Home Pregnancy Tests Questioned : University Study Raises Doubts About Reliability of Kits
Home pregnancy test kits--one of the most widely used and conspicuous forms of do-it-yourself medicine--may give wrong results far more often than manufacturers would like women to believe, a new study has charged.
The study by a researcher at Milwaukee’s Marquette University found that in three of the most widely used test kits, error rates ranged from 11% to more than 54%--figures that substantially exceed the accuracy claims made by manufacturers. Positive results were slightly more reliable than negative results, the Marquette study found. For the three kits, makers claimed overall accuracy rates of between 96% and 99%.
In the extreme, erroneous findings may lead to undetected tubal pregnancies that endanger women’s lives and to possible emotional devastation for women suffering from infertility problems who may be wrongly led to believe they are pregnant.
While the Marquette study has quickly become controversial, experts questioned by The Times agreed that doubts about the accuracy of home pregnancy test kits and the wisdom of using them are widespread.
Several experts--including two in Los Angeles--said they routinely counsel women not to use home pregnancy tests at all because of the potential for wrong results, either positive or negative. These doctors say the advent of especially accurate urine tests that can be done in a doctor’s office in a few minutes make reliance on home tests unnecessary.
In-Office Tests
This new generation of in-office tests is also priced competitively with many of the home kits, which retail for between just over $7 and just under $15. In Los Angeles, the new in-office test costs $15, too, doctors using it told The Times. The sophisticated new test can detect pregnancy before a menstrual period has been missed.
Manufacturers of the home test kits have quickly disputed the Marquette figures and the U.S. Food and Drug Administration says its own data have so far not shown error rates even approaching those reported in the new study. In fact, the American Journal of Public Health, which published the new study, also ran a brief letter from the FDA objecting to certain aspects of the Marquette research.
However, the FDA said it has never tested the kits in the way they were examined in Milwaukee--in a program where researchers evaluated kit accuracy under home-use conditions, as opposed to those of the laboratory.
The Marquette findings added new detail to research published in 1982 in which a University of Cincinnati team found that women using the kits failed to follow directions nearly a third of the time. A quarter of all results in that study told women they were not pregnant when, in fact, they were. Even a minor deviation from instructions furnished with the kits can lead to either a false-positive or false-negative reading.
Some experts believe the worst accuracy rates may be experienced by teen-aged girls, very young women and women with little education--three groups most likely to accept a home-test result and fail to seek confirmation of pregnancy from a doctor.
By postponing professional care, a woman may wait to seek an abortion or other options until her pregnancy is advanced. Such women may also continue drinking, smoking or taking drugs potentially damaging to the fetus--including aspirin--and expose their babies to high degrees of risk.
In question are e.p.t., Answer and Daisy 2 home pregnancy test kits. The products claim to be able to determine if a woman is pregnant in some cases as early as three days after a missed period and, in each case, by the time nine days have passed.
The pervasive skepticism among specialists in women’s health care has apparently not cooled what has become a hot market for the home kits, of which half a dozen brands are widely available. In 1985, women used between 4 million and 7.2 million kits, for which they paid between $38 million and $48 million, according to estimates by Frost and Sullivan, a New York-based marketing research firm, and Biomedical Business International, an Orange County research organization.
There were 3.7 million births last year nationwide, according to federal government figures.
While demand has been essentially unchanged in recent years, marketing experts believe that, by 1989, use may increase to more than 8.9 million kits, according to Frost and Sullivan. The first generation of home pregnancy test kits entered the marketplace about 10 years ago.
The new study was done by researcher Mary Doshi of Marquette, who distributed e.p.t., Daisy 2 or Answer test kits to 109 Milwaukee women, ranging in age from 18 to over 30. A quarter of the women were nine days or less late for a menstrual period. Doshi has since left Marquette.
While more than half the women had college educations, 10% had not finished high school. Sixteen percent of the subjects were not white. The women were selected from private doctors’ offices and public health clinics. All subjects used the kit tests in their homes, relying solely on printed instructions. Afterward, they all had laboratory pregnancy tests.
Manufacturers have long emphasized that accuracy of the home kits can be compromised if a urine specimen other than the first voiding of the morning is used (because the first urination is richest in pregnancy-related hormones), if less than the amount of time specified elapses before the test result is read, and if the test kit is subjected to any movement or vibration while the chemical reaction is going on, a period of as much as two hours.
Under in-home conditions, the 109 women experienced overall accuracy rates with the kits of 77.1%--significantly below the average manufacturers’ claims of 97.4%. For tests administered nine days or less after the missed period, the average kit accuracy dropped to 65.5% and for tests after nine days, the rate was 81.3%.
