Frank Davis

Banging on about the Smoking Ban

Previous Entry Share Next Entry
In classic Sherlock Holmes detective stories, a crime is committed, and the detective sieves through the evidence, and eventually, after following a number of false leads, identifies the culprit. But there's another sort of detective story in which the culprit is known from the outset, and the job of the detective is to piece together the evidence to show how the culprit did it. The crime fiction TV series Columbo is an example of this latter genre.

The show popularized the inverted detective story format; almost every episode began by showing the commission of the crime and its perpetrator. Thus there is no "whodunit" element. The plot mainly revolves around how the perpetrator, whose guilt is known, would finally be exposed and arrested. The show's creator once referred to it as a "howcatchem".

In a typical Columbo episode, the audience knows who committed the murder, but so in a sense Lieutenant Columbo also knows, always homing in unerringly on the killer almost from the moment he sets eyes upon them.

In real life detective work, it's almost always the classic format that is followed. The detectives don't know who the culprit is at the outset. But this isn't always the case. Sometimes the police have a prime suspect right from the outset, and will devote most of their efforts to building a case against the one suspect. Occasionally, so convinced are they of his guilt, that they will even conceal or fabricate evidence in order to secure a conviction, and end up framing someone for a crime they did not commit.

Exactly the same applies in epidemiology. In classic epidemiological studies, there is an outbreak of some disease, and a search is begun for the unknown cause of the disease, and very often it is finally identified, after years of patient study, as a virus or a bacterium.  But there is also an inverted form of epidemiology, just like there is an inverted form of detective work. And in this inverted form of epidemiology, the cause is believed to be known from the outset, and the epidemiologist's task is to link this to some particular disease.

Most epidemiology follows the classic format of starting with the disease and gradually isolating its cause. It was a method first employed by the London physician John Snow when he discovered that almost all the victims of an outbreak of cholera in 1854 had drunk water from a single public water pump. When Snow removed the pump handle, the epidemic came to an end.

But there is a case to be made that antismoking epidemiology adopts the inverse method. It starts with a single cause - tobacco - and sets out to link it to a variety of diseases. Which is exactly like the police arresting someone, and then setting out to prove that they carried out some number of crimes.

The London Hospitals Study

In Britain, the first such study was the Doll and Hill London Hospitals study, which was intended to find a cause for the rising epidemic of lung cancer. A set of lung cancer patients was asked to fill in a three page questionnaire, and a control group of hospital patients without lung cancer were asked to fill in the same questionnaire. Richard Doll's Times obituary continues the story:

Doll thought that the increasing incidence of the disease might owe something to the hundreds of tonnes of tarmac being laid down across Britain at this time, but soon discovered that in 649 lung cancer cases there were only two non-smokers. Doll himself gave up the habit two thirds of the way through the research.

Also part way through the study, Richard Doll wrote to Dr Green of the MRC to say:

'Incidentally the investigation has gone much better than I expected and it looks as if smoking will be incriminated to a major extent!'

By the end of the study, both Richard Doll and Bradford Hill were of one mind:

we were able to reach the conclusion, without as it happened any firm laboratory evidence, that 'cigarette smoking is a cause and an important cause of carcinoma of the bronchus'

On the face of it all this appears perfectly plausible. But there are some disturbing questions. Why did Richard Doll give up smoking before the study was completed? Why was it that he seemed pleased that the study looked set to incriminate smoking?

While it's true that nearly all of the lung cancer patients were smokers, it also happens to be true that nearly all of the control group were smokers as well. Pretty well everybody in this study smoked. And so it's hardly surprising that nearly all the lung cancer patients were smokers. Far from showing that smoking was a significant cause of lung cancer, the study could equally be construed to have shown that smoking had almost no effect at all.

In addition the eminent statistician Sir Ronald Fisher, who was able to eventually obtain the raw figures from the study, was able to use them to show that patients who inhaled tobacco smoke into their lungs were at significantly less risk of contracting lung cancer than non-inhalers, and that therefore inhaling tobacco smoke had a protective effect.

