I asked regular commenter Tony Ward earlier this month if he would like to write a guest post on lung cancer, and he very kindly agreed, and so I now have great pleasure in publishing his contribution:
Has it all been “a catastrophic and complacent howler” as Professor Sir Ronald Fisher, the father of modern statistics, suggested when writing about the “Alleged Danger of Cigarette Smoking” in the BMJ, back in 1957?
60 years on perhaps it is time to find out.
Fig 1 (see note 1 for sources).
From 1950 onward, medics such as Doll and Hill in the UK, and Wynder and Graham in the USA, ran surveys that found a striking correlation between smoking and lung cancer.
They became convinced that lung cancer was a preventable disease and that the way to prevent it was to stop people from smoking. Their theory was that lung cancer was caused by smoking tobacco (particularly cigarettes) over decades. Typically stated as a time delay of thirty years or more.
Politicians, initially in the UK, were persuaded that smoking tobacco was the cause of lung cancer and began a crusade against smoking and smokers. It’s a crusade that everybody is all too familiar with. From health warnings in the 1960s complete with TV, Radio and Newspaper campaigns, to advertising bans beginning in the 1970s and completing in 2001, smoking bans on private property by 2007 and medical pornography on packaging soon after, the message about smoking causing lung cancer has been ubiquitous. It has been ramped up year by year and decade by decade. In the 21st century, denying that smoking is the cause of lung cancer became the ultimate heresy.
The crusade initially spread to the US and then to other western countries. Over time it has become a global industry with an income estimated at up to $15 billion annually. It’s been hailed as a great medical success story and regular press releases get printed about the number of lives ‘saved’.
The objective of the crusade was to save lives by dramatically reducing or even eliminating the disease of lung cancer. But because of the time delay, the prevalence of smoking had to be used as a proxy measurement of success or failure.
In 1948, an estimated 63% of UK men and 40% of women, smoked cigarettes (note 2). The crusade dramatically reduced smoking rates and by 2016 the figures for men and women were roughly equal at about 20% (or even less according to anti-smoking claims).
Cigarettes were an estimated four times stronger in the past than today and were smoked almost everywhere. Almost everyone was exposed to tobacco smoke, almost everywhere. The method of smoking was different too. Andy Capp style, cigarette constantly between lips, rather than the more restrained approach common these days.
So on the basis of the proxy measure of smoking rates, the crusade has certainly been a huge success.
But more than half a century on, it is time to examine the success or failure in terms of the true objective. The obvious place to begin is the place where the crusade originated. So has the incidence of lung cancer in England declined in step with smoking prevalence or even at all?
The chart at the top (Fig 1) shows the very latest figures from the official UK Office of National Statistics (ONS).
Reductions in male lung cancer have been relatively small, at 14%, compared to the three fold reduction in smoking.
What is very clear though is that not only has female incidence not fallen in line with the halving of the smoking rate, it has grown enormously. Specifically, an increase of 226%.
To be fair, the population has increased (by 20% – see note 5) over the period and the age distribution has altered with more people in the cancer prone age groups. Fortunately the UK ONS provide age and population adjusted figures.
These figures are only marginally less shocking, showing male cases down by 43% and female cases up by 151%.
Fig 3 (note 3)
To repeat. Male cases down 43%, female cases up 151%.
This flatly contradicts the smoking/lung cancer theory.
Just to clarify, they didn’t say ‘reduce the smoking rate by a factor of 3 and we predict a small reduction in lung cancer’. No. They claimed, typically, that 90% of lung cancer was caused by smoking. And they certainly didn’t say ‘cut smoking rates in half and we reckon lung cancer incidence will increase by 150%’.
It is tempting to suppose that the female rate will stop rising once it matches the male one but that is wishful thinking. The male rate has followed a fairly classic pattern of disease with a gentle rise, a peak and, so far, a slow decline. The female rate appears to have lagged behind. Nobody knows where it will peak or when, if ever, it will decline.
The smoking/lung cancer experiment has run for more than half a century and the results are now in. As Richard Feynman put it:
“It doesn’t matter how beautiful your theory is, it doesn’t matter how smart you are. If it doesn’t agree with experiment, it’s wrong.”
I believe Fisher has been proved absolutely correct and it is to the everlasting shame of the medical establishment that his warnings were shouted down.
Back to Fisher for a moment :
“Your annotation on “Dangers of Cigarette-smoking” leads up to the demand that these hazards “must be brought home to the public by all the modern devices of publicity”. That is just what some of us with research interests are afraid of.”
I suspect he was, correctly, predicting that huge damage would be done to the credibility of medical research, statistical analysis and even to science and mathematics.
It has certainly skewed and probably hindered research into the real cause of lung cancer and potential treatments:
“Lung Cancer Causes 32% of All Cancer Deaths. Why Does It Get 10% of Research Funding?” (Fortune)
But much other harm has resulted too. Families have been broken up, friendships damaged beyond repair, businesses trashed and whole communities destroyed.
All in an utterly futile crusade.
The worry now is that there is such a large industry with attached gravy train that it will be difficult to put it out of everyone’s misery.
