Skip Navigation


August 3, 2008

HIV Incidence in the United States: A Recount

Each year, for more than 15 years, the Centers for Disease Control and Prevention has issued annual reports that the HIV incidence rate in the U.S. is holding steady at 40,000 new infections per year. An eagerly anticipated study, led by CDC researchers and published today in JAMA, calls into question the accuracy of those earlier estimates in a report that puts the U.S. HIV incidence rate in 2006 at 56,000 new infections—a 46 percent increase over the long-accepted yearly estimate of 40,000.

The research team, including Ron Brookmeyer, PhD, professor with the Johns Hopkins Bloomberg School of Public Health’s Department of Biostatistics and a co-author of the study, arrived at the new estimate by employing improved laboratory and statistical tools that can more accurately detect when a person became infected with the HIV virus and also provide more details about infection trends in specific groups.

Brookmeyer and David Holtgrave, PhD, chair of the Health, Behavior and Society Department at the Bloomberg School and an expert in HIV prevention, offer some insight into what the new numbers and methodologies mean and how the data can guide public health officials and policymakers in making informed decisions about where to direct HIV prevention efforts and resources.

Jackie Powder Frank, a writer with the Office of Communications and Public Affairs, spoke with Ron Brookmeyer and David Holtgrave about the paper, “Estimation of HIV Incidence in the United States.”

JPF: What does this research add to the understanding of the AIDS epidemic in the U.S.?

RB: What this paper is saying is that there are about 56,000 Americans who became newly infected with HIV in 2006. It’s saying that we have not won the war against AIDS, by any means, that these new infections can be prevented and we have a lot more work to do.

DH: The new estimates tell us that HIV incidence is higher than the previously estimated 40,000 infections per year. The old statistics suggested that there was a new infection roughly every 13 minutes in the U.S. The new numbers suggest that there is a new infection approximately every 10 minutes in the nation. 

JPF: Is it accurate to conclude that the CDC has been underestimating HIV incidence since the early 1990s?

RB: I think that’s a simple but perhaps naïve interpretation. The problem is there’s a bit of apples and oranges involved and you can’t really do a direct comparison. It looks like it went up, but the underlying methodology has changed and the tools we had before to come up with 40,000 were actually quite crude and didn’t give us a good, precise estimate. Now we’re saying it’s closer to 56,000.

Is it the tool, the measuring stick, that’s changed? Or has the epidemic really gone up? I don’t think it would be right to conclude that the rate’s gone up by 50 percent because I don’t think there was the precision in the original 40,000 number.

JPF: Explain how the new methodology used in the study is more precise than earlier models.

RB: This new method is based on a biomarker, called the BED assay, and new statistical analyses. We give a blood test and can tell if this person was infected very recently (on average, within about 150 days) or if the infection is long-term. We use this in combination with statistical models to convert the information to numbers that are useable from a public health point of view.

BED builds off of a 1995 paper that I wrote with Tom Quinn, MD [director of the Johns Hopkins Center for Global Heath], to use lab assays to distinguish who was infected recently from who was infected a long time ago, and then to use that information in statistical models to figure out HIV incidence rates.

JPF: How does more accurate data on HIV incidence numbers help public health officials strengthen HIV prevention efforts?

RB: HIV incidence is the leading barometer of what’s happening in the epidemic now. If we know what HIV incidence is doing in different subgroups and demographics, then we have a better idea of who to target with prevention efforts. How many new infections are occurring is one of the best metrics of measuring the current growth of the epidemic.

DH: In order to prevent HIV infection, we need to know where the epidemic is going and to get ahead of it with effective prevention programs. Having a much more precise estimate of HIV incidence will give us a better sense of who needs services most urgently. And going forward, the new methodology will give us a better sense if our national HIV prevention effort is intensive enough to meet national prevention goals. 

JPF: What else do the new data suggest about HIV incidence?

DH: The detailed analyses suggest that the epidemic among men who have sex with men (MSM)  is increasing very rapidly [of the 56,000 new infections in 2006, the study reports that 53 percent were among MSM ] and there needs to be especially urgent attention paid to providing HIV prevention services to MSM. Further, the paper documents tremendous racial/ethnic disparities in HIV infection, especially in African-American communities.

JPF: Looking ahead, how can public health officials best use this new information?

RB: The point is it’s not just about a single number. It’s all about developing a coordinated response. We now have better HIV surveillance tools, and we will be in a much better position to track trends. The epidemic is a moving target, but statistical and lab methods are also moving and improving. They can help in understanding what’s driving the epidemic, who is at greatest risk and which prevention efforts can make the biggest difference. 

Public Affairs media contact: Tim Parsons at 410-955-7629 or