Garrety, K. (1997) 'Social Worlds, Actor-Networks and Controversy: The Case of Cholesterol, Dietary Fat and Heart Disease', in Social Studies of Science 27: 727 - 73.
[A massively researched account, apparently based on a PhD thesis. The basic argument is that cholesterol came to be seen as a major cause of heart disease as a result of not only clinical trials but because particular social coalitions wanted to advance it as a major factor. These coalitions are then further analysed in terms of actor network theory and symbolic interactionism. There is an excellent summary of the argument right at the end, on the last two pages:
Knowledge and policies about diet and its links with health disease were constructed. Different bodies and organizations played different parts at different times. First of all, heart disease had to be seen as a high priority and major social problem. Then various bodies with views about diets, including food producers and their claims, had to be managed by various medical authorities in the interests of continuing their dominance. Attempts to derive scientific proof for the relation played only a part. Several interventions by various bodies, including Congress, where necessary to judge the rival claims, and as a result, 'the worlds of policy and science merged'. Eventually, the scientific basis of the claim ceased to be relevant in its popularity. ]
A large and expensive random clinical trial took place in 1984 attempting to settle once-and-for-all whether or not saturated fat and cholesterol was linked to coronary heart disease (CHD). The technique was to splits people into a control group and an intervention group and see if the number of heart attacks varied. The intervention involves reducing cholesterol. In order to make the differences clear, both groups were recruited from high risk subjects [so that lots of heart attacks would be observable], and cholesterol was to be reduced using a specific drug, rather than relying on longer term effects of changed diet.
After one year, there was less heart disease in the intervention group. However, a controversy arose about how statistically significant this difference might be [the footnotes tell an interesting story about the technical decisions involved -- to settle for a level of significance a 0.05 after having abandoned an earlier level of 0.01, and choosing a one tailed test of significance rather than a two-tailed one -- 'It is much more usual to use a two-sided test in clinical trials. A one-sided Test assumes that the result of the experiment can only go one [positive] way... a two-sided test also takes into account the possible deleterious effect of drug treatment' (n3 758). Nevertheless, the results were reported as conclusive, and the results were extended to include a claim that dietary reduction of cholesterol would benefit everybody [again, this was heavily criticised].
Reductions in dietary fat and cholesterol were therefore recommended strongly by the scientists involved. One quarter of the population was defined as being at risk, and a substantial educational programme was launched. It is now common knowledge that cholesterol and fat is linked to heart disease. A substantial 'medical surveillance and intervention effort' (729) ensued.
The issue for sociologists is how these 'facts' came to be accepted, and there are several possible ways to investigate the issue. The particular interest here is that the extent to which sociological factors such as solidarity and power influenced the debate. Actor network theory has been influential in suggesting that interests get translated as a result of membership of networks, but Latour in particular sees no reason to privilege social factors as against, say, non-human factors. Symbolic interactionism, on the other hand studies how social groups form, including occupational groups like scientists. In this approach, 'social worlds' explain how science gets constructed. Such social worlds can merge with other worlds and form alliances based around cooperation and conflict and boundary maintenance. Science can help to legitimize the operations of particular social groups like this, especially in demarcating boundaries by invoking the 'cognitive authority of science' (731).
First though, a controversy has to become known to different social worlds. In this case, heart disease began as a minority interest, but soon became a major social problem. This was partly due to the activities of heart specialists. In their campaign is to gain popular attention, they often speculated about the links with 'obesity, high blood pressure, "cultural conflicts" , aggression and overwork' (732). Some Washington lobbyists also targeted Congress for funding for research into cancer and heart disease, and increased the amount of funds considerably.
Cholesterol itself emerge as a factor from certain early laboratory experiments, including those designed to test the effects on rabbits. These experiments were controversial, and their ability to apply to humans was much in doubt [because cholesterol is not normally found in vegetarian rabbits]. However, this work did produce a number of easy tests for cholesterol levels. Later work found raised cholesterol levels in humans with heart disease, but early hopes for a strong and measurable relation were not supported. Instead there seemed to be 'an overlap in individual cholesterol measurements. Some heart disease victims had normal or even low cholesterol levels, while some apparently healthy people had high cholesterol levels' (734). As a result, cholesterol levels gave way to investigations of other elements.
By the mid-1950s, however, cholesterol was identified more firmly with scientific studies, and the evidence of an effect was seen to be adequate. Particular physiologists went public with the recommendations of a low-fat diets, although there were still some controversy. There was sufficient money to permit a 'highly influential worldwide epidemiological survey in populations whose diets varied in fat content' with apparently positive correlations being achieved (735). This work became immediately popular.
