PSA redux: The USPSTF finally recommends NOT getting it!

.

The US Preventive Services Task Force (USPSTF), the independent group of physicians and scientists who make recommendations to the government, medical community, and American people on the value of screening tests, recently came out with a new recommendation on the use of laboratory tests for Prostate Specific Antigen (PSA) in screening for prostate cancer.  It recommended AGAINST it --in their terms, a “D” recommendation. Previously, USPSTF had recommended against PSA screening for men over the age of 75, but had not taken a position for or against screening in younger men (an “I” recommendation, insufficient evidence to recommend for or against screening).

My belief is that this is a good, appropriate, and very overdue recommendation which will come as no surprise to those who have read this blog for some time. I, and guest authors, have addressed this issue several times (PSA Screening: What is the value?, Mar 21, 2009; PSA Screening: “One of Medicine's Great Success Stories"?, Oct 27, 2009 (by Robert Ferrer);, Men’s Health? Women’s Health? Valid screening opportunities or “Hallmark Holidays”?, Mar 15, 2011). In addition I have often linked to and cited the work of Kenny Lin, MD, who writes the Common Sense Family Doctor blog, and resigned from the Agency for Healthcare Quality and Research (AHRQ) as a member of the USPSTF support team in November, 2010, over his perception that these recommendations were being delayed by political considerations. Dr. Lin has also written about PSA testing often  (including  "It is time to stop this [PSA] June 21, 2011, PSA testing: will science finally trump politics? Feb 28, 2011) and has recently addressed the new recommendations on Oct 7, 2011, Shannon Brownlee on the pros and cons of early cancer screening.

Of course, a lot of people do not think that this recommendation is a good thing. Two large groups, in particular, oppose the new recommendations: urologists and others who earn their livings treating prostate cancer and “advocacy” groups, supported by many high-profile (as well as just regular folks) men who have survived prostate cancer. Many of these men are quoted in Gardner Harris’ NY Times article “US panel says no to prostate screening for healthy men”, October 7, 2011. One of those who is quoted (actually not in the published NY Times piece, but in another version of Harris’ article published in the Seattle Times, is my colleague Brantley Thrasher, MD, Chair of the Department of Urology at the University of Kansas Medical Center, who said, "It appears to me that screening is accomplishing just what we would like to see: diagnose and treat the disease while it is still confined to the prostate and, as such, still curable."

I like Brant Thrasher, I think he is a good and knowledgeable doctor and great surgeon, but I strongly disagree with him on this one.  As much as we would like, and believe me as a family doctor I would like, and Kenny Lin would like, a test that could find disease early while it was still curable and make a difference in people’s live, PSA is not that test and, at this point prostate cancer is not that disease. These are two separate issues, so let’s take them separately.

PSA is not a good test. Yes, it is often, maybe usually, elevated in men with prostate cancer. Of course, in some men with prostate cancer it is not above the “normal” cutoff. This has led some advocates of PSA screening to suggest use of “PSA velocity”: check it yearly and watch the rate of rise rather than the absolute value. But the bigger problem for PSA as a screening test is that it is often elevated in men who do not have prostate cancer but just have a big prostate (“hypertrophy”, almost universal in men above a certain age), or even DO have cancer, but the very-slow-growing-that-is-not-going-to-kill-you-before-you-die-of-something-else kind, which is by far the most common variety. These men are subjected to ultrasounds, biopsies, and treatments that cause significant morbidity (impotence, incontinence of urine, and “radiation proctitis” of the rectum and anus, developing congestive heart failure from hormone treatment, to name a few) with no benefit.  Baylor physician and panel chair Virginia Moyer notes in the Times article that “This test cannot tell the difference between cancers that will and will not affect a man during his natural lifetime. We need to find one that does.” In 2010, Richard Ablin, PhD, who discovered the a prostate specific antigen (but not the PSA test) in 1970, called use of the test “a public health disaster” and “not much better than a coin toss.” (“The Great Prostate Mistake”, NY Times, March 9, 2010.