The Best Overall
The e.p.t. test recorded the best overall accuracy rate, 83.3%, while Answer and Daisy 2 were accurate only 72.9% and 75% of the time, respectively. All of the tests rely on the same basic chemistry. They use modern, artificial agents to analyze the content of urine to determine concentrations of human chorionic gonadotropin, or HCG, a key hormone marker of pregnancy.
The overall accuracy rates for each kit were averages for all women using them. Among some women using them less than nine days after a missed period, the three brands together yielded accuracy rates as low as 45.7%. On the other hand, some women more than nine days after a missed period achieved average accuracy rates for the three brands of 89.1%.
“Manufacturers should be encouraged to re-evaluate their claims,” Doshi concluded. “Accuracy would be improved by increasing the number of days a woman should wait.” She noted that a false-positive result could potentially lead a woman to undergo an unneeded abortion procedure. In addition to the prospect of permitting a pregnancy to continue without suitable medical care, a false-negative finding, experts questioned by The Times agreed, could permit a tubal pregnancy to remain undiagnosed.
“Manufacturers should be urged to emphasize the vast sources of error in product literature, as well as to include control samples (confirmatory double-check tests) with each test kit,” Doshi urged.
Test kit makers were quick to criticize Doshi’s research. Dr. Joseph Clark, vice president for clinical research and regulatory affairs of Warner-Lambert Co., maker of e.p.t., contended that the Marquette study apparently used a version of e.p.t. that is no longer being manufactured and has been replaced by a new test that Clark said is significantly more accurate and reliable.
The new test, called “e.p.t. plus,” is gradually replacing e.p.t., but The Times found both versions still widely available in retail stores in Los Angeles.
Size of Study Group
Clark contended that the size of the study group in Milwaukee was too small to produce reliable results for any one of the three tests used. He said Warner-Lambert bases its contention that even the now-obsolete version of e.p.t. is 99% accurate on studies of as many as 400 women who used the test either in a doctor’s office or in their homes with only minimal instruction other than the printed materials supplied with the kit.
Clark conceded that it is a “good question” whether the company tests may have experienced skewed results because they relied exclusively on women who had sought medical care--as opposed to those who may rely on the home test alone, without resorting to advice from doctors. “We think that going to the physicians women come to when they think they are pregnant is the only way you’re going to get a good cross section,” Clark said.
“We’ve been criticized in the past when we’ve done studies in specialized clinics that these are not representative of the general population. That’s why we now use women who are garden-variety women.”
Clark said changes in the chemistry of e.p.t. in its most recent version have eliminated much of the test’s sensitivity to variables including even bumping lightly into the table on which the kit has been placed. “The product we have out now is essentially foolproof,” he said.
Ortho Pharmaceutical Corp., maker of Daisy 2, likewise defended its product, insisting that accuracy rates claimed for the test are correct. “We know beyond the shadow of a doubt that, if used correctly, the test is 98% accurate for positive results and 100% accurate for negative results,” an Ortho spokeswoman said. She said Ortho believes directions furnished with the Daisy 2 to be “very clear.”
Carter-Wallace Inc., the maker of Answer, did not respond to inquiries about the new study and its products’ reliability.
FDA scientist Wendy Johnson said the Marquette findings were at odds with data submitted to the FDA by manufacturers and with laboratory evaluations of the tests conducted by the agency itself. She criticized Doshi’s study for saying manufacturer-supplied accuracy rates were official FDA data and for implying the FDA does not regulate test-kit manufacturing, which the agency does.
“We have not had a large number of complaints” about accuracy, Johnson said. “Consumers may not know where to complain.” But Johnson also conceded that performing a sophisticated chemical test in the home environment is significantly different from doing the identical test in a laboratory. “The quality control is not there (in the home). It’s not the same,” she said.
Failed to Follow Directions
Barbara Valanis, professor of nursing and associate professor of environmental health at the University of Cincinnati, voiced concerns common among experts questioned about home pregnancy test reliability. Valanis led the 1982 study in which, among 144 women from private doctors’ offices, public health clinics and a Planned Parenthood clinic, 28.5% had used home test kits, but 32% of those who did failed to follow directions completely.
In the Cincinnati study, women who used the kits had false-negative rates of 33% and false-positive rates of 21% if they waited to perform the tests more than nine days after a missed period.
The study noted that younger women using the kits have a higher risk of getting a wrong result because their hormone balances have not been fully established and their menstrual cycles are not as regular as those of older women.
“Our findings suggest that if you do the test fairly precisely, you can get an accurate result,” Valanis said in a telephone interview. “But you do not know who is doing the tests and how they are doing them. In the general public, there are people with different reading levels and under different amounts of stress.”
