No one left to track.

DAVID WAS FREAKING OUT. IN A MOMENT COLORED IN EQUAL parts loneliness, horniness, and Special K (also known as ketamine, an animal tranquilizer gaining popularity as a recreational drug), he had unsafe sex with someone he, had just met. On the sober morning after, he realized there was a good chance this partner was HIV-positive, and so for weeks David went through all the emotions people tend to go through afterward: Fear. Regret. Self-recrimination. Lots more fear.

And finally, nagging curiosity. He wanted to find out if he had been infected. After all, there are treatments these days, right? It's important to know, right? Absolutely. But just when David decided to take the test, something happened that made him pull back. So far back, he still hasn't taken it--and isn't sure he ever will.

An article appeared in The New York Times reporting that the Gay Men's Health Crisis, his local AIDS service group, was leaning toward a new policy mandating that everyone who tests HIV-positive should be reported, by name, to the state health department. David wanted to know if he was positive, but he didn't want his name on a list. And so he didn't take the test.

Well, it turns out the report was wrong, sort of. GMHC was not really in favor of name reporting. Instead it was leaning toward a system whereby those who test HIV-positive are reported to their state health department using something called a "unique identifier, " a number or a code in which the infection is reported but not the name of the person infected.

Unfortunately, David isn't tremendously comforted by this. He's still afraid that today's unique identifier might end up being tomorrow's name on a list. And his story points out a serious quandary that's likely to grow more serious as time goes by. For years doctors have been required to report patients who have full blown AIDS to their state health departments, but not those who are merely HIV-infected. As a result, there is no way of knowing how many Americans have HIV, whether the numbers are rising or falling, where the hot spots of the epidemic are, and all sorts of other data we need to understand the nature of the epidemic, fight it, and provide services where they are needed most.

Since doctors have long been required to report the names of those who have syphilis, gonorrhea, and other sexually transmitted diseases, the HIV policy has sometimes been called "AIDS exceptionalism," and the reason for the exception is pretty obvious: There's a huge stigma against AIDS and lots of discrimination. Health authorities long feared that if name reporting and its cousin, partner notification, were applied to HIV, it would scare the Davids of the world away from HIV testing and thus away from the health care system.

Now many public health advocates believe that those concerns are fading. The stigma against AIDS is declining, the theory goes, and so is discrimination. In addition, new treatments extend the lives of people with HIV, but only if they find out they are infected and come in for treatment. As a result, there is now a major push in many states to end AIDS exceptionalism and establish a system of mandatory name reporting and even, in some cases, partner notification. Many states are indeed considering and implementing these policies.

But there are big questions about this traditional approach, questions nobody seems in a hurry to answer. Such as, does it even work? It is true that for years such techniques were used against traditional sexually transmitted diseases, but it is equally true that nobody ever evaluated them to see if they were effective. Name reporting and contact tracing were implemented in the 1930s, long before modem techniques of evaluation were developed. All we know is that throughout the decades, epidemics of syphilis and gonorrhea raged on anyway.

But if we don't know if the old techniques work, we do know that David's story isn't unique. In one study over 60% of those who were tested anonymously said they would not have been tested if their names were going to be put on a list. So while the upside of traditional techniques is uncertain at best, the downside seems very real: frightening the Davids of the world away from the health care system.

It is vitally important that we figure out ways to track the epidemic, and unique identifiers seem a logical way to do that. But it is just as vital to protect people not only from discrimination but also from fear of discrimination. If not, the supposed solution could end up, quite literally, making the problem worse.