According to a media release issued late yesterday by the American Urological Association, Governor Christie has signed legislation in New Jersey that asks “the Congress of the United States to seek the withdrawal of the United States Preventative Services Task Force recommendation against prostate-specific antigen-based screening for prostate cancer for men in all age groups.”
The full AUA media release is available on line.
It should be noted that this does not mean that the draft USPSTF recommendation will not be effective in New Jersey if it is approved. It simply means that all members of the state legislature in New Jersey have unanimously agreed to tell the US Congress that they don’t think this recommendation should have regulatory impact. This issue has now clearly become politicized.
Search for new and
ongoing trials on the
CTAG PCa web site
and in a mindless, dichotomous, uninformed, unhelpful kind of way. Bottom line: By state legislative authority: No more nooky for Snooki!
On one hand I am quite proud of the prostate cancer advocates. They show that they can have a significant voice in everything prostate cancer-related when they come together. The same thing happened with the mammogram debacle 3 years ago and I think the USPSTF recommendation there has gone largely ignored. On the other hand is the argument that we should leave everything to science as shown by Dr. Chodak’s comments in his recent video. But I think we need to include the advocacy arguments in consideration of what the final recommendation should look like. After all, it was science that gave us the PSA test and in doing so lacked the foresight needed to measure the ramifications on the population when used as a screening tool. In a statistical measurement of the general population, the current USPSTF anti-screening recommendation is reasonable. But for individual cases, such as my own, it’s potentially ludicrous and extreme, and will likely result in increased prostate cancer-specific death rates. Besides, I think this recommendation was already political before it was ever announced … and it’s destined to be more so.
It would appear that politics is now trumping science based on the fact that Gov. Christie has signed New Jersey’s legislation against accepting the USPTF recommendation on screening and the recent statement by two congressman.
No doubt, every politician wants to avoid angering potential voters and supporting screening is one way to do that. The comments about the “thousands” of lives saved ignores the fact that several million men have received unnecessary treatment and subsequent harms during the time period that screening has been conducted. These political statements are being delivered, in part, because the recommendation might affect coverage of PSA under the new Affordable Care Act.
People should realize that this act is only meant to cover health interventions with either an A or B recommendation by USPTF. Even if the latest recommendation about PSA were to be changed, at most it would only be to a “C” rating but cert-inly not to an “A” or “B” rating. That would still mean it should not be covered. That is what the science supports and nothing more. Although the AUA letter correctly addresses concerns about high-risk men, we still have no proof that they are more likely to benefit or that the benefits greatly outweigh the harms. For them, however, a more cautious recommendation might be warranted that expresses the complete uncertainty of the impact of routine testing. Regardless, it is very upsetting to see that science is being thrown in the back seat while politics does the driving! In the long run, that is not good for patient management or for proper use of government dollars.
NEW JERSEY AND OTHER FRIENDS: IT’S NOT A “RECOMMENDATION” — IS IT TOO MUCH TO ASK FOR A LITTLE PRECISION?
I’ve read the AUA statement which quotes the New Jersey legislation. Unfortunately, the legislation wrongly terms the USPSTF document a “recommendation” when it is not. It is a “DRAFT recommendation,” subject to comments, as we all should know. There is a great difference! Most of us are aware that there have been many comments and that the final USPSTF position is under consideration at this time. I remain hopeful that the position will be withdrawn entirely or drastically revised into a sensible position that links screening and active surveillance.
Many media stories also wrongly term the USPSTF position as a recommendation. We should oppose such sloppy language and sloppy thinking.
DR. CHODAK: “SCIENCE IN THE BACK SEAT”?
While the USPSTF draft recommendation did briefly acknowledge major flaws in the PLCO and ERSPC trials, it did not seriously address the adequacy of follow-up time in those studies. Many of us are convinced that follow-up time was grossly inadequate in both initial reports (2009 NEJM), that this remains a critical flaw, and that this argument has not been rebutted.
Indeed, at the IMPaCT conference last spring (March 10), following your presentation (thank you) at the Meet the Experts sesson on screening, presenter Dr. James Mohler of Roswell Park, during Q&A, stated he’s convinced that follow-up of 20 years is needed to get a good view of the value of screening. I have also heard other experts advocate for 20 years follow-up, and that estimate is credible to many of us based on our understanding of the typical long survival with prostate cancer and our own cases. (I am now in my 13th year, doing well — third vacation on intermittent ADT3 as sole therapy, after an initial PSA of 113.6, Gleason 4 + 3 = 7 (Epstein), stage 3 with a “rock hard” prostate, all cores positive, most 100%. With this diagnosis, stemming from my first PSA at age 56 at my insistence despite my primary care doctor’s reluctance, is my situation not comparable in key respects to that of the men diagnosed in the control arms of the studies? Are not the vast majority of them likely to have a much better chance than I of cruising past the 12-year point as I am, let alone surviving the disease?) (I’m not addressing here the other serious flaws in the two studies, flaws which both you and Dr. Mohler touched on in the IMPaCT screening session.)
