Email to the Gazette
Jul. 22nd, 2004 12:45 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
electriclandThe graphs you published to accompany the op-ed story "George W. Bush: AIDS warrior," by Sebastian Mallaby of the Washington Post (Thursday, July 22, p. A23), are at best uninformative and at worst dangerously misleading.
Let's start with the top graph, captioned "Condom availability doesn't mean fewer cases of HIV - African countries where condoms are most readily available often have the highest rates of infection." Readers might be forgiven for inferring from this that condom use increases rates of AIDS transmission. There are two problems with this interpretation:
1. The graph shows HIV prevalence (the percentage of cases in the population), not incidence (the percentage of NEW cases per year). Condom use can only reduce the number of new cases of HIV -- it can't cure existing ones.
2. Which figure is the cart, and which is the horse? It makes perfect sense that countries with high rates of HIV in the population -- like Botswana, Zimbabwe, and South Africa -- encourage condom use to prevent further transmission.
In short, your facts in this graph are correct, but so what? The relevant number -- new HIV infections vs. condom availability -- isn't shown here.
Now, on to the lower graph, "AIDS cases are declining in abstinence-friendly Uganda." Does this graph show prevalence or incidence? It doesn't say. Since the latest figure is 6%, which matches the prevalence rate for Uganda shown in the upper graph, one is led to assume that it's prevalence. Great -- except that since AIDS is incurable, the most likely interpretation for those declining numbers is that people infected with AIDS have been dying off, reducing the numbers of infected people in the population and thus the prevalence rate. Nothing to do with condom use; everything to do with the nature of the disease.
Furthermore, these numbers look great, but they're presented in isolation. As a reader, I'd like to see a comparison with another country that encourages condoms. It would have been more fair to your readers to show graphs for incidence of HIV infection over time for Uganda and a couple of other more "condom-friendly" (or "abstinence-unfriendly," if you prefer) countries -- say, Botswana and Zimbabwe. Or maybe Tanzania, which apparently has over two and a half times Uganda's HIV prevalence with the same rate of condom availability.
Maybe Mr. Mallaby's argument -- that focusing on condoms alone is not enough to reduce HIV rates -- is well founded. Maybe not. From the graphs you published, there's no way to judge his arguments, but it's very easy to draw mistaken conclusions -- and that does a real disservice to your readers.
Sincerely,
me.
(If I get a moment I'll scan in the graphs and annotate them fully. Crap! Total crap! The story they're illustrating was originally here: AIDS Activists Misfiring. I'm not too impressed with it either -- it's very light on hard evidence -- but honestly, these graphs were just too much. Watch The Top Of ELL's Head Blow Off First Thing In The Morning.)
Let's start with the top graph, captioned "Condom availability doesn't mean fewer cases of HIV - African countries where condoms are most readily available often have the highest rates of infection." Readers might be forgiven for inferring from this that condom use increases rates of AIDS transmission. There are two problems with this interpretation:
1. The graph shows HIV prevalence (the percentage of cases in the population), not incidence (the percentage of NEW cases per year). Condom use can only reduce the number of new cases of HIV -- it can't cure existing ones.
2. Which figure is the cart, and which is the horse? It makes perfect sense that countries with high rates of HIV in the population -- like Botswana, Zimbabwe, and South Africa -- encourage condom use to prevent further transmission.
In short, your facts in this graph are correct, but so what? The relevant number -- new HIV infections vs. condom availability -- isn't shown here.
Now, on to the lower graph, "AIDS cases are declining in abstinence-friendly Uganda." Does this graph show prevalence or incidence? It doesn't say. Since the latest figure is 6%, which matches the prevalence rate for Uganda shown in the upper graph, one is led to assume that it's prevalence. Great -- except that since AIDS is incurable, the most likely interpretation for those declining numbers is that people infected with AIDS have been dying off, reducing the numbers of infected people in the population and thus the prevalence rate. Nothing to do with condom use; everything to do with the nature of the disease.
Furthermore, these numbers look great, but they're presented in isolation. As a reader, I'd like to see a comparison with another country that encourages condoms. It would have been more fair to your readers to show graphs for incidence of HIV infection over time for Uganda and a couple of other more "condom-friendly" (or "abstinence-unfriendly," if you prefer) countries -- say, Botswana and Zimbabwe. Or maybe Tanzania, which apparently has over two and a half times Uganda's HIV prevalence with the same rate of condom availability.
Maybe Mr. Mallaby's argument -- that focusing on condoms alone is not enough to reduce HIV rates -- is well founded. Maybe not. From the graphs you published, there's no way to judge his arguments, but it's very easy to draw mistaken conclusions -- and that does a real disservice to your readers.
Sincerely,
me.
(If I get a moment I'll scan in the graphs and annotate them fully. Crap! Total crap! The story they're illustrating was originally here: AIDS Activists Misfiring. I'm not too impressed with it either -- it's very light on hard evidence -- but honestly, these graphs were just too much. Watch The Top Of ELL's Head Blow Off First Thing In The Morning.)
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2004-07-22 12:03 pm (UTC)no subject
Date: 2004-07-22 12:59 pm (UTC)I thought it was time someone called them on their crappy graphics, and it might as well be me... although I did try to sic my former colleague who's in Public Affairs at the MUHC on them too. We'll see. Ahahahaha.
no subject
Date: 2004-07-22 01:04 pm (UTC)especially since I was just having an argument with someone on this very topic.
no subject
Date: 2004-07-22 01:12 pm (UTC)no subject
Date: 2004-07-23 01:49 am (UTC)Remind me never to piss you of - you're obviously part lawyer. ;)
Of course there's a major moral agenda in pushing abstinence over safe(r) promiscuity.
Bear in mind also that condom availability has little to do with utilisation. From what I vaguely recall there's a big machismo issue, and to use a condom implies distrust or that you already have the disease.
And AIDS isn't just a sexual disease - the greatest number of new infections in most of the African states is from mother to new-born baby.
no subject
Date: 2004-07-23 07:00 am (UTC)I hate misleading graphs.. As part of my honours course a couple years ago, we had to find bad statistics and present them in class. Such fun!
no subject
Date: 2004-07-23 08:15 am (UTC)Bear in mind also that condom availability has little to do with utilisation. From what I vaguely recall there's a big machismo issue, and to use a condom implies distrust or that you already have the disease.
And AIDS isn't just a sexual disease - the greatest number of new infections in most of the African states is from mother to new-born baby.
So so true.
And even the highest "condom availability" figure was pretty pitiful -- something like 10 per male aged 15-59 per year. (Although I have nothing to compare it to, so maybe that's a perfectly acceptable figure, but it doesn't sound all that great to me. Of course, maybe there are lots of Africans with sex lives just as pitiful as mine and it averages out.)
no subject
Date: 2004-07-23 08:17 am (UTC)It gets to me because it's my job to present complicated medical information in a way that's clear and easy to grasp, and I hate seeing it done badly. Especially something as loaded as this. (I mean, I mocked the totally useless "graph" they had on the front page of the business section a few months ago -- April 23rd, I still have it -- but it didn't outrage me because it wasn't all that important.)
no subject
Date: 2004-07-23 08:37 am (UTC)Not that they printed it today. Hmph.
no subject
Date: 2004-07-23 08:37 am (UTC)*sigh*
Date: 2004-07-23 02:08 pm (UTC)*hugsELL*
Re: *sigh*
Date: 2004-07-23 02:10 pm (UTC)Re: *sigh*
Date: 2004-07-23 02:22 pm (UTC)*smacksELLupsidetheheadlovingly*