How We Do Harm (34 page)

The American Cancer Society guidelines are, for the most part, solid and are getting better.
The ACS prostate-screening guideline recognizes that the decision to screen could be appropriate if reached within the doctor-patient relationship.
However, the guideline in its current form specifically discourages participation in screening health fairs of the sort that triggered Ralph’s cascade of misery.

Most good guideline-writing organizations specifically
exclude
the subspecialties that provide care from membership on the guideline-writing panel.
For example, the USPSTF panel deciding on recent recommendations on breast-cancer screening specifically excluded radiologists, the specialists who provide mammography.
The objective is to rise above medical politics, self-interest, and self-delusion.
Should the person who performs the procedure in question be heard when guidelines are formulated?
Absolutely.
Should this person be involved in writing the guidelines?
Absolutely not.

A doctor who performs mammograms or colonoscopies could be sincere in weighing the evidence.
His or her biases don’t have to stem from greed.
They could come from performing a lot of the same procedures and not being in a position to assess harm.

Many doctors don’t know statistics or epidemiology, and all of us have conflicts of interest, maybe just from being enthusiastic about procedures and therapies we were taught to perform in medical training.
If you are a urologist and you want to perform a lot of surgery, you can find aggressive guidelines that tell you that it’s just the thing to do.
After that, you can proudly say, “I am following evidence-based practice guidelines.”

If you want to determine overall usefulness of a test such as PSA, you need to perform a comprehensive search of peer-reviewed medical literature.
Comprehensive review means going through every single study and grading it all based on strength of evidence.

Could Ralph have learned what he needed by surfing the Internet and reading everything that pops up?
The answer is simple: No.
You can’t figure it out as a patient.
It’s hard enough to figure it out as a doctor.

A rigorous discipline called quantitative clinical epidemiology has been developed to deal with the development of guidelines.
The word
epidemiology
means “upon people.”
Classical epidemiology taught in schools of public health deals primarily with origins of disease.
It poses questions such as “What causes AIDS?”
and “Do power lines cause cancer?”
Quantitative clinical epidemiology asks questions about diagnosis, prognosis, and response to therapy—things that clinicians need to understand to measure the impact of choices of medical services, such as the cascades that can be set off by an abnormal PSA test.

Clinical epidemiology is about cleansing yourself of prejudice, asking questions broadly, and getting reliable advice from people who have no financial stake in the process.
(In this case, urologists have a direct stake in the process, as do radiation oncologists, as do the good folks who make Viagra and Depend.) More than anything else, clinical epidemiology is about being ruthless with evidence as you separate what you know from what you don’t know and what you believe.

The USPSTF position on prostate-cancer screening is simple: “Current evidence is insufficient to assess the balance of benefits and harms of the service.
Evidence is lacking, of poor quality or conflicting, and the balance of benefits and harms cannot be determined.”

It’s an honest acknowledgment of uncertainty.
The task force recommends against screening for prostate cancer in men age seventy-five years or older.

“Older men, African-American men, and men with a family history of prostate cancer are at increased risk for diagnosis and death from prostate cancer,” the USPSTF guideline continues.
“Unfortunately, the previously described gaps in the evidence regarding potential benefits of screening also apply to these men.”

At this writing, USPSTF is in the process of lowering its letter grade for PSA screening from “I,” as in insufficient evidence, to “D,” as in don’t do it.
This battle has just begun.
A reader would do well to check the task force Web site,
www.preventiveservicestaskforce.org
, to see how it comes out.
It would be a crying shame if this Web site goes dark.

*

I
helped shape the American Cancer Society’s prostate-cancer screening guidelines long before I joined the society.
Our first prostate-cancer screening guideline recommended PSA for all men over fifty.

The words of Father Polakowski resonate in the guideline’s most recent iteration:
“The American Cancer Society recommends that men make an informed decision with their doctor about whether to be tested for prostate cancer.
Research has not yet proven that the potential benefits of testing outweigh the harms of testing and treatment.
The American Cancer Society believes that men should not be tested without learning about what we know and don’t know about the risks and possible benefits of testing and treatment.”

In March 2009, the American Urological Association (AUA) had to alter its guidelines.
The New England Journal of Medicine
had published the results of two studies that showed very different results on PSA screening.
A U.S.
trial showed that screening doesn’t work.
A European trial reported a marginally positive result.

Responding to this publication, AUA decided to strengthen its recommendation.
Men should now be screened at forty, a full decade earlier than the association recommended previously.
This struck me as a non sequitur, given the vastly different interpretation of trial results published in
The New England Journal of Medicine.

The guideline presented at the AUA 2010 annual meeting states: “The future risk of prostate cancer is closely related to a man’s PSA score; a baseline PSA level above the median for age 40 is a strong predictor of prostate cancer.… Such testing may not only allow for earlier detection of more curable cancers, but may also allow for more efficient, less frequent testing.”

Faced with data demonstrating harm and questionable benefits, the AUA doubled down.
The recommendation places U.S.
urologists in the position of advocating for the most aggressive screening measures at a time when two of the most important clinical trials of PSA ever done point to the potential for doing harm.

Turning to the list of panel members who drafted the AUA statement, I was puzzled by the outcome.
Smart people were involved.
Where in the world did that recommendation come from?
Looking over the justification, I saw that the recommendation to lower the screening age was based on one small study that focused on diagnosis of cancer.

As their livelihood was threatened, urologists gravitated to a relatively small body of literature that suggests that the PSA at the age of forty is somewhat predictive of whether one is at risk of prostate cancer later on.

