As more infants survive their birth, and as we all grow and develop better, live longer and healthier lives, and maintain various physiological functions longer and longer into old age, we become more anxious about our health. The absence of external stresses always produces introspection - suicide rates fall in wars. The undoubted physical benefits of modern life allow us to indulge ourselves in worry; indeed to wallow in it if so inclined. This leads to a situation where instinctive concerns replace a reliance on evidence, overthrowing the benefits of the Enlightenment. If we accept that the basic tenet of the Enlightenment was a belief in the superiority of reason over superstition, this is a great loss.
We are thus inclined to look for trouble and to invent mantras, like the precautionary principle, that may ward it off. The scientifically untrained have a tendency to overvalue single steps in a chain of causation - the dependence of a hypothesis on a complete chain of confirmed steps is counter-intuitive to many. It is not easy to provide instant certainty with science, and Bertrand Russell’s dictum that ‘what man desires is not knowledge but certainty’ is relevant. So if you have rheumatoid arthritis, a miserable disease that fluctuates in severity, the latest action taken to alleviate your difficulties will often be invested with a beneficial function it does not deserve if its application coincides with remission.
Now, as a doctor, it is simple and practical to rejoice in anyone feeling better for whatever reason. There can be no objection to any therapy that is not harmful if it confers relief. The assertion that cannot be supported is that the disease process is affected, unless there is data to show it. The downside is that treatments that allow further damage to occur while the disease progresses are irresponsible. Abundant data exist to show that those who have high blood pressure need to have it lowered. Any therapy that claims to do this must do it in a way that can be documented, and that documentation must be of an adequate scientific standard. It is perfectly simple to devise proper programmes in investigation of the effectiveness of homeopathy - that its effectiveness cannot be demonstrated is a challenge for it, not for orthodoxy.
There is another aspect to this part of the problem. Postmodern philosophers have replaced the Enlightenment’s reliance on reason with a conviction that ideological motives colour all deliberations. So in any scientific debate, an opponent’s reasons for holding a particular viewpoint or belief depend on his motives rather than his knowledge base. If you are not prepared to engage in a rational debate you can avoid changing your mind by attributing your opponent’s stance to his background and class interests, say. I find this offensive; as a pathologist often responsible for the production of uncomfortable information (a positive biopsy for cancer, say) my concern is for a clinical collaboration that results in an evidence-based intervention. The idea that ‘beliefs’ will colour this is absurd. Unnecessarily radical procedures went on being performed for carcinoma of the breast for more than 20 years because no one had evaluated their worth - it was ‘obvious’ that all you had to do was to get Ahead of the Tumour.
A state of anxiety is not a good substrate for rational thought. We are all anxious when ill, and perhaps more anxious about others when they are ill. But unfounded anxiety is damaging. In the controversy about the MMR (measles, mumps and rubella) vaccine, a set of indifferent data was made much of by the uninformed. Even had the hypothesis presented been true, the suggested hazard (there were no data to describe a ‘risk’) should have been balanced against the established risks - measles is highly contagious and outbreaks occur in communities with immunisation rates much below 75-80 per cent. The illness will be accompanied by ear infection in 1 in 20 cases; by pneumonia or bronchitis in 1 in 25 cases (often leaving permanent damage); by convulsions in 1 in 200 cases; meningitis or encephalitis in 1 in 1,000 cases; death in 1 in 2,500-4,000 cases; and the terrible problem of sub-acute sclerosing panencephalitis in perhaps 1 in 8,000 children. Data existed to refute the allegations - refusal to accept it damages the group we seek to protect.
But there is a better documented example of precautionary advice being damaging. In the years between 1986 and 1988 there were around 1,500 deaths per year described as belonging to the Sudden Infant Death Syndrome (SIDS) in England and Wales. In 2004 there were 313. What had happened?
Perhaps as a result of the view that the immature brain stem function of infants made them vulnerable to certain stimuli affecting the airways, it was assumed that it would be sensible to sleep infants on their front or side, in the way it was accepted that it was better to nurse unconscious or vulnerable adult patients. At that time there was also an increasing use of intensive care methodology for premature infants. This practice was precautionary - there were no data. After a great deal of investigation, some absurd hypotheses, and irritation with funding bodies (such as the Medical Research Council) that they were not doing enough to investigate the problem, a series of observations, mainly Australasian, demonstrated that this position was dangerous. The ‘Back to Sleep’ campaign resulted in a drop in SIDS that has continued (0.65 deaths/1,000 live births in 1996, 0.55/1,000 in 2000 and 0.43/1,000 in 2004). Blair (2003) estimated that the change in policy had saved 10,000 infant lives in the last decade; my own estimate is higher. It is important to note that the main epidemiological characteristics of these cases has not changed (marital status, maternal age etc. - see Leach et al. 1999), although a change in practice by some coroners in the description of SIDS versus an ‘unascertained’ cause of death in death certification may have altered the figures in a very small fraction of cases.