Moreover, I’m confident we all agree that the key follow-up metric is not time from registration in these prostate cancer screening studies as healthy men but instead time from diagnosis. Why then do so many medical commentators use the reported enrollment follow-up instead of noting that the time from diagnosis is years shorter? If time from diagnosis is applied to the studies (based on reasonable estimates from clues in the studies — haven’t seen actual figures), it is clear that follow-up time is well short of 10 years (even for the recent update of the PLCO trial). Is it not true that at 10 years from diagnosis survival of prostate cancer for low- and intermediate-risk men is virtually 100%, and that even high-risk men, like me, are registering survival of 95% at that point? If so, how can the PLCO trial tell us anything meaningful at this point about the benefits of screening? I’m thinking that’s true of ERSPC also, though I’m not that familiar with European survival statistics and treatment practices; perhaps less aggressive and advanced treatment resulted in the modest survival benefit for the screening arm, as reflected in the initial report, and in updates, including the Goteborg paper, with the benefit appearing to grow with lengthening follow-up.
Therefore, if the USPSTF fundamentally relied on the PLCO and ERSPC trials (cutting through the chaff of the meta studies), failing to reconcile its critical comments about those studies with the preponderant text coverage and thrust of its position, how can we credit the USPSTF with making a sound scientific assessment? Indeed, until the follow-up-from-diagnosis argument is credibly rebutted, many of us will remain convinced that the USPSTF adopted a scientifically unsound draft position. Many of us will continue to view premature PLCO and ERSPC trial-based-statistics about numbers needing screening to prevent one death as meaningless and unthoughtful.
We are not saying that there is strong evidence for the value of screening; we are saying that there is no proof that screening is of little value.
LONG SURVIVAL, SCIENTIFIC EVIDENCE, AND NOT ACHIEVING AN “A” OR “B” RATING UNDER “THE PATIENT PROTECTION AND AFFORDABLE CARE LAW”
Dr. Chodak’s post at 1/20 10:38 makes the point that at best, even if the USPSTF reversed itself and came out with a neutral or mildly favorable view of screening for prostate cancer, the recommendation would only be rated C, not enough for mandating coverage under the health care reform law.
Unfortunately, the wonderful good fortune of very long survival that the vast majority of us prostate cancer patients enjoy, compared to survival for all other major cancers, means that clinical trials to assess the value of screening are very difficult to execute. Moreover, ethical and cultural issues impinge on randomization that is standard in such probative trials, thereby creating additional serious obstacles. (We are certainly not willing or able to force people into accepting their randomized assignments, and we Americans are not overly keen on accepting the assignment requests of trial administrators.)
The USPSTF voting members and staff, like many of their medical colleagues, are accustomed to dealing with evidence from other cancers that have far shorter typical survival durations, short enough to enable practical testing with clinical trials featuring survival endpoints, thereby generating what passes the conventional wisdom test for solid, gold-standard type evidence.
While as a patient I like the fact that we are blessed with a time bomb with a very long fuse, I appreciate that this blessing creates a major problem with screening coverage policy, which depends on assessment of evidence. It appears we may never get the usual gold-standard evidence for prostate cancer. Therefore, why should we not advocate for a different standard of evidence for prostate cancer, recognizing that we must find ways around the “problems” of long survival, ethical considerations and cultural preferences? If we do not, will we not be forever trying to pound the round peg of flawed and unreliable clinical trial results into the square hole of USPSTF requirements for ratings A and B?
Politics needs to trump science when it is bad science. As a lay person, a recently treated prostate cancer patient (it’s probably too soon to call me a survivor), and Vietnam veteran, when I read the USPSTF report and comments, I see serious flaws in how the studies were conducted. My skepticism about all things government causes me to react that this is all about saving money for the government. My wife observed that the committee chair is a female pediatrician. Elsewhere I read that no one on the committee routinely dealt with prostate cancer. If the committee is truly concerned with preventing men from undergoing unnecessary treatment, they could surely have come up with recommendations that address that issue without condemning other men to painful deaths.
Doesn’t the American Cancer Society’s annual report show a continuing decline of deaths due to prostate cancer?
