from industry. Numerous resolutions in the Commu- niqué [17] referred to research and a detailed section on 3) Outcomes and policy formulation surveillance and monitoring proposed a funded collabo- The final Communiqué to government was to include a rative centre of excellence in research, prevention and "factual foundation" and recommendations and resolu- management. tions for future action. The purpose of this component of Page 6 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Much of the evidence put forward by advocates for a ban on food advertising focussed on the amount of food advertising in children’s programming and the relationship between food consumption and television viewing: “There are studies showing a clear correlation between the consumption of foods advertised on television and the amount of time spent watching television.” (expert/advocate, Day 1, pg 27); “More than 80% of ads are for foods that are outside what we call the core food groups.” (expert/advocate, Day 1, pg 28). The Lancet article by Ebbeling et al (2002) was also cited as evidence for a ban: “Lancet writers considered … that we should prohibit food advertisements and marketing that is directed at children.” (expert/advocate, Day 1, pg 24) . Advocates documented the number of food advertisements in children’s programming: “Australian children 5-12 years watch on average two and a half hours of television daily and this includes up to 52 minutes of advertising” (advocate, Day 2, pg 21) . Industry contested these data: “there is no legal way they could be exposed to 52 minutes of advertising as the limits per hour are 13 minutes per hour in general programming and 10 minutes in C time.” (industry, Day 2, pg 43). They also responded by citing a report by the Ministry for Agriculture, Fisheries and Food in the UK (1996) as support for the notion that parents are more influential in food choices than advertising. The advertising industry also reported that “media research shows advertising levels transmitted to children aged 5-12 of confectionary, snack foods, soft drinks and fast food in fact decreased by 19 per cent from 1996-2001” (industry, Day 2, pg 44). The example of Sweden, where food advertising to children has been banned for 10 years, was also given as evidence by industry of the failure of this approach to reducing obesity levels (30-32). They quoted a Swedish government public health report in 2001 that stated: “The proportion of overweight people has increased in all socio-economic groups since the beginning of the 1980s” (industry, Day 2, pg 58). Industry also cited the Premier’s opening remarks focusing on evidence based policy as support for recommendations being evidence-based. They did not accept that a ban on food advertising had been shown to promote health gain: “after rigorous examination of the research literature, we concluded that there is no serious and sound evidence which shows that food advertising leads to an increase in the consumption of ‘whole categories of foods’”. (industry, Day 2, pg 57) Advocates of a ban highlighted the cost of advertising: “A single 30 second ad can cost a million dollars” (researcher and advocate, Day 2, pg 54) and “the industry spends billions of dollars to make and broadcast ads for less healthy foods” (researcher and advocate, Day 2, pg 54) as evidence that it works to influence food choices. They also refuted the relevance of the Swedish case “the rate of obesity in Sweden is extremely low” (advocate and expert, Day 3, pg 3). One advocate also stated “I really do not want to see us deflected into more evidence being required before we ban food advertising .. it is up to the food industry to prove that food advertising does not have an effect.” (researcher and advocate, Day 2, pg 54). This debate about food advertising to children demonstrated how the concept of ‘evidence’ was embraced by the food industry and used expertly to undermine the evidence presented by the advocates of an advertising ban. In particular, the industry used the Premier’s comments about evidence-based policy (type 2 evidence) together with data from Sweden (type 1 evidence) to support their position that a food advertising ban was not ‘evidence-based’ and would fail to deliver the intended outcomes. While the Swedish data was disputed by the advocates of a ban, the industry arguments were sufficient to water down the resolutions agreed at the Summit – a review rather than a ban on food advertising to children. Contesting the evid Figure 1 ence: food advertising and obesity Contesting the evidence: food advertising and obesity. Page 7 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Limited attention was devoted to the financial and logis- However, lack of compelling evidence for interventions is tical feasibility of the resolutions – this was apparent by likely to have been a factor in the failure of government to the number of resolutions that required intervention at a commit significant new funds and to agree to controver- federal rather than state level. However, the preliminary sial recommendations around food advertising given response from the government in December considered strong industry opposition. The only contentious resolu- what was feasible in the current financial and political tion taken up by government was for mandatory guide- context: lines for school canteens: this appealed to many community groups and parents who attended the Summit "it wasn't really evidence-based, it was the feasibility of what- and government is likely to have perceived strong public ever strategy they had suggested..." (FGD) support for this intervention. Other commentators have questioned the soft policy options adopted in response to In the final Summit address by the NSW Health Minister the obesity epidemic in Australia following the NSW Sum- [16] the two resolutions specifically mentioned and mit and raised questions about the way public health strongly supported were the recommendations on school issues, such as obesity, are framed in public discourse canteens and a collaborative centre for excellence for over- [31]. weight and obesity research. These two initiatives were subsequently publicly announced in December as the key The food and advertising industries who were represented response to the Summit by the government [19]. The at the Summit used the lack of well supported 'scientific' advisers to the Minister and NSW Health were concerned evidence to oppose controls on advertising. In contrast, to ensure that the Summit resulted in some "announcea- the debates and resolutions around physical activity using ble" interventions – and the two chosen seemed "doable", anecdotal evidence, expert opinion and common-sense of value, and in some respects least contentious (FGD). solutions garnered widespread support as there was no industry that stood to suffer financially from the action proposed. Where strong interests and powerful groups Discussion This paper has sought to present key elements of the use oppose policy direction, the evidence base required for of data and evidence in the NSW Childhood Obesity Sum- government action, if it is to proceed, needs to be substan- mit. There are other dimensions of policy-making which tial. It is also possible that the more prominent role of the deserve attention and other interpretations of the process federal government in food advertising regulation and possible. As indicated by Ham and Hill "It is rarely possi- control worked against the agreement of concrete resolu- ble to agree on one version of events: the most that can be tions around food advertising to children. achieved is a plausible interpretation" [2] (p xi). Economically important industries have been seen by oth- Empirical evidence of the magnitude of the obesity prob- ers as critical in the preparedness of governments to sup- lem and the economic cost to the health system were crit- port controversial public health initiatives [32] and calls ical to generating publicity and framing the case for action