The New York Times recently reported that using Depo-Provera, one of the most popular contraceptives in eastern and southern Africa, may increase a person’s risk of transmitting HIV. I fear this is a case for The Guardian‘s Ben Goldacre… where a study gets a bit (understatement) too much spin in the media. I first became aware of this while in Uganda and saw the following headline in the local paper: “The injectable contraceptive that could double the risk of women contracting HIV.” That sure sounds like the shot itself does something. Or could this instead be a by-product of behavior change? Huge difference if you are deciding what birth control to use!
The Times article cited a study recently published in The Lancet, which showed that women using hormonal contraception—primarily the injection more commonly known in the U.S. by its brand name, Depo-Provera—were twice as likely to acquire HIV from their infected partners, and twice as likely to transmit the virus to their HIV-negative partners.
Whether this is because of biophysical changes or behavior changes is not obvious. And the paper acknowledges this. Several studies have questioned whether hormonal contraceptives have physiological effects beyond pregnancy prevention, perhaps by altering hormone levels in genital tissue or vaginal lining. To support that possibility, this study tests for the concentration of the HIV virus in women’s vaginal fluid, and finds that women using hormonal contraceptives have higher concentrations of the virus than women who are not using the same birth control method, even though the level of HIV in their blood is no different.
But this test still does not disprove that women may be acting differently from the ones using other forms of birth control. Rather than the hormone itself exacerbating the HIV epidemic, this might just be another example of the Peltzman effect: drivers wearing seatbelts feel more secure, and therefore drive less carefully; the number of traffic accidents goes up, and the benefit of wearing a seatbelt is offset. Women using hormonal contraception may feel more protected against pregnancy, and so take fewer precautions against other risky sexual behaviors, leading to increased transmission of HIV and other sexually transmitted diseases. Or, women lose the socially-acceptable reason for wearing a condom (“I don’t want to get pregnant”) and are stuck having to resort to a trickier message to an impending sexual partner (“I don’t trust you to not have HIV.”) So they end up having more unprotected sex.
In The Lancet study, all contraceptive use was self-reported, including hormonal contraceptives and condoms. This is obviously not easy data to collect, in a truthful way (see earlier post on List Randomization for one technique). Still, there is some evidence that people using hormonal contraceptives engaged in riskier sexual behavior. For women with HIV, the ones using hormonal contraception were slightly—and statistically significantly—more likely to have sex without a condom with their HIV-negative partners (12.9 percent) compared to those not using hormonal contraception (10.1 percent).
The World Health Organization is now reconsidering its recommendations on contraceptives, while international health organizations like FHI 360 were quick to respond that more rigorous evaluation is needed. Clearly a case of unintended consequences… personally my hope (hopefully not falsely optimistic) is that the lesson is to integrate lessons about risks of HIV transmission; or perhaps the answer is more private administration of Depo-Provera? Certainly at a minimum this is a call for a better understanding of the behavior change.
And what’s all this have to do with hockey? This recent paper by Alberto Chong and Pascual Restrepo finds that hockey players got more penalty time after switching to helmets with visors, ie, ones that protect them a bit more.
While we’re on the topic of hockey, can someone more informed than I explain why they don’t make Sumo wrestlers goalies? Or a Sumo goalie and one Sumo defender? Is it because it would shift the equilibrium too fast and everyone would do it, and then just cause rule changes to prohibit it? Or do existing rules prohibit it in some way? Or, is it just too hard to find one who can ice skate?