Skip Cancer Screens

4 Trackbacks

1. [...] Beware Cancer Med By Robin Hanson · May 23, 2011 4:20 pm · Discuss · « Prev · Next » [...]

   By Overcoming Bias : Beware Cancer Med May 23, 2011 at 4:35 pm

2. [...] A post by Robin Hanson on cancer screenings is getting some play. Robin notes that screening detects more cancer, but doesn’t necessarily save more lives. Ezra Klein thinks that Robin goes a bit too far in his conclusions. I think Ezra goes a bit too far here (emphasis mine): That’s why, if you want to control health-care costs, you somehow need to convince, incentivize or otherwise conscript doctors into doing it for you. Robin Hanson can write as many blog posts as he wants, but as long as doctors are telling scared and uncertain patients that they need to get screened, they’re getting screened. The moment they stop telling patients to get screened, screenings will plummet. In health care, doctors are really the relevant decision-makers. And right now, they don’t have the evidence to make good decisions nor the incentives to make cost-effective decisions. [...]

   By There’s more to the problem than a lack of evidence | The Incidental Economist May 25, 2011 at 7:32 am

3. [...] is from Robin Hanson, who also provides a brief literature [...]

   By Why Screen for Cancer? | John Goodman's Health Policy Blog | NCPA.org May 27, 2011 at 3:30 pm

4. [...] Robin Hanson: skip cancer screening as it doesn’t reduce mortality. [...]

   By Google Reader Flush « Rhymes With Cars & Girls August 27, 2011 at 4:27 pm

41 Comments

1. Lord

   May 23, 2011 at 5:00 pm | Reply

   The problem is if a cure is found, it won’t be found by failing to screen.

2. Yvain

   May 23, 2011 at 5:09 pm | Reply

   I’ve only read the abstracts of these studies and not the full article; maybe the full article addresses these complaints. However…

   If I understand these studies right, they’re saying they find a reduction in cancer-related mortality but not in all-cause mortality. One explanation is that deaths from complications of cancer treatment counterbalance the decreased deaths from cancer. But another explanation is lack of study power.

   For example, imagine one million people die in the US each year, and ten of those deaths are from lightning strikes. Divide the US into two groups: Group A, who are advised not to play underneath tall trees in thunderstorms, and Group B, who are advised playing underneath tall trees in thunderstorms is totally okay and they should do it all the time.

   At the end of a year, Group A may have 511,021 deaths, zero of which are from lightning strikes, and Group B may have 510,405 deaths, eight of which are from lightning strikes.

   In this case, scientists would say that there’s evidence that avoiding tall trees during thunderstorm causes a significant decrease in lightning-related mortality, but not in all-cause mortality. Under your criteria, this would make it acceptable to declare that people should ignore advice against playing under tall trees during thunderstorms.

   Although I can’t say for sure that’s what’s going on without having access to those papers, I think next time it would be a good idea to check and rule it out, or at least mention the possibility, before giving advice that could potentially kill people.

   As always, looking for the medical consensus held by the people who have read all the studies in a field remains more informative than cherry-picking results from a couple of studies and ignoring the rest. I urge everyone interested in this topic to look for summaries and explanations of such consensus: for example, http://en.wikipedia.org/wiki/Breast_cancer_screening#Whether_it_works for breast cancer.

   AFAIK, there are a few types of cancer for which screening continues despite a medical consensus that it’s not worthwhile: prostate cancer is in this category, and a lot of women get pap smears until long after the age when they have a reasonable chance of getting cervical cancer. But other types of screening are definitely believed to do more good than harm.

   • Yvain

     May 23, 2011 at 5:15 pm | Reply

     Sorry, that comment should be limited to the colorectal cancer study, the one that found the decreased cancer but same all-cause. As for the others:

     - Prostate, as mentioned before I agree.
     - Breast: could be true, but also looks like they exclude a lot of studies and indicators that don’t come to the conclusion they want for reasons that I can’t tell if they’re legitimate or not without reading the paper.
     - Lung: could be true.

