Feb 11, 2004 in COMMENTARY

Deconstructing Reporting on HIV and AIDS

Public health professionals and veteran reporters outline common mistakes in coverage and discuss how to avoid them.

by Bob Roehr


Note: Information and related links can be found here.

Skepticism is the reporters stock in trade. For health reporters, it rises automatically when a pharmaceutical company touts its latest wonder drug. They know the company stands to make big bucks and so they dig deep to try and figure out for their readers exactly where the elixir lays on the scale of placebo to panacea.

Public health can be another matter. The single biggest trap for reporters is that people who do public health for a living are a lot like us. It is awfully easy to get close to them and feel like our goals and their goals are in sync, says Maryn McKenna, who covers the Centers for Disease Control and Prevention (CDC) for the Atlanta Journal-Constitution.

Marilynn Marchione has worked the health beat for the last dozen years at the Milwaukee Journal-Sentinel and echoes that concern. You share a common goal of trying to communicate risks to the public, but you must have a very clear vision of when your interests are the same as the people you are covering and when they are different.

For prize-winning author and Newsday reporter Laurie Garrett, The real task for the science reporter is to know what questions to ask, and to understand how data can be collected and massaged.

That task becomes particularly difficult when the subject is laden with social taboos, as is the case with sexually transmitted diseases like HIV. There is an ick factor, a reticence to discuss associated issues, let alone read about them over breakfast in a family newspaper.

These maladies appear to disproportionately affect the economically disadvantaged, which often means people of color, contributing a further overlay of cultural and sometimes linguistic complexities that can challenge even the best journalist capacity to understand and explain. So it should be no surprise that mistakes do occur.

It is a rare story on AIDS where I feel theyíve got all of the data right. It is very frustrating, it is either over sensationalized or misses the story, says Jennifer Kates. She is an HIV policy analyst with the Kaiser Family Foundation, which seeks to be an unbiased source of information on health issues.

The most frequent errors that I see reporters make, in both print and broadcast, are on data, and misusing and misinterpreting data. HIV and AIDS infection rates, prevalence, incidence they are almost invariably screwed up, says Michael Cover. He trained as a journalist and worked for two small newspapers in Ohio prior to becoming director of communications at the Whitman-Walker Clinic in Washington, DC, one of the largest AIDS services organizations in the country.

He grimaces over an article in the Washington Post that was published in the summer of 2003. I spent an hour and a half on the phone with [the reporter] on a Friday night. I took him through our web site. And still he screwed up the incidence and misattributed data to the CDC.

Looking At The Numbers The starting point to covering public health is a solid comprehension of epidemiology and how research is conducted (see sidebar for Learning Resources). Understanding when a number is significant, when it is not, what small cells mean, what a 200% increase is when it went from 5 to 15, all of those kinds of things, are critical, says Terje Anderson, executive director of the National Association of People With AIDS (NAPWA).

But there is more to it than just an analytical foundation. Anderson uses the example of data on getting tested for HIV. Who are the people who come forward for testing for HIV, who doesnít? What are the reasons why they come forward for testing? What are the reasons for people who donít? How representative is that sample? How much can you infer from that? That is just the simple stuff, but you have to address that before they start to see where biases emerge.

One mantra of HIV, which is true of other sexually transmitted diseases (STDs), is that every epidemic is localoit plays out differently among risk groups depending on the site. For example, HIV primarily affects gay men in San Francisco and Denver; the rural South has seen a recent surge in heterosexual transmission among African Americans; while New York City remains a mixing bowl of every possible ethnicity and transmission risk category.

Not only is there a difference between HIV and AIDS the former indicates infection with the virus, the later is an advanced stage of disease measured principally by a nadir of fewer than 200 CD4 cells there also is a very significant difference between how the national numbers for each are gathered.

AIDS has long been a reportable condition in all states and the CDC numbers are very solid.

