Which may be why Walsh was so distressed when criticisms began to appear, first in the medical journals and then in the popular press. On the day I visited him, he took two calls from newspaper reporters, growing increasingly vexed as he answered their questions. It was possible to discern the surgeon's temper beneath his affability. He replaced the receiver with some vehemence. The criticisms, he said, were "misleading and biased." They were "really a step backward." They were "irresponsible." The press coverage was all too often "confused." Men are dying of this disease, he said, and many of them can be helped. "I can't believe they're suggesting that we do nothing," he said. His hand hovered over the desk as though he were thinking of banging it on the wood. "I find that incredible."
Clear Costs, Unclear Benefits
A test that lets surgeons treat a fatal disease - who could possibly object to such a thing? But that is precisely what a growing chorus is doing. The attack gained volume last May, when the Journal of the American Medical Association published two articles by the Prostate Patient Outcomes Research Team. Taking the PSA test, the team's John Wasson said, would be pointless if no effective treatment existed for the prostate cancer it diagnosed, or if the treatment were worse than the disease. A wellknown treatment for prostate cancer, radical prostatectomy, does exist; deciding whether it is effective turns out to be surprisingly complex. This means, Wasson argued, that the case for the widespread use of the test is muddy at best.
As any doctor knows, even the best treatment is not perfect. Aspirin, one of the safest and most effective drugs ever discovered, nonetheless managed to kill fortyfive Americans last year (at least twentyfive of the deaths, however, were suicides). The trick is to ensure that the benefit of the treatment outweighs the cost. In the case of aspirin the decision is obvious. If taken by the right people, this inexpensive drug could prevent perhaps a quarter of all first heart attacks, preventing thousands of premature deaths.
Radical prostatectomy for prostate cancer is a much harder call, because the cancer it removes may not be fatal, and because it takes a serious financial and personal toll: biopsy, surgery, hospitalization, possible major side effects, even death. A big, wellrun, randomized clinical trialthousands of people given surgery after a positive diagnosis, an equal number left untreated, the outcomes compared years laterwould go far toward resolving the issue. Nobody has completed such a trial, although the Department of Veterans Affairs is considering undertaking one and another has recently begun in Sweden and Finland. These studies will produce no results for at least a decade. Thus the critics are right: the increase in screening for prostate cancer means that this country is embarking on a vast experiment without being certain that the outcome will be beneficial.
In the absence of results from a clinical trial, researchers have tried to estimate the consequences of measuring the PSA of every eligible man in the country. The value of such projections is limited by the lack of data and the need for simplifying assumptions - for instance, that PSA tests are never given without an accompanying rectal exam. But they give an idea of how health professionals try to establish the desirability of medical procedures that may affect the lives of millions of people. According to the 1990 Census, there are about 23.4 million American men from fifty to seventyfive, a time of life that has been described as the optimal screening years. Not every one of these men should be screened for prostate cancer, because some have diseases that are likely to be fatal before prostate cancer would be, such as heart disease and other types of cancer. Figuring out how many men have such diseases is difficult, but one can make a rough approximation by looking at the most widespread of them - heart disease. A longterm study based in Framingham, Massachusetts, has by extrapolation shown that at least 2.8 million men aged fifty to seventyfive are afflicted with severe cardiac problems. Subtracting these people from the total number leaves 20.6 million potential candidates for screening.
To estimate the results of administering PSA tests to this many men, researchers have used data from pilot experiments. In the biggest screening project yet undertaken, William Catalona and three colleagues from Washington University tested the PSA levels of 9,629 older men, all at least fifty but some over seventyfive, whose physical characteristics were reasonably similar to those in the population at large (reasonably but not entirely similar - the researchers did not screen anyone with a history of prostatitis). Of these men, 9.4 percent had elevated PSA levels, suggesting a problem. Further diagnostic tests were performed on these men, and about one in three had cancer. If Catalona's findings could be extended to all 20.6 million eligible men, a nationwide PSA test would pick up almost two million people with high PSA levels. After biopsies about 1.3 million of them would be told not to worry (although some of them would have cancer that would be picked up on subsequent screening). About 640,000 would learn they had cancer. Of these, some 235,000 would have advanced cases that many doctors would regard as inoperable. Thus universal PSA screening would find, according to this rough estimate, about 405,000 cases of cancer treatable by prostatectomy.
In the past most of these smaller carcinomas would not have been discovered, and hence would have remained untreated. After ten years about two thirds of them would not have spread outside the prostate, if one is to believe the results of the largest study to date on the subject, in which six Swedish researchers tracked for an average of ten years 223 men with untreated early-stage prostate cancer. (This is standard practice in Sweden, where staterun medical programs are chary of procedures that have not been proved useful.) These results suggest that a significant proportion of the 405,000 tumors that would be detected in a national screening program would not spread in the patient's lifetime.