     It should also be mentioned that there’s no one “right answer” to screening – different groups will have different harms and benefits. Screening non-smokers for lung cancer would be a waste of time, but screening people with a long family history of lung cancer who smoke two packs a day and hang around asbestos factories might not be. Guidelines tend not to be of the “screen everyone” or “don’t screen anyone” type, but more about which groups are or aren’t worth screening.

   • g

     May 23, 2011 at 9:05 pm | Reply

     I think next time it would be a good idea to check and rule it out, or at least mention the possibility, before giving advice that could potentially kill people

     Robin Hanson’s medical advice isn’t about helping people get better medical outcomes.

   • Douglas Knight

     May 24, 2011 at 11:46 am | Reply

     One explanation is that deaths from complications of cancer treatment counterbalance the decreased deaths from cancer.

     Yet another is that being in the study biases classification of one’s death.

3. EH

   May 23, 2011 at 6:34 pm | Reply

   Saying skip cancer screens is a bit disingenuous, especially as at least one of those screens (lung cancer x-rays) has not been standard of care in my lifetime. Of the other three, fecal occult blood testing is no longer the preferred method of screening for cancer (although fobt is good, colonoscopy is better) and prostate cancer screening is falling out of clinical practice (it mostly continues to be offered for legal reasons). As for breast cancer, the vast majority of literature points towards a small benefit which in all likelihood doesn’t really justify the screen. Conspicuously absent from the list is cervical cancer screening.

   Whi

4. Robin Hanson

   May 23, 2011 at 8:38 pm | Reply

   Lord, you can’t win at the casino if you don’t play.

   Yvian, these analyses together include over a million people. Any benefit that random fluctuations hides in that many people must be pretty small, so you can safely skip it.

   EH, when after a decade or two one finally has a disappointing randomized trial result, docs complain that they don’t do the treatment the way the trial tested anymore, so the result is irrelevant.

   • Cyan

     May 24, 2011 at 1:32 am | Reply

     Regarding the reply to Yvain: yes, an efficacious intervention on a subset of a population must eventually show up in all-cause mortality, but the size of the benefit that can be hidden by random fluctuations in the entire population is of order of the square root of the sample size, which is not really that small even when the sample size is one million. So I think it’s illegitimate to ignore the proportion of the population that is in the relevant subpopulation and hand-wave off an actual cost-benefit analysis.

   • Unnamed

     May 24, 2011 at 2:43 am | Reply

     Yvian, these analyses together include over a million people. Any benefit that random fluctuations hides in that many people must be pretty small, so you can safely skip it.

     Can you put some numbers on this? How small is safe to ignore? How comfortable are you in dealing with these sorts of numbers?

     For instance, consider a screening that prevents 1% of all deaths (from all causes). Maybe 5% of all deaths in the relevant population are due to a certain illness, and the screening prevents one fifth of those deaths. That seems to me like a meaningful improvement? For someone with a 1/100 annual mortality risk, the screening would increase their chances of surviving each year by 1/10,000. If it’s annual screenings at $100 each, or a screening every 10 years at $1,000 each, that’s $1 million per life saved, which seems like a good deal.

     Now, how big a sample size would you need to find that difference in all mortality? Let’s say that you’re doing a study over 10 years, and you’re looking at a population where 10% of people die over 10 years. For people who get the treatment, only 9.9% will die.

     If you have 1.4 million people in the study (half who get the screening), and things go exactly according to the percentages, you’ll end up with 70,000 deaths from the 700,000 people in the control condition and 69,300 deaths in the screening condition. This would be just barely statistically significant (p = .048, by a two-tailed chi square test). That means that statistical power is about .50, since half the time (when the error happened to go in the same direction as the effect) the study would turn out statistically significant, and half the time (when the error went in the opposite direction) it would not be statistically significant. This 10-year study of 1.4 million people, on a treatment that prevents 1 out of every 100 deaths (and would be expected to save 700 lives during the study), only has a 50-50 shot of being statistically significant.