for more research have been presented as tactics to delay on childhood obesity in the lead up and during the Sum- policy change [33]. However, creative and clear commu- mit. This evidence was never contested and became a part nication of the evidence has been instrumental in other of the factual foundation of the Summit Communiqué areas, notably the successful efforts to ban tobacco adver- [17]. It is clear that the combination of Type I data, which tising in Australia and in many other countries around the was largely epidemiological in origin, and Type 3 data world despite powerful industry opposition [32, 33]. about the economic costs of the problem was persuasive. There is also more scope for interaction and collaboration with the food industry than with the tobacco industry as The lack of empirical evidence for many of the influencing food as a product is not inherently harmful [33]. The food factors and related interventions, for example in the area industry can have an important role in supporting a range of physical activity, did not hamper agreement of resolu- of policy initiatives that promote healthy eating as was tions at the Summit and was instrumental in funding a evident in the NSW Childhood Obesity Summit, but are research centre to collect better data and evidence for what likely to remain adversarial where industry profits are, or works. Health officials who recognised the lack of an evi- appear to be, at stake. dence base for interventions sought to promote those that seemed most logical and appropriate, along with a con- Conclusion cern to ensure subsequent careful evaluation. The Summit The NSW Childhood Obesity Summit played a role in demonstrated that policy action will move forward in the promoting an agenda for action to address childhood absence of strong research evidence if government sees the obesity. It raised awareness in the public and political need to respond to public concerns. arena and provided a public forum for debating research evidence. The Summit demonstrated that while it is not Page 8 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 authors would also like to thank the reviewers of the original submitted necessary to have all the evidence in place to agree actions, manuscript for their constructive and considered comments. The School of that more radical policy change is much more difficult to Public Health and Community Medicine, The University of New South achieve in the absence of established and detailed evi- Wales, and The Centre for Chronic Disease Prevention and Health dence, given the interests of important stakeholders, nota- Advancement, NSW Health, are collaborating on a project, funded by the bly the private sector. The process and the outcomes of the Centre, to foster reflection, analysis and writing between academics, prac- Summit suggest that in the absence of strong Type 1 data, titioners and policy-makers, and aimed at contributing to bridging the aca- and where Type 2 evidence is contested, that policy-mak- demic-policy divide. ers may opt for the path of least resistance: a call for more and better research and support for the systematic evalua- References 1. Walt G: Introduction to health policy. Process and power. tion of interventions. While beneficial to researchers, London, Zed Books; 1994. direct and short term health gain may be limited. 2. Ham C, Hill M: The policy process in the modern capitalist state. (Second edition). Hertfordshire, Harvester Wheatsheaf; The lack of an agreed evidence-base provides politicians 3. Rist RC: Influencing the policy process with qualitative with a freer hand in choosing actions which have wide research. In (Editors). In Handbook of Qualitative Research Edited by: Denzin NK and Lincoln YS. London, Sage Publications; appeal and are less controversial, rather than those which 1994:545-557. may produce greatest health benefit. The Summit's success 4. Bowen S, Zwi AB: Pathways to "evidence informed" policy and in generating a set of resolutions should not be dis- practice: a framework for action. Public Library of Science Medicine 2005, 2:e166. counted even if large resource allocations were not forth- 5. Lin V, Gibson B: Evidence-based Health Policy: Problems and coming. Tobacco control initiatives have taken decades of Possibilities. South Melbourne, Oxford University Press.; 2003. concerted effort to realise [33] and obesity control efforts 6. World Health Organisation: Obesity: preventing and managing the global epidemic. Report of a WHO consultation. WHO are likely to face the same challenges around evidence and Technical Report Series 2000. action. The prospects of controlling obesity in the future 7. Bennett SA, Magnus P: Trends in cardiovascular risk factors in Australia: results from the National Heart Foundation's Risk will be amplified if researchers and public health advo- Factor Prevalence Study, 1980-1989. Medical Journal of Australia cates enhance their understanding of the policy process, 1994, 161:519-527. the interests and tactics of the different stakeholders 8. Dunstan D, Zimmet P, Welborn T: Diabetes & Associated Disor- ders in Australia: The final Report of the Australian Diabe- involved, and the role different types of evidence can play tes, Obesity and Lifestyle Study (AusDiab). Melbourne, in influencing public debate and the decisions of policy- International Diabetes Institute; 2000. makers in time-limited yet high profile events such as 9. Magarey A, Daniels L, Boulton TJC: Prevalence of overweight and obesity in Australian Children and adolescents: reassess- Summits. Further research is needed to increase our ment of 1985 and 1995 data against new standard interna- understanding of the role of Summits in the broader pol- tional definitions. The Medical Journal of Australia 2001, 174:561-564. itics and processes of policy-making. 10. World Health Organisation: Global strategy on diet, physical activity and health. World Health Organisation; 2004. Competing interests 11. US Department of Health and Human Services: The Surgeon Gen- eral's call to action to prevent and decrease overweight and Elizabeth Develin was involved in the original work obesity. Washington, US Department of Health and Human described leading up to and including the NSW Child Services.; 2001. Obesity Summit and is employed by NSW Health. All the 12. National Health & Medical Research Council: Acting on Australia's weight: a strategic plan for the prevention overweight and other authors are part of the School of Public health and obesity. Canberra, Commonwealth of Australia; 1997. Community Medicine, UNSW, which is in receipt of mod- 13. NSW Health: NSW Childhood Obesity Summit Draft Pro- gram. Edited by: Committee NSWCOSO. Sydney, NSW Depart- est funds to collaboratively develop this reflection and ment of Health; 2002. paper by NSW Health, the body which organised the Sum- 14. NSW Health: NSW Childhood Obesity Summit: Report of mit described. proceedings of the first day. [http://www.health.nsw.gov.au/ obesity/adult/summit/hansard/hans.html]. 15. NSW Health: NSW Childhood Obesity Summit: Report of Authors' contributions proceedings of the second day. [http://www.health.nsw.gov.au/ obesity/adult/summit/hansard/hans.html]. This paper was conceived jointly by all authors; all con- 16. NSW Health: NSW Childhood Obesity Summit: Report of tributed to analysis and writing. SN prepared the first draft proceedings of the third day. [http://www.health.nsw.gov.au/ with significant contributions by ED and comments from obesity/adult/summit/hansard/hans.html]. 17. NSW Health: NSW Childhood Obesity Summit Communi- AZ and NG. All authors contributed to writing successive que. [http://www.health.nsw.gov.au/obesity/adult/summit/commu drafts, feeding in literature, analysis and insights. nique.pdf]. 18. NSW Health: Prevention of Obesity in Children and Young People: NSW Government Action Plan 2003-2007. Sydney, Acknowledgements NSW Department of Health; 2003. The authors wish to acknowledge insights and feedback from Bill Bellew 19. NSW Minister for Health: Youth Obesity - Five Year Plan. In (Director, Centre for Chronic Disease Prevention and Health Advance- Media Release , NSW Health; 2002. 