I am concerned that adoption of this recommendation will cause a lot of men to be overly concerned that they are at risk for prostate cancer and will ultimately produce personal tragedies akin to Ralph’s.

*

INITIALLY,
the impetus for writing guidelines came from academia.
The entire endeavor was intended to make medicine more rational.
It was, in essence, anticommercial.
Yet, in the mid-1990s, commercial interests usurped the language of clinical epidemiology, making it impossible even for an educated person to distinguish a real recommendation based on science from a thinly disguised advertisement for medical services.

Seeing guidelines become the instrument for doing harm is particularly painful to my friend David Ransohoff.

Being a decade older than I, Ransohoff was at the forefront of what amounted to an insurrection of young doctors against the tyranny of venerable professors.
In the sixties and seventies, these young rebels started to think of medicine in terms of algorithms while waiting eagerly for computer technology to replace their beloved slide rules, making it possible to model medical decisions and gauge the outcomes of medical interventions.

These doctors sought to change the way medicine was practiced, making it possible to replace intuition and “common sense” with science.
The literature they were developing was entirely prescriptive.
Their colleagues weren’t demanding it, and some thought it was either threatening or absurd.
Indeed, how can anyone rely on algorithms—if-then statements—to practice medicine?
Are they trying to make doctors into robots?
How can algorithms—no matter how advanced—replace a doctor’s clinical intuition developed over decades?
Even now, doctors are free to ignore science, which they do, wrapping themselves in the language of pseudoscientific guidelines.

Ultimately, to get meaningful, enduring change in the system, we will need to return to the essence of the dreams of these young geeks and make the rest of the medical profession respect evidence.

That’s just one aspect of the solution.
Yet, change will have to come from below, which can happen when the American people realize that it will take popular action to make doctors practice medicine.
Supply would have to meet demand.

Before this can happen, let’s consider the evolution of science-based medicine, to enable the reader to distinguish it from medicine of self-interest.
After that, we will turn to an example of one patient advocacy group that came to understand the value of evidence-based medicine and started to demand it.

Chapter 23

Algorithms for Judgment

FOR RANSOHOFF,
the fundamentals of quantitative clinical epidemiology run deep.
He grasped the principles embedded in good guidelines at age five.
In 1951, William Ransohoff, his father, started taking David on house calls.
The boy would sit in a beat-up, old Chevrolet, waiting for his father to emerge from the house call and present a brief outline of the case.
This was a lesson in clinical judgment and understanding evidence.

“Dad would come out and tell me about a patient and say something like ‘The patient is having pain, but the real problem is that he is not getting along with a spouse, or that he lost his job,’” Ransohoff recalls.
“And Dad would want to understand the real problem and actually try to fix it.
Patients like it when doctors give them penicillin for sore throats, and doctors like taking credit for making them better.
But patients get better on their own.”
He thought there was something wrong with this system, and he wanted to know what worked and what didn’t.

The Ransohoffs are a German Jewish family that had been producing doctors who practiced in Cincinnati since the early nineteenth century.
David was conceived to join this tradition.
The elder Ransohoff had training in psychosomatic medicine, a movement in medicine that stressed the mind-body connection.
Encountering a patient complaining about a sour stomach, a physician like William Ransohoff will wonder whether this is due to poor diet or a sour marriage.
Sometimes, the elder Ransohoff would also see people as a psychiatrist.
Instead of rushing to make a diagnosis, he remained open-minded about what the real problem was, never going beyond evidence.

Where I rely on the guidance of Father Polakowski and his maxim, Ransohoff stands on his father’s shoulders.

*

AS
a medical student, Ransohoff was fascinated to rediscover the question he encountered while waiting for his father to return from house calls: what is clinical judgment and how might you teach it to people who are not doctors?
The need to break down clinical judgment into its elements was urgent.
In the late 1960s and 1970s, the country was experiencing a shortage of doctors, and physician assistants and nurse practitioners were stepping in to provide care that had traditionally been provided by doctors.

As a med student at Case Western Reserve, Ransohoff became aware of the work of a faculty researcher, Larry Weed.
In the late sixties, Weed developed a tool called the problem-oriented medical record.
His basic idea was to understand the evidence in an individual patient’s history by breaking it down into four components.

First, the subjective data: what the patient told you.

Second, the objective data: physical examination and laboratory data.
This may be stronger evidence.

Third, assessment: this is where you put the evidence together, synthesizing the subjective and the objective.

Fourth: the treatment plan.

“This was very attractive to me, because it was basically a set of rules about how to organize evidence: subjective, objective, what’s going on here?
What’s the differential diagnosis?
Have I been broad enough?
And in the plan, how do I narrow down the differential diagnosis?”
Ransohoff says.

Organization and rigor were key to cutting through the complexity and making the right decisions.
“One of Larry Weed’s cardinal concepts was never to state a problem at a higher level than you understand it,” Ransohoff says.
“Don’t go beyond the evidence.
If the problem is stated as ‘chest pain,’ don’t state it as ‘rule out myocardial infarction,’ because that is just one part of a plan, and labeling it that would narrow your thinking, and you might overlook an aortic dissection, a serious problem that requires specific thinking and tests to find.”

In the seventies, doctors interested in medical decision-making started finding each other.
Most had an interest in mathematics and a determination to move beyond Weed, sometimes creating quantitative models, such as decision analysis, to understand medical choices and their consequences.
Many of these doctors would go on to spectacular careers.

As an intern at Dartmouth, Ransohoff and one of the attendings, Hal Sox, would compete over which of them could be more compulsive in taking care of patients.
Sox would later become the head of the US Preventive Services Task Force and, after that, the editor of the
Annals of Internal Medicine.