     If the study can accurately attribute the cause of death, and 5% of all deaths are due to this illness, then (according to the percentages) it would find 3,500 deaths due to the illness in the control condition and 2,800 in experimental condition, p less than .00001. The relative risk ratio would be 0.8, with a 95% confidence interval of (.76, .84). In other words, this study which is barely on the border of being able to identify these lives saved out of all-cause mortality can come up with a very precise estimate of the effectiveness of the screening if you focus only on mortality caused by the illness.

     tl;dr A 10-year study of 1.4 million people, evaluating a treatment that prevents 1% of all deaths, would have about a 50-50 shot of finding a statistically significant reduction in all-cause mortality, but it could provide a very precise (and highly significant) estimate of the reduction in mortality due to a single cause.

   • Lord

     May 24, 2011 at 1:05 pm | Reply

     That trivializes the fact that knowledge advances. You may not choose to be the guinea pig, but it is doubtful it can advance without guinea pigs. People may underestimate how much they are guinea pigs and the alternative of doing nothing may be better, but many would accept the uncertainty of winning for the uncertainty of loss, and many would do so to advance knowledge even if it doesn’t benefit them.

   • Silas Barta

     May 24, 2011 at 2:00 pm | Reply

     Not playing at a casino constitutes winning in my book…

5. skin_cancer

   May 23, 2011 at 9:47 pm | Reply

   How about skin cancer?

6. skin_cancer

   May 23, 2011 at 9:52 pm | Reply

   Ignore weird splotches on your skin, noticing them may increase your odds of dying in a car crash.

   • billswift

     May 24, 2011 at 4:58 am | Reply

     The point is not to ignore symptoms; the point is that screening of non-symptomatic people is worthless.

     • skin_cancer

       May 24, 2011 at 10:46 am |

       billswift: actually I don’t see any examination of skin cancer screenings done whatsoever in anything cited in this blog post, yet the admonition is to avoid *all* cancer screenings. Idiotic.

     • Dan Weber

       May 24, 2011 at 12:30 pm |

       Cancer screenings, by definition, are done on people who are asymptomatic.

       Asking the doctor about the splotches on your skin is not screening. Having some skin cells scraped off and put through a lab analysis to find out “just in case” you might get skin cancer is screening.

7. Dave

   May 23, 2011 at 10:32 pm | Reply

   “EH, when after a decade or two one finally has a disappointing randomized trial result, docs complain that they don’t do the treatment the way the trial tested anymore, so the result is irrelevant.” True but irrelevant.
   All cause mortality is the last thing to fall when medical progress is made. It is really very tricky to prove by statistical studies that a new screening method is successful. Insisting on this is much too contrarian.
   It took many years to prove the cervical cancer screening was effective. Now cervical cancer deaths are rare in women following screening guide lines. It was soon found out that other types of screening were not as good,especially for lung cancer. For years no screening was recommended for this. Now more sensitive means are being tried. No results yet.
   Colonoscopy appears to be successful in preventing many colon cancer deaths. Unless you have evidence that it kills more people than it saves,it is more likely that the results of statistical games are less reliable than direct observations derived prospectively.

   • Robin Hanson

     May 23, 2011 at 10:59 pm | Reply

     Where are the randomized clinical trials showing cervical cancer screening reduces all cause mortality?

     • Dan

       May 24, 2011 at 4:53 pm |

       It’d be considered unethicial to not screen for cervical cancer at this point considering the strong observational data.(ex: http://www.ncbi.nlm.nih.gov/pubmed/9328198)

       Almost all of the women who get cervical cancer in the screening era are either from developing countries without screening programs, or women from developed countries who do not regularly getting screened.

