How We Do Harm (33 page)

As a clinical-trials kind of guy, I wonder what happened to the two trials that were not reported.
I wonder why the data from seven very different trials was pooled.
I am used to seeing such things and would expect each trial to be presented and then what is known as a meta-analysis of the seven trials.

My greatest concern is that most of the trials had control arms that were not actively followed.
Some trial participants may not even have known they were in a trial.
The screened group, of course, knew they were being screened.
This creates a situation in which if the screened group was diagnosed they would be treated in one of the centers running the trial.
These centers are staffed by physicians who are interested in prostate cancer, many with specialized training in prostate cancer in the United States.
The control group, if diagnosed due to symptoms, was treated in the community.
European doctors generally are much less aggressive, and in some cases downright nihilistic, in their treatment of prostate cancer.

The lay media and some medical and advocacy groups have presented the ERSPC as a comparison of a screened group versus an unscreened group.
It’s not.
The ERSPC should be considered a comparison of a group of men who were screened and got American-style prostate cancer treatment if diagnosed versus a group who were not screened and got European-style treatment if diagnosed.
It is definitely not a comparison of screening versus nonscreening.

While I am critical of ERSPC and aspects of how it was portrayed in the media, I still think it an important study that suggests something that we are doing in prostate cancer—be it screening and/or some of our modern treatments—is beneficial.
It needs to be continued and needs to be watched carefully.
All nine studies need to be reported separately and as a collective meta-analysis.

*

HOW
can we identify the cancers that are a threat to a man’s life and need to be treated from those that are not a threat and need to be watched?

In the 1990s, many urologists believed that almost all prostate cancers in men under seventy-five needed treatment.
The publication of studies that showed that areas with high rates of screening had the same prostate-cancer death rates as areas of low cancer screening suggested that some prostate cancer did not need to be treated.

The Prostate Cancer Prevention Trial discussed earlier proved to many that there can be a diagnosis of cancer that does not kill.
Gradually, the prostate-cancer physician community is realizing that some kinds of cancer don’t need to be treated.

One problem is that few realize that our definition of prostate cancer was given to us by pathologists in the 1840s in Germany.
They used light microscopy and biopsies from patients who had died of prostate cancer that had spread throughout the body.
Technology has progressed over the past 160-plus years.
Now we can find that pathology localized to the prostate, and we assume that pathology is going to leave the prostate and kill the patient.

Yet sometimes that pathology is genetically programmed to stay in the prostate and never leave during the man’s remaining lifetime.
We can end up “curing” a patient needlessly.
In others, the pathology is genetically programmed to leave the prostate and kill or has already left the prostate.
What we need is a 2012 genetic definition of prostate cancer.

*

WHAT’S
a man to do?

I believe that a man should know what we know, what we don’t know, and what we believe about prostate cancer.
I have been concerned that many patients and physicians have confused what is believed with what is known.

Over the past year since the two randomized trials reported, I have started my screening talks by asking how men would feel about a hypothetical pill: if I had a pill that, taken daily, by men beginning in their fifties would double the risk of a diagnosis of prostate cancer but might decrease risk of death by 20 percent, would you take it?

Because doctors use relative risks and absolutes to their advantage in medical argument, meaning whichever sounds most impressive, I will give the relatives as well.

If the pill definitely doubled lifetime risk of prostate-cancer diagnosis from lifetime risk of 10 percent to lifetime risk of 20 percent, yet might decrease risk of death from 20 percent from a lifetime risk of 3 percent to a lifetime risk of 2.4 percent, would you take it?

I have posed this question a dozen times to several thousand urologists in all.
Few are willing to take the pill.

I then give my talk on the principles of prostate-cancer screening and conclude by telling the audience that while we do not have a pill with those efficacy numbers, those same numbers are drawn from the European screening trial for a man who chooses to be screened for ten years.
I find it interesting that no one wants to take it as a pill, but everybody pushes an annual blood test with those statistics.

I have long tried to get men to accept reality.
It’s important to understand that life is a crapshoot.
There is risk in all that we do.
We do not know that a man diagnosed and treated for prostate cancer benefits from the treatment.
We know that some do not because they did not need to be treated or because they die anyway.
We hope some benefit in that their lives are saved.

Men who understand the risks and the unknowns and who want to be screened anyway should be screened and not be criticized.
Men who don’t want to be screened should not be criticized, either.

*

IT’S
December 2009, just before Christmas.
I get a call from Ralph’s daughter.
My friend has been taken to the hospital with a fever, she tells me.
It’s a urinary tract infection.
Ralph’s daughter and I stay in touch.

Unlike Ralph’s many previous infections, this one progresses to sepsis, a widespread bacterial infection in the blood.
The battle is lost.
On the fifth day of hospitalization, Ralph dies.
The death certificate reads that death was caused by a urinary tract infection.
It doesn’t mention that the urinary tract infection was due to his prostate-cancer treatment and a radiation-induced fistula.

When causes of death are counted for cancer statistics in 2009, Ralph’s will not be considered a death due to prostate cancer, even though his death was caused by the cure.

He was seventy-seven.

Chapter 22

False Guidelines

MEDICAL SUBSPECIALTIES HAVE A TOOL
for making themselves seem indispensable.
It’s called
evidence-based practice guidelines.
These weighty words suggest that the patient is getting care based on rigorous science and formulated by luminaries whose judgment is not in the least bit motivated by self-interest.