Reporting HIV infections has only recently become mandatory and trend analysis is based upon data from 25 states that began collecting the information a decade ago. But that dataset is missing some big players three of the four largest states in terms of infections, California, New York, and Florida and so it represents only about a third of the total number of people living with the virus with holes in the data for particular at-risk groups.

The CDC readily acknowledges the limitations of their data but reporters often fail to understand the differences or make them clear in their articles.

Trying to think about the real meaning of some of the numbers in some of the categories is a challenge, Anderson readily admits. We come to use shorthand, not always accurately. People talk about a rise in infections in gay men, they look at the statistics of MSM (men who have sex with men) and the headline becomes: Syphilis in gay men up. Maybe syphilis is up in gay men. Maybe syphilis is up in non-gay-identifying men who have sex with men. Maybe it is guys on the west side.

I think that there has been more emphasis simply on demographics in a way that ignores some other factors. It ignores geography, social networks and community.

Just saying that it is Latino womenÖ until you understand that it is Latino women living on the south side of town, in a poor neighborhood, not documented -- whatever it is that gives a whole picture of what it is that is making these folks vulnerable youíre not getting the whole story, says Anderson. What are the social circumstances surrounding infection? What are the circumstances in people is lives that create this phenomena?

You can pick datasets out of Alabama and Mississippi and Florida and they might be similar, but you canít compare them with Los Angeles, says Frank Beadle de Palomo, director of the Center on AIDS and Community Health at the Academy for Educational Development. There is a huge difference between Mexicans, Puerto Ricans, and Cubans in this country. Much of it is tied to access to resources and insurance.

Minorities often are underrepresented in clinical trials, public health data generally has exactly the opposite bias. People of color, young adults without health insurance, people with AIDS on disability, those who are more dependent upon public health services are more likely to have their health information recorded by state agencies.

In contrast, physicians in private practice seldom are compensated for filling out public health reporting forms for sexually transmitted diseases and seldom take the time to do so. Often those cases never make it into the state registry.

The data on AIDS cases is fairly reliable because most of the patients are on expensive therapy that is seldom used to treat other illnesses. This is not the case with other STDs where treatment generally is short-term, with inexpensive broad spectrum antibiotics. If a patient is symptomatic, say with a drip or canker, many doctors will simply treat without ordering a confirmatory diagnostic test, so there is no lab report, which is another cornerstone of public health data.

Throw in the fact that with some STDs, especially chlamydia, many physicians understand the difficulty of getting the male partner of a diagnosed female in for evaluation and treatment. So they simply send a second packet of drugs home with the woman.

NAPWA Terje Anderson illustrates the spotty character of the data with an example of statistics on hepatitis C drawn from his experience within a public health department, running the HIV program for the state of Vermont.

The reports from community health centers ranged all over the map, he says. That had nothing to do with the incidence or prevalence of hepatitis C among the patient population, it was what the doctors were looking for. They found patients with hepatitis C because they began to look for them.

Asking The Right Questions You have to ask I am willing to sound stupid when I ask these kind of questions about how they chose the study population. Who did they exclude or not include and why?" says McKenna.

She turns to the massive CDC Behavioral Risk Factor Surveillance System as an example. It was conducted by telephone and excluded the incarcerated. But the methodology also excluded the homeless and likely underrepresented those without phones. Those are three really big groups that have a lot of public health problems, and those survey weaknesses should be pointed out whenever it is cited in an article.

With behavioral research, especially when dealing with sex or illegal activity such as drug use, the question asker can be is as important as the question itself. Reporters often forget that, warns Garrett.

We all know as journalists that we are capable of asking questions in a way that ends up winning our [interview] subjects over to what they perceive to be a solidarity with them. Many people doing questioning for public health are capable of creating the same ill-conceived bias.

One striking example came to light only through a fluke. Researchers studying women at high risk for HIV, asked a series of questions about sexual practices. Six months later a computer-assisted interview program was ready that allowed women to self-administer the survey. New enrollees in the study only used the computer while the initial group of women was asked to retake the survey using the computer.