Yes, prostate cancer is more challenging because of the long natural history of the disease. But implicit in this whole controversy is a very basic issue about health and science. Which is better for society: Embrace and promote health initiatives that might be good for people until there is clear proof that they really do not work or help, or wait to endorse new efforts until science clearly shows us that more people are helped than harmed?
Think about the number of drugs, etc.,that have not been approved because proper studies showed they did not work or caused too much harm. If those treatments had been available to the public without conducting those studies, would society have been better off? I believe that society needs good science to guide us and if we endorse things in the absence of good proof then our society will be worse off, not better. While PSA screening has probably helped some unknown number of men, it is now quite clear that several million have been treated unnecessarily since its discovery and a significant percentage of them have been harmed. To borrow a quote from Spock in Star Trek, “Does the good of the few outweigh the good of the many?”
EVIDENCE FAVORING SCREENING, COURTESY OF THE PLCO TRIAL!
I have now studied the complete paper covering the 13-year update by the PLCO research team. I think I’m seeing tentative evidence that screening is making a favorable difference. Here’s the background and what I’m looking at.
As discussed above, follow-up is still far too short to be able to detect a difference in prostate cancer mortality in the PLCO trial. However, follow-up is more than ample to see a difference in stage at diagnosis, and that data is reported in Table 2, though with emphasis on treatment types and not on differences in confined versus non-confined cancer, which is not discussed. The table breaks out stages in a number of sub-categories, clearly displaying figures for Stage III and Stage IV. If screening makes a favorable difference, we should see a smaller percentage of men in the “organized screening” arm of the study diagnosed with Stage III and Stage IV disease compared to men in the “usual care” arm.
Staging was not available for a small percentage of men, 21 (0.5%) in the organized screening arm and 38 (1.0%) in the usual care arm. I deducted those numbers from the totals presented in Table 2 for each arm, yielding 4,229 men diagnosed in the screening arm and 3,777 diagnosed in the “usual care” arm. Stages III and IV account for 154 patients (3.7%), in the organized screening arm versus 176 patients (4.6%), in the “usual care” arm. In other words, 20% fewer patients in the organized screening arm had Stage III or IV disease at diagnosis!
We need to remember that the numbers of patients in each arm result from the randomized assignments to those arms, rather than from actual screening or non-screening. In fact, screening in the “usual care” arm was 52% during the screening period, and 45% of men entering the trial had had at least one PSA test in the 3 years before randomization in the trial. It would be valuable to get a display like Table 2 for men who had been screened during the trial versus those not screened, regardless of the arm to which they were assigned. I believe the results would show an even greater benefit than the 20% reduction in Stage III and IV disease at diagnosis. It would be possible to go further and move men screened in the 3 years before randomization but not during the trial protocol period from the unscreened tally to the screened tally. I suspect this would result in an additional increase in benefit. It is clear these data exist, and executing this analysis should be an easy task.
Has such an analysis been done?
If so, has the USPSTF or the PLCO team considered the analysis? I read the full USPSTF draft recommendation and the two PLCO papers, and my strong impression is there was no such discussion.
Jim:
No such analysis has ever been published, even if it has been done. It is not clear to me whether such an analysis is actually justifiable based on the study design. One can always carry out this type of analysis post facto, but whether it is justifiable is a quite different question.
Jim,
I have been a member of the data and safety board overseeing the PLCO trial, and every type of analysis has been looked at.
It is a curious but also surprising observation that despite the higher number of stage IV cancers in the control group, it did not translate into a survival difference. Given the relatively short survival of those men, a survival difference should have been seen by now.
Another unrelated point about the long natural history of this disease is that so many people have referred to the drop in the prostate cancer death rate starting in 1993 as being a direct result of PSA screening that began in earnest in 1989-90. If screening was really the reason for that drop, then the 600 extra cancers diagnosed in the screened group during the first few years of PLCO should have definitely led to a drop in mortality, which has not happened. The only real conclusion then is that the drop in mortality seen in 1993 cannot be a direct result of PSA screening.
DON’T WE NEED DIFFERENT SCIENCE STRATEGIES?
Sitemaster, Dr. Chodak, and other Link buddies,
Thank you for your replies of 1/26 8:30 am and 1/26 11:42 am to my post of 1/25 5:47 pm and attention to these issues. I am grateful, Dr. Chodak, that you are sharing your experience with the trial as part of the data and safety board. Indeed, one value I see in the results — one valid reason for early publication — is additional documentation of some mortality risk, real though very small, from treatment. I have some follow-on thoughts and questions.