But the bigger issue is that there is no good evidence that treatment of any kind – surgical, radiation, hormonal – makes any difference in the outcome of prostate cancer. Surgeons like Brant Thrasher think it does, and he may be some day proven correct , at least in some circumstances, currently there is much more evidence supporting that it doesn’t than that it does. If you have the common, less-aggressive kind of prostate cancer, you won’t die from it, with or without treatment. If you have the rarer, highly-aggressive kind, you will probably die from it, with or without treatment.  The Times article notes that  “…advocates for those with prostate cancer promised to fight the recommendation. Baseball’s Joe Torre, the financier Michael Milken and Rudolph W. Giuliani, the former New York City mayor, are among tens of thousands of men who believe a P.S.A. test saved their lives.” They may believe it, but they are probably (I obviously don’t have access to their medical records) wrong. The test diagnosed prostate cancer, they were treated for prostate cancer, and they are alive. QED. But it’s false logic, an association that doesn’t demonstrate cause. If they are alive now, they would be alive (at least as far as the prostate cancer is concerned) without the treatment. And they wouldn’t have those “little” problems like incontinence and impotence that seem like a small price to pay for not dying of cancer, but are a big price if the treatment didn’t make any difference. The famous folks who have died of prostate cancer, like Frank Zappa, died despite treatment.

The Times quotes Thomas Kirk, of Us TOO, the nation’s largest advocacy group for prostate cancer survivors, saying “The bottom line is that this is the best test we have, and the answer can’t be, ‘Don’t get tested.’” He’s wrong. That is the answer. We not only need a test that can distinguish the “bad” kind of prostate cancer that will kill you from the kind that probably won’t, we need treatments that evidence shows makes a difference in survival and quality of life if you do have the bad kind. In the meantime, getting tested is likely to create more harm than benefit.

 .

Healthful Behaviors: Why do people adopt them? Or not?

While I am not a psychologist or psychiatrist, I am both a family physician and a person. As such, I have observed human behavior for a long time. I have noted some psychological behaviors seem to be very common in the people that I have met both personally and professionally.  I won’t say that they are “human nature”, since this phrase almost always refers to something that the speaker believes in or finds dominant in his/her environment, and is usually very culturally bound. However, they are common. One of these is the tendency to deny the magnitude of risk inherent in the risky things we do (or the risks we take because of things we don’t do). At the same time, we magnify the degree of risk inherent in the things we pride ourselves on not doing (or doing, when we see doing them as if beneficial, and not doing them as risky). As a corollary, we are likely to criticize those who adopt the risky behaviors that we do not, or do not adopt the beneficial behaviors that we do. That is, judging others is easy.

Perhaps because many of these potential risks are to our health and safety, these attitudes are common in health care and public health workers. Health professionals who do not smoke, and have never smoked, often severely condemn those who do. But alcohol? A little wine is good for you, right? Maybe, but it depends on who you are. If you have a tendency towards alcoholism, or are pregnant, or are going to drive, it is not good for you. Or for others. Public health workers can strongly advocate for wearing bicycle and motorcycle helmets, and using infant car seats, but it is just possible that once or twice they were late for something and drove too fast or too carelessly. And hopefully didn’t have an accident, but could have, and certainly increased their risk for it. From a risk/benefit point of view (fire trucks and ambulances and police aside), being late for work is NEVER a reason to drive faster or more carelessly; in fact, because there is a natural temptation to do so, conscious governance of that temptation is the beneficial behavior.

The utility of adopting a healthful, or not adopting an unhealthful, behavior is complex. It depends on the likelihood of something bad happening, how bad that thing is, and how many people it affects. So eating unhealthful food and not exercising is bad, but mainly for the person (and their immediate family) if they get sick or die. Smoking in public places, and even more, driving less than carefully or under the influence of alcohol or drugs potentially affects more people. Not immunizing your children because it allows you prevent a common but unpleasant effect (getting a lot of shots) and possibly a bad but extraordinarily rare long-term effect (whether real, like Guillain-Barre from swine flu shots or not, like autism[1] must be balanced against both the risk of their acquiring the disease and its sequelae, as well as the impact on the overall population that results if lots of children, not just yours, are unimmunized.

Not long ago I saw a patient in her early 30s who was pretty obsessed with getting breast cancer. She had no particular risk factors (no first-degree female relatives with it), but had previously talked a physician into ordering a mammogram when she was just 28 (it was normal), and wanted another one. We discussed the risk, but she was pretty fixated on breast cancer. We also talked about other risks, of much more concern to me than to her: smoking 2 packs of cigarettes per day, having 3 different sexual partners and rarely using condoms, and having untreated hypertension. I suggested, strongly but I hope appropriately, that all of these were much greater risks to her health than was breast cancer. I don’t know that I got through.