20. Patton MQ: Qualitative Research and Evaluation Methods. ment, NSW Health) and Phillip Vita (former Manager, Nutrition & Physical 3rd edition edition. California, Sage Publications Inc.; 2002. Activity Branch, NSW Health, now Executive Officer, NSW Centre for 21. Mays N, Pope C: Assessing quality in qualitative research. BMJ Physical Activity & Health) and the work of Kate Hawkins and Sarah Yallop 2000, 320:50-52. from NSW Health who conducted the key informant interviews. The Page 9 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 22. Denzin NK, Lincoln Y: Handbook of Qualitative Research: 2nd Edition. California, Sage; 2000. 23. Ebbeling CB, Pawlak DB, Ludwig DS: Childhood obesity: public health crisis, common sense cure. Lancet 2002, 360:473-482. 24. Goodman S, Lewis PR, Dixon AJ, Travers CA: Childhood obesity: of growing urgency. The Medical Journal of Australia 2002, 176:400-401. 25. NSW Government: NSW Drug Summit. [http://www.dru ginfo.nsw.gov.au/drug_summit/ government_response_to_the_drug_summit]. 26. NSW Health: NSW Childhood Obesity Summit Background Paper. [http://www.health.nsw.gov.au/obesity/adult/summit/ bground_paper.html]. 27. The Australian: Going off the scales. In The Australian 1 July 2002 edition. Sydney, ; 2002. 28. Daily Telegraph: Dishing out blame. In Daily Telegraph Sydney, ; 29. Campbell K, Waters E, O'Meara S, Kelly S, Summerbell C: Interven- tions for preventing obesity in children. The Cochrane Database of Systematic Reviews 2002. 30. Lin V, Robinson P: Australian public health policy in 2003-2004. Australian and New Zealand Health Policy 2005, 2:. 31. Chapman S, Lupton D: The fight for public health: principles and practice of media advocacy. London, BMJ Publishing Group; 32. Yach D, McKee M, Lopez AD, Novotny T: Improving diet and physical activity: 12 lessons from controlling tobacco smoking. BMJ 2005, 330:. Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 10 of 10 (page number not for citation purposes) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Australia and New Zealand Health Policy Springer Journals

An Australian childhood obesity summit: the role of data and evidence in 'public' policy making

Loading next page...
 
/lp/springer-journals/an-australian-childhood-obesity-summit-the-role-of-data-and-evidence-3aMP0Ulht1
Publisher
Springer Journals
Copyright
Copyright © 2005 by SA et al; licensee BioMed Central Ltd.
Subject
Medicine & Public Health; Public Health; Social Policy
ISSN
1743-8462
eISSN
1743-8462
DOI
10.1186/1743-8462-2-17
pmid
16029512
Publisher site
See Article on Publisher Site

Abstract

Background: Overweight and obesity in Australia has risen at an alarming rate over the last 20 years as in other industrialised countries around the world, yet the policy response, locally and globally, has been limited. Using a childhood obesity summit held in Australia in 2002 as a case study, this paper examines how evidence was used in setting the agenda, influencing the Summit debate and shaping the policy responses which emerged. The study used multiple methods of data collection including documentary analysis, key informant interviews, a focus group discussion and media analysis. The resulting data were content analysed to examine the types of evidence used in the Summit and how the state of the evidence base contributed to policy-making. Results: Empirical research evidence concerning the magnitude of the problem was widely reported and largely uncontested in the media and in the Summit debates. In contrast, the evidence base for action was mostly opinion and ideas as empirical data was lacking. Opinions and ideas were generally found to be an acceptable basis for agreeing policy action coupled with thorough evaluation. However, the analysis revealed that the evidence was fiercely contested around food advertising to children and action agreed was therefore limited. Conclusion: The Summit demonstrated that policy action will move forward in the absence of strong research evidence. Where powerful and competing groups contest possible policy options, however, the evidence base required for action needs to be substantial. As with tobacco control, obesity control efforts are likely to face ongoing challenges around the nature of the evidence and interventions proposed to tackle the problem. Overcoming the challenges in controlling obesity will be more likely if researchers and public health advocates enhance their understanding of the policy process, including the role different types of evidence can play in influencing public debate and policy decisions, the interests and tactics of the different stakeholders involved and the part that can be played by time-limited yet high profile events such as Summits. Page 1 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Within this environment characterised by public policy Background Any policy-making process is complex – it deals with inertia, the issue of childhood obesity, and the need for human and political dynamics, the use of resources, and effective interventions, was brought to the forefront of power [1]. The development and implementation of pol- one Australian state government's agenda through the icy in a democracy seeks to meet multiple objectives [2]: NSW Childhood Obesity Summit (hereafter referred to as addressing major health and social policy problems, using 'the Summit') in 2002. While obesity had already been public resources wisely, satisfying a range of stakeholders, identified as a problem, how to respond was unclear. avoiding conflict, and ensuring that political and eco- With little evidence available to guide Government nomic objectives are met. Research is only one influence responses to the issue, the state health department's (here- in the ongoing process of policy-making [3]. In setting the after referred to as NSW Health) articulated purpose of the agenda, formulating policy, and implementing and evalu- Summit was to i) create better understanding in the com- ating it, various forms of evidence are sought and utilised. munity; ii) inform Members of Parliament; iii) hear and While conventionally such evidence is conceived as being consider the views of families, parents and young people; derived from "scientific and objective" research, it is iv) examine existing approaches and consider new ideas increasingly clear that a much wider range of sources and in a bipartisan forum; v) consider evidence; vi) identify forms of evidence are influential [4]. There has been sig- ways to improve existing strategies and services; vii) build nificant debate in Australia about the interface between community consensus about future directions, and viii) evidence and policy-making [5], but little detailed analy- recommend a future course of action so that the best avail- sis of the way evidence shapes the process of policy-mak- able strategies, both long and short term, would be imple- ing. This paper examines the role of data and evidence in mented to overcome the childhood overweight and public policy-making in response to childhood obesity in obesity problem [13]. This paper examines the role of evi- an Australian state, New South Wales (NSW). dence and data in entrenching childhood obesity on the policy agenda, in shaping the Summit debate and inform- Overweight and obesity (O&O) in Australia, as in many ing the outcomes and the policies that were subsequently countries, has risen at an alarming rate over the last 20 adopted. years. Overweight is classified as a body mass index (BMI) of 25 and above, and obesity as 30 and above [6]. Obesity Methods in men in Australia rose from 9.3% in 1980 to 17.1% in Data collection 2000 and for women from 8.0% to 18.9% [7,8]. O&O in Data were collected from the transcripts of the Summit children and young people has also increased markedly. proceedings [14-16], media articles, the Summit Commu- From 1985 to 1995 the level of combined O&O in chil- niqué [17] which outlined the agreed resolutions, the Gov- dren more than doubled in all but the youngest age group ernment Action Plan [18] published after the Summit and of boys whilst the level of obesity tripled in all age groups the announcement by the NSW Health Minister in and for both sexes [9]. December 2002 [19]. 