Regarding Sitemaster’s comment, I too do not see how a post-hoc analysis of Stage III and IV diagnoses could be “justifiable based on the study design.” Certainly, whatever claims the original reports might have to sound use of inferential statistics would be absent in such an analysis. However, given the very long follow-up “period since diagnosis” needed to develop sound estimates, and given the great obstacles (such as “contamination”) that confound even the best plans, I’m thinking we may not have lost much in true meaning and may enhance insight by putting our Sherlock Holmes skills to work. After all, what fun is it to leave it all to the statisticians? I am convinced we need to work with what is possible, rather than what is ideal but impossible. Given the realities, especially the reality of long survival typical of prostate cancer, it just makes no sense to me to insist on elegantly planned and executed Phase III clinical trials for prostate cancer, ideally very large, very long lasting (and expensive in this eara of budget austerity!), randomized, double-blind and placebo controlled, with survival as the primary endpoint (disease specific and/or overall), for men with no diagnosed disease at outset (screening) or early disease. Such trials would work well for pancreatic cancer, lung cancer, and many other cancers, and of course for late-stage prostate cancer patients, but the vast majority of us prostate cancer patients have the blessing of very long survival.
It is possible there may soon be even greater challenges to execution of ideal trials given the emerging opportunity to detect, target, and treat with curative intent oligometastatic disease in recurring patients and those with advanced cases, a development, though tentative and far from proven, with potential to extend survival even further. (I am now in the process of qualifying for that option in my 13th year of survival with a challenging case.) One expert has come to believe that a great proportion of men with challenging recurrences and cases may be oligometastatic and potentially curable; if so, that could have a significant effect on survival averages in the near future, and the impact on research planning would be that follow-up in trials would need to be even longer, perhaps as long as more than 20 to 30 years after diagnosis.
Even at present survival durations, long trials run the likelihood that technological advances, such as the recognition of widespread and apparently often treatable oligometastatic disease, will affect their relevance. We don’t have to look far to see this. For instance, in the PLCO trial, the recent JNCI report breaks out figures for men diagnosed in each stage by type of treatment. Substantial numbers of PLCO patients were treated with radiation, but, as PLCO prostate cancer enrollment commenced in November 1993 and 92% of participants had been followed for 10 years by the end of 2009, with 57% followed for 13 years by then, most diagnosed men electing radiation were likely radiated in the eras when external beam doses were used that are now recognized as often inadequate for cure, and when undetected cold spots were a too-frequent problem with brachytherapy. Were they to have been treated with modern doses and techniques, cure rates would almost certainly have been higher, as would survival rates. (Granted, those are issues for research, but this illustrates the issue, does it not?)
I am so impressed with the large scale of the PLCO study (76,685 men enrolled) carried out for such a long period (already at nearly 19 years for the first enrollees). How extraordinary that this daunting research has been carried on for so long! While the limitations are clear, it is also clear that a monumental treasure trove of data has been accumulated. Why should we not mine that data? Dr. Chodak, would you not like to see an analysis of the effect of real screening, not intent-to-screen, versus real non-screening on diagnosis of Stage III and IV cancers? Granted, the evidence would have limitations that would qualify any conclusions, but I have a hunch we would achieve useful insight. (In addition to what was sketched in the post of 1/25 5:47 pm, it might be wise to try to filter out Phase III and IV cases that existed when patients enrolled. That makes sense to me as the objective would not be survival comparison but a comparison of Phase III and IV totals as affected by screening. I suspect that quite a few such cases were pre-existing.) It is possible that analyzing groups sorted into true screening or non-screening would show a disadvantage for screening. Though I think that quite unlikely, such a finding would be quite a challenge to what many of us have accepted as wisdom. Either way the outcome would go, I’m eager to know it.
Thank you for sharing your observation about the lack of a survival difference in the Phase IV cases. That’s food for thought. I’m speculating that the answer might lie partly in the numbers of men in each group who had Stage IV cancer at enrollment who were detected early in the study — men whose disease would not be affected by any benefit or harm from screening. Could another possible influence be a disproportionate number of men who were not compliant with the protocol, such as men in the “usual care” group who were experiencing symptoms that motivated them to be screened early? Your last conclusion makes good sense to me, but I’m thinking that part of the continuing drop in mortality in later years is due to screening.
I remain puzzled about one point in the discussion sections of the 2009 and recent PLCO reports: why wasn’t the short follow-up from diagnosis emphasized as a “limitation”? (Along the same line, why wasn’t there a table column on “time from diagnosis” for the various sub-groups?) That point seems as critical as it is obvious. To many of us, the authors ignored the elephant in the room, an elephant that appeared to be stomping the unfavorable implication for screening — the implication so bandied about by the media and central to the draft recommendation of the USPSTF — into dust.