|Turned up to eleven: Fair and Balanced|
Sunday, April 27, 2003
So, I embarrassed myself yesterday in Matt Yglesia's comment section, ripping on some guy for saying there were 3.1 million HIV deaths per year, when I KNEW that couldn't be the case. Well, after going to the WHO site and poking around, I found out I was wrong. The reason? The last time I looked at mortality statistics was probably 1997 or 1998, when I was a grad student studying for my oral examination, and the situation has changed, it would appear. So, forty lashes with a wet noodle for me. But, in the course of learning about this, I found some serious discrepancies in the counts, and I figured out why. On this "fact sheet" (from July 2000!!!), they say 5 million deaths per year from HIV/AIDS, Malaria, and TB. Now, here is a global map of HIV/AIDS deaths worldwide, showing the estimated 3.1 million deaths per year, and here is the fact sheet on TB, showing 2 million deaths per year. Now, I know that Malaria kills about 3000 African children per day, which means that unless no one else in the world dies of Malaria, more than 1 million people die from this disease (I would guess the number is considerably higher).
Double counting, funding, and confusion
Now, to be very clear, I am forced by this data to reassess my position on HIV/AIDS v. other infectious disease. While I always thought that all of the pandemic diseases were a big problem, I didn't realize that the AIDS pandemic had grown so much in the past 5 years (boy, that has gotten a lot of media coverage, hasn't it?). But, there is a serious question for epidemiologists and microbiologists, which is, how do you decide what someone died from? Now, in many, many cases this is quite easy, but suppose a man comes into a hospital and dies, and when you test his blood, you find HIV, the malaria parasite, E. coli, leishmania, and several things you've never seen before (and this is just the blood smear!). So, what did he die of? Now, of course, there aren't literally thousands of WHO people out there counting all the bodies, autopsying them, doing blood work, etc. They "estimate" the numbers, presumably relying on hospital and public health reports from the countries (hmm, that could be a big source of systematic error), and epidemic models to extrapolate from that. Now, here's a cynical thought; if you want money to bolster your public health system, do you list the guy above as dying from gram-negative sepsis (E. coli), malaria, or HIV/AIDS. I think you know the answer.
Slow moving infections, latency, and confusion
There is a benign explanation, also, and I am sure that the real thing is a combination of the benign and malign. The reason is, you really cannot die from the direct effect of HIV. You will, eventually, definitely die (with a very few exceptions), from a disease that was able to kill you as a result of the weakening of your immune system by HIV. Because of this, anyone who dies of AIDS actually succumbs to an infection by some other microbe besides HIV, which means that that microbe is the proximate cause of death. Now, when the bug in question is something that never causes disease in HIV negative people, like the fungus Pneumocystis carinii, this is no problem; that is an "AIDS related cause". But what about TB? Schistosomiasis? Trypanosome diseases? Malaria? There are synergies, I am quite certain, between the many chronic infectious agents prevalent in the tropics, but when trying to "count" the dead, it seems at best naive and at worst cynical to pick a cause out of this particular hat. No one wants to hear "20 million people died of malaria, AIDS, TB, Schistosomiasis, diarrheal disease, and respiratory infections", that doesn't seem very helpful (and it's not). But the real story is too complicated for a "fact sheet", I guess.
Take home message
What is the take home message? I suppose it is this; There are a number of infectious diseases, chronic ones especially, which are a major, major problem, particularly in sub-Saharan Africa, but also in Southeast Asia, and to a lesser extent, South America (note to self: why not so much South America?). These diseases, taken together, constitute a huge burden on the people, societies, and economies in those regions. But they are not easily separated into individual infectious disease foes, and must be tackled with an eye toward reducing all of their incidences. The figures used to describe them are, when scrutinized, not all that accurate, but taken together describe the overall picture. A grain of salt should be applied to the individual disease mortality rates. I want to stress here that I am not trying to minimize the plight of these people, suffering under this enormous burden, but merely to discuss the complexity of infectious disease surveillance.
UPDATE: Thanks to Gregg Easterbrook (not that he knows about it) for the "science-like substance" line, a staple in his terrific TMQ column (here is the most recent example of his brilliance).