Presenting and Interpreting Research Results: HIV Data

"Study Calls HIV in D.C. A 'Modern Epidemic': More Than 80 Percent Of Recent Cases Were Among Black Residents"
Susan Levine, The Washington Post, November 26, 2007.

The lead for the article linked above is:

The first statistics ever amassed on HIV in the District, released today in a sweeping report, reveal "a modern epidemic" remarkable for its size, complexity and reach into all parts of the city.

I wonder what the authors mean by the term "modern"? Is the HIV epidemic bigger, more complex or with broader range than the black death, the Spanish Flu pandemic, or the epidemic of diseases of aging (cancer and heart disease) we are not experiencing? Perhaps the author means it is modern in that it is smaller, or better understood by contemporary public health officials.

Why does the title focus on the race of the victims? The District population is about three-fifths black, so the fact that four-fiths of the newly infected people are black indicates that they are at risk. But the population is about equally divided between males and females, and 70% of the new cases are males. But neither being black nor being male is a "risk factor" as normally understood.

Males are presumably more at risk than females and blacks as compared with non-blacks because more people in those large categories have high risk. The important risk factors are unprotected (anal) sex with infectious persons and sharing needles with infected people, are they not? Of course in the District's racially divided society, people tend to associate more with others of their own race, so all other things being equal the group with higher prevalence will have higher incidence of the disease. However, all other things are not equal, and the group with the larger percentage of intravenous drug users or with the larger percentage of people engaging in unsafe sex will tend to have the higher incidence of the disease.

In presenting epidemiological data a basic rule is to present not only the numbers of infected, but also the numbers at risk. Thus the data in the figures to the right indicate that more cases of HIV infection were heterosexually transmitted than homosexually. It seems likely, however, that the risk per person involved in homosexual transmission is still much higher than that for those only participating in heterosexual sex. While we now know that AIDS is not only a disease of homosexual men, and that efforts to prevent heterosexual transmission of the disease are needed, there should still be priority accorded to preventing transmission during heterosexual relations.

It is especially important to use epidemiological information for planning public health responses. The indication that the District has the highest incidence of HIV among large cities in the United States leads me to conclude that the District should spend proportionately more on HIV control in its public health budget.

The report that there is a large number of new-borns with HIV infection, combined with the knowledge that transmission can be blocked and that pregnant women are reasonably available for screening and births are attended in hospitals suggest an immediate priority for preventing transmission to new infants.

According to the District's press release:

The District accounted for 9 percent of all pediatric AIDS cases in the United States during 2005. Between 2001 and 2006, there were 56 children ages 13 or younger diagnosed with either HIV or AIDS in the District of Columbia.

Thus the District in 2005 had 18 times the national rate of pediatric AIDS cases!

Prevention efforts should be directed to those who would most benefit, usually those at highest risk. I would assume that in addition to pregnant women, they should be directed to those in the high risk categories identified above. Blacks are at only slightly higher risk than other groups, and if one can effectively reach the high risk groups within the black (and other) populations, that should suffice. Still, there are some general HIV/AIDS education programs, and knowing that 80% of the incidence of HIV is in black populations may have some modest benefits in directing that general education.

Why does the graph reproduced indicate that while the AIDS incidence parallels the HIV incidence it is always higher? Why does the article present such counter-intuitive information without explaining it?

One might argue that the newspaper should not be expected to present such information in the most useful form for its readers, who as citizens affect public health policy, and as individuals are those at risk of being infected. After all. reporters and editors are not epidemiologists. I think, however, that good reporting of this kind of evidence should be held to a very high standard; the editors should get editorial advice from epidemiologists. This is especially true in the District of Columbia where we have easy access to the best in the land! At least the WP put the story on its front page, above the fold.

I offer a link to the report itself, which the WP appears not to have had the courtesy to do.

Click here to go to the Report itself.