THE effort would be very expensive. PSA tests and rectal exams cost about $80 together; biopsies cost perhaps $350. If 20.6 million men had PSA tests and rectal exams and two million had biopsies, the bill would be in the neighborhood of $2.3 billion. The average cost of a prostatectomy is about $20,000, according to Kit Simpson, a healthpolicy specialist at the University of North Carolina at Chapel Hill; 405,000 of them would cost $8.1 billion. Treating the impotence or incontinence caused by surgery can cost $10,000 per patient. The total price of remedying such side effects would depend on how often they occurred, but in 1990 an economic study by a statistician and a urologist at the U.S. Army medical complex in Fort Sam Houston, Texas, projected that the extra expense for side effects might reach $2.1 billion. At the time, the two men pointed out, the national bill for prostate cancer was $255 million. An immediate nationwide program of screening all men in the proper age category, they said, would drive the price up to $28 billion - a hundredfold increase. To critics, such figures suggest that a screening program could sentence hundreds of thousands of people to immediate and costly suffering in the name of avoiding a disease that will kill only a small number of them many years in the future.
Nobody I have spoken with thinks that the United States could examine and treat 20.6 million men cheaply and painlessly. But proponents of mass PSA screening argue that estimates like these are based on inaccurate data. Claims that side effects from operations in the future will be at the level of those in the past ignores evidence from surgeons like Walsh and Catalona that the nervesparing technique is an important advance. Instead of a death rate of one in a hundred, these surgeons report death rates of zero to one in 600. Walsh and Catalona say that instead of seven out of a hundred becoming completely incontinent, and many more having problems, not one of their first 1,200 patients lost all urinary control, and only six to eight out of a hundred had any problems whatsoever with incontinence. Although these figures should not be taken as national averages, using them in calculating the cost of treating side effects would change the projections drastically. Indeed, it was the publication of the pessimistic estimate by the Prostate Patient Outcomes Research Team, which made little allowance for the improvement in surgical techniques, that caused the calls from the newspapers which so exercised Walsh during my visit.
Proponents also accuse critics of skewing their figures by overestimating the costs (such as saying that many seventyfiveyearolds will have the operation) and underestimating the benefits (failing to include the avoided years of pain from bone cancer, for instance). A prostatectomy today may cost $20,000, but treating that same cancer at a later, more advanced stage may cost $70,000. Most important, although many men diagnosed as having prostate cancer will never feel any effects from the disease, failing to treat the rest would sentence thousands of people to a death of rare awfulness. "Every fifteen minutes an American man dies of prostate cancer," Catalona says. "And when that happens, he's been through hell. None of them have a pleasant death. There's a long period of pain and agony with broken bones, urinary obstruction, constipation from all the morphine - it's horrible."
"The screeners say, How can you deprive all these people of a valuable treatment?" says Gerald Chodak, a urologist at the University of Chicago School of Medicine. "The other side says, How can you do all these things to people and not know if they really have any benefit?" It seems impossible to resolve the issue rationally, because the evidence is of such poor quality. The American Board of Urology and the American Urological Association have no idea, for instance, how many surgeons have been trained in Walsh's techniques. Surgeons like Catalona and Walsh assert that many of their colleagues know the new methods; outside observers like Wasson suspect that they don't.
Similar uncertainties dog the larger question of whether radical prostatectomy in fact saves lives. Most of the reports in favor of radical prostatectomy come from surgeons like Walsh and Catalona, who collect data on their patients. Alas, these articles suffer from "publication bias" - the tendency for people with good results to publish their data more often than those with bad results. "You won't see articles written by people with poor results," Chodak says. "They're put out by honest people who are superb surgeons, but the problem is that they may be such excellent doctors that they are totally unrepresentative of the average doctor that the average Joe is going to face." The rosy reports, he believes, may lead patients without access to topquality medical care to a falsely optimistic picture of the odds they face.
In the view of Brent Blumenstein, a medical statistician at the Fred Hutchinson Cancer Research Center, in Seattle, publication bias is merely part of a larger problem created by the lack of emphasis on research methodology during the surgicalresidency period. Unschooled, surgeons fall into many logical and statistical traps when they do research. Often, he says, they do not list the criteria they used in selecting patients for a study, invalidating their work from the outset. Several researchers have told me that data that call into question the use of surgery tend to come from studies with a disproportionately large number of old patients, who could be expected to have more side effects, and that data favoring it tend to come from studies with too many young patients. Carefully interpreted, such partial reports could be useful if they were put together with forethought. Unfortunately, they rarely are. Last May the Prostate Patient Outcomes Research Team reviewed more than 1,600 journal articles published since 1966 on the treatment of prostate cancer. Only 144less than a tenthwere detailed enough to permit independent reexamination of their data. Of these only one clearly described the patients' ages, to what point their cancers had progressed, and other medical conditions, and also gave the number of people who failed to attend followup appointments, although such information is supposed to be standard in medical research.
To be fair, determining the value of prostate surgery is a knotty matter. Because a majority of the men suffering from prostate cancer die of something else, the benefits of surgery, whatever they may be, for the most part accrue to the minority who would actually die of the disease. Even for them surgery may have little value. "If you have cancer that would kill you at sixty-nine, and you have treatment and then die at seventy of heart disease, you didn't gain much," Chodak says. Gains in lifespan are therefore most likely for younger, healthier patients - men under the age of sixty or seventy. Unfortunately, these are the very people who are most likely to live for many years without symptoms anyway. The benefits will show up only years later, assuming that the tumor would indeed have become lifethreatening.
One way of cutting through the fog of competing statistics is to estimate the overall benefit from PSA screening on the basis of the most optimistic assumptions about the use of radical prostatectomy. If it is small, one might question the value of the test. Kit Simpson presented such an analysis at the annual meeting of the American Urological Association last May. Unsurprisingly, she found that the benefits were unevenly distributed, with most of them accruing to the fifty to sixtyyearold men whose cancer would have spread quickly and whose lives would therefore presumably be lengthened by surgery.