How We Do Harm (11 page)

Neither Lilla nor millions of other cancer patients—particularly those with breast and head-and-neck cancer—had any reason to suspect that the drugs they received for counteracting anemia were doing real, measurable harm.

Lilla was simply taking the treatment her medical team suggested would ease her situation.

*

IN
2003—the year Lilla first received Procrit—my skepticism about that drug was in part instinctive and in part based on evidence.
I simply was not convinced that these drugs were worth the potential risks, and J&J’s aggressive advertising campaign only increased my skepticism.
Direct-to-consumer ads during the Super Bowl and the evening news makes my prescription pad retreat deeper into my pocket.
As a physician, I need to see the data, not advertising.
Ideally, I need to see a comparison of the results of multiple patients receiving Treatment A and an equal number of patients receiving Treatment B or placebo.
That’s the gold standard in the hierarchy of evidence.

When pressed, I can make do with trials that don’t have a comparator arm (by which I mean trials that don’t have patients that are receiving Treatment B).
They are less reliable, but better than nothing.
The opinions of my esteemed colleagues don’t hold much water with me because we doctors are routinely proven wrong.
The anecdote—the experience of a single patient—is less valuable still.

Direct-to-consumer ads are the least valuable way of describing the true value of a treatment.
Television spots don’t even pretend to reflect the experiences of real people.
They are scripted and staged to convince cancer patients to pressure their doctors to prescibe drugs that the patients and their insurers will pay for.

I saw no point in asking my patients whether they were feeling fatigued.
It’s the sort of question that suggests—indeed predetermines—an affirmative answer.
You ask twenty-five healthy people if they are tired, and the majority will say yes.
And so J&J’s Super Bowl advertising was selling a problem as much as a solution.
In fact, the company manufactured a medical condition: cancer fatigue.

What cancer patient isn’t going to report feeling fatigued?
Their bodies are being undermined by both the disease and interventions seeking to control it.

Though I didn’t specifically follow all the literature on anemia drugs, I did read the thirty-plus-page insert, a document that summarizes clinical data and other materials on drugs approved by the FDA.
(Few doctors read the entire label before prescribing a drug.
I do.) In 2003, the data in the Procrit label did nothing to calm my misgivings.

*

I
don’t claim to be uncommonly insightful, but I am a good student, and I am consistent in interpretating evidence.
It’s always about the balance of what I know, what I don’t know, and what I believe.

The man who taught me the most about evaluation of evidence, Father Richard Polakowski, had never been to medical school.
He taught eleventh-grade English at the University of Detroit Jesuit High School.
Father Polakowski had a maxim that no student who took his class could miss: “Say what you know, what you don’t know, and what you believe—and label it accordingly.”

When I took his class in 1975, Polo (Father Polakowski’s nickname) was about fifty.
A slight man, he retained a childish, lisping manner of pronunciation.
Polo taught a heavyweight curriculum: Chaucer, Shakespeare, Elie Wiesel, Chinua Achebe, Yasunari Kawabata, Isaac Bashevis Singer, Graham Greene, Richard Wright.

By year’s end, all of us became experts in applying Polo’s maxim to literary criticism.
Use your intellect, gentlemen.
Start with knowledge, find its boundary.
Do not stop!
Save room for belief, but examine it fearlessly, for genuine examination knows no limitation.

Polo concluded his multicultural blitz with
A Portrait of the Artist as a Young Man,
James Joyce’s reflection on his Jesuit education.
“Some people read about it, gentlemen,” Polo said.
“You are
living
it.”
You could see how proud he was to teach a criticism of the institution he was a part of, the institution he loved.
It takes a special combination of humility and pride to invite a group of would-be young artists—or at least young contrarians—to tear apart their teacher, their school, the religious order that runs it, and, if they so choose, God Himself.

Examination thus practiced can be deeply unsettling.
A priori, you have to accept the consequences of acting on your conclusions, whether these include a crisis of faith, a bleeding ulcer, an excommunication, or an insurrection.