       It’s a shame this wasn’t tested in a RCT before it became standard practice, but I doubt you could find a medical researcher or ethicist that would allow women to be in a placebo arm now. I think similar points would be made about colonoscopies. We need further advocation for evaluating screening tools with RCTs before they become standard clinical practice though considering the issues we have with the other cancers (partartically prostate)

       Overall, I think you’re making an important point but your sweeping generalization that all types of cancer screening is off base and (unfortunately) likely discredits the whole argument in many reader’s minds.

8. Jmack

   May 24, 2011 at 12:06 am | Reply

   No one in medicine or public health thinks that lung cancer screening is advantageous. There aren’t good enough treatments to justify it. Very very very few think that prostate cancer screening is advantageous. More often than not, it is better off left alone.

   The breast cancer link review included only studies when screening begins at age 40 or earlier. That is not the standard now.

   And really, colonoscopy is the standard for colon cancer screening.

   The statement that cancer screening doesn’t lead to less cancer mortality is false.

   I can’t imagine why you would want to use all-cause mortality as a measure? Especially in a meta-analysis.

   • Robin Hanson

     May 24, 2011 at 7:04 am | Reply

     But there are no randomized trails of colonoscopy screening.

     • Dave

       May 24, 2011 at 10:56 am |

       No, that would be considered immoral at this juncture. And that is a problem. Many studies are terminated early when there are either excessive bad results or excessive good results in one arm of the study. Shades of the Tuskegee study! By the way did this study show an increased all cause mortality in the test group? I don’t know. It should have. Studies like this will probably never be done again.

9. AaronM

   May 24, 2011 at 11:29 am | Reply

   I would not expect Robin Hanson’s opinions on health care to garner the support of many physicians. But I am one, at least, who fully agrees that the marginal health care dollar has zero utility, that harmful overtreatment abounds, and that disease screening is far more difficult that it seems.

   As a future oncologist, this post is of particular interest to me. While I am generally one of the first to point out the difficulty of effective cancer screening that’s not to say it is altogether useless. Unfortunately, some of these screening tests are not as cut-and-dry as reading the post would have you believe.

   Breast Cancer: Omitted was a particularly relevant bit from the cited paper’s overall conclusions, “Screening is likely to reduce breast cancer mortality. As the effect was lowest in the adequately randomised trials, a reasonable estimate is a 15% reduction corresponding to an absolute risk reduction of 0.05%.” Absolutely, mammography is not silver bullet that the mass media would lead us to believe. But that’s not to say there is no benefit.

   Additionally, if you get through to the data and analysis section, when they segregate the women by age, they find that the reduction in breast cancer mortality is greater for women above 50 (28% vs 11% reduction at 7 years, 23% vs 16% at 13 years). There is currently a heated debate within the medical establishment regarding such findings as this, but the new thinking is to target the screening to older women where there is more likely to be benefit, and less likely to be overtreatment.

   Also, it should be noted that self-breast exams and clinical breast exams have generally been shown to be useless for breast cancer prevention.

   Colorectal Cancer: As others have pointed out, the cited paper is talking about stool blood testing, NOT colonoscopy (which is now the standard method). On the other hand, Robin is quite correct in pointing out that there are no randomized, controlled trials to prove that colonoscopies actually save lives. Ideally, there would be, and, hopefully, there will be at some point. But for now, we have to do our best with lower-quality data, which is quite suggestive of a benefit from colonoscopy screening.

   Prostate Cancer: I am in complete and full agreement. There is nothing to show that either digital rectal exam or PSA testing is useful as a screening tool. These practices should be ended as soon as possible.

   Lung Cancer: Yep, X-rays do no good for preventing lung cancer. This has been repeatedly shown. However, we are starting to make inroads – it was recently found that using specialized CT scans in people at high risk (former smokers) probably does save lives (http://www.lungusa.org/press-room/press-releases/on-the-national-lung-cancer.html). Not something that can be done for the general non-smoking population, but it does show that screening can work in principle, even if it is often very difficult in practice.