I imagine that it is pretty easy for health professionals to agree with me about the relative risks for this woman. Why she was so concerned about breast cancer rather than her real risks is another question. Some obsessive neurosis? Excessive effectiveness of breast cancer awareness advertising? I’d suggest that in large part it is about personal responsibility, about whether she would have to take action to prevent a bad outcome. If she were really worried about the risk from blood pressure, from smoking, from unprotected sex with multiple partners (and she should be), she would have to do something, take some action to change her life, to take medicine, to give up an addiction. This would be hard. On the other hand, since there are no clear behaviors she would need to change to avoid breast cancer, this is a safer – that is, less challenging – thing to be concerned about, to be fixated on.

Are the rest of us so different? Even those of us who have almost no dangerous or risky habits or behaviors (are there such? If we apparently have none, there is a fair chance that we might be suffering from obsessive-compulsive disorder, also a potential risk!) Besides, some of us may always take care to wash our hands when using the restroom (and even use our elbows to turn off the water, as I saw a very young man do in a public place the other day), but take the risk of riding our bicycles on public thoroughfares. Or we may practice what we believe to be healthful eating, and may regularly ingest herbs and give our children vitamins that there is little or no data to support doing, but not give them immunizations.

Reducing health risk is also impacted by societal memory, or the lack thereof. This has been examined in the case of abortion rights, where younger women who have grown up during a period when abortion was legal (if increasingly unavailable, largely resulting from the campaign of terror from violent anti-abortion forces) do not see the urgency of fighting to continue it. It also often true in the case of HIV/AIDS, where young people who did not grow up seeing all their friends die of the disease before effective treatment was available may find themselves adopting the same high-risk behaviors. Or for those who never saw the devastation of epidemics of pertussis or diphtheria, or of measles, or of awful outcomes from Hemophilus influenza infections, to not see immunizing their children as critically important.

In addition, when we as individuals have good outcomes (or don’t have bad ones) we may tend to think it is deserved rather than attributing it to good fortune. We haven’t had car accidents because we are good drivers, not because we are lucky. We think we are healthy because we bike to work, or “eat right”, not because we are young and in a low-risk group. When we are older, we may believe that we are less ill than our friends because we do healthful things like yoga or take certain herbs, not because we lucked out in not getting cancer (or being born into a family with resources who could feed us well and educate us and provide us with other advantages) See also Social Determinants, Personal Responsibility, and Health System Outcomes, Sept 10, 2010).

I am not going to say “let s/he who is without sin cast the first stone”. I would, rather, ask all of us to recognize that an honest appraisal of our own risk behaviors is a first step to understanding those of others, and to helping them, and helping our society, achieve greater health.

---

[1] Data on vaccines presented at the recent American Academy of Family Physicians (AAFP) meeting suggest the chance of an adverse vaccine outcome is approximately equal to the chance of winning the lottery, and that of dying from a vaccine about equal to spontaneously having quadruplets.

Risk, Primitive Reactions, and Human Health Behaviors

.
NPR’s “All Things Considered” recently reviewed several scary events from 2010 (“The year in fear: fright or fallacy?”). Reporter Jon Hamilton spoke with Dr. David Ropiek, Director of Risk Communication at the Harvard Center for Risk, about what made these events (Toyota’s acceleration problems, the Deepwater Horizon oil spill and the use of chemical dispersants, etc.) particularly frightening. Dr. Ropiek said that people tend to make decisions, and react positively or negatively, based on very simplistic (and usually unconscious) criteria, rather than on careful critical analysis of the relative benefits of one course of action over another, with the most important criterion being “is there immediate danger?” The reason he gave was that our basic neurobiology was unchanged over human history while our culture and society was remarkably more complex than when quick decisions were mostly about achieving immediate results (fight or flee). “We use a risk-perception system that evolved in simpler times, when the risks were bad guys with clubs, and the dark, and wolves. It's quick. But quick isn't necessarily the best for the complicated stuff we face in modern society.” Thus, for example, even though the evidence would show that the human and environmental danger of the oil spill was in the oil much more than any risk from chemical dispersants, “Just the word chemicals in your listeners' minds is currently setting off a little organ in their brain called the amygdala, which is the 24/7 radar in our brain that says - is there danger in that data?”; that is, our fears are triggered by the word (chemicals) which we have come to associate with danger.