Polo’s maxim quickly turned into my favorite device for exploring the universe.
Over the years, I’ve applied it to everything I do, from epidemiology to the design of clinical trials to the discussion of treatment options with my patients.
Much of what we do in oncology requires balancing benefits against harm.
Harm can include conventional metrics such as blood clots, but also cost to the patient and society.
Sometimes we treat one hundred people to benefit ten.
All one hundred will be subjected to harm, and ninety people will pay the price for the lucky ten.
Patients need to be informed about uncertainty in order to decide to roll the dice or sit out the game.
If you truly respect the patients you treat, you will not obscure the line where your knowledge stops and your opinion begins.
It’s the only decent thing to do.

I consider myself religious, but not overly religious.
Polo convinced me that God expects us to work for social justice, and the best way to serve Him is through caring for others.
Some people praise God by going to church on Sunday.
I seek to do the same daily by helping those in distress—and by telling the truth.

*

FROM
the outset, the things I knew and the things I didn’t know about Procrit—and another anemia drug, Aranesp—made me believe that I was observing a systemic failure in medicine.

The evidence that led to Procrit’s approval had been remarkably—spectacularly—unconvincing.
The FDA approved the drug for the treatment of anemia in cancer patients in 1993, ten years before it was offered to Lilla.
The approval was based on data pooled from six remarkably small studies that altogether enrolled a total of only 131 patients.
Generally, the FDA doesn’t approve drugs based on pooled studies, which are also called meta-analyses.
This is because a meta-analysis can be only as good as the studies it pools.
In the case of Procrit, the agency rolled the dice.
The nation’s blood banks were threatened by the AIDS epidemic, and the prospect of finding a hormone that would induce the body to produce its own hemoglobin was enticing.

Yet, even if the Procrit data
had
come from a single study, the total number of patients would have been unconvincing.
That’s because the patients in these studies had a variety of tumors, which presented another problem.
There was no way to measure how Procrit behaved in this mixed bag of diagnoses.
Would its impact on a patient with early-stage breast cancer differ from its impact on advanced colon cancer?

The six minuscule trials from which the data were pooled had not asked the questions I needed answered before I would reach for a prescription pad.
The studies asked only whether Procrit had the ability to prevent blood transfusions.
(Procrit accomplished this well enough to get on the market.) However, before I would be convinced to start giving this drug to my patients, I needed to know about Procrit’s toxicity as well as its impact on survival, or at least on the tempo of cancer’s progression.
Did people who received Procrit face a higher risk of blood clots?
Did their underlying cancer recur?
Did they die sooner?
Was their anemia corrected?
Did the answers to these questions differ from cancer to cancer?
These questions were unanswered.

Reading Procrit’s label critically, I saw that not a shred of data said anything about either “fatigue” or its opposite, “strength.”
Those words—which were central to the direct-to-consumer ads—were simply not on the label.
Yes, anemia is associated with fatigue, but would correction of anemia—pushing the hemoglobin levels into a higher range—alleviate fatigue?
There was no way to know, certainly not based on data.
The claims of giving patients “strength for living” were completely, blatantly unsupported.
I couldn’t understand why the FDA had not pulled the plug on the J&J ads.
(The answers would emerge later, when a congressional investigation would unearth evidence that FDA scientists attempted to stop those ads.
However, pro-industry attorneys appointed by the Bush administration to run the agency’s top legal office blocked the scientists’ efforts.)

The manner in which many of my colleagues prescribed these drugs made me increasingly uneasy.
The problem I saw was overtreatment, and overtreatment equals harm.

If you accepted that Procrit was equivalent to a blood transfusion (which we had no evidence for), it would make sense to use it in a way that mirrors the use of blood transfusions.
Normally, we transfuse patients who are suffering from significant anemia.
No blood bank will release blood for a patient whose hemoglobin is above 8.5 g/dl.
However, an ever-increasing number of my colleagues considered it justifiable to give Procrit to patients whose red blood cells were at much higher levels.
Lilla got it at 10 g/dl.
Generally, doctors thought it was reasonable to give the drug to patients whose hemoglobin was sliding below 12 g/dl.
Some didn’t bother to check what the patent’s hemoglobin was and erred on the side of giving Procrit every time they gave chemo.
This was spectacularly efficacious for the doctor’s wallet, but harmful to the patients and the health-care system.

