The STI Argument for Circumcision Has a Hole In It: Risk Compensation

The single most powerful argument for circumcising healthy boys is a number: the randomized trials in Orange Farm, Kisumu, and Rakai reported that adult male circumcision cut female-to-male HIV transmission by roughly 60%. From that finding grew a continent-wide campaign — voluntary medical male circumcision, or VMMC — and a talking point that now travels far beyond Africa: circumcision prevents disease, so the trade-off is worth it. It is worth looking closely at that number, because the people who cite it most rarely mention what it leaves out.

The claim, stated fairly

We will not pretend the biology is nothing. Reviews by circumcision's strongest advocates compile real associations: lower rates of HIV during heterosexual intercourse, and reduced risk of some other infections — high-risk HPV, HSV-2, syphilis. We catalogue those papers in our reference library, including the other side's most thorough version of the argument, so the case can be engaged at its strongest rather than a strawman. The question is not whether circumcision does anything. It is whether that something justifies cutting a child — and whether the prevention math is even complete.

The variable the trials forgot

It is not. In 2007, writing in PLoS Medicine, Kalichman, Eaton and Pinkerton named the hole in the model: risk compensation. When people believe a technology has made them safer, a predictable share of them take more risk to match — fewer condoms, more partners, earlier sexual debut. The circumcision trials, conducted under intensive counselling and free condom supply, never fully accounted for what happens when that protective belief meets the real world without the counsellors.

The consequence is an irony the campaigns do not advertise. A condom blocks essentially every sexually transmitted infection and protects both partners. Circumcision blocks some infections, partially, in one direction. If the belief in circumcision's protection nudges a man away from the condom, the math can invert: he is now more exposed to the gonorrhea, chlamydia and other infections the condom would have stopped — and so is she. A partial, one-directional shield purchased at the cost of the complete, two-directional one is not obviously a bargain.

What the evidence actually shows

Here honesty cuts both ways, and we will be honest. The evidence on whether risk compensation actually materializes is mixed. Some cohorts — including post-trial follow-up in South Africa — found circumcised men did not reduce condom use. A Cape Town study even found that men who had heard circumcision was protective used condoms slightly more, not less. Other work, in Kenya and elsewhere, found exactly the compensation Kalichman warned about. The truthful summary is not "circumcision increases STIs." It is narrower and harder for the other side to dismiss: the protective claim is incomplete, the real-world effect is unsettled, and a confident "it prevents disease" oversells a contested, partial, condom-inferior benefit.

The deeper blind spot

Step back and a pattern appears in how these arguments are built. Researchers optimize a single variable — a virus, in a population — and report the number that variable produces. What rarely enters the model is the human being downstream of it: the man whose reduced sensation may change how he uses protection, the second-order behavior, the psychological weight of having been permanently altered as an infant for a statistical edge he never chose. A circumcision that lowers one risk while raising another, and removes healthy tissue from someone who could not agree, is not the clean win a single hazard ratio makes it look like.

Why it doesn't matter as much as they think

And here is the part that makes the whole STI debate, for us, beside the point. Suppose the prevention figures were airtight and risk compensation never happened. The case for infant circumcision would still fail, because adults who want the protection can choose it for themselves, with consent, as adults — and a child cannot. We don't remove healthy appendices or breast tissue from infants to lower their lifetime cancer odds, even though it would. Bodily autonomy is not a number to be outweighed by a better number. It is the thing the numbers were never entitled to decide.