   Cervical Cancer: It was also pointed out that this one was omitted. I haven’t looked into this specifically, but I have heard it said that “there are no randomized, controlled trials proving the benefits of Pap smears.” Even if true, this is akin to saying “there are no randomized, controlled trials proving that smoking causes lung cancer,” or that “…rubella vaccines prevent neonatal blindness.” Sometimes, the observational data is clear enough.

   • Dan Weber

     May 24, 2011 at 12:32 pm | Reply

     Pap smears do have good solid evidence for improving outcomes. They are very rare among screening technologies for this.

   • BlueStreak

     May 25, 2011 at 4:04 pm | Reply

     Re: Prostate cancer testing, you say that “these practices should be ended as soon as possible.” In March 2004, at age 54, my primary physician thought my prostate felt “asymmetrical”, my urologist concurred, a biopsy came back positive and I had my prostate removed. So far, so good. I understand that many prostate cancers never fully develop, and some are extremely long-developing, but are you saying that I should NOT have had testing done? Are we just supposed to roll the dice and hope for the best? What if I didn’t test and I turned out to be too dead to be writing this response? That’s not really an acceptable outcome in my opinion.

     • retired urologist

       May 26, 2011 at 5:18 pm |

       Statistically, 10-year survival rates in huge cohort studies of prostate cancer indicate no treatment has a better survival rate than “no treatment”. Going in, you couldn’t know which individual in the study you would be. You still can’t. The choice you made was for a published 10-year survival rate, either with the surgery you had and it’s attendant side-effects, or without it. This is a bitter pill for both the patients and their doctors, the latter of whom have a vested reason to ignore the statistics. Happily, although thousands of Americans die from prostate cancer each year, they represent a very small portion of those with the disorder.

     • AaronM

       May 26, 2011 at 9:30 pm |

       @BlueStreak:
       Is it possible for prostate cancer to be detected (and prostate cancer deaths prevented) through digital exam? In principle, yes. I truly hope that this was true for you, and that the medical care you received was beneficial for your health.
       However, based on what we think we know from the best studies we have, most likely it was not. When considering disease screening, it helps (at least for me) to think about the small percentage of screened patients who we actually end up helping…
       1. Most patients screened will have a negative result, with only a small minority having evidence of disease (so a lot of unnecessary tests).
       2. Of those with a positive screen, only a few will go on to have cancer on biopsy (so a lot of unnecessary biopsies).
       3. Of those with positive biopsies, we have to subtract those patients whose cancer never would have hurt them anyway (and with prostate cancer, we know this is a very high % of cases), because those people won’t benefit from any treatment they go on to receive.
       4. Of those with dangerous cancer, subtract all those who would not benefit from EARLY treatment – maybe at some later point the asymptomatic cancer detected on exam would start producing urinary obstruction, at which point it may still be curable (and therefore the patient gains no benefit from having found it earlier).
       5. Also subtract people for whom it is already “too late” for helpful treatment at the time of screening (this is true of a great number of young men who get prostate cancer).

       For prostate cancer in particular, the remaining number who will actually benefit from having been screened is quite small (if present at all). Furthermore, if we make the assumption that there ARE a few people in there who have a potentially lethal cancer cured, then in light of the all-cause mortality data (which shows no benefit to screening), then the logical conclusion would have to be that the small number we help is being offset by an equal number of people being harmed in the course of aggressive medical treatment.

       When also considering the discomfort, cost, complications, and side effects of the screens, biopsies, and surgeries, the deal looks less and less appealing.

       You ask if we are “just supposed to roll the dice and hope for the best”. I really wish I had a better answer to that question than “yes.” If we ever end up discovering a different, better test that actually helps save lives, I will be the first to start offering it to as many of my patients as I can. In the mean time, given the current state of prostate cancer screening, my own personal choice would indeed be to avoid it.