'Factiva', a searchable archive of print media, was used to identify articles that referred to Despite rising obesity, the policy response has been lim- childhood obesity in the four main NSW statewide news- ited and hampered by a lack of evidence concerning effec- papers and one national newspaper (the Daily Telegraph, tive interventions. The World Health Organisation Sunday Telegraph, Sydney Morning Herald, The Sun Her- (WHO) has highlighted "globesity" and released the Glo- ald and the Australian) in the three months prior to the bal Strategy on Diet, Physical Activity and Health [10]. Ear- announcement of the Summit in July 2002 until the first lier, the United States (US) Surgeon General's Call to public response from government in December 2002. Action emphasised the need to create supportive environ- There were 127 articles retrieved from this search. ments which provide accessible and affordable healthy food choices and convenient opportunities for regular Seven semi-structured key informant interviews [20] and physical activity [11]. one focus group discussion (FGD)[20] with three health staff involved in the Summit's organization were also con- Australia was one of the first countries to produce an inte- ducted. The key informants included NSW health staff grated national strategy for the prevention of O&O. The and experts in human nutrition, physical activity, and National Health and Medical Research Council population health. The interviews and focus group discus- (NHMRC) report 'Acting on Australia's weight: a strategic sion used a guide to elicit opinions on the stimulus for, plan for the prevention of overweight and obesity' [12], was and organization of, the Summit and its outcomes. The released in 1997, but its recommendations, which focus group discussion was transcribed for analysis and included strategies such as promoting physical activity, the interviews were used as background material. dietary monitoring, and encouraging the development of school canteen policies remained largely unaddressed. Page 2 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Data analysis Results The transcripts of the Summit proceedings [14-16], media Three phases were discernible in the process of policy articles and other key documents were reviewed and con- making that occurred as part of the NSW Childhood tent analysed [20] to examine what type of evidence was Obesity Summit: 1) building and maintaining the used, by whom (eg. experts, industry, advocates) and for momentum 2) summit debate and 3) outcomes and pol- what purpose. Evidence that was valued or contested in icy formulation. the Summit debates and the media coverage received par- ticular attention. The type of evidence used was catego- 1) Building and maintaining the momentum rised into three types based on a model adapted from Obesity had been recognised as a longstanding and Bowen & Zwi [4] who outlined five types of evidence. The increasingly important public health problem. Ebbeling categorisation used in the current study were empirical et al (2002) pointed to publications decades earlier high- research (Type 1), such as randomised controlled trials, lighting the issue and the need for a policy response[24]. case control and cohort studies, time series analyses, Media interest in the issue of obesity in Australia was stim- observational studies, case reports and qualitative studies; ulated by available data highlighting "the doubling and tri- ideas and opinions (Type 2) which incorporated the two pling" of rates of obesity and concerns around the "second categories of 'knowledge and information' and 'ideas and fattest kids in the world" (FGD). Obesity was seen as "the interests' outlined by Bowen & Zwi, and included evi- new tobacco" – the public health issue which was being rec- dence such as the results of consultation processes, opin- ognised as demanding attention. Articles published in the ions and views of "experts", interest groups and peer review literature around this time [24,25] were trig- community members; and economic data (Type 3) which gers for media coverage and interviews with key inform- focused on economic evaluation, finance and resource ants and the focus group with NSW Health staff all implications. emphasised the importance of media coverage in bringing the issue to public and policy attention: Rigour Rigour was addressed through triangulation, clear exposi- "It [media coverage] was partly driven by data...the MJA tion of methods and reflexivity [21]. [Medical Journal of Australia] also carried some data on child- hood obesity and ... reinterpreting existing data sets. ...so that Triangulation is the use of different approaches, such as put it on the radar, that doesn't mean you've got [a] Summit interviews and document analysis, to answer the same happening yet... the data is essential – it is necessary, but not question which strengthens the rigour of a study and the sufficient." (FGD) interpretations made [22]. In the current study, inter- views, analysis of transcripts from the summit debate and "The doubling and tripling was the most used [news] grab eve- related documents and media coverage were used to rywhere, in every article, and it is still used." (FGD) answer questions posed in relation to the role of evidence in the NSW Childhood Obesity Summit. Why was NSW Health interested? The issue was shown to be important to the public. It provided the opportunity to It is important to consider the ways in which researchers divert attention away from other health issues which are and authors' past and present experiences may have considered solely the responsibility of government, for shaped the way data was collected and interpreted – often example, health care service provision. NSW Health also referred to as reflexivity [22]. All the authors of this paper wanted to show leadership in an area where there was are involved, at some level, in public health advocacy and arguably Federal Government inaction. In New South support a range of initiatives to address public health Wales there was a clear perception that "prior to the Summit problems, including childhood obesity. The paper arose there was a national leadership vacuum" around childhood from a desire by the authors to better understand and obesity (FGD). An earlier government summit on illicit reflect upon the role of evidence and its use by the differ- drugs [26], had mobilised massive public attention and ent stakeholders in the Summit debate and how the resources and it was hoped by NSW Health that a child- debates around evidence were seen to influence the reso- hood obesity summit would draw in funds and resources lutions agreed. The involvement of all authors in the anal- to address this public health problem. A summit was seen ysis and interpretation of the data presented in this paper, as providing scope to debate interventions in an area data triangulation, clear exposition of methods, conduct where there was no scientific or political clarity at the of a focus group with some of the key actors involved, and time: reflection on alternate ways of viewing the data were all important in enhancing the rigour of the study and the "there was interest, we were asked to do things, write things, credibility of the interpretations made [22,23]. pull things together... there were lots of false starts...we had things in train that were going to take another 5 or 10 years Page 3 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Table 1: Number of media articles and evidence type used Number of articles Month Total Articles Type 1 Type 2 Type 3 April 5550 May 7630 June 9952 July 15 740 August 10 6 3 0 September 40 15 16 2 October 11 450 November 11 8 8 2 19 12 12 3 December Announcement of Summit Month in which Summit was held Month in which government initial policy response was announced and they said they wanted a solution today... a summit was sug- tions [16]. Once the Summit was underway, Type 2 data gested as a way forward."