Flawed HIV Transmission Models Beget Flawed Strategies

A researcher called Jeremy R Magruder has written an interesting thesis about 'marital shopping' or 'spousal search' patterns being behind high and medium prevalence HIV epidemics in African countries. According to his model, when people are searching for a partner, they have unprotected sex with several different people in a relatively short space of time.

The factor that really increases the probability of HIV transmission is that people who have recently been infected are most infectious. Their HIV positive status may not even be detected by tests, so regular testing would be unlikely to help much. Therefore, this brief period of rapid partner change is likely to result in more transmissions than other scenarios.

This model is interesting because Magruder suggests that advising people to use condoms for the first three months of a new relationship could provide very high levels of protection. Better still, evidence has suggested that people respond better to behavior change strategies that are more feasible.

Unsurprisingly, being told to abstain for an indefinite period, even if it is until marriage, is a lot less feasible than abstaining until the end of secondary school. And using condoms for a few months into a new relationship is a lot more feasible than using condoms, not just until marriage, but throughout marriage.

Magruder also notes that efforts to reduce sexually transmitted infection prevalence and male circumcision are unlikely to have much effect on transmission rates. He even dismisses the benefits of looking at antiretroviral treatment as prevention, something some technology fanatics have suggested as their magic bullet of choice.

But the model becomes less credible when you start to ask why prevalence in some African countries is only 1%, in others it is over 5% and in the worst affected countries it is over 25%. Prevalence is over 40% in some demographic groups and many people who never have sex, or who always use protection, are also infected.

Within countries, prevalence also varies greatly, with less than 1% in some areas and figures that are 10 or 20 times higher in other areas. Variation between tribes can be equally high and the ratio of male to females in some tribes range from 17% to 120%. Could this really be explained by differences in spousal search patterns?

Magruder could explain these differences by reference to differences in levels of 'unsafe' sexual practices or in levels of protective behavior. But this is one of the great mysteries of sexually transmitted HIV: transmission rates can be higher among those who practice 'safe' sex. Attested levels of safe and unsafe sexual practices don't correlate with HIV prevalence.

Also, is Magruder suggesting that spousal search patterns changed some time in the 1970s and 1980s, a phenomenon which might explain why the virus was spread so rapidly during this period and not before? And is he suggesting that this search pattern changed again some time in the 1990s, when HIV incidence, the yearly rate of new HIV infections, peaked and declined in most African countries?

Magruder has assumed the truth of a version of the behavioral paradigm, the view that HIV is almost always transmitted sexually. He simply ignores the possibility that some, perhaps a lot of HIV, is transmitted non-sexually.

But Magruder doesn't subscribe to the view that African people engage in inordinate amounts of sex with little concern for the consequences for themselves or their families. And that is refreshing. But it may not help him gain acceptance for his thesis. The HIV orthodoxy is very committed to its racist and sexist views of HIV transmission.

Magruder may have shed some light on how sexually transmitted HIV could be spread by people who are only human and can only engage in humanly possible levels of sexual behavior. But until some light is also shed on what proportion of HIV is transmitted sexually and what proportion is transmitted non-sexually, his strategy will bring only limited benefits.