Furthermore, doctors routinely prescribed the drugs for uses in which it had not been studied—such as anemia caused by cancer itself, as opposed to anemia caused by chemotherapy.
I know of a case where a patient was sent to hospice with a supply of syringes filled with these drugs.

At the time Lilla got her first injection of Procrit, J&J and a competitor, Amgen Inc., were pursuing studies aimed at pushing the cancer patients’ hemoglobin to “near-normal range,” which they defined as 14 g/dl and above.
These were exceedingly puzzling clinical trials, in part because no one can say with certainty what the normal level of hemoglobin should be.

Fatigue is in the eye of the beholder.
It’s hard to measure it objectively, and asking people whether they are feeling tired will not do the trick.
I know that many of my patients are not aware that their hemoglobin is at 8 g/dl.
Healthy pregnant women do extremely well at 7.5 g/dl.
A Denver resident can have a hemoglobin level a full point above that of her cousin in Wichita, Kansas.
This happens because the air in the Mile High City contains less oxygen than the air on the Great Plains.
Would this justify putting Procrit in Wichita’s water supply?

No scientific study has shown that a particular level of hemoglobin is important.
As a clinician, you have to consider these levels based on the patient’s characteristics.
A patient with coronary artery disease usually experiences a shortness of breath as she walks up a flight of stairs when her hemoglobin drops to 8.5 g/dl.
However, for a fifty-year-old woman in good shape, the 8.5 g/dl level is no problem unless she decides to run a mile or two.

Hemoglobin’s importance isn’t lost on athletes, which is why these same hemoglobin-building drugs are often used as an illegal doping agent.
We need oxygen to perform better.
We also know that our bodies adjust to variations in oxygen levels.
This is why football teams often go to Denver a couple of days before playing there.
Practicing at that altitude for three or four days helps players adjust (though it takes two weeks or so to become completely acclimated).

But Lilla wasn’t competing in the Tour de France or playing football in Denver.
No data suggested that her self-reported “tiredness” had any clinical significance or that increasing her hemoglobin level would make her feel better.
Since nothing was known about the potential for improvement, there was no way to weigh it against potential costs, which include side effects and the cost of treatment.

No matter how I looked at it, I couldn’t justify prescribing these drugs to a patient like Lilla, who was not presenting with any symptoms of anemia and was doing reasonably well.

Why was she prompted to say that she was “fatigued” only to be given a drug that was not proven to have the ability to correct fatigue?
Why were these doctors treating a number on a chart instead of treating Lilla Romeo?

*

LILLA
believed she received Procrit roughly four times a year.
Each time, she needed at least two shots to get her hemoglobin to 12 g/dl.

In 2004, she was told that the hospital had switched from Procrit to another drug, Aranesp, which supposedly had a longer-lasting effect and therefore required fewer shots.
Lilla could tell the difference between the two drugs.
Aranesp caused a burning sensation under her skin at the injection site.

“I remember saying that it stings, and I would much prefer Procrit,” Lilla recalled.
However, she was told that Procrit was no longer available, as the hospital had shifted to the Amgen drug.

As a marketing expert, Lilla would have been interested in the nasty warfare between J&J, the company that sold Procrit, and Amgen, the company that sold Aranesp.
She might have been even more interested in examining the intricate financial incentives Amgen had created to induce oncologists to prescribe more Aranesp.

In oncology, it’s rare to find a drug that is used to treat a variety of tumors.
Since areas of cancer treatment are becoming increasingly specialized, it’s unlikely that we will see another Taxol, a drug used for lung, ovarian, breast, and head-and-neck cancer, as well as advanced Kaposi’s sarcoma
.
It’s also unlikely that we will see many drugs that can be used to treat all patients with a particular tumor, such as all breast cancers, all colon cancers, or all lung cancer.
Many of today’s drugs are developed to treat minuscule numbers of patients, and as development efforts become more focused—say, on diseases associated with particular genetic characteristics—it will be common to see drugs that may help a thousand or two thousand new patients a year.