Chris Snowdon reports on the inexplicable failure of Maltese heart attack rates to fall after their smoking ban was introduced:
Malta smoking ban fails to stub out heart disease
We are shocked over smoking ban results
The smoking ban may have led to a decrease in the heart disease death rate and hospital admissions in every country where it was introduced, but it has had absolutely no impact in Malta, a medical study has revealed
I’m sure that some ad hoc explanation will be found for it. Perhaps that Malta is a small island on which only two or three people live, and so can hardly be expected to replicate results found elsewhere among higher populations.
The other explanation, of course, is that all the other results have been, as Chris Snowdon suggests, the product of cherry-picking, and that the Maltese result is actually the only accurate one. For the others were studies designed to generate headlines trumpeting the success of smoking bans. They were press release science, intended to be reproduced uncritically in the mass media, which of course they duly were. They have since become established scientific ‘facts’. Indeed ASH’s Deborah Arnott was repeating the claims on radio only a week or two back.
It is seldom remarked that the purpose of the famous British Doctors study that began in 1951, and ran for 50 years, was also to generate headlines – as a brief announcement in the BMJ a few years ago revealed:
Doctors’ smoking cohort study ends
Martin Gaba – BMJ. 2001 December 1; 323(7324): 1270.
After running for 50 years, the cohort study on the smoking habits of 40,000 British doctors, which helped to established the link between smoking and lung cancer, has ended with a valedictory thank you letter to the surviving doctors who were recruited in 1951.
Professor Sir Richard Doll, emeritus professor of medicine at Oxford University, who wrote the letter and was involved in the study since its inception, said it was devised by Sir Austin Bradford Hill to achieve maximum publicity for the critical relationship between smoking and lung cancer. The link was first established by him in a trial in 1947-49, but “rejected by the Department of Health cancer committee,” and “not believed by a public, in which 80% of men smoked.”
Dr Maurice Gaba, 89, a former GP in Glasgow, was one of those recruited at the beginning of the study. Dr Gaba, who now lives in Jerusalem, said: “I was a forty a day man, when I received a letter in 1951 from a professor asking me about my smoking history, ending with a request to view my death certificate. I thought this doctor cares more about my health than I do, and I have never smoked since.”
So there you have it, straight from the horse’s mouth: The study was devised to achieve maximum publicity for the link between smoking and lung cancer.
And very successful it was too. The long-running study released regular reports every 5 or 10 years for over 50 years, garnering publicity every time, in a drumbeat that slowly hammered into everybody’s head the ‘link’ between smoking and lung cancer, with the ‘link’ always being supposed to be causal.
The brief article also reveals that the ‘link’ being publicised was found in an earlier study by the same Doll and Hill duo. This was the London Hospitals study which was published in 1950, in which it was found that over 99% of lung cancer cases were found among smokers, with the inference that the one caused the other. From Richard Doll’s Telegraph obituary:
The report, published in the British Medical Journal in 1950, was based on a survey of lung cancer patients in 20 London hospitals; it found that smoking was the only factor overwhelmingly implicated in lung cancer, and that it was rare for non-smokers to suffer from the disease.
BBC obituary of Richard Doll:
His seminal 1950 study, which he wrote with Austin Bradford Hill, said smoking was “a major cause” of lung cancer.
What isn’t quite so often mentioned is that in this study pretty well all the patients, in both the control group and the study group, were smokers (the more precise figure is 98% of them), and that even if smoking had no causal link whatsoever with lung cancer, 98% of lung cancer patients would have still been found to be smokers. Much in the same way that quite possibly 98% of the same lung cancer patients would also have been found to be Londoners. Or tea drinkers.
It was objections of this sort that led to the study being rejected, and widely disbelieved. But Richard Doll remained a true believer. As did a number of other doctors, including George Godber, who was to help launch the passive smoking scare in 1975. And the subsequent British Doctors study (in which 87% of the doctors were smokers) rammed their message home over the next 50 years until absolutely everybody knew that smoking caused lung cancer, and nobody questioned it any more – if only because most of the early 1950s critics had died in the intervening years.
Bradford Hill, who was co-author with Richard Doll in both the London Hospital study and the subsequent British Doctors study, devised the Bradford-Hill Criteria
…otherwise known as Hill’s criteria for causation, are a group of minimal conditions necessary to provide adequate evidence of a causal relationship between an incidence and a consequence,
the gist of which are that if some consequence follows on from some incidence of something, with sufficient regularity and plausibility, then the one must be the cause of the other. The causal link between smoking and lung cancer is perhaps the most famous example of the application of the Bradford-Hill Criteria. Millions of people started smoking cigarettes in the early 20th century, and a few years later an epidemic of lung cancer started. Cause and effect.
But it’s not the only example. For the same causal claim is being made about the link between the rise in carbon dioxide in the atmosphere and the subsequent rise of global temperatures. There seemed to be no other obvious explanation for the rise in temperature, even if the exact mechanism wasn’t fully understood (just as the exact mechanism by which smoking causes lung cancer has never been understood).
The problem of causality is a philosophical problem of some antiquity, and was discussed by, among others, David Hume in the 18th century. It is also closely related to the Problem of Induction, of deriving general laws from particular cases. It might be said that the dispute between the inductivists (as exemplified by Doll and Hill, and by global warming alarmists) and sceptical deductivists is the continuation of a very long-running debate. The former believe that scientific laws can be generated by statistical analysis of data to show causal links between one event and another. The latter tend to believe that causation can only be demonstrated by showing deductively how one event necessarily leads to another.
Malaria, for example, was inductively linked to malodorous swamps, and ‘bad air’ was for a long time regarded as the cause of ‘malaria’. Subsequently, it was found that swamp-breeding mosquitoes carrying the Plasmodium parasite introduced it into humans they bit, giving rise to the symptoms of malaria when the parasites subsequently multiplied. In this manner an entire chain of cause and effect was deductively established, showing that neither swamps nor ‘bad air’ were the true cause of malaria, and allowing a variety of interventions (insecticides, net curtains, quinine, etc, at various stages in the chain.)
It might also be said that whenever some scientific law has only been inductively established, it entails an element of belief or ‘faith’, quite simply because the deductive path that ties cause to effect has not been established. And therefore, in the case of global warming, ‘deductivist’ sceptics want to know exactly how carbon dioxide in the atmosphere can cause global temperatures to rise to the extent forecast by alarmists, and will not be fully persuaded until they have had all the steps described in detail. And this may also be why ‘deductivists’ are also more readily persuaded by hypotheses such as those of Henrik Svensmark, with his detailed account of how a surfeit or deficit of cosmic rays striking the atmosphere might increase or decrease terrestrial cloud, and cause either cooling as more clouds reflect more sunlight, or warming as fewer clouds reflect less sunlight.
Another way of thinking about it might be to regard inductive science as the kind of preliminary science by which general laws are first empirically derived, and deductive science as the kind of science by which such laws are subsequently understood. Thus an example of an inductive law might be the gas law whereby the pressure of a gas is experimentally found to vary linearly with its temperature, with nobody knowing quite why this should be so. But subsequently other scientists were able to explain mathematically how particles of gas bouncing around inside a chamber would exert a higher pressure on its walls the faster they were moving (i.e. the hotter they were). These theoretical physicists were able to deduce the gas laws that their predecessors had discovered by experimentation.
Nevertheless the argument between inductivists and deductivists looks set to run and run.