And what happened to the other questions in the questionnaire, unrelated to smoking? What were the questions? What did Doll and Hill discover from the responses to them - for example about exposure to tarmac, or proximity to gas works? We will never know. The final paper only dealt with smoking. There was supposed to be a further paper dealing with the other responses, but it was never written. And so the upshot was that, in effect, the cancer patients were only asked about their smoking habits. Whatever other habits or characteristics they may have shared were disregarded or discarded. Or, at least, nothing was ever published about them, which amounts to the same thing.

The British Doctors Study

The following year, in 1951, Doll and Hill began the British Doctors study, sending out a questionnaire to Britain's doctors. And this time, rather than three pages of questions, there was just one question: were the doctors smokers, ex-smokers, or had they never smoked at all? If they were smokers or ex-smokers there were asked some additional questions about the amount they smoked. When the doctors eventually died, the researchers would be notified of their cause of death, and over a period of years data would accumulate.

One has to ask at this point what the purpose of this new study was. The London Hospitals study had (ostensibly) looked at various possible causes of the lung cancer epidemic. Clearly the new study wasn't going over the same ground as the previous one. And perhaps the authors were no longer particularly interested in what caused lung cancer, because they believed the previous study had shown smoking to be the culprit. The purpose of the new study was to strengthen the link between smoking and lung cancer, but also to see what other diseases might be linked to smoking. It was a fishing expedition, to see what else smoking might cause. Not only lung cancer was considered, but all other sorts of cancers, respiratory diseases, coronary thrombosis, cardiovascular disease, and every other kind of disease. In 1954, when the first results were published, smokers were shown to be at slightly greater risk in more or less every single disease category. But once again, as with the London Hospital patients there was a very high percentage of smokers among British doctors. 87% of the doctors that responded to the questionnaire were smokers.

Unlike the London Hospitals study, the British Doctors study ran on for another 50 years until 2004, issuing updates every 5 or 10 years. If nothing else, this kept it in the news for 50 years. And perhaps this was the entire purpose of the study. In a valedictory letter in 2004, Sir Richard Doll wrote that:

it was devised by Sir Austin Bradford Hill to achieve maximum publicity for the critical relationship between smoking and lung cancer.

Really? Is that all? The entire 50-year British Doctors study was nothing but a publicity stunt? It had never been intended to be a serious investigation at all. And as a publicity stunt it had been a tremendous success. Like a mighty siege gun firing year after year, the periodic reports that were issued served to link smoking and lung cancer for year after year, and decade after decade. And the same reports also served to link smoking with all other diseases as well.

On this basis, the only substantive research that was done was in the London Hospitals study. This is the really important study which showed that 'cigarette smoking is a cause of lung cancer'. The subsequent British Doctors study was designed to publicise this finding.

But why were Doll and Hill so readily convinced by the single London Hospitals study that smoking was a cause of lung cancer? The answer perhaps is that, quite fortuitously, several other studies in the same year, 1950, had drawn much the same conclusion. But there is also another explanation. And this was that a considerable number of epidemiological studies had been published in Nazi Germany prior to and during World War II.

The Nazi Precursors

Research linking smoking and lung cancer is usually regarded as having started in 1950 with Doll and Hill in the UK, and Wynder and Graham in the USA. In reality, it had been going on for 20 years beforehand in Nazi Germany. [1][2][3][4][5][6][7][8][9][10] Were the post-war American and British researchers aware of the earlier German research? Clearly Richard Doll knew about them, because he mentions one German 1939 study (Muller 1939) in his list of references in the London Hospitals study. In addition, as a medical student, Richard Doll had attended lectures by SS radiologist Hans Holfelder in 1936. Ernst Wynder's links to Nazi Germany were even closer. Also a doctor like Doll, he grew up in Nazi Germany, and emigrated to the USA shortly before WWII.

The Nazi attitude to tobacco and smoking followed the inverted epidemiological model. Hitler certainly possessed a firm conviction that tobacco was an evil. The Scientific Institute for the Research into the Hazards of Tobacco opened in April 1941 with a message of support from Hitler, who sent 'best wishes for your work which will liberate mankind from one of its most dangerous poisons’, and a personal donation of over 100,000 RM. The very title of this institute indicates that its purpose was to discover hazards in tobacco, which was already known to be a 'dangerous poison'. It was just a matter of finding out what ill effects the poison had. This is exactly the same as the sort of case where the police first arrest the culprit, and then set out to discover, Columbo-style, how he did it.

