Tuesday, June 17, 2008

The Davis-Weller Study on HIV Transmission Misrepresented Again

In 1999, Davis and Weller published a paper in Family Planning Perspectives entitled "The Effectiveness of Condoms in Reducing Heterosexual Transmission of HIV". Their meta-study combined results from other studies about couples where one partner was HIV-positive and the other not, with respect to how often a condom was used. Unfortunately, their conclusions (and the conclusions of an earlier study by Weller alone) have been systematically misrepresented by conservative Christian activists.

I've written about this misrepresentation before. In 1993, Harold Albrecht, a local MP, misrepresented the earlier study by claiming that "An analysis by researchers at the University of Texas estimates that when condoms are used, the risk of acquiring HIV from an infected partner is 31 per cent over a year's time."

Now the work of Davis and Weller has been misrepresented again. Our local paper, the Kitchener-Waterloo Record, has a "Community Editorial Board", where local residents are tapped to write a series of opinion pieces. (I was on the Board in 2000, and you can see my columns here.) Yesterday the Record carried this column by Harriette Mostert, who is described as a "part-time teacher and a longtime community volunteer".

Mostert claimed that this NIH report says that "HIV/AIDS carries a 15 per cent risk of transmission even with a condom". However, the NIH report was referring to the Davis-Weller study, and it is being misrepresented again.

The Davis-Weller study found that using a condom reduces the risk of HIV transmission by 85%. Now, you might think that Mostert's 15% figure is just 100%-85%, and so she's correct. But you would be wrong.

Davis and Weller were studying the reduction in risk obtained when using a condom; Mostert incorrectly labels this the "risk of transmission". They're not the same at all. I think most people would interpret "risk of transmission" as meaning "the chances that you will get the disease in a single encounter", and indeed, that's the interpretation I got when I asked several people what they thought it means. Or maybe it means "the chances that you will get the disease in a year"? The lack of units should raise warning bells in the mind of any educated person.

The answer is that Davis and Weller found that in one year, 6.7 infections per 100 person-years occurred when a condom is not used, and 0.9 infections per 100 person-years occurred when a condom is always used. The reduction in risk is therefore (6.7-0.9)/6.7, or approximately 85%. Contrary to Mostert, the "risk of transmission" when using a condom is less than 1 infection in 100 person years. There's no way this can be characterized as "15%".

Mostert uses this bogus figure to argue for "chastity or monogamy", and smugly concludes "Interestingly, this is also consistent with the ideals set out in many faith communities." She fails to note that the very study she cites concludes "These data provide strong evidence for the effectiveness of condoms for reducing sexually transmitted HIV."


Anonymous said...

in the first paragraph, you mean "systematically misrepresented"?

Jeffrey Shallit said...

Thanks for catching that! I've fixed it.

Unsympathetic reader said...

What is a 'person-year'. Is that equivalent to having sex every day (365 person-days = 1 person-year)?!!! Or once a year?

And for couples, aren't we talking about two people and therefore, two person-years? If that's the case, then could one person-year mean that there was only one person having sex? What are the odds of catching HIV from exactly one partner, if that partner is yourself? Pretty minuscule, I'd bet and a clear case where using a condom would provide absolutely no reduction in the risk.

I can see the headlines now in Wingnut Daily: Condoms provide absolutely no protection from HIV*

* ...when used with no partners.

Amy© said...
This comment has been removed by the author.
ctyri said...

Paraphrasing a quote (I've forgotten who to attribute it to...):

She uses statistics like a drunkard uses a lamppost, for support rather than illumination.

(Found your blog looking up "science, math, politics, religion" blogs... by the way.)

David Swart said...

Jeff, regarding a version of this blogpost in the letters to the editor of our local paper The Record.

My coworker wishes me to pass on thanks. We both believe there will be a very real benefit to all who read it.