It is Sweet and Beautiful to Die for One's Scientists

A little while back, I blogged about research that identified HIV prevalence 'hotspots', where high figures were clustered together. This research, carried out in Lesotho, looked at hotspots for males and females and found that they were spatially distant, which may seem anomalous for what is said to be a mainly sexually transmitted disease.

Similar work carried out in South Africa was a bit disappointing. Because, even though it is well known that fewer men are infected than women, often far fewer, this research by Handan Wand and Gita Ramjee only looked at hotspots of high HIV prevalence and incidence for women [Apologies for providing the wrong link yesterday.] Excluding men from such research is odd if it is assumed that women are mainly being infected by men.

But it clearly is assumed that all, not just almost all, transmission is through sex. "At all visits, all participants received counselling on risk reduction and as many male condoms as desired. Counsellors emphasized that condoms are the only known method to prevent HIV and sexually transmitted infections (STIs), and that condoms should be used for every act of sex."

There is something small but crucial wrong with this statement: condoms are the only known method to prevent sexually transmitted HIV. But using condoms during sex will not protect against non-sexually transmitted HIV, which could result from unsafe health care or unsafe cosmetic practices.

Indeed, failing to inform people about non-sexual risks means that they are unable to protect themselves from them and this could be a reason why HIV prevalence is so high in some sub-Saharan African countries. Pretending that HIV is always (or even almost always) sexually transmitted flies in the face of all evidence and is probably responsible for a substantial proportion of all transmission.

As for this research, it found some significant correlations between being HIV positive and certain types of behavior considered to increase the risk of being infected with HIV. 'Significant' in the strict statistical sense of the word, but not particularly large correlations. A non-statistician might think the data needs to be re-examined in the light of such apparently minor differences but not our intrepid authors.

For example: "The proportion of women who reported being legally married was significantly higher among those outside the hotspots than within them (16% vs. 12%, p = 0.001). Significantly more women in the geographical hotspots reported being Christian (94% vs. 90%, p < 0.001) and speaking Zulu at home (91% vs. 86%, p < 0.001) compared with those in non-cluster areas."

The article goes on: "The spatial clustering of HIV cases was found to be related to certain demographic and risk behaviours. Number of male sexual partners was not collected in this study; however, being single, combined with high frequency of sexual acts, gives strong evidence for those women having multiple partners, as well as possibly engaging in transactional sex."

'Strong' evidence? How does this 'high frequency of sexual acts' compare to countries with low HIV prevalence? The same question regarding 'multiple partners'? And would the authors come to the same conclusion of 'possibly engaging in transactional sex' if the risk factors in a low prevalence country were similar? I don't think so.

The authors may be leaving the door open to further research about the safety of health care facilities, but then again, they may not: "These results may be due to fundamental differences between the communities with regard to health care centres, population density and other socio-economic factors. These data provide new evidence to support the need to investigate potential sources of infection and to study transmission patterns in the community in order to apply relevant interventions for prevention of this devastating disease."

Let's hope that 'relevant' interventions means interventions that prevent non-sexual as well as sexual transmission, but it doesn't look as if these particular researchers will see things that way.

It is concluded that "Information on the spatial distribution of populations and services is essential to understand access to health services." But if some HIV is being transmitted nosocomially, as a result of inadequate health facilities, this also needs to be established. Otherwise increasing access to health services could result in higher rates of HIV transmission.

The authors don't just need to "determine and target the specific communities that are most in need of education, prevention and treatment activities", they also need to determine exactly what sort of education and what sort of prevention activities are required. Otherwise, at best, things could remain as bad as they are and atworst, they could get a lot worse. Neither of these would be good for South Africa.