Exactly the same kind of inverted reasoning informed Nazi racial science. Jews and Gypsies and the like were regarded from the outset as a 'racial poison', and policy flowed from this conclusion. Nazi racial science set out to discover in what ways Jews were so dangerous, and find exactly how the Jewish World Conspiracy fiendishly plotted to take control of the world. The Nazis had the culprit, and it was just a matter of showing how he had committed the crime. In the end, they simply executed him anyway.

The historian Robert Proctor has argued that not all Nazi science was like its indefensible racial science. He cites the V2 rocket as an example of 'good' Nazi science. But as I've shown elsewhere, the V2 rocket was a technological innovation whose development proceeded very largely despite the Nazi regime, and despite Hitler's indifference. The hallmark of Nazi science was that it proceeded along an inverted path of reasoning, starting from a conclusion and working backwards. Thus it was concluded at the outset that the Jews were a 'racial poison'. In the same manner, tobacco was also identified from the outset as a 'dangerous poison'. There was no equivalent fixation upon rockets as a potentially war-winning weapon: Hitler only came to recognise their potential quite late in the war.

The Inverted World

Another characteristic of inverted reasoning of this sort is that it always rapidly jumps to a conclusion. While normal, classical scientists are still filled with doubts and uncertainties, inverted scientists race ahead with rapid 'breakthroughs'. And the result is that inverted science always gains a head start over normal, slow, plodding science. It may take decades or centuries for classical, rational science to catch up. In the meanwhile, the dogmas of inverted science may become established conventional wisdom, which 'everybody knows' and about which 'the debate is over'. Lieutenant Columbo solves the murder mystery in days, but Sherlock Holmes may refute him a year later. By which time, the wrong man may have been tried and executed for the crime.

A further characteristic of inverted reasoning is that, once the culprit has been identified, there is little or no further progress. In the case of classical epidemiology, such as John Snow's cholera study, the culprit is narrowed down to the Broad Street pump. Further studies would narrow down the culprit further to a bacterium - Vibrio cholerae -. Classical epidemiology brings a steady sharpening of focus. In time, the culprit may well be narrowed down to a single strand of DNA inside that bacterium.

Inverted epidemiology, by contrast, goes in the opposite direction, becoming ever less focused. The culprit becomes ever more diffuse and amorphous. Antismoking epidemiology started out by fingering cigarette smoking as the culprit behind the lung cancer epidemic. The focus never narrowed any further. Instead within a decade or two it had widened to include all tobacco products. And a decade or two later, secondhand environmental tobacco smoke also came to be regarded as a significant threat. Now even thirdhand tobacco residues on clothes and furniture is being postulated as a further threat. Tobacco smoke has become an ubiquitous menace, growing larger and larger with every passing decade of widening pandemonium.Yet no firm laboratory evidence has ever showed how smoking might cause lung cancer.

And at the same time, the quality of the research gets worse and worse. At the outset, it has the appearance of scholarly rigour. But gradually the rigour melts away, and the conclusions drawn become as vaporous as the threat. Research results come out as press releases before the studies have been completed. In the case of thirdhand smoke, there isn't even any research at all. The threat of thirdhand smoke is merely a rumour. By contrast, classical scientific research uses ever better tools and techniques as its focus narrows.

Lastly, inverted epidemiology ends up demonising people. In the case of Nazi racial science, it ended up demonising and murdering entire peoples. In the case of antismoking science, it has now begun to demonise smokers, marginalising and excluding them, firing them from jobs, evicting them from houses, refusing them medical treatment. And this sort of demonisation is inherent in an inverted methodology which starts out by identifying something or other as an 'evil' or a 'poison'. In time, as the threat becomes ever greater, and ever more ubiquitous, harsh measures have to be adopted. Witches are burned. Jews are gassed.

Inverted science is also superficially attractive. It offers quick results. Its researchers display a plausible certainty about their results. It very often brings its practitioners considerable wealth and status, as funds pour in.