(FGD) were more widely reported and ideas from experts, com- munity members and key stakeholders concerning the Table 1 shows the number of media articles by month way forward, were presented in the media. In putting for- between April 2002 and December 2002. Within each ward their views, these stakeholders called on common month the percentage of articles that used Type 1, Type 2 sense understandings, research studies or pointed to a lack or Type 3 evidence are identified. All the articles drew on of conclusive evidence to support inaction. more than one type of evidence. Peak months of coverage were July when the Summit was announced (n = 15), Sep- Food advertising to children was a case in point. Prior to tember when the Summit was held (n = 40), and Decem- the Summit, debates about evidence in the media focused ber when the Health Minister announced the preliminary on taxing 'high fat foods' and banning food advertising to government response (n = 19). In the months prior to the children. The soft drink industry spoke about the lack of announcement of the Summit, childhood obesity was good evidence for the effectiveness of such initiatives and covered 1–2 times per week in the newspapers studied. the negative economic impact of a "fat tax". Physical activ- ity and the role of parents as an influence on obesity were In the lead up to the Summit, most of the articles cited evi- highlighted by the advertising and food industries as dence of at least one type concentrating on Type 1 evi- being the major influences on childhood obesity. Results dence focussed on the magnitude of the problem, backed from Sweden which were stated by the food and advertis- up by expert opinion (Type 2). In the month before the ing industry as showing obesity rising despite an advertis- first announcement by government in December eco- ing ban were used to demonstrate that "there is no evidence nomic data (Type 3), always referring to the cost of obesity that advertising makes children eat more fatty foods" (The Aus- to the health care system, were also reported. tralian Newspaper, 1 July 2002)[28]. It became clear from the media coverage during the Summit that a ban on food Prior to the Summit and throughout the study period, advertising was the critical concern for industry who were Type 1 evidence was widely reported and largely uncon- calling a 'clear link' between harmful childhood behav- tested, quoting authoritative sources such as the Lancet iour and commercials, with editorials suggesting that [24] and the Medical Journal of Australia [25] concerning instead "parents are the dominant influence on food choices" the magnitude of the problem. Media representations (Daily Telegraph, 12 September 2002) [29]. drew on such data to present 'sound bites' to stimulate debate. The most commonly reported statistics were that 2) Summit debate either one in four, or one in 5 children in Australia was The Summit was held in September 2002 at Parliament overweight or obese and that overweight and obesity had House in NSW. An across-government organising com- doubled between 1985 and 1995. These data from mittee oversaw delegate selection and sought to ensure Magarey et al (2001) [9] were also contained in the back- balanced representation including: i) children and young ground document prepared for the Summit [27] and people; ii) families, parents and community perspectives; included in the factual preamble to the Summit resolu- iii) experts; iv) relevant peak bodies; v) special population Page 4 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 perspectives, such as the socially disadvantaged, people data were provided. Data from the US concerning changes from culturally and linguistically diverse communities, in levels of physical activity were presented: "most people in Aboriginal and Torres Strait Islanders, rural and remote my generation walked to school, today less than a third of chil- communities, and people with disabilities. dren in the United States walk to school" (US expert, Day 1, pg 12) and Australian data on sedentary activity: "97% of The Summit provided an opportunity for delegates to our adolescents watch television ... between 60 and 80% play present their case for action during plenary sessions. Dur- computer or video games." (Australian expert, Day 1, pg 16). ing the Summit, nine working groups (WGs) were con- Anecdotal observations about changing societal behav- vened: i) Early Childhood, ii) Family and Community, iii) iours and environments were widely cited and seen as School Education, iv) Health, v) Sport, Recreation & Fit- important factors to address despite the lack of reliable ness, vi) Local Government, vii) Commercial Food Indus- trend data and research evidence: try, viii) Media, and ix) Transport and Planning. "we do not yet have evidence that any single one of these factors The WGs were requested to put forward 10–15 resolutions is driving the epidemic" (US expert, Day 1, pg 13) for the Communiqué to be presented to government. The importance of evidence for the resolutions was made clear "we know very little in any, firm solid way about the factors that by the NSW Premier on the first day of the summit when influence young people to be active or sedentary – all we have he referred to the NSW Drug Summit [26], which had to work with over the next three days are some recently been held in May 1999. informed guesses and some far less well-informed speculations" (Australian expert, Day 1, pg 12) "The Drug Summit emphasised looking at evidence, basing pol- icies on evidence ... I would like that to be your guide too." The views of young people, were an integral part of the (NSW Premier, Day 1 pg 33) Summit process and provided an emotive appeal to take action. Young people's stories were shown on video and The case for action to tackle childhood obesity was uncon- they addressed the Summit. However, there appeared to tested from the outset. In opening the Summit the NSW be little attempt to draw these views together, articulate Minister for Health referred to strong empirical evidence common threads or examine whether and how such views of the magnitude of childhood obesity: related to other empirical data and expert opinion. A young person opening the Summit stated that "It is genu- "In Australia the level of combined overweight and obesity inely important that our voice be heard"(young person, Day 1, between 1985 and 1995 has more than doubled... Today in pg 2). The FGD participants saw young people's stories as New South Wales one in five children aged between seven and being powerful in stimulating action: 15 are classified as being either overweight or obese." (NSW Minister for Health, Day 1, pg 5) " we had to have lots of consultation processes that included the voice of the child... engaging the children... it was the most Experts from government departments, academic institu- powerful thing." (FGD) tions and the health service put forward similar statistics that highlighted the magnitude of obesity. Many outlined A young person at the Summit, however, expressed frus- the consequences of overweight and obesity for type 2 dia- tration about the focus on evidence: betes in particular. The use of simple statistical concepts such as "doubling in rates" and "one in five of our children" "Unfortunately we have been bombarded with statistics. They were commonplace. Economic evidence highlighted the have been repeated over and over again ... we are almost scared cost of the "obesity epidemic" to society – "it costs us a com- to put up a decent suggestion." (young person, Day 2, pg 30) munity $830 million a year" (Minister for Health, Day 1, pg 7) and individuals – "in one year the personal cost to individ- In contrast