Original URL: https://www.theregister.com/2007/06/15/cancer_survival_and_mortality/
Winning the (propaganda) war on cancer
Where do you bury the survivors?
Posted in Science, 15th June 2007 14:02 GMT
A couple of years ago, the American Cancer Society gaily reported that cancer survival rates have been improving dramatically and steadily. A few weeks ago, Cancer Research UK announced a similarly sunny prognosis for the public at large: UK survival rates have doubled in 30 years. Yes, we are all "surviving" longer with cancer, the experts can assure us.
What fabulous news this seems to be: people diagnosed today are twice as likely to survive for 10 years than people diagnosed 30 years ago. As the press release says: "While survival varies widely between different types of cancer, on average, a patient with cancer now has a 46.2 per cent chance of being alive 10 years after diagnosis. This compares with 23.6 per cent 30 years ago".
Now, what do you suppose that means? Do you imagine that treatments have improved so much in 30 years that we are actually able, on average, to slow the disease's rate of progress by half? That we have actually doubled the life expectancy of cancer patients? That certainly is one message that we might take from the press release.
Well, those who actually bother to read medical literature know that what is being claimed here is a bit less fabulous than that. In the field of medicine, words like "survival" and "cure" are professional jargon with specific meanings and implicit qualifications. Researchers and doctors know what these terms mean, and they use them very carefully, assuming the implied qualifications are understood. Unfortunately, you and I might understand something quite different from what is meant.
So, what do professionals mean when they speak of survival? Typically, they speak of five-year survival in a way that gives patients a decent guess of their chance of still breathing five years after diagnosis. The doctor considers the specific form of cancer, its state of progress at the time of diagnosis, the patient's overall health, age, sex, and numerous other factors, and then a comparison is made between the individual patient and a group of similar patients.
So, it's common for a practitioner to compare your current condition to a large group of others previously diagnosed in similar circumstances, and tell you, for example, that among those similar to you when they were diagnosed, 60 per cent survived for five years.
Thus you would have a 60 per cent chance of surviving for five years. You might also be told that you have, say, a 30 per cent chance of surviving disease free (i.e., in remission), and a 45 per cent chance of surviving progression free (disease is detectable but not worsening), for five years, based on comparisons with others whose conditions were similar to your own at the time of diagnosis. Thus, "survival", in the professional sense, is relative.
But if survival is a relative measure, what can it mean when Cancer Research UK says that survival rates have improved overall? What, exactly, is being compared to what in that case? Does it mean that we have got so good at treating cancer that the disease typically progresses at half the rate it once did? Does it mean that patients are actually living twice as long as they did 30 years ago? I do wish I could say this is so, but then I would tell a lie.
A meaningless boost
Repeatedly, we are reminded that early diagnosis and early treatment are major reasons for the improvement. We have better diagnostic tools and screening methods than ever before; more mass screenings for the public are scheduled, often by charities serving populations with poor health services; family practitioners are more aware of the disease and its early symptoms; and more patients are now aware of the symptoms that might warrant a trip to the quack. Of course, the treatments themselves have improved as well, and this certainly is having some effect.
However, the item that we need to question most carefully is the effect of early diagnosis and early treatment. Treatments are starting earlier these days, and we are all "surviving" longer with cancer. But just how fabulous is that news, really?
The cancer treatment industry desperately wants us to think it's beyond fabulous: cancer drugs, and related diagnostic and treatment devices, are among the most expensive, and most profitable, known to medicine. There are those who profit from our assumption that early diagnosis and treatment extend patients' lives. And profit they do: think of the hospitals, drug makers, device makers, diagnostic technicians, radiotherapists, surgeons, and oncologists. Cancer treatment is a vast, complex industry, and like any industry it flourishes best in a field of semi-educated consumers.
And so we are subtly shepherded toward the belief that our deaths from cancer will be postponed significantly because of early, and very costly, medical intervention.
Unfortunately, this is not always easy to prove. And the word "survival" itself is tricky, and normally used in a relative context, as mentioned previously. Indeed, you can "survive" longer than another patient with the same cancer, but not actually live any longer, because the survival clock starts running on the day of diagnosis. Thus, everyone who is diagnosed early automatically survives longer, independent of any other variable. Conversely, if you are diagnosed late, you are not going to "survive" for long, although you might already have lived for quite a long time with your cancer.
So it is hardly time to break out the noise makers and Champagne merely because the survival numbers have doubled. Let's go to an illustration: imagine a doctor with a patient who has got Stage IV (metastatic) cancer at the time of diagnosis. Now, one might ask the doctor a few questions, such as, how long did the patient take to progress from Stage I to Stage IV? Or, just when did carcinogenesis actually occur - six months ago? Six years ago?
Of course, no one can say with certainty; patients are individuals, and diseases work differently in different people (as do drugs and other treatments, by the way). But because our example patient is diagnosed at Stage IV, we'll learn only how long the disease needs to kill them when treatment starts at that point.
Now, compare this to a patient with the same disease, only Stage II at the time of diagnosis. If there is good treatment and follow-up, we'll get a detailed picture of how long the disease needs to kill this early-treatment patient. But because we don't know the late-treatment patient's disease history, we can't say much about how long that patient has lived with their cancer, and thus it is difficult to compare the progress of these two hypothetical patients.
