Computer Checks of Pap Tests Worth It?

A device that uses computer technology to double-check women’s Pap smears after they’ve been read by technicians can pick up a few additional abnormal test results, but it’s more expensive than having a human technician recheck the cells on the microscope slide.

That’s the conclusion of a new study of PAPNET, a computerized system for rechecking Pap smears, the test performed on cells from a woman’s cervix to detect and prevent cervical cancer. The report appeared recently in the Journal of the American Medical Assn.

The study by researchers at the Armed Forces Institute of Pathology in Washington reached a different conclusion from several studies cited by PAPNET’s manufacturer, Neuromedical Systems Inc., which suggest that the technology is reasonably cost-effective. PAPNET has been marketed aggressively to doctors, patients and laboratories as a way of offering added quality assurance in the evaluation of Pap smears. The charge to a patient for having a smear rechecked by PAPNET is about $40 and often is not covered by insurance or managed-care plans.

Congress tightened the regulation of clinical laboratories that read Pap smears in 1988, after news stories highlighted some labs’ failure to detect abnormal results and thereby prevent cases of cervical cancer. Public health experts consider the Pap smear to be one of the most successful and least expensive tests to prevent cancer, but despite its effectiveness it remains underused.

To perform a Pap smear, a doctor lightly scrapes a sample of cells and mucus from the patient’s cervix--the opening of the uterus--during a pelvic examination and smears them on a glass slide. The slide is then sent to a laboratory where specially trained technicians and pathologists examine the cells under a microscope, looking for changes that often precede the development of cervical cancer.

Since the Pap smear was introduced after World War II, the number of cervical-cancer deaths in the United States has dropped by 70%, according to the College of American Pathologists. The disease was formerly the No. 1 cause of cancer death in women. There are an estimated 14,500 cases of the disease in the United States each year, and 4,800 deaths, many of them preventable.

Medical organizations unanimously recommend regular Pap smears, although some favor an annual test while others say that for most women a test every three years is adequate.

In a Gallup poll commissioned by the College of American Pathologists last year, 40% of women 18 and older sampled said they had not had a Pap smear in the year preceding the survey. Women who were 50 or older were the age group least likely to have had the test, although they are the group at highest risk of being found to have cervical cancer. Today, 80% of U.S. women who die of cervical cancer have not had a Pap smear for five years or more.

Timothy J. O’Leary, chairman of the department of cellular pathology at the Armed Forces Institute of Pathology and principal author of the new study, contended that resources spent on computer-assisted devices to improve laboratories’ accuracy in reading Pap smears might be better spent on programs to provide Pap smears to women who aren’t currently getting them.

“The Pap smear, done at intervals of even as long apart as three years, is capable of eliminating . . . up to 90% of potential invasive cancers,” he said. “No [rechecking] method, whether it be manual or computer-assisted, helps the woman who doesn’t get the Pap smear to begin with.”

In the study, AFIP researchers used PAPNET on 5,478 Pap smear slides, all of which had already been examined twice by technicians. Out of the slides that PAPNET flagged for further review by pathologists, six were found to contain abnormal cells that could be precancerous. None was found to contain cancer cells.

In an analysis using the average PAPNET charge of $40 per patient, the team calculated that routinely using PAPNET to try to identify additional cases of cancer would cost more than $100,000 for each case found.

It would be cheaper and nearly as effective, O’Leary said, to have each slide checked twice by a technician. Currently in most laboratories, the majority of Pap smears are checked only once. If a technician finds an abnormality, the slide is reviewed by a pathologist. Federal law also requires labs to select 10% of normal Pap smear slides at random to be rechecked.

Most laboratories also make an effort to double-check any Pap smear belonging to a patient who is considered to be at high risk for cervical cancer, but doctors said laboratories differ in how they define that group. Risk factors for cervical cancer include a previous history of abnormal Pap smears, a history of infection with the human papilloma virus or other sexually transmitted diseases, a history of having had many sexual partners (or having sex with a man who has) and cigarette smoking.

Laurie Mango, vice president and chief medical officer of Neuromedical Systems, said the greatest value of PAPNET is its ability to improve the accuracy of Pap-smear interpretation, especially in labs that do a rushed or sloppy job. She said patients and doctors usually can’t judge the quality of a lab and that managed-care plans often contract with the cheapest labs with the fastest turnaround time.

“Unfortunately, the faster you turn it around, the more likely it is that it wasn’t looked at that carefully,” she said.

Mango cited data from a multicenter trial submitted to the Food and Drug Administration that found that rechecking Pap smears with PAPNET identified about seven times as many abnormal smears as having them rechecked by a technician.

Another study, which used a mathematical model, calculated the cost-effectiveness of PAPNET testing every two years and compared it with that of other tests and treatments, using figures derived by various researchers on how much money must be spent on a test or treatment in order to save a year of someone’s life. The study concluded that the cost-effectiveness of PAPNET was comparable to that of mammography for women in their 40s.

If the makers of the computer-assisted laboratory devices get their wish, the new technology may increasingly replace human beings as evaluators of Pap smears. A competing company, Neopath Inc., has asked the FDA to approve its device, the AutoPap (R) System, to perform the initial screening step. The device is already used by some labs to select the 10% of normal slides that are to be rechecked.

Under Neopath’s proposal, the AutoPap device would become the first to examine all smears arriving at a lab. It would select 25% as the “most normal.” These would never be checked by a human being. The other 75% would be examined by technicians or pathologists, with special attention to the 15% flagged by the computer as the “most suspicious,” said Stacie Byars, a Neopath representative.