Ways in which the anti-smoking industry will handle this issue.
1. Ignore it and hope it goes away.
2. Threaten any news outlet or politician that mentions it. Bear in mind that their billion dollar annual income is purely for advertising and lobbying. The industry does not produce any tangible goods. They also have controlling influence over medical budgets both private and public, and can lean on Government departments to support their boycotts. Boycotts that can destroy any media outlets that dare to dissent.
3. They might try to claim that what matters is the mortality rate (see note 4) rather than disease incidence.
Male lung cancer mortality peaked in 1985 with a slight fall to 1,231 in 1994 followed by a more substantial fall to 694 in 2016.
Female mortality rose and then flattened with 456 cases in 1994 and 474 in 2016.
You could argue that even these figures are fairly damning for the anti-smoking industry although the male rates look encouraging.
But the anti-smoking smoking claim has never been about lung cancer treatment or survival rates. They’ve always claimed that smoking causes lung cancer. They’ve never, as far as I’m aware, claimed that non-smokers survive lung cancer whereas smokers don’t. So what matters for the crusade is the number of cases diagnosed. Here are the latest mortality figurers for England and Wales just for interest.
Fig 4 (note 4)
4. They may claim that lung cancer is only one of many diseases caused by smoking tobacco. Heart disease being the main one. The studies/surveys mentioned above found a correlation (relative risk) of up to around 1.7 between smoking and heart disease. (compared to ~20 for lung cancer). Perhaps surprisingly, this is regarded as very weak and epidemiologists generally advise that relative risks of less than 2 or 3 are not generally worth reporting on let alone making causality assumptions about. Public propaganda about smoking and heart disease only starting appearing in the late 1970s. Since then they’ve been making all manner of ridiculous claims about smoking causing everything under the sun. All without any justification.
But it was always lung cancer that was the only credible, all important and overriding claim.
Why didn’t anyone sound warnings that the smoking causing lung cancer claim was at least potentially unsafe?
The answer is that they did:
Although not as damning as these latest figures from England, this is not the first time that data has cast doubt on the smoking/lung cancer theory. Many articles have shown mismatches between observed and expected lung cancer rates around the world.
And here’s an earlier one that sounded the warning about English rates back in 2012:
What’s more, none of this should have been a surprise. Several serious, large and hugely expensive scientific studies have been run to test whether anti-smoking interventions were beneficial for health or not. All of them found no health benefit despite great success in reducing the proxy of smoking prevalence. These results were published over the course of the 1980s although the final, final report on the most important one wasn’t published until the 1990s
Dr Johnstone and Professor Finch wrote about this at length:
Several eminent scientists have also put up very strong, reasoned objections. All were shouted down.
Professor, Sir Ronald Aylmer Fisher FRS – aka the ‘Father of Modern Statistics’
Professor Philip Burch
Professor Hans Eysenck
Professor Phillipe Even (on so called ‘passive smoking’)
What about court cases and justice?
(Note that the “Master Settlement Agreement”(MSA) was an agreement, and a profoundly corrupt one at that, rather than a court case. I won’t attempt to cover it here.)
Over the years, the anti-smokers have won a few court cases where judges accepted the anti-smoking claims at the outset, although these were generally overturned on appeal. Whereas it seems they always withdrew their legal action in other cases whenever the defendant was permitted to mount a defence.
However, there has been one case, possibly the only one in the world, where the anti-smokers tried to make their case in a court of law.
The McTear vs ITL court case has a wealth of information about the key players and gives an insight into the nature of the debate or lack of one. The anti-smoking industry was annihilated to put it mildly.
1. Unadjusted Cancer registrations
Most of the data was copied from Table 1 of the spreadsheets linked to here:
For 95-97, google “Registrations of cancer diagnosed in 1995–1997, England”
2. Smoking Prevalence
Here are some ONS figures for comparison:
And some further figures from Doll’s Doctors study:
The 1951, an estimated male cigarette smoking rate of 73% derived from here:
12.3% non-smokers in the population giving 87.3% smokers (from Table I). Note that that figure only applies to subjects aged 35 and over and includes an unknown number of ex-smokers.
Of the smokers, 83.3% smoked cigarettes (Table II). So cigarette smoker prevalence in the study population was 83.3% of 87.3% =72.7%
And a 1951, 50% female rate here.
3. Adjusted Cancer registration figures
3a.These are copied from two datasets covering 1971-2015.
Table 2 in ONS (.xls)
Table 11 in ONS (.xls)
Both these datasets can also be obtained via the link in Note 1
(3b.There is a second sequence of adjusted figures from 1986-2011 which appears to have been calibrated with a (male) starting value of 100 (see spreadsheets linked by note 1). There appear to be some slight inconsistencies in this sequence and it is far shorter. It is broadly in line though with: Male –44% Female +25% )
I chose to use sequence 3a which is far longer, up to date and is consistently age standardised based on 2013 EU stats.
4. Mortality Figures
Table 5 in ONS (.xls)
This dataset can also be obtained via ONS:
5. Population Data