However, the work can be criticised. One problem is that it was not a longitudinal study measuring the effects of cholesterol over time. Another that 'diagnostic and medical practices varied' (735). Finally, there are many other variables which should have been taken into account -- 'other illnesses, the level of physical activity, rates of smoking, other components of the diet', and exceptions 'such as the French (who have a low heart disease rates despite a high-fat intake)' were not included (735).
The popularity of the argument remained, while criticism appeared only in 'specialist journals and textbooks' (735). The scientist concerned was also an active network builder and good at involving the media. Journalists neglected all the caution and asserted that there was a connection. The connection became a popular press story, and journalists increasingly suggested there was a scientific consensus about cholesterol and its link with heart disease.
Medical journals also published studies reporting some positive correlations (736), but controversy still remains, especially about the effects of dietary change on heart disease itself [not just on cholesterol]. Not all scientists agreed with the popular view. There was a widespread suspicion that nutritionists were not scientific. Some of this scepticism appeared in the press as well, attacking "food faddism" (737) and substantial intervention in the American diet. The sceptics managed to get their views heard in the American Heart Association (AHA).
Other social worlds with an interest also developed. Some members of the public got involved, and, in the late 1950s, the food industry did as well. Food companies had already become interested in the work of the Nutrition Foundation, 'a research body funded by the Food Industry', and were keen to use science to legitimate their products (739). Thus Mazola corn oil began to market itself as a preventive against heart disease, because it contained polyunsaturated fatty acid.
Butter and margarine manufacturers joined in. They had already been involved in a 'long-standing feud' (739) involving prices and taxes. The new twist was to make margarine from polyunsaturated fats, so it could then claim health benefits. Scientists responded by insisting on more research, insisting that causal relationships between cholesterol and heart disease had not been proved, despite an apparently strong public demand for healthy products identified by manufacturers of low cholesterol foodstuffs.
The early division between science and food faddism was weakened by a new AHA policy,'a new strategy for regaining cognitive authority' (740). Interventionists had been steadily gaining influence in the AHA. However, there was still a great deal of caution in the statement. This caution was exploited by rival commercial interests to support their own claims. Another national body (the American Medical Association -- AMA) made similar cautious recommendations about diet, and this one was also used by rival claimants. The AMA insisted on laboratory tests and medical control, but this did little to calm exaggerated claims and public fears. Finally, both bodies issued general statements advocating a reduction of fat intake.
None of this drew from any more recent or authoritative research. There was some epidemiological support, but generally, the decisive factor was 'a new construction of the vague but suggestive relationship between cholesterol levels and heart disease risk' (742). The epidemiological work involved in grouping men according to cholesterol levels and then calculating rates of heart disease. The rates were then standardized so they could be compared. Relative risk for different cholesterol levels could then be calculated. According to one study, risk increases considerably from low to high levels of cholesterol. Such statements appeared in the medical literature and became increasingly authoritative, although there were still caution about a lack of "definitive proof" (743).
There was a series of dietary experiments, but the methodologies were still suspect -- dealing with small groups, high drop-out rates and a lack of double-blind allocation (see 743). There was an attempt at a large scale definitive experiment in the 1960s, but methodological problems seemed insuperable:
'Biostatisticians calculated that for a trial using "normal" healthy people, a five-year study of 100,000 men would be necessary to detect, with a satisfactory level of statistical significance, a 20 per cent decrease in the incidence of heart disease in the group on the cholesterol - lowering diet' (743).
After much debate the definitive experiment had still not been conducted. An alternative design involved a "multiple risk factor intervention trial" designed to see if intervention on smoking, high blood pressure and high cholesterol levels could reduce deaths by heart disease. This also 'failed to produce the desired result. There was no difference in the death rate from heart disease between the intervention and control groups' (744). However, this was explained away as an effect of the trial itself in persuading all the men involved to change their habits.
Another large trial was conducted [the one discussed earlier]. Scientists wanted journalists and the public to see this as providing a decisive final answer. In the process, they legitimated drug treatment of cholesterol levels. Pharmaceutical countries saw 'an opportunity for large profits' and continued to attempt to develop a cholesterol lowering drug. Following the results of the trial, sales of these drugs 'have skyrocketed' (745).
The trial received considerable press attention, and appeared to take the official view that the link between cholesterol and heart disease was "now indisputable" (745). However, objections to the conduct of the trial ensued [as we saw]. Some critics argued that science had lost its objectivity and had become affected by public opinion and policy.