Research into lung cancer in Germany during the Nazi era very rapidly fixed upon smoking as a likely cause. These studies largely employed retrospective questionnaires of lung cancer patients. After the war, research in the UK and USA adopted much the same methodology, also focused on smoking, and made the same link between smoking and lung cancer. It's not entirely clear what links there were between the pre-war Nazi research and the post-war British and American research. It's clear that there were some links. The result was that what had begun as Nazi research in the pre-war era continued uninterrupted in the post-war era, shorn of its overtly Nazi overtones. Most importantly, the focus remained fixed invariantly upon tobacco. There were hardly any further successful developments in the field of lung cancer. The principal cause of lung cancer in 2000 remained exactly what it had been in 1950 - smoking. Throughout this period, lung cancer deaths continued to mount, even though the popularity of smoking plummeted. Smoking came to be regarded as more and more dangerous, as fewer and fewer smokers apparently caused more and more lung cancer.

Recently, however, there are some signs that classical epidemiological science may at last be beginning to catch up, as Human Papilloma Virus (HPV) has been identified as the prime cause of cervical cancer, and HPV has been discovered present in 25% - 80% of lung cancers as well. It may be that Sherlock Holmes is catching up with Lieutenant Columbo.

  • 1
I heard someone say, some while ago, that if Hitler hadn’t given up smoking the very worst atrocities of the Nazi regime might never have taken place. And given the spiteful, venomous and utterly hateful comments of many ex-smokers, both public figures and private individuals, I can’t help thinking that there just might be a grain of truth in that statement ……


Here is the quote you are looking for. Cut and paste that you can shove it in the faces of the anti-smoking zealots with the greatest of pleasure!

“I am convinced that if I had been a smoker, I never would have been able
to bear the cares and anxieties which have been a burden to me for so long.
Perhaps the German people owe their salvation to that fact.” Circa 1942

Re: Columbo

Wow! Many thanks for that. I never realised it was The Man Himself who gave rise to the idea. What a coup! Yey! I shall keep it safely, and also commit it to memory to quote at the next "comparing-us-to-the-Nazis-is-totally-unreasonable" anti that I come across!!

"But there's another sort of detective story in which the culprit is known from the outset..." If it's of interest to any one, the literary term for this is 'dramatic irony'; when the audience has some knowledge that the characters don't possess.

During the 19th century, scientists studied the volume of human skulls by stuffing them with either lead shot or seed. This was done to demonstrate that whites had larger brains than people of other races. Of course, it was later discovered that these studies were full of all kinds of biases; skulls were thrown out of being too large or small, data was thrown out that didn't conform to the preconceived conclusion, and the scientists themselves forced more shot into the skulls of deceased whites.

This is detailed in Stephen J. Gould's book "The Mismeasure of Man". Anyway, I bring this up because Gould goes to pains to point out that he doesn't believe that the scientists of the time were engaging in any kind of fraud. Rather, he concludes that the scientists were so convinced of what the outcome should be that they simply dismissed anything that didn't conform to their expectations.

The research was only being done to prove that whites had bigger brains. It wasn't being done to prove that whites were superior, because, to the scientists, that was already known. What mattered to them was that their findings on brain size would conform an already established expectation.

Without explaining at greater length, I hope you see the parallels to what you've written above. Social perceptions and expectations can have an insidious effect on scientific research. WS.

I've not read the Mismeasure of Man (although I've read a number of his books), and I've not heard that story, but it's essentially the same thing: people believe they know the answer already. I agree that it's not deliberate fraud.

It may be that this sort of process is quite natural. We all think we 'know' how everything works. And most of this 'knowledge' is unquestioned. It is what 'everybody knows'. Learning anything first entails unlearning this 'knowledge'. But before it can be unlearned, a stout defence of that conventional wisdom first takes place, before it is surrendered to some newer and better hypothesis.

On the brain size question, it's usually supposed that the bigger the brain, the greater the intelligence. And yet, as our 'thinking machines' - computers - get more powerful, they get smaller and smaller. The laptop on which I'm writing this is far more powerful than a computer of 50 years ago that filled several rooms. That's led me to wonder whether it's those creatures with the smallest brains which are actually the smartest, and houseflies are geniuses, and humans are the most stupid creatures on this planet.