The second time around, the same women were nine times more likely to report engaging in anal sex when they used the computer than when they had been interviewed by another human. The researchers were shocked, their best explanation of this huge differential was that the social stigma attached to anal sex played out in face-to-face interviews but less so with the anonymity offered by the computer.

First Report Often I find that when a reporter comes to me, the reporter already has a preconceived notion about the story. Even if the story isnít physically written, it is written in the reporter mind, says Phill Wilson, executive director of the Black AIDS Institute, located at UCLA. So all they are doing is searching for these quotes to plop into the preconceived spots in their minds, and often people are perfectly willing to give them a quote.

I canít tell you the number of calls I get from reporters saying, I need an African American mother of three who has got two or three positive babies, her husband left her, and she relapsed, says Cover in exasperation.

Before working with an unfamiliar journalist, Cover does his research, reading what they have written in the past, sometimes even calling sources to ask if they were quoted correctly, if the reporter has an agenda. I give every reporter one chance. But there are times when he decides that it simply is not worth his time and resources to work with a reporter.

Wilson says, If you are serious about doing the story, you should not come with the finished concept and thesis, you should come to really report. Reporting means actually finding out what is going on.

He uses the example of a July 2003 New York Times Magazine article on the down low, black men who have sex with men and also women, but identify as neither gay or bisexual, and keep that activity hidden from their female partners. While he appreciates the depth and humanity of the article, made possible in part by its length, he is still concerned that there is this kind of rush to find some generalized answer to this question.

Is this a phenomena that existed or is this, particularly the labeling, the result of marketing? Or did the media help to create it? asks Wilson. To the degree that phenomena exists, it is tied to age, and youth, and hip-hop culture. It may be another way that hip-hop culture manifests itself, just like it manifests itself in music and in fashion and in other behaviors, this is the way it manifests itself in a homosexual context.

An additional problem is that general audience science editors often lack training in the field and are stretched so thin, from astronomy to zoology, that they seldom are in a position to offer backup in the critical area of context and history. Garrett adds, there also is a tendency to ignore the follow up studies that may contradict the initial one.

Sourcing Public health is kind of an inbreed field, says Marchione. A lot of these people know each other, many of them went through the EIS [Epidemic Intelligence Service] program at CDC. You have to be careful not to get into cronyism where they recommend their friend or someone who shares the same kind of global view of the world.

The problem goes beyond one of training and the inculcation of a cultural perspective toward public health, CDC also funds the lion share of academic research and the disease prevention programs run by state and local public health departments.

Potential self-censorship [by sources] is a big problem, says McKenna. Partly it is the money issue [of funding from CDC], partly it's that under an aggressively Republican administration, a lot of the public health world feels like they kind of have to hang together, and that they wonít air their dirty laundry in public, if there is any.

That perspective seeps over into community advocates as well, and often for the same reasons. I donít want to be the one that says It is not bad, because we are all trying to build the case that there is a need for more resources, explains Anderson.

Much of the time, in order to balance a story with views that are critical of how public health is conducted, McKenna says she has to go to people who politically, journalists tend to be suspicious of, often those on the far right of the political spectrum. She calls it an ethical and a professional challenge for journalists, but one that they must confront.

I look for the outlier, somebody who is going to say something different, says Marchione. She uses Lexis-Nexis or just plain old Google, to try and get out of the Midwest mentality that a lot of issues donít hit us here.

Marchione has found Profnet to be very useful. She makes liberal use of the filters to screen in and out the types of expertise she is seeking. I find that e-mail works great. Iíll probably e-mail four times the number of people I end up interviewing because I like to see who is biting. I like to see who is willing to talk about an issue.

McKenna suggests interviewing a lower ranking person in an organization. Generally they will have a little more time than a section head and often they still have an enthusiasm for explaining things that bureaucracy and repetition hasnít yet beaten out of them.

I donít think that anyone has ever asked me, as a part of an interview, is there anyone else I should talk to?, says Wilson. He urges reporters to make the question a standard part of an interview and develop broader networks of sources, particularly within communities of color and other underserved populations.