However, advocates of dietary change would only acknowledge "gaps in knowledge' (745). They went on to suggest that science should not be separated from policy and should reduce its standards of proof given the seriousness of the matter.
Different social worlds continue to offer different interpretations. Cholesterol is now a popular factor in heart disease, and most official organizations support dietary change 'Over the decades, the sceptical scientists were marginalised and the saturated fats industries managed to adapt to the new nutritional knowledge' (746).
The American diet is now lower in eggs and dairy. Meat industries tried to fight back and persuaded Congress to delay or even support their efforts. They also tried to create new healthier versions, low in fat and cholesterol. However, attempts to discredit the cholesterol hypothesis were less successful, and legal challenges prevented them from making them.
Health policy makers liked the idea that healthy diets could reduce disease, especially since that focused on individual prevention. Critics of the food industry also welcomed the cholesterol lobby effort. A congressional committee finally produced dietary goals suggesting lower fat and cholesterol, but the 'egg, meat and dairy lobbies' managed to soften their recommendations (748). The affluent middle classes became interested in healthy lifestyles and keeping fit [which has been much discussed].
Commercial interests seem to have compromised by selling both healthy and unhealthy products, and attacking the cholesterol lobby is 'no longer worth the effort'. The controversy exists only as an occasional technical commentary on the lack of 'facts' behind nutritional policy: however, policy makers and the public 'have never required unequivocal proof of efficacy' (749).
It is possible to see a number of alliances and connections between journalists, research lobbyists, scientists, the food industry and policy makers. There was also conflict between social worlds and controversy. In the first case, the public began to develop their own views about diet, building on our long interest in 'lay health movements' (750). However, this led the medics to oppose them and to attempt to assert their own authority by drawing boundaries between quacks and proper medics. However, boundaries are difficult in the case of food, and much effort is devoted to maintaining them.
Specialist bodies like the AHA did not want to be seen to be doing nothing about heart disease. Cholesterol diets at least had a more scientific basis than other fads. Pro-cholesterol researchers managed to gain influence and were able to reject accusations of being quacks. Medical cautions carried less weight in the ensuing policy, and were quietly minimised.
Food lobbies were less successful in claiming legitimacy -- after all, they were associated with 'deceptive advertising' and 'the pursuit of profit', while medical associations could pose as disinterested professionals. Eventually, the food lobbies recruited sceptical scientists, but they were in a minority and associated with 'an earlier and obsolete era of nutritional knowledge' (752).
Eventually, in 1980, doubts resurfaced in a report by the Federal Nutrition board, but this time they were seen as defying 'the orthodox scientific position', and a wealth of policy documents recommending dietary change. The critics were seen as linked to food industries and public criticism led most of the critics to resign from the board. Thus 'by 1980, it had become very costly to argue that "gaps in knowledge" warranted a cautious attitude towards dietary change' (753). Similarly, it would have been impossible to announce another 'long, expensive trial... most people were unaware, or did not care that the "definitive proof" was still lacking. The scientific "evidence" had simply become irrelevant' (753).
Some aspects of this story supports actor network theory and the way in which networks translate interests. However, a significant issue is why so many people came to support the cholesterol hypothesis. Actor network theory tends to focus on uncontroversial applications of scientific knowledge, where disputes are settled by 'non humans, or "nature" itself' (753). In these cases, sociological explanation seems unnecessary. In controversial areas though, 'nature' is less decisive. The cholesterol hypothesis had to become accepted through persuasion and through the activities of organizations such as the food industry, advertising, lay diet enthusiasts, and various other social movements.
The social worlds theory seems to be better in allowing a larger role for human agency. Concepts such as boundaries and 'boundary objects' require the acknowledgement of multiple meanings. Boundary objects inhabit both social and scientific social worlds -- in this study, boundary object are 'a bundle of knowledge claims which are linked dietary fat and cholesterol to CHD' (755). The facts had multiple interpretations. An analysis of power is required to establish why some definitions were imposed and legitimized. Symbolic interactionism might not be suitable here, although an analysis of power can be accommodated. In this example, is clear that some of the participants had more 'cognitive authority and resources' than others (756). Symbolic interactionism already analyses how professionals develop and maintain their power to define the situation -- however, this power is never absolute and has to be negotiated. This case study shows how proponents of dietary change were able to do this, through various structures and opportunities.
Overall, the construction of knowledge and policies about diet and heart disease is complex. It was first necessary for heart disease and its solution to be highlighted, and then for various alliances and bodies to be managed [and the story is summarized 757 - 8].