to the research evidence supporting the magni- uals who are obese is $19 billion" (expert, Day 1 pg 16). Such tude of the problem and the influencing factors, evidence data were uncritically and widely accepted during the supporting calls for action were mostly opinion and ideas Summit. with some reference to overseas efforts. Nonetheless, much was made of the need for evidence-based strategies, On the opening day, experts, parents, community groups with a US expert claiming three strategies that were "defen- and industry talked anecdotally about societal changes sible, but not conclusive" (US expert, Day 1, pg 13): breast- over decades and their impact on physical activity and feeding, limiting television viewing and the promotion of food consumption. Statistics and studies were referred to physical activity. A Cochrane systematic review [30] cov- in support of these observations, such as an increased reli- ering 1985–2001 and encompassing 14,000 studies was ance on carbonated sugared drinks, although no actual reported (researcher, Day 1, pg 37). It found 11 studies of Page 5 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 a high enough quality to examine the effectiveness of the the Summit was to: "Frame evidence-based solutions within a intervention and it found only small or no effects with community-based 'reality check' perspective."(Day 3, pg1) those interventions that were most effective focussed on reducing sedentary behaviour. A few delegates questioned Evidence of the magnitude of overweight and obesity was the need for evidence from primary prevention trials, included with little debate. Statements about the influenc- pointing to broader experiences that tell us "what works". ing factors were carefully worded to reflect agreement on They highlighted other successful public health cam- importance and available evidence: paigns such as in tobacco control as evidence for the suc- cess of a range of strategies, including advertising controls "Although physical activity trend data is lacking, it is apparent and taxes: that children and adolescents are less physically active" (Day 2, pg 72) "we do need evidence, we do need to work at what has been shown to be the most effective, but that should not inhibit us "An increase in television viewing is associated with an increase from acting now. There have been a number of successful public in obesity in children. An increase in sedentary behaviour is health programs that have been introduced without definitive associated with an increase in obesity in children. Experts have evidence." (expert, Day 1, pg 39) advised that television viewing needs to be one of the targets for obesity control efforts" (Day 2, pg 78) The FGD and comments by Summit delegates highlighted the need for action coupled with thorough evaluation: Exposure to advertising messages was included in the fac- tual preamble referring to the range of potential influ- "There needs to be a recognition of the sense of urgency...that ences on food selection behaviours. The resolutions about policy won't wait for the data." (FGD) food advertising to children generated the most debate concerning the evidence-base for such interventions (see "This is about promising interventions, we have to just go with Figure 1) and the relationship between food choice and promising interventions, make sure they do no harm and just television viewing. This debate illustrates the use of differ- evaluate the heck out of them, and then maybe in ten years ent types of evidence by industry representatives as one time, if they weren't the best things to do, well at least we did means of opposing calls for a ban on food advertising to something" (FGD) children. "we need periodic surveys to tell us how we are doing with A resolution to ban food advertising to children was not respect to implementation of strategies ... we need causal mod- agreed. In its place agreement was reached to have an els, that is, longitudinal studies which allow us to link risk fac- independent review by the Federal government of the reg- tors like change in the food supply with changes in the ulatory arrangements for food advertising "in recognition prevalence of obesity." (US expert, Day 1, pg 13) that food advertising is one of the contributing factors to the prevalence of eating habits that may promote obesity"(Day 3, "we would like to see a regular – maybe five yearly – national pg 9) in addition to a review of a voluntary code to be nutrition, physical activity and health survey." (industry, Day undertaken by industry. Attempts by the Food Industry to 2, pg 2) have this statement deleted from the Communiqué were not successful. A systematic review of the impact of food "Certainly, we need to take action, but at the same time we advertising on diet, physical activity and childhood obes- need to be doing research. We cannot continue to act in an evi- ity was also recommended. dence vacuum." (expert, Day 2, pg 20) All other resolutions passed with minimal debate, includ- The most contentious issue centred on the role of food ing those addressing physical activity, school education, advertising to children (see Figure 1). The intensity of the transport and planning. Most of the resolutions agreed at debate between food industry representatives and the the Summit and taken up in the subsequent Government advocates of a ban on food advertising to children clearly Action Plan [18] were focused on physical activity and illustrates the way different types of evidence are drawn nutrition education. Mandatory guidelines for school upon to articulate a particular position or undermine that canteens also passed as a resolution despite some opposi- of opposing perspectives. tion from industry. Numerous resolutions in the Commu- niqué [17] referred to research and a detailed section on 3) Outcomes and policy formulation surveillance and monitoring proposed a funded collabo- The final Communiqué to government was to include a rative centre of excellence in research, prevention and "factual foundation" and recommendations and resolu- management. tions for future action. The purpose of this component of Page 6 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Much of the evidence put forward by advocates for a ban on food advertising focussed on the amount of food advertising in children’s programming and the relationship between food consumption and television viewing: “There are studies showing a clear correlation between the consumption of foods advertised on television and the amount of time spent watching television.” (expert/advocate, Day 1, pg 27); “More than 80% of ads are for foods that are outside what we call the core food groups.” (expert/advocate, Day 1, pg 28). The Lancet article by Ebbeling et al (2002) was also cited as evidence for a ban: “Lancet writers considered … that we should prohibit food advertisements and marketing that is directed at children.” (expert/advocate, Day 1, pg 24) . Advocates documented the number of food advertisements in children’s programming: “Australian children 5-12 years watch on average two and a half hours of television daily and this includes up to 52 minutes of advertising” (advocate, Day 2, pg 21) . Industry contested these data: “there is no legal way they could be exposed to 52 minutes of advertising as the limits per hour are 13 minutes per hour in general programming and 10 minutes in C time.” (industry, Day 2, pg 43). They also responded by citing a report by the Ministry for Agriculture, Fisheries and Food in the UK (1996) as support for the notion that parents are more influential in food choices than advertising. The advertising industry also reported that “media research shows advertising levels transmitted to children aged 5-12 of confectionary, snack foods, soft drinks and fast food in fact decreased by 19 per cent from 1996-2001” (industry, Day 2, pg 44). The example of Sweden, where food advertising to children has been banned for 10 years, was also given as evidence by industry of the failure of this