10. David

    May 24, 2011 at 1:53 pm | Reply

    It seems there are a few reasons screenings won’t help reduce mortality too much:

    1) Other causes of death are just as likely to strike
    2) False negative leads to someone not getting necessary treatment
    3) Most people just don’t have cancer
    4) Many screenings come too late to treat
    5) Screenings are not actually a treatment

    Am I missing anything else?

    • Glen Raphael

      May 24, 2011 at 2:14 pm | Reply

      Two big ones you’re missing:
      6) False positive leads to someone getting unneeded treatment that causes harm
      7) True positive leads to someone getting treatment that does more harm than good

      Iatrogenic illness is one of the largest causes of death in the US – we should all rather err on the side of too little care than too much.

    • Dan

      May 24, 2011 at 5:05 pm | Reply

      The screening tools with the most evidence (pap smear/hpv testing and colonoscopies) are treatment in a way. they remove pre-cancerous lesions.

11. Joseph Hertzlinger

    May 24, 2011 at 11:08 pm | Reply

    According to a quote (with favorable comment) on Overcoming Bias last year: “The evidence strongly suggests that students learn better when they are not graded and certainly not when they are graded on a curve.” Using similar reasoning, we can conclude that being part of a randomized controlled study makes medical care less effective.

    • daedalus2u

      May 25, 2011 at 9:20 am | Reply

      Huh? Could you describe the reasoning process you use to make that conclusion? I simply don’t see it.

12. Steve Sailer

    May 25, 2011 at 5:10 am | Reply

    Having had cancer myself, I recall that I spent a lot of time considering moving to cities with outstanding cancer clinics. Ultimately, I found what I needed to survive at home in the Chicago area, but if I had relapses and thus needed a last ditch stem cell transplant, I probably would have gone to a state with a state of the art cancer center, and I probably would have died in that state.

13. Dr. Liberty

    May 25, 2011 at 8:53 am | Reply

    Even if a type of cancer screening is shown to improve cancer mortality, or even all-cause mortality, even THAT would not be evidence enough to do it. It would also have to be a better expenditure of resources than the next best use of those resources. So if the benefit is small or the cost huge, the statistically significant mortality benefit may not be enough.

14. thomas

    May 25, 2011 at 6:18 pm | Reply

    Robin,

    In general I agree completely that screening is overrated — my teaching soundbite on it is that screening is the opposite of treatment, since it takes healthy people and makes them sick.

    However, I think two of your examples are wrong at least in part.

    There’s good evidence that the Cochrane Breast Cancer was wrong to discard the trials it claimed had inadequate randomization. See the very detailed analysis by Freedman, Pettiti, and Robins. The authors aren’t pro-screening ideologues — two of them are statisticians and so predisposed not to believe in screening, and Diana Pettiti was co-chair of the panel that, last year, recommended against routine mammography before age 50.

    As someone pointed out above, even though ordinary x-ray screening doesn’t work for lung cancer, there is now good evidence for CT-scan for lung cancer in smokers and former smokers. I was surprised by this when it came out, but the results are pretty clear.

15. Jane

    May 25, 2011 at 7:51 pm | Reply

    There’s a reason that GIs build stand-alone colonoscopy suites and book back-to-back procedures in assemby-line fashion: They’re cash cows.

    Colonoscopies are standard treatment because they make the most money for assorted actors (GIs, hospitals, equipment manufacturers), not necessarily because they find cancer in its early stages or improve patient outcomes. There no evidence of these benefits.

    Within the past two years, meta-analysis (albeit with some methodological flaws) has show very little benefit of screening colonoscopies, especially for right-sided cancers. Left sided-cancers are better detected, and sigmoidoscopy is perhaps an adequate screening tool given the meta-analysis. But, that procedure costs about $1500 less than a colonoscopy though it is arguably just as effective.

    Perhaps worse for patients, routine colonoscopies can lead to a false sense of security wherein patients ignore signs of colon cancer because they’ve recently been given a clean bill of health.

    Screening scopes are almost certain to be a thing of the past as companies improve and certify DNA testing for colon cancer, as has been done in Germany and other European countries.

16. Donna

    May 26, 2011 at 2:14 pm | Reply

    Don’t forget that cervical cancer screening has been scaled WAY back due to the understanding of the relationship between HPV, abnormal Pap smears, and cancer.

    For decades, young women who were newly sexually active were subjected to further diagnostic procedures which can damage the cervix and cause problems in subsequent pregnancies. It turns out that most HPV infections resolve without treatment within 2 years.

    Also, many women oddly perceive a normal Pap as a certificate of good health. After all, if that’s what we continually test for, that’s obviously the most worrisome health concern, right?

    If the amount of time, effort, and money devoted to screening for cancer in the US was instead directed toward weight loss, we would be healthier.

    • Doug S.

      May 26, 2011 at 11:04 pm | Reply

      Weight loss interventions are even more useless than cancer treatments. As far as I am aware, the only weight loss intervention that has been clinically proven to result in long-term weight loss is stomach surgery. Every single diet-based approach anyone has ever tried has consistently failed.

17. Kittykitty7555

    May 27, 2011 at 2:28 pm | Reply

    We shouldn’t be too quick to credit colonoscopy for a reduction in incidence and mortality from colon cancer. incidence and mortality of stomach cancer has plumetted something like 80 percent in the last 50 years, and there is no screening for it. Docs who do colonoscopies have found an absolute gold mine, but it’s completely unclear whether their patients are helped at all.

18. hud

    June 1, 2011 at 6:39 pm | Reply

    Can it also be assumed that a certain percentage of people who seek screenings (perhaps a large percentage) are taking action and getting screened specifically because they know that cancer runs in their family. And those who don’t get screened probably aren’t worried because it does not run in their families.

    So, screened people who ended up with cancer mortality versus non-screened people who ended up with cancer mortality seems misleading to me.

    Maybe there should be a study on the effectiveness of screenings, specifically for people who have cancers that run in their family and a separate study for the same thing specific to people who don’t have cancers that run in their families.

    My guess is that if you looked at the data that way, both data sets would show that screenings helped.

19. Elizabeth (Aust)

    June 6, 2011 at 9:05 pm | Reply

    I know many believe the pap test is the most effective screening test in history. I can’t agree…
    There are no RCT’s for pap testing and also, this was always a rare cancer, in decline before screening started and those factors are still having an effect on the incidence and death rate. Dr Gilbert Welch points out things like better hygiene and less sexually transmitted disease in his book, “Over-diagnosed”. I’d add more hysterectomies (1 in 3 US women will have one by age 60!), fewer women smoking, better condoms etc…
    It’s true that stomach cancer has also fallen by a similar or greater margin with no screening at all, yet you can bet if there was a screening test, they’d be grabbing the credit.

    The thing that writes off cervical screening for me: vast over-detection and potentially harmful over-treatment. I also hate the way cancer screening is regarded as compulsory for women, elective for men. Women don’t get balanced and complete information and our legal right to give informed consent for cancer screening is dismissed…in the States and Canada women are routinely coerced into cancer screening and often, over-screening and reckless screening. (screening women not yet sexually active from age 21)
    I know American doctors refuse women birth control pills, HRT, migraine meds and even all medical care until they agree, to not only cancer screening, but completely unnecessary and potentially harmful breast and gyn exams. (these exams are not recommended in asymptomatic women here and in many other countries)
    There is a double standard in cancer screening.

    In Australia we also over-screen (not as badly as the States) and there is zero respect for informed consent (for women) – our GP’s even receive an UNDISCLOSED target payment for pap testing (about to be raised to 65% of eligible patients)…this is unethical as it creates a potential conflict of interest. These payments were recently changed in the UK, thanks to the work of some amazing advocates for informed consent for women.

    Here we screen women from 18 (sometimes earlier) knowing this is of no benefit, but exposes these women to harm and great worry.

    “No country in the world has shown a reduction in the incidence of or the mortality from cervical cancer in women under 30, irrespective of cervical screening. Many countries do not perform cervical screening on women under 30″…”Cervical cancer screening” a handout for doctors (not women!) in “Australian Doctor” 2006 by Assoc Prof Margaret Davy and Dr Shorne.
    Still we continue to test young women….

    Finland has the lowest rates of cc in the world and sends the fewest women for colposcopy/biopsies (fewer false positives) – they offer 5 to 7 tests, 5 yearly from age 30. (even this schedule sends 35%-55% at some stage for follow-up)
    Australia often boasts about having the lowest mortality rate in the world, but we keep our shameful over-detection and over-treatment rates under wraps. This low rate is achieved at a terrible cost paid by the healthy population of women – the more than 99% who’d never have an issue with this cancer and all with no informed consent.

    Dr Angela Raffle, UK screening expert showed that 1 in 3 pap tests will be “abnormal” in women under 25 – false positives caused by normal changes in the maturing cervix and from harmless and transient infections. Cervical cancer is rare, very rare before 30 and screening doesn’t change the incidence or death rate in young women anyway.
    It is unethical to test women under 30 and definitely before 25 and to test any woman without informed consent.

    We also test 2 yearly and this means lots of false positives for no additional benefit – it’s over-screening.
    77% is the lifetime risk of referral for colposcopy and usually some sort of biopsy to cover a 0.65% lifetime risk of cervical cancer. Take out false negative cases and considering other factors are affecting the incidence and death rate – fewer than 0.45% of women are helped (assuming any woman is helped) The States it’s even higher – lifetime risk of referral – 95%…(DeMay article at Dr Sherman’s site)
    This amount of over-detection and over-treatment IMO, makes this testing unsuitable for population screening. I don’t believe it would be approved today.
    Now we have a program that is highly lucrative, very political and highly emotive….it helps few and harms (to some degree) many….
    Some women end up worse off after unnecessary biopsies and procedures like LEEP – cervical stenosis, (infertility, infections, endometriosis and may need surgery if the cervix is scarred shut) cervical incompetence – miscarriages, high risk pregnancy that may require cervical cerclage, more c-sections, premature babies and psych/psychosexual issues.
    No one seems to care about these women and there are lots of them – the entire focus is on screening and the incidence and death rates for this rare cancer.
    My own younger sister had a cone biopsy after a false positive pap test – a devastating experience.

    It’s the same with mammograms – we don’t get honest and complete information and there is no respect for informed consent.
    Thankfully, the UK has some great people prepared to speak up and warn women of the risks. Professor Michael Baum, UK breast cancer surgeon gave another informative lecture at UCL recently, “Breast cancer screening: the inconvenient truths” – at the Medphyzz site.
    It’s also possible to get to the facts behind cervical screening – Dr Joel Sherman’s patient privacy site under women’s privacy issues has lots of great references, including research by Angela Raffle. “1000 women need regular smears for 35 years to save one woman from cc” (BMJ:2003/4) plus over-detection and over-treatment rates.

    It’s hard to understand why we spend a small fortune “fighting” this rare cancer when there are FAR greater threats to our lives out there…it says to me that these programs have little to do with our health. The priorities are skewed by pressure/lobby groups, political and vested interests. ($$$)
    I think cancer screening is a great threat to our health, rights and lives.
    I made an informed decision, as a low risk woman, not to have pap testing more than 25 years ago and more recently, rejected mammograms. I didn’t get the information I needed from my doctor.
    More than ever before we need to do our own reading, spread the word and make informed decisions about our health care. I don’t give a damn about screening targets or my GP’s incentive payment…
    Is this really in MY best interests?