I see that Hank’s still touting his “Padian found ZERO seroconversions among 176 discordant heterosexual couples over 6 years” line, even after that study was extensively analyzed here (and his ideas about it were shown to be mistaken). At least he’s dropped it from 10 years down to 6 years–progress, I suppose. He’s now discussing a new MMR report that documents 88 HIV transmission events over a 17-year period in Georgia prisions. Despite the fact that Hank doesn’t dispute that these sexual transmission events occurred (noting that the “infectivity is low, low, low”), Hank still ends his post with the question: “How can you have a STD, that ain’t transmitted by sex?”
Of course, what he neglects to mention is what they actually did in the study.
First, a bit of background. The seroprevalence for HIV, as noted in their paper, is about 5 times higher in the prison population than in the public at large (2% in prisons vs. about .4% nationwide); hence, prisons are a good place to carry out studies looking at HIV. The paper stems from a convenience: the Georgia Department of Corrections began mandatory HIV testing of inmates entering their facilities in 1998, so the serostatus of all inmates who entered the prison since that time was known. Additionally, they began offering voluntary HIV testing to inmates at that time, increasing it to annual voluntary testing in 2003 (“or if clinically indicated,” meaning likely if they had some form of unexplained immunosuppression or an AIDS-defining disease) and from 2005 to present offers testing on request.
Note several things in that paragraph: the “voluntary” follow-up testing. This introduces a giant bias in their study. Testing on admission is mandatory–testing after that is voluntary. The study isn’t inclusive of all prisoners for this reason, and there’s no data presented on how many prisoners were actually voluntarily tested.
So, as far as HIV seroconversions, they found:
During July 1988–February 2005, a total of 88 male inmates were known to have had both a negative HIV test result upon entry into prison and a subsequent confirmed positive HIV test result (i.e., seroconversion) during incarceration. Of these 88 inmates, 37 (42%) had had more than one negative HIV test result before their HIV diagnosis.
However, contrary to what Hank states (and similar to the Padian study in the link above), this wasn’t the end-point of their study. The point wasn’t to measure the rate of transmission of HIV, and despite what Hank and news write-ups of the research claim, they can’t tell from their data that “few prisoners contract HIV,” because we don’t actually know the numerator data (in other words, how many prisoners are actually currently seropositive to HIV). We only know how many prisoners, out of those who voluntarily asked for an HIV test, have been infected since they were incarcerated. These could represent 50% of the actual HIV seroconversions, or 5%. Because the test was voluntary, we simply don’t know how many HIV seroconversions were missed during the 17 years since the introduction of the test upon admission. Of course, that doesn’t make for as good of a story, now, does it?
The real topic they were investigating here were risk factors for HIV transmission. These they could look into, because they knew that, in their sub-population of people, they had become HIV-seropositive during their time in prison, and they could try to tease out what put them at a high risk of becoming so. They compared these, then, to control inmates who were re-tested and found to still be negative for HIV antibodies. They then interviewed these men (the cases and controls), asking about “sex, drug use, and tattooing behaviors during the 6 months before incarceration and during the incarceration period.” They then modeled this several different ways, and in their final model found that male-male sex in prison, receipt of a tattoo in prison, BMI of <25.4 kg/m2 on entry into prison, and black race were all factors that increased the likelihood of HIV seroconversion.
Additionally, of those interviewed, they found that 72% reported engaging in consensual sex; are these representative of the prison population? Relatively few reported rape; might those who are raped (and potentially contract HIV that way) be less likely to come forward for voluntary testing? We don't know, as this isn't addressed, and again, it shows why we should be careful extending too much significance to the "only 88 seroconversions in 17 years!" summary that Hank gives the paper. Like the Padian paper, this study was meant to determine risk factors for transmission of the virus, not the overall seroconversion rate. In the case of the Padian paper, it was risk factors among couples; in the MMR paper, it’s risk factors within a prison population.
So, what we end up with here is a paper showing 88 documented HIV seroconversions in a state prison system, and showing that being black, entering prison with a BMI on the lower side, getting a prison tat, and having homosexual sex increased the likelihood of seroconverting while in prison. And even in the (IMO pretty poor) Washington Post article, they note that this was transmitted “…chiefly through same-sex intercourse.” From this, Hank concludes from this that this means that HIV “…ain’t transmitted by sex.”
CDC. 2006. HIV Transmission Among Male Inmates in a State Prison System — Georgia, 1992–2005. April 21, 2006 / 55(15);421-426. Available online at http://www.cdc.gov/mmwr/preview/mmwrhtml/mm5515a1.htm