approach to reducing obesity levels (30-32). They quoted a Swedish government public health report in 2001 that stated: “The proportion of overweight people has increased in all socio-economic groups since the beginning of the 1980s” (industry, Day 2, pg 58). Industry also cited the Premier’s opening remarks focusing on evidence based policy as support for recommendations being evidence-based. They did not accept that a ban on food advertising had been shown to promote health gain: “after rigorous examination of the research literature, we concluded that there is no serious and sound evidence which shows that food advertising leads to an increase in the consumption of ‘whole categories of foods’”. (industry, Day 2, pg 57) Advocates of a ban highlighted the cost of advertising: “A single 30 second ad can cost a million dollars” (researcher and advocate, Day 2, pg 54) and “the industry spends billions of dollars to make and broadcast ads for less healthy foods” (researcher and advocate, Day 2, pg 54) as evidence that it works to influence food choices. They also refuted the relevance of the Swedish case “the rate of obesity in Sweden is extremely low” (advocate and expert, Day 3, pg 3). One advocate also stated “I really do not want to see us deflected into more evidence being required before we ban food advertising .. it is up to the food industry to prove that food advertising does not have an effect.” (researcher and advocate, Day 2, pg 54). This debate about food advertising to children demonstrated how the concept of ‘evidence’ was embraced by the food industry and used expertly to undermine the evidence presented by the advocates of an advertising ban. In particular, the industry used the Premier’s comments about evidence-based policy (type 2 evidence) together with data from Sweden (type 1 evidence) to support their position that a food advertising ban was not ‘evidence-based’ and would fail to deliver the intended outcomes. While the Swedish data was disputed by the advocates of a ban, the industry arguments were sufficient to water down the resolutions agreed at the Summit – a review rather than a ban on food advertising to children. Contesting the evid Figure 1 ence: food advertising and obesity Contesting the evidence: food advertising and obesity. Page 7 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 Limited attention was devoted to the financial and logis- However, lack of compelling evidence for interventions is tical feasibility of the resolutions – this was apparent by likely to have been a factor in the failure of government to the number of resolutions that required intervention at a commit significant new funds and to agree to controver- federal rather than state level. However, the preliminary sial recommendations around food advertising given response from the government in December considered strong industry opposition. The only contentious resolu- what was feasible in the current financial and political tion taken up by government was for mandatory guide- context: lines for school canteens: this appealed to many community groups and parents who attended the Summit "it wasn't really evidence-based, it was the feasibility of what- and government is likely to have perceived strong public ever strategy they had suggested..." (FGD) support for this intervention. Other commentators have questioned the soft policy options adopted in response to In the final Summit address by the NSW Health Minister the obesity epidemic in Australia following the NSW Sum- [16] the two resolutions specifically mentioned and mit and raised questions about the way public health strongly supported were the recommendations on school issues, such as obesity, are framed in public discourse canteens and a collaborative centre for excellence for over- [31]. weight and obesity research. These two initiatives were subsequently publicly announced in December as the key The food and advertising industries who were represented response to the Summit by the government [19]. The at the Summit used the lack of well supported 'scientific' advisers to the Minister and NSW Health were concerned evidence to oppose controls on advertising. In contrast, to ensure that the Summit resulted in some "announcea- the debates and resolutions around physical activity using ble" interventions – and the two chosen seemed "doable", anecdotal evidence, expert opinion and common-sense of value, and in some respects least contentious (FGD). solutions garnered widespread support as there was no industry that stood to suffer financially from the action proposed. Where strong interests and powerful groups Discussion This paper has sought to present key elements of the use oppose policy direction, the evidence base required for of data and evidence in the NSW Childhood Obesity Sum- government action, if it is to proceed, needs to be substan- mit. There are other dimensions of policy-making which tial. It is also possible that the more prominent role of the deserve attention and other interpretations of the process federal government in food advertising regulation and possible. As indicated by Ham and Hill "It is rarely possi- control worked against the agreement of concrete resolu- ble to agree on one version of events: the most that can be tions around food advertising to children. achieved is a plausible interpretation" [2] (p xi). Economically important industries have been seen by oth- Empirical evidence of the magnitude of the obesity prob- ers as critical in the preparedness of governments to sup- lem and the economic cost to the health system were crit- port controversial public health initiatives [32] and calls ical to generating publicity and framing the case for action for more research have been presented as tactics to delay on childhood obesity in the lead up and during the Sum- policy change [33]. However, creative and clear commu- mit. This evidence was never contested and became a part nication of the evidence has been instrumental in other of the factual foundation of the Summit Communiqué areas, notably the successful efforts to ban tobacco adver- [17]. It is clear that the combination of Type I data, which tising in Australia and in many other countries around the was largely epidemiological in origin, and Type 3 data world despite powerful industry opposition [32, 33]. about the economic costs of the problem was persuasive. There is also more scope for interaction and collaboration with the food industry than with the tobacco industry as The lack of empirical evidence for many of the influencing food as a product is not inherently harmful [33]. The food factors and related interventions, for example in the area industry can have an important role in supporting a range of physical activity, did not hamper agreement of resolu- of policy initiatives that promote healthy eating as was tions at the Summit and was instrumental in funding a evident in the NSW Childhood Obesity Summit, but are research centre to collect better data and evidence for what likely to remain adversarial where industry profits are, or works. Health officials who recognised the lack of an evi- appear to be, at stake. dence base for interventions sought to promote those that seemed most logical and appropriate, along with a con- Conclusion cern to ensure subsequent careful evaluation. The Summit The NSW Childhood Obesity Summit played a role in demonstrated that policy action will move forward in the promoting an agenda for action to address childhood absence of strong research evidence if government sees the obesity. It raised awareness in the public and political need to respond to public concerns. arena and provided a public forum for debating research evidence. The Summit demonstrated that while it is not Page 8 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 authors would also like to thank the reviewers of the original submitted necessary to have all the evidence in place to agree actions, manuscript for their constructive and considered comments. The School of that more radical policy change is much more difficult to Public Health and Community Medicine, The University of New South achieve in the absence of established and detailed evi- Wales, and The Centre for Chronic Disease Prevention and Health dence, given the interests of important stakeholders, nota- Advancement, NSW Health, are collaborating on a project, funded by the bly the private sector. The process and the outcomes of the Centre, to foster reflection, analysis and writing between academics, prac- Summit suggest that in the absence of strong Type 1 data, titioners and policy-makers, and aimed at contributing to bridging the aca- and where Type 2 evidence is contested, that policy-mak- demic-policy divide. ers may opt for the path of least resistance: a call for more and better research and support for the systematic evalua- References 1. Walt G: Introduction to health policy. Process and power. tion of interventions. While beneficial to researchers, London, Zed Books; 1994. direct and short term health gain may be limited. 2. Ham C, Hill M: The policy process in the modern capitalist state. (Second edition). Hertfordshire, Harvester Wheatsheaf; The lack of an agreed evidence-base provides politicians 3. Rist RC: Influencing the policy process with qualitative with a freer hand in choosing actions which have wide research. In (Editors). In Handbook of Qualitative Research Edited by: Denzin NK and Lincoln YS. London, Sage Publications; appeal and are less controversial, rather than those which 1994:545-557. may produce greatest health benefit. The Summit's success 4. Bowen S, Zwi AB: Pathways to "evidence informed" policy and in generating a set of resolutions should not be dis- practice: a framework for action. Public Library of Science Medicine 2005, 2:e166. counted even if large resource allocations were not forth- 5. Lin V, Gibson B: Evidence-based Health Policy: Problems and coming. Tobacco control initiatives have taken decades of Possibilities. South Melbourne, Oxford University Press.; 2003. concerted effort to realise [33] and obesity control efforts 6. World Health Organisation: Obesity: preventing and managing the global epidemic. Report of a WHO consultation. WHO are likely to face the same challenges around evidence and Technical Report Series 2000. action. The prospects of controlling obesity in the future 7. Bennett SA, Magnus P: Trends in cardiovascular risk factors in Australia: results from the National Heart Foundation's Risk will be amplified if researchers and public health advo- Factor Prevalence Study, 1980-1989. Medical Journal of Australia cates enhance their understanding of the policy process, 1994, 161:519-527. the interests and tactics of the different stakeholders 8. Dunstan D, Zimmet P, Welborn T: Diabetes & Associated Disor- ders in Australia: The final Report of the Australian Diabe- involved, and the role different types of evidence can play tes, Obesity and Lifestyle Study (AusDiab). Melbourne, in influencing public debate and the decisions of policy- International Diabetes Institute; 2000. makers in time-limited yet high profile events such as 9. Magarey A, Daniels L, Boulton TJC: Prevalence of overweight and obesity in Australian Children and adolescents: reassess- Summits. Further research is needed to increase our ment of 1985 and 1995 data against new standard interna- understanding of the role of Summits in the broader pol- tional definitions. The Medical Journal of Australia 2001, 174:561-564. itics and processes of policy-making. 10. World Health Organisation: Global strategy on diet, physical activity and health. World Health Organisation; 2004. Competing interests 11. US Department of Health and Human Services: The Surgeon Gen- eral's call to action to prevent and decrease overweight and Elizabeth Develin was involved in the original work obesity. Washington, US Department of Health and Human described leading up to and including the NSW Child Services.; 2001. Obesity Summit and is employed by NSW Health. All the 12. National Health & Medical Research Council: Acting on Australia's weight: a strategic plan for the prevention overweight and other authors are part of the School of Public health and obesity. Canberra, Commonwealth of Australia; 1997. Community Medicine, UNSW, which is in receipt of mod- 13. NSW Health: NSW Childhood Obesity Summit Draft Pro- gram. Edited by: Committee NSWCOSO. Sydney, NSW Depart- est funds to collaboratively develop this reflection and ment of Health; 2002. paper by NSW Health, the body which organised the Sum- 14. NSW Health: NSW Childhood Obesity Summit: Report of mit described. proceedings of the first day. [http://www.health.nsw.gov.au/ obesity/adult/summit/hansard/hans.html]. 15. NSW Health: NSW Childhood Obesity Summit: Report of Authors' contributions proceedings of the second day. [http://www.health.nsw.gov.au/ obesity/adult/summit/hansard/hans.html]. This paper was conceived jointly by all authors; all con- 16. NSW Health: NSW Childhood Obesity Summit: Report of tributed to analysis and writing. SN prepared the first draft proceedings of the third day. [http://www.health.nsw.gov.au/ with significant contributions by ED and comments from obesity/adult/summit/hansard/hans.html]. 17. NSW Health: NSW Childhood Obesity Summit Communi- AZ and NG. All authors contributed to writing successive que. [http://www.health.nsw.gov.au/obesity/adult/summit/commu drafts, feeding in literature, analysis and insights. nique.pdf]. 18. NSW Health: Prevention of Obesity in Children and Young People: NSW Government Action Plan 2003-2007. Sydney, Acknowledgements NSW Department of Health; 2003. The authors wish to acknowledge insights and feedback from Bill Bellew 19. NSW Minister for Health: Youth Obesity - Five Year Plan. In (Director, Centre for Chronic Disease Prevention and Health Advance- Media Release , NSW Health; 2002. 20. Patton MQ: Qualitative Research and Evaluation Methods. ment, NSW Health) and Phillip Vita (former Manager, Nutrition & Physical 3rd edition edition. California, Sage Publications Inc.; 2002. Activity Branch, NSW Health, now Executive Officer, NSW Centre for 21. Mays N, Pope C: Assessing quality in qualitative research. BMJ Physical Activity & Health) and the work of Kate Hawkins and Sarah Yallop 2000, 320:50-52. from NSW Health who conducted the key informant interviews. The Page 9 of 10 (page number not for citation purposes) Australia and New Zealand Health Policy 2005, 2:17 http://www.anzhealthpolicy.com/content/2/1/17 22. Denzin NK, Lincoln Y: Handbook of Qualitative Research: 2nd Edition. California, Sage; 2000. 23. Ebbeling CB, Pawlak DB, Ludwig DS: Childhood obesity: public health crisis, common sense cure. Lancet 2002, 360:473-482. 24. Goodman S, Lewis PR, Dixon AJ, Travers CA: Childhood obesity: of growing urgency. The Medical Journal of Australia 2002, 176:400-401. 25. NSW Government: NSW Drug Summit. [http://www.dru ginfo.nsw.gov.au/drug_summit/ government_response_to_the_drug_summit]. 26. NSW Health: NSW Childhood Obesity Summit Background Paper. [http://www.health.nsw.gov.au/obesity/adult/summit/ bground_paper.html]. 27. The Australian: Going off the scales. In The Australian 1 July 2002 edition. Sydney, ; 2002. 28. Daily Telegraph: Dishing out blame. In Daily Telegraph Sydney, ; 29. Campbell K, Waters E, O'Meara S, Kelly S, Summerbell C: Interven- tions for preventing obesity in children. The Cochrane Database of Systematic Reviews 2002. 30. Lin V, Robinson P: Australian public health policy in 2003-2004. Australian and New Zealand Health Policy 2005, 2:. 31. Chapman S, Lupton D: The fight for public health: principles and practice of media advocacy. London, BMJ Publishing Group; 32. Yach D, McKee M, Lopez AD, Novotny T: Improving diet and physical activity: 12 lessons from controlling tobacco smoking. BMJ 2005, 330:. Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 10 of 10 (page number not for citation purposes)

Journal

Australia and New Zealand Health PolicySpringer Journals

Published: Jul 20, 2005

References