Similarly, people can grasp the specific, and feel the pain, for an individual more easily than for a large, amorphous population. Thus, the outpouring of concern for “Baby Jessica” falling down a well in 1987, or for the child dying of leukemia, is much stronger than that for thousands of people, especially those in other countries, dying of war, disease, or even more abstract, structural violence. It is not just the one versus the many; it is the suddenness of it. We feel for the trapped Chilean miners, or the victims of a bombing; Ropiek says “… a chronic risk doesn’t ring our alarm bells the way a catastrophic, all-at-once one does. Because it concentrates the mind to see a bunch of the tribe all whacked at once.” So a particularly gory battle or atrocity is horrifying, but when there are chronic, repeated bombings and battles (as in Iraq or Afghanistan), even though they lead to much more death, we feel less.

We can see a murderer as a bad person, but it is harder to identify the members of the “grifter class” (coined by Matt Taibbi, “Griftopia”[1] ) who are responsible for the financial system that has visited so much evil on all of us. When people hear about something they know little or nothing about, especially if it is very complex and hard to understand, they often deal with it by putting a “frame” around it, tying it to something that seems similar enough (at least in one dimension) that they feel they can hang their hat on the analogy and judge it. For example, “chemicals=bad” in the example above is such a frame; so is dealing with universal health insurance by framing it as “socialized medicine”=”socialism”=”bad”. Unfortunately, the world is far more complex than this, and more unfortunately unscrupulous politicians and opinion-makers (my frame = “selfish evil people”) take advantage of this to obscure complexity and buy into often nonsensical self contradictions (taxes=bad, deficits=bad; let’s not have either!)

When it comes to health and medicine, the same issues come into play. People perceive immediate distress with acute problems (e.g., cough, fever, and most especially pain!) and know how much they would appreciate relief. The impact of conditions that do not cause appreciable symptoms right now but will cause really bad outcomes (death, morbidity, poor quality of life) if untreated in the future, are much harder to get people to make high priorities. The doctor sees untreated hypertension in terms of a future outcome (stroke, kidney failure), but this is more difficult for the patient. Even when s/he believes and understands it intellectually, it is much less likely that the treatment of a largely asymptomatic condition will rise to the top of life’s many more urgent priorities (food, clothing, housing, childcare, work) than if it were, say, pain.

The problem is even greater for public health, as I discussed in Public Health and Changing People's Minds (Saturday, May 15, 2010) where populations are huge, timelines are long and risk is relative. Public health addresses risks for populations, not me, or my family; translating population risk into individual prior probability is fairly difficult. For most people, even the concept of risk – that a given event will not definitely have or definitely not have a particular result, but will be somewhere on the continuum between them – is something they are not accustomed to thinking about, although they use it all the time (deciding whether to cross on a red light, for example). Consciously comparing the relative risk of different actions is very difficult, especially when the results have very different timelines. A definite immediate benefit (have that tasty fried or sweet food; throw a wrapper out the window, get a big gas-guzzler, have unprotected sex) has a lot more weight than the possibility of a bad long-term outcome (besides, next time, in the future, I’m going to go on a diet, give up smoking, use condoms). Dr. Ropiek notes that because events that cause “a bunch of the tribe to be all whacked at once” happens relatively rarely, “…we tend to downplay chronic risks like car accidents, diabetes, heart disease and the flu.” Sometimes public health officials can create that fear and mobilize the attention of the populace, as with concern about the swine flu of 1976, but that is also an example of how, when predicted risk of bad outcomes doesn’t happen, it reinforces the tendency to downplay those chronic risks.

In making decisions about medical care, this sort of perception can cut either way, depending on how a person looks at it based on personal and familial experience, cultural beliefs, and the way they “frame” medical interventions, as well as how urgent or important a solution is. Some people do not trust doctors or medicines, based on these criteria, and prefer to not take medicines or advice, even when an analysis of the relative risk shows the treatment to be definitely beneficial. Others have unrealistic expectations of what medicine can do (fueled, of course, by both doctors and direct-to-consumer drug advertising), and are angry when the doctor cannot cure their viral illness, make their back pain disappear, or compensate for all of the other parts of life that are bad and make them happy. At times of serious illness, where both treatment and non-treatment have real risks, or at end of life when people are not ready to accept that it is the end of life, even a professional evaluation of relative risk/benefit is difficult, so it is hardly surprising that people return to simpler methods of decision making (will I be able to live another day? Will it end my/his/her pain?).

Hamilton ends the interview segment with: “So Ropiek says we need to acquire a new fear - the fear of getting risk wrong.” I wish us luck on that.


[1] Taibbi, M. Griftopia: Bubble Machines, Vampire Squids, and the Long Con That Is Breaking America. Random House. New York. 2010
.