Although this tragic loss of precious lives is the price of precaution without information, there is a more important issue for science as a methodology. The ‘background noise’ of these deaths had obscured a significant number of deaths caused by overlying in those sleeping with their infants. Further advice from the Department of Health has further reduced unexplained deaths in infancy from a cause that any pig farmer would have anticipated.
There are plenty of other exploded certainties relating to both therapies and diet, some documented by Ruth Gilbert and her colleagues’ review of the SIDS issue (Gilbert et al. 2004). These are illustrated by the anti-oxidant story - these compounds are effective in reducing cell damage from oxidative processes in cell and tissue culture studies but not effective in trials of food additives in Man (one trial in smokers had to be abandoned because of an increase in cancer incidence in those treated) and by increasing difficulties with the ‘five-a-day’ fruit and vegetable mantra - intervention studies, often large and of good quality, have failed to show benefits (for carcinomas of the breast and colon, for example).. Here again, inadequate science may obscure the real value of a concept. There are good data on the benefits of some types of diet for populations, but their failure to confirm their promise in trials may be explicable. I have examined a PhD thesis by Dr Barbara Thompson, a food scientist from New Zealand, which may explain why; it is possible that the advice given in relation to fruit and vegetable intake will need modification.
Apart from problems with the precautionary principle and its essentially non-scientific nature, its erratic application is a major difficulty. Why is the precautionary principle applied to GM crops but not organic food? I know of no regular monitoring scheme for mycotoxins in these foods, yet food-related mycotoxin toxicity is a well-established phenomenon and fungicide treatment has been demonstrated to prevent it. A number of papers have demonstrated the consequences of failure to treat (notably well documented in the case of root crops). Why are some ‘natural’ products not subjected to precautionary regulation when we have the REACH initiative? I suspect that we have a mindset problem; it is obvious to some that particular things are dangerous. Professor Edzard Ernst, in his survey of 95 British Complementary and Alternative Medicine (CAM) organisations, found that few understood the concept of adverse reactions and some said that adverse events were only connected with mainstream medicine, but were inconceivable in their own practise. This despite L-tryptophan and the eosinophilia-myalgia story (with many deaths), germanium and selenium and renal toxicity, the loss of a transplanted heart to St John’s Wort and deaths from Asinasin (a new vegetable juice) and from a Chinese herbal dieting regime.
I have concentrated on the problems of bad science. To what extent can these be overcome? It is possible to estimate risks for most interventions and new technologies and thus to design monitoring studies that would enable an initiative to be modified if it results in adverse outcomes. In the first report on risk from the Royal Society (the second is not nearly so good) there was a category of ‘Risks not Foreseen’. There is no system to protect us from these - attempts to do so will be stultifying.
In discussing those who should rule the Republic, Plato was emphatic that the rulers must be an elite. There were three classes of citizens: the Guardians who ruled the city, the Auxiliaries who were guardians but who remained warriors, and the Producers (the rest). In order to ensure that the leadership of the Guardians was accepted by the citizens, a ‘noble lie’ was told about their origins - that they were all born of the same mother but that some had gold in their souls, some silver and some bronze, determining the role they would play in society. This myth was told ‘for the sake of those being ruled’. We are beginning to legislate in this way with regard to risk; regulators are in danger of destroying a trust by assuming they can decide what is good for people without listening to their views. The study of Trewby and his colleagues (2002) shows that we have done this with regard to data about cardio-protective drugs (they do not work as well as patients are led to believe they do). We may be able to give good advice about diet and lifestyles; but only if we have good data.
Sir Colin Berry is emeritus professor of pathology at Queen Mary, University of London.
Blair, P. (2003). ‘Sudden infant death syndrome’. Archives of Disease in Childhood 88: 1031.
Gilbert, R. et al. (2005). ‘Infant sleeping position and the sudden infant death syndrome: systematic review of observational studies and historical review of recommendations from 1940 to 2002’. International Journal of Epidemiology 10: 1-14.
Leach, C. et al. (1999). ‘Epidemiology of SIDS and explained sudden infant deaths’. Pediatrics 104: 43.
'Turn That Racket Off' [Opens in new window]
"The Battle of Ideas is like a huge intellectual fair where a bewildering number of thinkers set out their stalls."
Julian Baggini, editor, The Philosophers' Magazine