As for Columbo, it was always my impression, somehow or other, that Columbo also knew what the audience knew. He just could not come out and say, "I know you murdered him", but had instead to piece together just how he'd done it, to make a case in a court whose officers did not have this superior knowledge. This was never stated, but there always seemed to be a peculiar light in Columbo's eyes when he first met the culprit. But I've not seen all episodes of Columbo, and there may well have been one where he got it wrong, and another detective was able to show that he'd got the wrong man.

"That's led me to wonder whether it's those creatures with the smallest brains which are actually the smartest, and houseflies are geniuses, and humans are the most stupid creatures on this planet."

I was told in college that cetaceans have larger brains than humans, with more convolutions in their brains, and a greater portion of their brains are comprised of cerebral cortex. Make of that what you will, I guess. LOL. WS.

Sir Richard Doll - "it was devised by Sir Austin Bradford Hill to achieve maximum publicity for the critical relationship between smoking and lung cancer."
Sir Richard Doll from the McTear case - "Sir Richard said that Professor Bradford Hill, who taught him epidemiology, taught him that the scientist's responsibility was always to try to disprove his own work, not to try to look for evidence in support of it. He had the idea that if they learnt about the smoking habits of British doctors and divided them according to the amount smoked, they would see if they could disprove the relationship between cigarette smoking and lung cancer by following up the non-smokers and the heavy smokers. Of course, he said, the opposite happened: it greatly strengthened the conclusion."

So was it devised to "disprove" the London Study or devised "to achieve maximum publicity"? One would think that disproving should be the priority over publicity!

Fredrik Eich

Well, in the McTear case Sir Richard Doll was up in front of a judge, and attempting to portray himself as a responsible scientist.

The valedictory letter, however, was for the eyes of his co-workers over 50 years in the British Doctors study. The message for them was quite different.

I can't see in what way the British Doctors study can be seen as an attempt to disprove the link between smoking and lung cancer. It was instead a fishing expedition that set out to link smoking with many other illnesses in addition to lung cancer. No other possible cause of lung cancer was considered other than smoking. No other cause of lung cancer other than smoking was seriously considered in the prior London Hospitals study, given that they didn't even bother to write a paper showing why other possible causes (tarmac, coal gas) should be excluded.

The British Doctors study makes much more sense as a publicity device. It kept the link between smoking and lung cancer in the news for 50 years, as one report after another, saying much the same thing, rolled off the printing presses.

And when did Richard Doll ever manage to disprove any of his own hypotheses, whether they were about smoking or anything else? He was, throughout his life, a medical 'authority' (which is why he was wheeled into the McTear case), whose authoritative opinion about everything and anything was much sought after, and for which he was very handsomely paid.

Clearly the London Study is just measuring the number of smokers using lung cancer as a unit of measure! If smoking was a drug administered in the hospitals and used by a tiny percentage of the population then that study shows smoking probably causes lung cancer. But as ~70% of men and ~50% of women smoked at the time, the only thing that really stands out is the unexpectedly low number of lung cancers for never smokers. But seeing as the main point of the study is to answer the question; what has caused this rise in lung cancer up to 1950? And that the theory is that it is the rapid rise of smoking cigarettes that is responsible for the massive rise in lung cancer up to 1950 and not the smoking of pipes and cigars. Then one can, in the London Study, add the ~30 (5.7%) odd lung cancer cases in pipe smokers only and the ~70 (9.7%) controls to the non-cigarette users and the difference is about the same as for the women (clearly not many women smoked pipes!). So I think what is most powerful about that study is the graph showing the rise of tobacco use after about 1910 (which we assume to be mainly a rise in cigarette use) and the rise in lung cancer. If the theory goes that the rise is due to the advent of cigarettes then the Fig 1 should be comparing cigarette users to non-cigarette users in the context of explaining the rise! I remember Doll saying that something to the effect that if more doctors had understood the science that his message would have got through quicker but I sometime wonder if the reverse is true, they did not really understand it but believed it any way. After all most have them had given up smoking ~5 years later, one would think on the strength of studies such as the London Study!!

  • 1

Log in

No account? Create an account