Almost every medical travesty inflicted on my friend Ralph was based on evidence-based practice guidelines promulgated by medical subspecialties.
The only exception was the radiation treatment he received.
However bizarre and catastrophic, even that treatment choice was standard enough to warrant Medicare coverage.
(It was bad—shameful—but it didn’t create legal exposure for the radiation oncologist.
No one had the authority to challenge him, no one could stop payment.)
These doctors betrayed the oath and fundamental principle of medicine:
primum non nocere
—first, do no harm.
Their actions warrant double contempt because they benefited financially from doing harm.

According to practice guidelines promulgated by urologists, offering Ralph a PSA test represented good medicine.
The American Urological Association at that time recommended that screening with the PSA blood test and a digital rectal exam be “offered” to asymptomatic, well-informed men over fifty years of age with a life expectancy of at least ten years.

The word
offer
is important.
When a waiter at a party offers you a cold shrimp from a silver tray, you reach for it.
When someone you trust—your doctor, or a group of doctors—offers you something that’s supposedly in your interest, you accept.
That’s exactly what happened with Ralph.
He was offered screening at a health fair and he accepted.

With screening, you are redistributing risks and benefits across a population.
You have to perform a screening test on a certain number of people to detect a certain number of cancers in order to save a life.
You may be saving lives, but you accept the notion that you are also hurting people—possibly even killing some with interventions.

For example, recently published results of the National Lung Screening Trial—a $240 million study that compared low-dose helical computed tomography and standard chest X-ray in current and former smokers—found that screening decreases the chance of death by 20 percent.
The study, which was conducted by the National Cancer Institute and published in
The New England Journal of Medicine,
found that 53,454 people with the smoking history of at least thirty pack-years had to be screened to save sixty-two lives.
The patients were randomized into two groups.
Of the people who had abnormal findings in the CT group and went through diagnostic workup, sixteen died within sixty days of getting a follow-up procedure.
Of those sixteen, six had no disease.

Thus, follow-up to CT killed as many as sixteen people to save as many as sixty-two.
The number of deaths will likely turn out to be a bit higher as a result of radiation risks from CT screening.

At this writing, the guidelines for CT screening are yet to be written, but the general principles that will guide them are well established.
A person whose smoking history puts him at high risk may want to gamble on the potential benefits of screening.
A person who has never smoked, or a person who grew up in a household of smokers but never smoked herself, or a person who smoked for a short time and quit doesn’t have the same risk and doesn’t stand to benefit as much as a heavy smoker.
For people at lower risk of developing lung cancer, the chances of harm could exceed the chances of benefit.

When it comes to screening, a doctor who says, “Let’s err on the side of caution,” may actually err on the side of reckless ignorance and grave harm.

*

“SAYS
who?”
is the most important question to ask about a practice guideline.
What process was used in formulating the guideline?
Did the medical specialty that performs the procedure play a role in formulating the recommendation?

There is no shortage of guidelines out there.
Altogether, worldwide, at least three hundred organizations have issued over 2,300 guidelines.
For example, in 2011, the National Guideline Clearinghouse listed 555 practice guidelines for managing hypertension.
No doubt some of these guidelines are reasonable, and some are self-interested and harmful.
Many of these guidelines are commercial documents, yet no one promulgates good practices for their composition, no one rates them, no one regulates them.

“Many people pretend to be guideline makers,” says my friend David Ransohoff, author of a classic paper on methods to evaluate screening published by
The New England Journal of Medicine
in 1978.
“And there are no checks and balances.
It’s hard enough for a well-motivated physician to figure out.
It’s impossible for a civilian to figure out.”

When you look deeper at the way guidelines are written, the problem becomes more visible.
In 2000,
The Lancet
published an examination of the processes used to prepare 431 guidelines produced by specialty organizations.
Altogether, 67 percent of these guidelines didn’t describe the participants in the guideline making, 88 percent didn’t provide information on how relevant literature was identified, and 82 percent didn’t grade the strength of the evidence.
Only 5 percent met these three criteria.

“There are so many guidelines, and if they differ, they can’t all be right.
It’s like having several different judicial systems, some of which are not impartial and expert,” says Ransohoff, a gastroenterologist, epidemiologist, and cancer-screening researcher at the University of North Carolina.
“And these guidelines may render very different decisions.
How do you handle that?
That’s basically what is happening now.”

Some medical specialties are well-meaning.
Others are cold, calculating, pragmatic.
Self-interest, after all, is rational.
A doctor has to make a living.

Ransohoff is on a quest to help patients distinguish rational care from nonsense and huckstering.
His dream is to convince someone—perhaps a consumer protection group—to take on the monumental task of rating the practice guidelines.
So far, he has found no takers.
Admittedly, the task he is proposing is enormous, but it can be done—and should be done to prevent future medical disasters.

Ransohoff has a fantasy: he wants doctors to recognize when they are providing good science-based care, and he wants patients to know enough to demand nothing less.
I share this fantasy, and as the record shows, I am willing to fight to make it real.

*

THE
U.S.
Preventive Services Task Force, an independent panel that advises the U.S.
government, is arguably the most rigorous of the guideline-writing groups.
The USPSTF developed a detailed, rigorous, explicit, and transparent method to derive guidelines for all medical screening.
When you screen, you reshuffle the deck, altering the health risks of the people who go through testing.
Some of them could be helped by finding a deadly disease early.
Some will be harmed by cascades of medical interventions.
Some will be harmed by the mere knowledge of having an abnormal finding.
One guiding principle for judging such interventions is that “benefit must be shown to outweigh harm.”
Another is that strong evidence, such as the results from a randomized, controlled trial, is greatly preferred over weak evidence.
The USPSTF process for formulating screening guidelines takes years to complete.
This isn’t because it’s bureaucratic.
It’s because it’s appropriately complicated.