Did the Stage IV patient progress faster from Stage I than the early-treatment patient did? We might assume that they did, but we don't really know. Indeed, their own body might have fought the disease well, and they might possibly have lived longer with untreated cancer than the early-treatment patient lived with treated cancer. This might be unlikely, but the uncertainty here is nonetheless real.
Oncologists who see hundreds of patients over the years will get a sense of the disease's natural progress, and should be able to say with some confidence that when this or that type of cancer is left untreated, it usually progresses a lot faster than it does when it's treated. And that instinctive sense, learned from experience and observation, is probably reliable, although it should be noted that it isn't proof.
Handy assumptions
Among professionals, jargon like "survival" can be useful; it is generally used correctly, among others who will interpret it correctly. And yes, we can learn from studies where these terms are in play. We can certainly compare treatments among groups of early-treatment patients and see if one regimen is better than another, using five-year survival as a context. We can also compare late-treatment patients to early-treatment patients who have arrived at the same late stage, and see if early treatment has provided an edge in survival from that point forward.
But the idea that early treatment inevitably causes a patient to live longer is not always proven. Treatments make tumours stop growing, or shrink, or even go away. This is not the same as saying that the patients die significantly later than they otherwise would. We want to believe this, and there are people and corporations with an interest in encouraging us to believe it, but it is, as I said, very tricky to prove in some cases.
Now, here is a discouraging observation: there certainly are occasional breakthrough treatments, but generally, when we compare different regimens, we rarely see a dramatic difference in effectiveness among study groups, which suggests that the treatments probably aren't doing as much as we've been led to believe. Especially in the difficult cases.
Let's consider a few of the really tough cancers: stomach, pancreatic, lung, ovarian: these have very poor survival rates and treatment is usually not very effective. Why? Because symptoms rarely show until the disease is well advanced and metastatic. Catching one of these cancers early is usually a matter of luck.
Generally, there is little luck to speak of when you are diagnosed with one of these cancers. Your oncologist sits you down and gives you "the talk". It's advanced and spreading; we're going to do all we can; there are treatments and we will explore all the options; we learn more every day about treating this disease, etc. But you know that with a late diagnosis, there won't be a great deal of time before you die, and you think, my God, if only it could have been caught earlier: I would be able to live longer. You've heard all the glowing rhetoric about early diagnosis and early treatment, and you feel cheated of your chance to really fight the disease, and cheated of the most valuable thing of all: time to live.
But this might be total nonsense, and you might be suffering additional anguish for no reason. You see, with a late diagnosis, your survival is going to be short, but your life might not be much shorter than it would be if you'd been diagnosed early. As we discussed previously, you might already have survived quite a long time with your cancer; you simply haven't been aware of it. There is no reason to believe that you are not already an impressive cancer survivor.
Unfortunately, few patients will feel this way, and this is an incredibly cruel consequence of the survival propaganda that the media and medical industry disseminate.
Furthermore, to be "cured" of cancer usually means that you are in remission five years after diagnosis. So here's another problem: early diagnosis automatically increases the rate of so-called "cures", just as it automatically increases the rate of survival, independent of any other variable.
Personally, I think the notions of "survival" and "cures" should be abolished from public discourse, and should remain in the realm of research and treatment, where they are useful. Instead of five-year survival, patients and the public at large should be talking about mortality. "Has this treatment been shown to impede the disease's progress and delay death?" That really is the only question we should be asking, at least in the context of cancer survival (certainly there are other legitimate goals for treatments, such as palliative care, etc).
And as for "cures", this nonsense needs to be replaced with honest language as well: in the public sphere, "cured" should not be permitted to mean anything other than, "the patient died from a different cause". If you come out of your oncologist's office with a new diagnosis, and absent-mindedly step into the path of a city bus, well, that bus will cure your cancer. The fatal heart attack you have a month later will certainly cure your cancer. But, outside the context of medical research and calculating a patient's odds, it is preposterous to speak of a "cure" unless something else brings about your death.
This linguistic sloppiness makes it very difficult for us, as consumers of medical products and services, to know if early diagnosis and intervention do significantly more than enrich the medical industry. Certainly there are tumours that can be detected early, and when diagnosed at Stage I (in situ), may often be removed surgically and then legitimately called cured, because many patients with such experiences do in fact end up dead from other causes.
But there are other cancers that are extremely difficult to treat, and - if we were to dispense with the linguistic vagueness - are probably close to 100 per cent incurable, with very high (i.e., rapid) mortality rates. We might as well face the facts. But we cannot face them when journalists and interested parties fuel public debate with professional jargon that is often misinterpreted outside the medical industry.
If we in the press were to chuck "survival" and report on cancer mortality instead, and save "cure" for those who die of something other than cancer, the public might learn a few things that the industry would prefer them to remain foggy about. As patients, we might begin asking whether an expensive regimen has actually been shown to extend lives. As charitable donors, we might begin asking if our money is going into dead-end research where virtually no mortality advantage has ever been found - and whether it might be better spent on public screening programmes, or prevention initiatives, or better palliative care.
But, yes, the UK's cancer survival rates have doubled in 30 years. That certainly is true. Are cancer patients living twice as long? Certainly not. Some of the increase is due solely to the fact that early diagnosis improves survival numbers independent of mortality. Yes, some of the increase is due to improved treatment regimens, and to the fact that early treatment can improve mortality rates - at least to some degree, and at least for some patients. But how much, and for how many? Well, that is the question. So here's an idea: why don't we in the press start using language that helps people to answer it? ®