Friday, January 6, 2012

Male circumcision and risky sexual behavior

There is a well-known phenomenon that states that when an activity becomes safer, people will respond with riskier behavior. The classic example is the use of seat belts that has lead to more aggressive driving (although there is a claim that this example is wrong). Now switch to another risky behavior: sex. It is well established that male circumcision leads to a significantly lower transmission probability of AIDS. Is risk compensation also happening here?

Nicholas Wilson, Wentao Xiong and Christine Mattson look at recently circumcised males in a region of Kenya and find that their sex behavior is less risky than the uncircumcised control group. This is not a selection effect, as circumcision has been randomly assigned. The authors think that the circumcised ones have become less fatalist about future life prospects and thus changed their behavior for the better (this is also why you want to provide health insurance conditional on not dying from AIDS). I wonder though whether the circumcision has made them more aware of the risk of AIDS as well.


Allen Stairs said...

Didn't read the study, but I recall someone raising a question about it elsewhere: for at least a while after circumcision, we'd expect less risky behavior because we'd expect less sexual activity, period. I'm assuming (hoping) the authors took this into account, but a word or two about it would be useful.

Economic Logician said...

Good point. It is not mentioned. However, they measure sexual activity in the first six months after circumcision, and the following six months. Results are more significant in the second interval, where your point should not have bite.

Dani said...

The data on circumcision and AIDS prevention is flawed. While rates of AIDS went down in one region with increased genital cutting. In another it went unchanged. Although a lot of money is to be made with prolific medical procedure, yes? This seems sad to me. In my opinion the money should be spent on AIDS education and condom distribution.

Hugh7 said...

Did they factor in the additional counselling on HIV transmission and safer sex that the men assigned to circumcision get? Even sitting in a waiting room or a recovery room and being exposed to leaflets (while experiencing at least a nagging ache down there) could make a difference.

In 10 of 18 countries for which USAID has figures, more of the circumcised men have HIV than the non-circumcised. In Malaysia, 60% of the population is Muslim (almost the only circumcised people in that country) but 72% of HIV cases are Muslim. Shouldn't all that at least be explained before asserting that it is "well established that male circumcision leads to a significantly lower transmission probability of AIDS"?

The latest review damning the "circumcision prevents HIV" studies is in the (Australian) Journal of Law and Medicine last month. The ENTIRE basis of the claim is 73 out of 5,400 circumcised men who didn't get HIV in less than two years, who MIGHT have if they hadn't been circumcised, while 64 circumcised men DID get it.

G. Boyle and G. Hill unravel some of the many reasons that may not be cause and effect:
researcher expectation bias;
participant expectation bias;
inadequate double blinding;
lead-time bias;
selection and sampling bias;
attrition bias; and
early termination.

Contacts were not traced so we don't even know which if any of the men got HIV from women or even by sex. Non-sexual transmission is a large and unaddressed issue in Africa, and male-male transmission goes largely denied and underreported (especially in countries like Uganda where anyone who admits being gay risks death).

Mark said...

You said this:
"It is well established that male circumcision leads to a significantly lower transmission probability of AIDS."

This is simply not the case:

"There appears no clear pattern of association between male circumcision and HIV prevalence—in 8 of 18 countries with data, HIV prevalence is lower among circumcised men, while in the remaining 10 countries it is higher."

The South African National Communication Survey on HIV/AIDS, 2009 found that 15% of adults across age groups "believe that circumcised men do not need to use condoms".

From the committee of the South African Medical Association Human Rights, Law & Ethics Committee :
"the Committee expressed serious concern that not enough scientifically-based evidence was available to confirm that circumcisions prevented HIV contraction and that the public at large was influenced by incorrect and misrepresented information. The Committee reiterated its view that it did not support circumcision to prevent HIV transmission."

The one randomized controlled trial into male-to-female transmission showed a 54% higher rate in the group where the men had been circumcised btw:

ABC (Abstinence, Being faithful, and especially Condoms) is the way forward. Promoting genital surgery will cost African lives, not save them.

Kansan said...

It looks like the study is very careful to keep the circumcised and uncircumcised group informed in a similar way. In fact, they check that their beliefs about the effectiveness of circumcision does not differ across groups, and it does not, even after 12 months.

And after all, it is not the actual effectiveness that matters in this case, only the belief of effectiveness.

TLC Tugger said...

It is false to describe this study as a randomized controlled trial. There was no randomly assigned intervention group. ALL the participants chose to be circumcised, all were paid for their participation, and all knew the aim of the their benefactors was to demonstrate the effectiveness of circumcision.

The gold standard in research is the double-blinded trial. This is especially important when the subjects or the researchers themselves have reason to promote one or another outcome.

These men chose circumcision for cultural reasons. These researchers have a long history of promoting circumcision - for inexplicable reasons.

Anonymous said...

They all chose to be circumcised. But only some of them actually were, randomly assigned. Isn't that randomization?

Hugh said...

Anonymous: Men who choose to be circumcised are not a random sample of the population. Since they were paid to take part (and quite well by local standards) the sample would also be skewed demographically - they'd be more likely to be poor. This means their sexual behaviour and other HIV risk-factors may not be typical of the population either.

Simon said...

Hugh makes an interesting point:

"Since they were paid to take part...the sample would also be skewed demographically - they'd be more likely to be poor."

While many diseases infect poorer people more than richer people, this is often not the case with HIV in high prevalence countries, such as South Africa, Uganda, Tanzania, etc. Similarly, HIV prevalence is often higher among better educated quintiles.

Therefore, the lack of randomization could result in groups with lower HIV prevalence taking part in the trials (which could be a significant factor even if HIV positive people were excluded from the trial). This effect would be strengthened if the trials included a lot of rural dwelling people, among whom HIV prevalence is almost always lower.

Rwanda is one of those countries where HIV prevalence is clearly higher among circumcised than uncircumcised men. And while HIV prevalence is lower among Muslim men than any other religious group, prevalence is about three times higher among Muslim women.

daniel dwight said...

How can someone assure that being circumcised can prevent the spread of HIV and other transmitted disease? The increase in the widespread of the disease is very alarming.