By Bernadine Healy, M.D.
One of the most sobering pieces of information from a recent prostate cancer screening study—one of two studies released last month that involved the blood test for prostate-specific antigen—is that to save one life, more than 1,000 men would have to be screened and 48 treated. That's a lot of screening, expense, and potentially unnecessary treatment. Is it all worthwhile? As these PSA studies show, answering that question—the question at the heart of what medical researchers call comparative effectiveness—is not nearly as easy as we might wish. Yet comparative effectiveness is critical, not just in evaluating cancer screening tests but in achieving successful health reform.
The new prostate studies are relevant for patients trying to decide whether to get PSA screening. But they give neither a simple thumbs up nor thumbs down. That underscores the need for clinicians and patients to make many subtle judgments, based on the evidence available. For example, urologists have developed criteria to identify which prostate cancer patients could safely pursue watchful waiting, delaying treatment if not eliminating it altogether. But a one-size-fits-all approach to screening and treatment, which is a feature of socialized healthcare systems, does not always allow those nuanced judgments. Should we continue with the imperfect and keep plugging to make it better or suspend the good until the perfect arrives? These are not scientific decisions but ones of policy. They are sure to stir debate.
A growing chorus of health services researchers in this country have challenged the value of PSA screening to both public health and individual men. The U.S. Preventive Services Task Force, composed of health policy and preventive medicine experts who advise the secretary of health and human services, concludes that the evidence is not strong enough for doctors to make recommendations to patients. Inform patients of the pros and cons, it advises, and let them decide. In contrast, based on the same body of research, urologists and cancer groups recommend routine screening for men over age 50 and in the 40s for those at high risk, including African-Americans and those with a family history of prostate cancer. How about the 47-year-old boss of HHS and patient in chief Barack Obama: What should he do?
It's a general dictum of oncology, dating back to the ancients, that the best way to cure a cancer smoldering in an organ is to find it early and remove it before it invades surrounding tissue or spreads to distant sites. Most cancers hide when they are young (and most curable), grow silently, and produce symptoms only after they have taken a strong hold. That explains the push to screen for and find tumors before they cause symptoms. The logic behind screening does not always translate into effective medicine, however, when cancers that could safely be left untreated get removed.
But we have to care for patients now, with the imperfect science we have at hand. Prostate cancer mortality rates have plummeted in the United States over the past 20 years, coinciding with the widespread use of PSAs. (No such drop has occurred in Europe, where PSA screening, by policy, is uncommon.) This suggests—though it certainly doesn't prove—that PSA screening saves lives.
That's why the two just released studies, one from Europe and the other from the United States, have been so eagerly awaited. But despite their similar intention, they have turned out to be, by accident of place, very different, and confusing at first glance.
In the European trial, men were screened once every four years (not yearly, as is standard in the United States). Within a decade, mortality fell by 20 percent with screening. The study was conducted in several centers in Europe, and each came up with the same answer: Screening was associated with a decrease in death rate from prostate cancer—even though, because it's a slow-growing cancer, a decade is still too short a time to get the real picture. (Further follow-up is in progress.) The study also found evidence that those who were screened had significantly less spread of tumor into bone, which can become an excruciating problem.
This might seem at odds with the American study, which detected no effect from the screening performed within the trial. Not so. Through no fault of the investigators, the U.S. trial veered way off course from its original goal to compare men who were and were not screened by a PSA test. The cause: More than half of the men assigned to the group that was to have no screening ended up getting PSAs from their personal doctors, who were not involved in the study. What's more, some who did get their PSAs outside the trial may well have received better screening than that performed in the trial, which, frozen to 1993 standards, was less good at picking up early tumors.
Today, PSA screens are even better than those used in either study. Researchers have learned that there is no single threshold PSA level to be applied to all patients. (The thresholds used to trigger a biopsy were 4 nanograms per milliliter in the U.S. trial and 3 ng/ml in the European study.) For one thing, PSA increases naturally with age. And some men can have even highly aggressive cancer with levels below 3 ng/ml. According to Inderbir Gill, professor and chairman of urology at the University of Southern California, a sudden rise from 2.0 to 2.5 ng/ml over a period of six months to a year can signal a cancer, and the velocity of PSA change is more useful than any isolated measurement. Today's PSA screening incorporates these considerations into the decision to perform a biopsy.
To be sure, any ongoing randomized clinical trial can suffer if real-world medical advances get ahead of what's being studied. In that regard, comparative effectiveness research, which strives to integrate benefit, safety, and cost, will have to come as close as possible to measuring medicine as it's practiced now, not yesterday. And it must be nimble enough to assess innovations that can emerge quickly. That means embracing many forms of acquired knowledge, including basic research, scientific observation of large groups of patients, and so-called case-control studies, which compare patients who are alike in all respects except the one under evaluation.
Patients don't want to hear only about population averages; they want to know about people who are most like them. And they deserve guidance from doctors. Getting back to the patient in chief, I asked Dr. Gill for his recommendation. His response, without a moment's equivocation: "All 47-year-old African-American men must have PSA testing done, even if they are not president of the United States."
More from Heart to Heart: What to Make of the Prostate Cancer Screening Studies