Garrett laments the silly embargo system... If it is a really big story, hundreds of journalists are trying to call the same finite pool of sources; it is very difficult to get a complex analysis in. So all too often they read like a hastily written wire story, because there wasnít the time and access to do much more than that.

When he was at the National Council of La Raza in the early 1990s, Beadle says, All of the Latino data used to be just destroyed by reporters. He began to take their print articles and redline them and send them back and say, if you are going to work with us, these are the kinds of things you are going to have to correct.

And it helped. Folks started to get better at it. They were pissed off at first, they would say, who do you think you are? And we would say, if you are going to continue to ask us for support, this is what you need to do. We started to see a change in what we had to correct for them.

Writing Language colors the presentation of data by researchers, public health officials and journalists alike. Syphilis is a recent example of the three groups allied in common cause.

Syphilis-the word just sounds nasty with its twin sibilants, certainly much more so than gonorrhea or chlamydia (which have the added burden of being difficult to spell). Add in the demonization of illegal drugs, in this case crystal methamphetamine, and you have a winning combination for, if not the fact at least the reporting of, the latest epidemic-- - syphilis in gay men and the fear of increased transmission of HIV.

But Beadle says reporters missed the third element that really made ñthe epidemic all possible Viagra.

When you look at the dataset that came out a few years ago around circuit parties and the whole bare-backing phenomena, what they found was that it wasnít crystal meth, it wasnít heroin, it wasnít any other elicit drug that was the co-factor for increased transmission of HIV, it was Viagra.

Crystal meth inhibits erections and thus HIV transmission; Viagra counters the effect. But that was not the news story, says Beadle. The news story was the syphilis outbreak because then it becomes something that focuses on either demonizing or shamingopeople should know better. He says the tendency has only been reinforced by the Bush administration, Anything that you can do to demonize is fine with them.

Thus, in California, 1,200 cases of syphilis are promoted as an epidemic, while 24,000 cases of gonorrhea are just another statistic lurking in the background, even though both infections have similar medical and behavioral implications for the transmission of HIV.

Another underreported fact is that in most locations, half of the syphilis cases in gay men occur in those who are HIV positive and know it. Some have been infected more than once. This suggests that a limited pool of high-risk individuals are passing the infection round and round.

There is a mentality about syphilis, that it is something that is not suppose to exist because we can make it not exist, says Kates. Eliminating syphilis was going to be another notch in the belt of public health. But the disease and the people spreading it are not cooperating. At times it its hard not to feel that the CDC is taking it personally.

Part of the frustration is that journalism is not peer-reviewed scientific publication, which does have the room for the nuance, says Anderson. But to inform the public, you have to present them with the starkest information possible.

One has to pay attention to cultural biases, says McKenna. I donít think that there is an African American in the South who doesnít know about the Tuskegee project, the decades long study that followed but did not treat black men diagnosed with syphilis.

That still colors so much. There is very justified suspicion of government interventions in the health of minority communities. A lot of the suspicion of HIV being a biological warfare artifact or a form of genocide invented by whites, flows from the justified suspicion and anger of the aftermath of Tuskegee. She says, I have an intellectual understanding but not a visceral understanding of why the effects are so long lasting.

Marchione takes a bit of a divergent view, not because she dismisses minority concerns but because I donít think that is really unique to minorities. I see a tremendous distrust/mistrust of officialdom at all levels when it comes to health.

I donít think that people in general trust their doctors. The whole alternative medicine movement has spread on the desire of people to take their care into their own hands because they are distrustful of or dissatisfied with what traditional medicine offers them.

Welcome to the world of public health and the challenges of covering it fairly and accurately.

Bob Roehr is a freelance journalist and biomedical writer who specializes in HIV and infectious disease. He writes for consumers through more than a dozen gay newspapers across the country and for health care professionals, including conference coverage for Medscape; and for the National Institutes of Health.