When screening for HIV, how good should a test be?
Dr. Rochelle Walensky has devoted much of her career to championing the importance of routine AIDS testing. It is, she and other HIV specialists insist, vital to stopping spread of the virus and to getting people treated when powerful drugs can deliver the biggest punch.
But as faster, easier screening methods have grown in popularity, AIDS specialists have confronted a trade-off that sometimes leaves them uneasy: In the pursuit of screening as many people as possible, is it acceptable to tell a small number they might have the virus when they really don't?
That debate flared anew when a study last week showed that an oral HIV test given to emergency room patients at Brigham and Women's Hospital incorrectly told four out of every 100 they might be infected. More definitive tests were immediately performed, and patients knew, often on the same day, that they did not actually carry the virus.
"My goal here is to identify cases of HIV infection, to tell people who are uninfected that they're negative, and to find the truth without doing any harm," said Walensky, leader of the Brigham's testing campaign.
Oral screening is attractive, she said, because if patients are squeamish about blood tests and therefore won't be screened, "then you haven't met your goal. But I would also agree that if you tell too many people they're [preliminarily positive] when they're negative, you've done some harm
"What we have to figure out," she said, "is what the trade-off is."
Federal disease specialists estimate that 250,000 Americans are infected with the AIDS virus and don't know it. Since 2006, a national initiative has sought to identify them by making HIV tests as routine as cholesterol screening.
A big part of that push has involved tests that can tell patients in as little as 20 minutes whether they might have the virus. There are tests that use a lancet to draw a drop of blood from a finger.
And then there are tests that instead use a swab wiped along the gums. The oral tests allow agencies such as Tapestry Health in Western Massachusetts to perform HIV tests from a mobile van and at community health fairs.
To be sure, most people taking rapid HIV tests - whether oral or blood - are told on the spot they have nothing to worry about. But health agencies in New York and Minnesota have reported occasional clusters of oral tests that generate a rate of false-positives - incorrectly telling people they may have HIV - exceeding what the manufacturer had predicted.
From last November through this April, nearly half of 444 patients who received preliminary positive results at New York health department clinics wound up testing negative when subjected to more rigorous analysis.
"Although, yes, we had 213 instances of false-positives, we were able to get 30,000 people tested, and those that were negative, we could tell them they were negative," said Dr. Susan Blank, an assistant health commissioner in New York City.
Dr. Bernard Branson, a testing specialist at the US Centers for Disease Control and Prevention, said investigators have found nothing to explain why some sites have experienced elevated rates of falsely positive tests. There's nothing to suggest, for instance, faulty lab techniques.
"One caution I would raise is our expectations of HIV tests are extremely high," Branson said, adding that the false-positives in New York accounted for less than 1 percent of all tests.
At the Brigham, from February through October of last year, 849 adults took the test, called OraQuick, with 39 results suggesting patients might be harboring the AIDS virus. Of those patients, 31 underwent confirmatory testing at the Brigham. (The other eight presumably received follow-up testing by primary care physicians or opted not to be evaluated further, Walensky said.)
After the 31 patients underwent extensive blood testing - which included looking directly for the virus as well as disease-fighting cells that marshal in the presence of HIV - researchers found only five patients were truly infected.
The test's maker,
OraSure spokesman Ron Ticho said the Brigham results, which he described as a "relatively small sample size," were "clearly not indicative of the way our test is performing in the field on a national level." The company sent representatives to the hospital after Walensky expressed concerns.
The Brigham continues to use the oral screening, and Walensky said her intention in releasing results, published in the Annals of Internal Medicine, was not to deter use of OraQuick. Instead, she said, clinic workers need to explain the implications of a preliminary positive result before performing the screening and make sure confirmatory tests are conducted.
"What I didn't want the message to be," she said, "is that this is a rotten test."
Still, she acknowledged that the hours between a patient getting a preliminary positive result and a definitive negative result can be fraught with anxiety.
"There were phone calls to me, huge extended families showed up in infectious disease clinics, one of our providers spent the night waiting" with a patient for results to come back, she recalled. "I wasn't particularly happy I was stressing people out."
Stephen Smith can be reached at stsmith@globe.com. 
