I’ve come across a very well-written and incisive PDF…
Don’t be put off by the title (‘The Plain Truth About Tobacco’). It’s not about plain packaging and it doesn’t argue that smoking is not the main cause of lung cancer (in fact, any remaining deniers should be forced to read it),..
As I remarked in my reply to Rich, that was rather throwing down the gauntlet to the denialista likes of him and me. So, frogmarched to the paper by Chris Snowdon, and told to read the effing thing and wise up, I have indeed been forced to read it.
At 150 pages, it’s quite long. And since I only started reading it this evening, I’ve only got about a third of the way through it.
The first thing I’d like to say is that I entirely agree with Chris that it’s very well written. And I’d like to encourage my readers to read it too. Although I can’t say that I think they should be forced to read it. Nor would I even say that they ought to read it. I’d just like to say that it’s very well written, and it explains a number of things very clearly, in ways that might make it reference work for clear explanations of things.
I was sort of expecting, given Chris’ introduction to it, that it would start out saying on page one that Smoking Causes Lung Cancer, and Anyone Who Disagrees is a Pinhead. But it’s very far from that. In fact among the first things he says is:
The link between cigarette smoking and lung cancer is often cited as the strongest case in the vast research on factors influencing cancer risk amongst the population at large and this is a fair assessment. However, what has typically been made of that link by medical and public health practitioners and institutions, is absolutely and terribly flawed.
Now saying that smoking is linked with lung cancer isn’t the same as saying that smoking causes lung cancer. All sorts of things are ‘linked’. I could say that I’m ‘linked’ to Lana Del Rey. Because I have lots of links to her videos and articles about her. But it would probably be stretching things a bit to suggest that I am in any way the cause of her subsequent overnight global superstardom. It’s a bit more multifactorial than that, I suspect. Although if I get an email from her saying, “It was all thanks to you, Frank,” I may have to revise my judgement.
But then, in fairness to Chris Snowdon, all he said was that the author “doesn’t argue that smoking is not the main cause of lung cancer.” And this would indeed appear to be the case. But, as far as my reading has got, I’m going to be astonished if he does argue that smoking is the main cause of lung cancer.
If nothing else, he might be said to have introduced another hypothesis: that Irritation Causes Lung Cancer.
Generally, in terms of common sense:
To anyone who smokes cigarettes considerably, irritation of the lungs over time is self evident. If a person smokes heavily he will come to recognize his “smoker’s cough”. The cough is due primarily to irritation of the breathing passages of the lungs, and it is particularly cancers of these breathing passages, to which smokers are at heightened risk.
Though not really surprising in itself, that irritated tissues are more vulnerable to cancer, likewise appears evident based on wider observation.
So, if I follow his logic correctly, it’s not smoking that causes lung cancer, but irritation of the lungs as indicated by subsequent coughing. And anyone whose lungs are irritated in any other way, and coughs a lot, is just as likely to develop lung cancer as a heavy smoker. And so if you want to smoke and avoid lung cancer, then don’t smoke so much that you end up coughing a lot. This is a perfectly plausible hypothesis, and would go some way to explaining why quite a few non-smokers get lung cancer: their lungs were irritated in some other way.
And indeed one of the things he is very quick to point out is that non-smokers don’t have a zero risk of lung cancer, despite many doctors and pundits talking as if they do.
And he also goes into the difficulties of diagnosing lung cancer. A cancer is (or should be) only called lung cancer if the lung is the place where the cancer originates. If it starts somewhere else, like someone’s knee, then it is (or should be) called knee cancer. And he goes on to describe how difficult it is, in many cases, to ascertain where a cancer originated before metastasising (spreading) elsewhere.
In fact, it strikes me as being a quite impossible task. It’s a bit like finding a ticketless deaf-mute on a railway platform in, say, Barnsley, and wondering where they started their journey. Perhaps, after a while, you might ascertain that the person is Norwegian, and lives in New York, and has a wife in Australia. You might find an old bus ticket from Cologne in one pocket, and Tokyo bullet train ticket in another. How on earth can you ever know where he started out?
With all of this the opportunity for diagnostic bias is positively enormous
…the biased perspective tends strongly to perceive that, if the patient is a never smoker, the cancer cannot have begun in the lung, but if the patient is an ever smoker, the cancer surely did begin in the lung.
…strong diagnostic bias exists regarding lung cancer is clear indeed, simply from encountering clinicians’ common statements revealing colossal over-interpretation of the smoking / lung cancer link, and this has long been apparent.
He makes the point very clearly.
As Doctor Feinstein, an early critic on general terms of epidemiological and clinical overinterpretation, wrote on the overall subject of diagnostic bias in 1974: “It seems important to recall that in epidemiologic surveys of causes of disease, the investigators get data about the occurrence of diagnoses not the occurrence of diseases, and that the rates of diagnosis may be affected by bias …”
Doctor Feinstein summed up the particularly egregious bias surrounding smoking and lung cancer then in famously saying, “Cigarette smoking may contribute more to the diagnosis of lung cancer than it does in producing the disease itself.” Things have only grown worse over time.
In terms of statistics, of course, the plain result of this diagnostic bias is that clinicians under-diagnose “primary lung cancer” amongst never smokers (thereby statistically understating the base risk) and over-diagnose “primary lung cancer” amongst ever smokers (thereby statistically overstating the relative risk from smoking).
Well, I haven’t got to the end of the paper, but given what I’ve read so far I’m not expecting the author to come out and declare that Smoking Causes Lung Cancer and The Debate Is Over. And if he does, then he has provided an arsenal of reasons for doubting any such claim. Faster readers may wish to point out to me where he makes such a declaration.
In fact, if anything, he is lending further force to some of the sceptical denialista arguments I’ve been assembling over the past year. For it seems to me that if you don’t really know how many “cigarettes” anyone has smoked in the lifetime, because any number is just a guess, and furthermore you don’t really know whether their lung cancer originated in their lungs, then you really have no data, and whatever mathematical operations you perform on the numbers that you have will produce results as nebulous as the figures on which they are based.
As for plain packaging, H/T Belinda for this example: