Analysing the analysis of analysis of...

Today, I became aware that Tony at RollingDoughnut.com has published an analysis of my analysis of Intact America's propaganda.

Ok, let me roll up my sleeves and dissect Tony's remarks:

Intact America ran an open letter, as an advertisement, in yesterday's Washington Post urging the American Academy of Pediatrics not to recommend that American parents circumcise their infant sons as a strategy against HIV. [Full disclosure: I attended an event hosted by the organization and interact with some of its representatives because I support its cause.] It's a logical request, based on the necessary combination of science and ethics. A pro-circumcision advocate, Jake Waskett¹,

Tony includes a lengthy footnote at this point, justifying his assumption. I'll omit it, because he's correctly guessed my name. It's a shame that he mischaracterises me as a "pro-circumcision advocate", though (I'm pro-parental choice, not pro-circumcision).

has attempted a deconstruction of the letter, labeling it "propaganda".

"Labelling" seems a curious choice of word, implying that the choice of term is dubious. Propaganda is defined as "The systematic propagation of a doctrine or cause or of information reflecting the views and interests of those advocating such a doctrine or cause." Thus, it seems a perfectly appropriate choice of term for an advertisement created by an anti-circumcision organisation for the explicit purpose of promoting their cause to the AAP.

His support for that charge is preposterous, as any approach advocating the circumcision of healthy infant males must inevitably be, but his critique fails because he ignores the central issue involved. After a brief introduction, complete with an absurd assumption about Intact America's motives, Waskett quotes the opening paragraph:

(Quotation omitted. I think we can regard this as "noise"...)

I agree with this, [ie., IA's opening paragraph] although I'm not a fan of appeals to authority. As should be evident with the apparent intention of the CDC to recommend infant circumcision, it only takes one ill-conceived recommendation to distract from the core issue.

This last sentence is utterly incomprehensible.

Despite my misgivings, Intact America structures the argument correctly because it identifies that core: ethics demand not imposing medically unnecessary surgery on normal, healthy children, regardless of gender or potential benefits.

Tony is, of course, free to subscribe to whatever system of ethics he so chooses. However, to my mind he is setting an extraordinary requirement: that an intervention should not merely be medically beneficial, but must actually be necessary. If applied consistently, such a standard would mean, for example, that vaccinations are unacceptable, since they are rarely necessary.

Waskett assesses this with an odd bit of snark about people inventing fire

Not really "odd". IA are essentially arguing that medical organisations shouldn't recommend circumcision because no medical organisations have done so to date. If you apply the same reasoning to the invention of fire, you see how absurd it is.

before issuing a parenthetical aside suggesting that the national medical bodies of African nations now implementing mass circumcision programs implies approval. Perhaps this is the case, which circles back to my reservation about an appeal to authority. But assuming it is not a point of fact. Still, if he's granted the point, what does this prove about Intact America's ethical argument?

IA offer three reasons why (according to them) it is unethical: unnecessary, potentially risky, and not recommended by major medical organisations. As I point out, the last is an extremely weak argument in the context of what the AAP may recommend. I agree with Tony, to some extent, that it is also an appeal to authority.

The risk of female-to-male HIV transmission through vaginal intercourse is a significant problem in Africa. In America HIV transmission risk through sex overwhelmingly involves male-to-male transmission, from which the (voluntary) circumcision of (adult) males has shown no statistically significant reduction.

Tony's words are somewhat misleading here. There haven't been any controlled trials of voluntary circumcision in MSM yet. The American studies to date have mostly compared previously (and probably neonatally) circumcised men with uncircumcised men. Some studies have shown a statistically significant reduction, but others have not.

Even if this wasn't the case, the ethical issue of applying scientific research to healthy children through surgery centers on infant circumcision, not infant circumcision. That's the point Waskett ignores. His defense:

(Quotation omitted.)

(Apparently in relation to my remark that risks need to be weighed against the benefits) No, these risks need to be weighed against the need, or rather, the lack of need. The ethics of proxy consent require parents to choose a balance between the most effective and least invasive solution to remedy their child's malady. But there is no malady when the boy is healthy.

As Tony correctly observes, the situation we're discussing is not one in which there is an immediately pressing need for therapeutic intervention, hence the "most effective and least invasive" criteria for choosing that intervention do not apply. Instead, the situation involves a healthy child, much as with vaccinations. And as with vaccinations, we weigh the risks (adverse reaction) against the future benefits (reduction of risk of disease). Tony is of course free to apply his own ethical standard, but he should not be surprised that others choose not to follow him.

Setting the ability to chase potential benefits as the ethical standard opens the range of allegedly valid parental interventions to include any number of surgeries we recognize as offensive. The science becomes ungrounded by any concern for the individual child as an individual.

Unfortunately, Tony hasn't identified any of these "surgeries we recognize as offensive" that are valid when benefits and risks are properly weighed. I would be interested to learn of any that he - or anyone else - can think of.

Invoking the topic of vaccinations does not change this evaluation. There are similarities between circumcision and vaccination, based on potential benefits. However, the difference rests on how the problems the interventions are meant to prevent occur. For example, becoming infected with measles requires no effort other than participation in society, while acquiring HIV from an HIV+ female through vaginal intercourse requires a very specific action, an action not undertaken by infants. Comparing the two solutions as comparable for parental consent fails.

This is a nonsensical argument: it is absurd to analyse the issue as though children never grow up. Peter Pan is fiction. Children grow up to become adults, and yes, that includes having sex.

Add to this the fact that parents treat the same maladies circumcision is supposed to prevent with less invasive, non-surgical methods when they affect their daughters, and Waskett's argument misses the ethical case against infant circumcision because he's making the case for circumcision devoid of context and ethics. That's a case that works only if it's a voluntary decision by the adult male himself.

This paragraph makes no sense.

Next, Intact America requests that the AAP defend the ethics against infant circumcision rather than considering a revision in favor of the surgery since science necessarily involves ethics when applied to a person's body, particularly via proxy consent. Waskett calls this request "bizarre," despite having failed thus far to address the ethical argument made by Intact America.

If Tony had been paying attention, he would have noticed that I actually identified the three reasons why IA claimed that circumcision was unethical, and addressed each in turn.

[Regarding my statement that the sums saved in disease prevention are probably comparable to those spend on circumcision] Waskett's claim is based on speculation. Perhaps his analysis is correct, but he does not provide proof for his assumption here. We have statistics from other western nations demonstrating the incidence rates for the diseases to which he refers. Since we can analyze circumcision on these terms, "no doubt" is insufficient

Several studies have been published that have compared the costs of circumcision with that of non-circumcision. For example, Schoen et al. The exact findings have varied a little from study to study, but most have shown that, from a cost point of view, the figures are comparable.

(At this point I'll skip a paragraph, since it basically consists of Tony repeating his claims about what is and is not justified. Since this is clearly his opinion, it seems to require little response from me.)

Waskett seems to think that Intact America ignores the randomized controlled trials showing risk reduction in female-to-male HIV transmission from voluntary adult circumcision. The letter noted this fact in an earlier paragraph. Still, as I read the letter, Intact America is not making an argument about epidemiology. Rather, it is making an argument about populations and risk factors.

As a reminder, here is what IA claim: "Doctors have a responsibility to tell parents the truth: circumcision does not prevent disease. Most European nations, with circumcision rates near zero, have lower HIV/AIDS rates than the United States."

As I read that, the second sentence seems to be presented as evidence for the first. If that is so, IA appear to be saying that the most definitive evidence can be found in between-country comparisons.

The risk factors among America's population are similar to those of European nations, not African nations. Our risk is male-to-male transmission and shared needles during IV drug use.

If Tony is confident in his assertions, perhaps he will volunteer to have heterosexual intercourse with an HIV+ woman. Probably not, I suspect, because of course that's a risk anywhere. The main difference, of course, is that the probability of exposure changes dramatically. Put bluntly, if you sleep with a person then your risk of acquiring HIV depends on the probability that they are HIV+.

Circumcision protects against neither.

I wouldn't personally make such an assertion with any confidence.

Is that complete proof that infant circumcision in America, unlike the randomized trials involving adult volunteers in Africa, is irrelevant to the United States? No, and I don't think Intact America is suggesting otherwise. It is simply working from the central fact,

Correction: opinion.

which is that it is unethical to circumcise healthy infant males - who are not sexually active - to prevent a disease for which most of them will face minimal lifetime risk and for which less invasive, more effective prevention methods exist. Europe is an appropriate anecdotal case study that (infant) circumcision is not necessary to achieve the results health officials desire.

I believe that Tony is going beyond a generous interpretation of IA's words here. IA clearly state that "circumcision does not prevent disease". That's not a statement in the context of ethics. It's in the context of science: it poses a testable hypothesis.

The complications of circumcision affect individuals. Those individual have rights. We recognize this for female minors, legislating against parental proxy consent for medically unnecessary genital surgery on daughters for any reason. The ethical argument against infant male circumcision involves the equal rights concept that the same protection should be applied to males. Waskett hasn't yet made a coherent case for denying these rights to male minors.

It is not meaningful to compare female genital cutting to circumcision. Female genital cutting is a net harm, with no known medical benefits, immediate risks, and a considerable chance of permanent harm. Society passes laws to protect the vulnerable from harm, and so it makes sense to protect children from female genital cutting. But - applying the same principle - it doesn't make sense to legislate against circumcision, because there is no net harm. Most reasonable people, weighing the risks and benefits, come to the conclusion that it is neutral or beneficial.

(Tony's final paragraph omitted, as it is essentially repetition of his arguments.)

Some HIV updates...

A recently published study, Templeton et al., reports:

On multivariate analysis controlling for behavioural risk factors, being circumcised was associated with a nonsignificant reduction in risk of HIV seroconversion [hazard ratio 0.78, 95% confidence interval (CI) 0.42-1.45, P = 0.424]. Among one-third of study participants who reported a preference for the insertive role in anal intercourse, being circumcised was associated with a significant reduction in HIV incidence after controlling for age and unprotected anal intercourse (UAI) (hazard ratio 0.11, 95% CI 0.03-0.80, P = 0.041).

This study provides valuable new data for evaluating the effect of circumcision among men who have sex with men (MSM). A meta-analysis published in 2008 by Millett et al. reported:

The odds of being HIV-positive were nonsignificantly lower among MSM who were circumcised than uncircumcised (odds ratio, 0.86; 95% confidence interval, 0.65-1.13; number of independent effect sizes [k] = 15).

But, of course, meta-analyses depends entirely on their source data; there is a need to update this earlier meta-analysis to include Templeton's findings. In particular, that earlier meta-analysis included only three studies in meta-analysis of the effect of circumcision among MSM favouring an insertive role in anal intercourse:

A separate analysis (not shown in Table 2) of 4 findings from 3 studies29, 33-34 reporting HIV infection and circumcision status for MSM who engaged exclusively or primarily in insertive anal intercourse (n = 2238) was protective, but not statistically significant (OR, 0.71; 95% CI, 0.23-2.22; k = 4). The power for this analysis was 0.94. Although there was high heterogeneity among these 4 findings (I2 = 90%), too few findings were available for a stratified analysis.

On a related topic, there are some reports from studies piggy-backing on the African RCTs that provide some new and interesting information about the mechanisms by which circumcision is protective against HIV. First, we have Johnson et al., who report:

Epithelial inflammation was present in 4.2% of men with neither HIV nor HSV-2 infection; 7.8% of men with only HSV-2; 19.0% with HIV alone (P = 0.04); and 31.6% in HIV/HSV-2 coinfected men [prevalence rate ratio (PRR) 7.5, 95% confidence interval (CI) 2.3-23.8, P < 0.001]. [...] Foreskin inflammation is increased with HIV and HSV-2 infections, higher HIV viral load and presence of smegma. Foreskin inflammation may have implications for HIV transmission and acquisition in uncircumcised men.

Next, we have Auvert et al., who report:

When controlling for all covariates, HIV incidence increased significantly with HR-HPV positivity (aIRR = 3.76, 95% CI: 1.83 to 7.73, P < 0.001) and with the number of HR-HPV genotypes (adjusted-P linear trend = 0.0074). [...] One [of several explanations] is that HR-HPV facilitates HIV acquisition.

Pearly penile papules

A recently published study by Agha et al. reports that - as suspected by many - pearly penile papules are less frequently found in circumcised men.

The prevalence of PPP was 38.3% in <25 years, and 11.4% in >50 years (P < 0.001). [...] In the younger age group, the prevalence of PPP was 26.5% in circumcised participants, and 42.4% in uncircumcised participants (P < 0.05)

Pearly penile papules (PPP) are small raised bumps that are frequently found around the corona. They're sometimes mistaken for genital warts, and can cause significant embarrassment. Although frequently found in partners of women with cervical intraepithelial neoplasia, there is little evidence of a causative role, and PPP are frequently (it seems) considered harmless.
Agha's findings are consistent with those of Rehbein in 1977, who reported:

there was a significantly increased incidence of pearly penile papules in Negroes and uncircumcised men. [...] The incidence of papules was greatest in young adults and tended to decrease with increasing age.

Foreskin surface area and HIV risk

The African randomised controlled trials were amazingly productive. Not only were the direct results of the trials (namely proving the causative effect of circumcision in reducing HIV risk) important, but they also provided an environment for many interesting piggy-back studies.

This is an interesting new paper. Entitled "Foreskin surface area and HIV acquisition in Rakai, Uganda (size matters)", it is essentially an analysis of the relationship between the surface area of the foreskin and the risk of HIV infection among control-arm participants in two of the HIV RCTs.

Here's the key finding:

Mean foreskin surface area was significantly higher among men who acquired HIV (43.3 cm2, standard error 2.1) compared with men who remained uninfected (36.8 cm, standard error 0.5, P = 0.01).

This is logical, really: given that the foreskin increases the risk of HIV infection, one would expect more foreskin to further increase the risk. It might, in principle, be information that would be useful in targetting circumcision: men with longer foreskins could, for example, be warned that their risk is greater. But the information will prove most useful, I believe, in helping researchers as they continue to investigate the mechanisms by which circumcision is protective.

For those uncomfortable with metric measurements, 43.3 square centimetres is 6.71 square inches, and 36.8 square centimetres is 5.7 square inches. To my knowledge, this is the first study of foreskin surface area in the literature - I have never been able to identify any evidence for the "15 square inches" commonly claimed by the anti-circumcision lobby, which appears to be merely a guess, and apparently a rather over-inflated guess. So now that we have hard data, we can expect to see that figure disappear. I wish I wasn't joking - I've never yet seen the anti-circumcision lobby abandon any of their myths in the face of contrary evidence - but I might be pleasantly surprised.

Analysis of Intact America propaganda

I recently became aware of this. An obviously well-funded group called "Intact America" have placed what appears to be a full-page advertisement in the Washington Post. They must be terrified that the AAP will decide to recommend circumcision - the advert must have cost a fortune! Anyway, I thought it would be instructive and educational to analyse their advertisement, to see what propaganda techniques they're using. So let's dissect An Open Letter to the American Academy of Pediatrics:

American parents trust their pediatricians and rely on them for the best advice in caring for their children. As a matter of ethics, that advice cannot include neonatal male circumcision - a medically unnecessary, potentially risky surgery that no major medical authority in the world recommends.

Let's start with the last point: what Intact America are arguing, in effect, is that the AAP (a major medical authority) should not recommend circumcision because no major medical authorities have done so to date. Now, that's an ... interesting ... argument. It's a good thing that these people weren't around when people invented fire!

(As a point of fact, I'd question whether their assertion is even true. Many African countries are embarking on large-scale circumcision programmes in the fight against HIV, and it seems likely that they're doing so with the assistance and approval of their national medical bodies.)

So what do we have left? A "potentially risky surgery". Well, yes, it is. There are risks, of course, albeit small. But these need to be weighed against the benefits: a reduction in the risk of certain conditions.

Finally, "medically unnecessary". Again, yes, it is. But that's not an argument against it: something can be beneficial, even advisable, without being necessary. Take vaccinations, for example: they're not strictly necessary, but they're certainly advisable.

Their claim that circumcision is unethical seems to be on shaky ground.

That is why Intact America is asking the task force charged with reviewing the American Academy of Pediatrics' current neutral position on infant circumcision NOT to revise that position in favour of the surgery. Further, we ask you to take an ethical stand against the removal of a healthy, functioning body part - the prepuce, or foreskin - from non-consenting newborn babies.

Well, I guess they're entitled to ask, however bizarre the request may be.

The United States is the only western nation today where doctors routinely circumcise infant boys in medical settings. Although the rate has fallen from above 90 percent 30 years ago to below 60 percent today,

This is false. The recorded infant circumcision rate 30 years ago (ie., in 1979) was 64.3%. Figures for today are unavailable, but from the same source, the most recent data (for 1999) indicate a rate of 65.3%.

still, more than one million American babies undergo the surgery every year driving one billion dollars in health-care spending.

And, no doubt, saving comparable figures in disease prevention.

Now, based on studies conducted among adults in sub-Saharan Africa that found reduced transmission of HIV from women to men (though not from men to women, nor men to men), some are suggesting that the AAP - meeting this week in Washington - should recommend circumcision for all newborn boys in the United States.

"Some" is rather vague, but this seems basically correct. News reports over the past year or so have contained similar suggestions.

Doctors have a responsibility to tell parents the truth: circumcision does not prevent disease.

Which? Should doctors tell parents the truth, or should they tell parents that circumcision does not prevent the disease?

Most European nations, with circumcision rates near zero, have lower HIV/AIDS rates than the United States.

Are Intact America really so naive about epidemiology that they think that between-country comparisons constitute a decisive answer to such a question? Evidence-based medicine requires use of the best available evidence (usually randomised controlled trials), not the least (ecological analyses such as this are considered one of the weakest methodologies, and for good reason).

Circumcision rates in America do not correlate with HIV rates in any ethnic population or geographical region.

Yes, they do. The CDC cited several American studies in which such correlations were observed in their factsheet (see the section entitled "HIV Infection and Male Circumcision in the United States"). Are Intact America unaware of this, or are they intentionally making false claims?

Furthermore, circumcision has significant risks, including infection, bleeding, impairment of sexual function, and even death. Earlier this year, an Atlanta family was awarded $2.3 million because a physician accidentally amputated much of their infant son's penis during a "routine" hospital circumcision. A Canadian baby bled to death in 2004, after being circumcised in a British Columbia hospital. In 2008, a baby from South Dakota bled to death, and his parents have filed suit against the hospital where he was circumcised, as well as the doctor who performed the surgery.

Yes, accidents happen, and of course they're tragic. But let's be sensible. If we're going to consider the risks associated with circumcision, we also have to consider the risks associated with non-circumcision. Babies die of urinary tract infections - and circumcision reduces the risk. Adults die of penile cancer (again, the risk is reduced) and of HIV (and again).

Infrequent though complications may be, because the surgery is performed on healthy babies who have no need for it, each injury and death is utterly indefensible.

This is illogical. If you choose to avoid the risks associated with circumcision, you choose the risks associated with non-circumcision. It's impossible to avoid risk altogether.

Growing numbers of medical professionals and expectant parents are saying "No" to infant circumcision.

There's no convincing evidence that this is the case.

We urge members of the AAP's circumcision task force, and all pediatricians, to make the same decision on behalf of the babies who are their patients.

They're entitled to urge, but how convincing do they imagine they are? It's easy to put together a bunch of illogical arguments and false claims, but a persuasive argument requires factual accuracy and logical reasoning.

Exponential growth in interest in circumcision

A little bit of fun.

The following graph shows the level of interest in circumcision (as measured by the number of PubMed abstracts published in a given year) vs time. The line shown is the best fit for the data (found using least squares regression, the curve shown is exponential, with the formula being exp(-109.44761+0.05724*y), where y is the year).

What this means is that interest in circumcision is not only increasing, but the rate of increase is itself increasing. The interesting questions are: why? And will it continue?

For those who are interested, the graph is reasonably easy, though a bit tedious, to reproduce. Go to PubMed, and search for "circumcision 1970[dp]", without including the double quote marks. Note the number of items found. Then repeat for all years from 1971 to the present. I typed the figures straight into a CSV file, which I then loaded into R for analysis and plotting.

A little late, but...

Okay, I allowed a little more time to elapse between posts than I intended... Nearly three years, to be more precise, and - of course - a lot has happened in that time.

I'm calling today my 6 and a half year circumcision debate anniversary. I'm not sure if that's exactly right, and it could be a month or two either way, but I haven't kept records, and I think it's close enough. In mid 2003, I entered the public circumcision debate, and I've been involved on a daily basis ever since. I don't know how much impact I've had by myself (how can anyone identify the impact of a single person), but I figure it's more than zero. I hope so, anyway.

Anyway, at the moment I'm involved in three ongoing debates, and as a result of the events in these debates, I've been thinking about behaviour: how participants in a debate behave, and how that influences the debate. I'd like to try to codify some of the (informal) rules that I follow.

First, remember your audience. It's probably impossible to convince your opponent of anything, but other, perhaps more reasonable people are reading and they will be convinced by reasoned argument. If you're lucky, you'll hear from them privately, from time to time. So don't give up.

Second, be polite. This is the single most important rule, as few people pay attention to someone who is rude or abrasive, even if the rudeness isn't directed at them personally. An important part of this is avoid personal attacks, which includes subtle ones such as speculating about people's motives or thought processes. Remember: if someone seems like a crank to you, they probably seem like a crank to your audience, too.

Third, stick to the subject. If you're wrong, don't be tempted to change the subject. And if your opponent does, try pointing that out and stick to the subject yourself rather than joining him/her on the off-topic debate.

Fourth, if you mess up, admit it and apologise. Everyone's human, and everyone makes mistakes. It's much less annoying if you admit it.

Fifth, always provide evidence (or at least avoid making a point unless you know you can back it up with evidence).

Sixth, don't cherry-pick evidence. Present it all, both the bits that support your argument and the parts that do not. The key is to show that the weight of evidence supports your position, even if there are exceptions. And there probably are exceptions, because...

Seventh, remember that most evidence is imperfect. And some forms of evidence are better than others (see evidence-based medicine). The consequence of this is that you should expect a certain proportion of studies to be wrong, especially those of lower quality. In general, expect most studies to point in the right direction, but don't be surprised by a few that do not.

Eighth, not every battle is worth fighting. Don't spread yourself too thinly. Know when to walk away.