Brava! to Caitlin Gerdts and Divya Vohra at Daily Beast for a superb, and much-needed, dissection of the flaws in this week’s heavily hyped Lancet study by Heffron et al. The study purported to show elevated HIV risk associated with hormonal contraceptive use among women in parts of Africa (abstract here, subscription needed for full text). The NY Times ran a front-page story, claiming that
[t]he most popular contraceptive for women in eastern and southern Africa, a hormone shot given every three months, appears to double the risk the women will become infected with H.I.V.
Gerdts and Vohra add the essential context that was missed by the newsmedia: about a half-million women die during or because of childbirth each year, almost all of them in poor countries.
At RH Reality Check, Jodi Jacobson summarizes the main cautions about the Heffron study, and points to a Guttmacher Institute white paper. She takes into account concerns about high maternal and infant mortality in parts of Africa, the harms associated with complications of pregnancy and unsafe abortions, and, of course, the substantial possibility of vertical transmission of HIV in places where antiretroviral therapy isn’t universally available.
To claim that poor women should give more weight to the rather remote risk of acquiring a virus that might cause serious illness years down the road than to the dangers of pregnancy itself in the near term is to reduce real women to automata. Facing dire straits they might be, but they’re supposed to be reasoning machines, programmed to engage in the AIDS industry’s preferred calculus, risk.
A quick summary of the shortcomings of the Heffron et al. research: comparing users of hormonal contraception to nonusers, the difference in actual risk of acquiring or transmitting HIV was very small, amounting to 1 to 3 new infections per one hundred contraceptive users over and above the infection rate for nonusers. And it’s impossible to say that these excess infections were actually attributable to the contraceptive — because the study wasn’t a clinical trial. A great many aspects of social setting, relationships, health, and welfare of the study subjects would have been different between contraception users and nonusers, some of which would undoubtedly account for differences in rate of HIV transmission.
And since all of the subjects were in so-called discordant couples — one partner infected with HIV, the other not — it would be unusual to expect no HIV transmission at all.
Unless, of course, the subjects had been offered antiretroviral therapy, ART. But this the researchers did not do. They referred eligible subjects to HIV clinics. They seem not to have checked whether people who needed ART were getting it. They seem not to have offered ART to women who got pregnant, either. Certainly, their Lancet article makes no report of doing so.
I wondered if this was too much to expect of researchers — so I asked the students taking my course on global AIDS and human rights. Undergraduates, I find, generally have a clearer sense of ethics than most medical researchers.
Even the students who felt that the Heffron study was worth doing and basically sound were troubled by the researchers’ lack of curiosity as to whether HIV-infected subjects were getting the ART drugs they needed. And most of the students thought this was a disabling ethical fault, which should have caused human subjects committees to make the researchers redesign the study. One student pointed out that the Bill and Melinda Gates Foundation, one of the study’s funding sources, could easily afford to pay for antiretroviral therapy for all of the roughly 2,000 HIV-infected people in the study.
In the end, my students had the questions that Marcia Angell raised in her editorial in the New England Journal of Medicine in 2011 (PDF at this link angell editorial nejm 2000 ): Don’t physician researchers have the same responsibility to study subjects that they do to their own patients? And therefore, when their subjects lack resources to obtain effective therapy for treatable conditions, don’t the researchers have a moral obligation to make the therapies available?
Heffron et al. didn’t do this. They watched HIV-infected people transmit HIV to their partners (the researchers provided HIV testing and counseling about avoiding transmission — but they don’t make clear whether they notified uninfected partners that they might be in harm’s way). They did little to prevent transmission. Notably, they didn’t offer ART to people with low CD4 counts. Nor did they offer post-exposure prophylaxis to uninfected people who had had intercourse with an infected partner. They just watched.
On this account, the Heffron study wasn’t only flawed — it was so questionable on ethical grounds that the Lancet should be ashamed to have published it. And the funders — the NIH as well as the Bill & Melinda Gates foundation — censured.