When The Condom Breaks, The Cradle Will Fall…

Newsflash: I am an epidemiologist. One of the important dimensions of epidemiology is using indicators correctly. I stress to my students that it’s important to consider when to use a relative measure vs. an absolute measure.

For example, there were 6 cases of cholera in the USA in 1999 (MMWR, 1999). Every case of cholera makes public health officials very nervous, because it may signal a potential outbreak. But 6 cases constitute a prevalence rate that is vanishingly small.

In 1998, there were 5 new cases of cholera. The additional case in 1999 meant an increase of 20%! A 20% incidence increase in any disease sets off public health alarm bells. But, really, there were only 6 cases in total. No biggy.

The relative measure was a 20% increase. The absolute measure was a single new case. Depending on which statistic you look at, you get a whole different picture of the cholera burden of the USA.

Now, my public health colleagues are going to have a fit over what I’m going to talk about next…

A recent study concluded that 30% of women who date online have had sex on the first date with gentlemen they’ve met online. (Zod bless those fine, upstanding women). Moreover, the study found that 77% of these women had had unprotected sex in those encounters. Of course, public health officials are outraged. The STI specialist quoted in the linked articled said,

“When you have unprotected sex with people you are meeting online, you are playing russian roullette [sic] with your health. It’s not a matter of ‘if’ you’ll get a sexually transmitted disease, but rather ‘when’ and ‘how many’.”

Well, I’m here to report that, epidemiologically, this is a bit of hyperbole.

There is no question that STDs/STIs are bad. There is no question that globally, and among some sub-populations, they exist in epidemic proportions. There is no question that one should always err on the side of protection, and wear a condom to protect oneself from pregnancy, STIs and overall ooginess.

However, that is no reason to conflate the risks suggested by the actual numbers! Let’s break it down..

In Canada, chlamydia is overwhelmingly the most common STI. The most recent reliable stats we have are from 2004. In that year, there were a total of nearly 63,000 cases detected. Chlamydia is a reportable disease, which means that the government is notified of every detected case; almost none fall through the cracks. Undiagnosed cases are, of course, unknowable, but are often tracked down if the infected person spreads it to others. In other words, 63,000 is probably a fairly accurate estimate of the chlamydia burden of Canada. (Compare this to about 9000 cases of gonorrhea and maybe 1000 cases of syphilis… very roughly.)

The Canadian STI surveillance report states that STIs are on the rise. Public health officials are worried about this trend. I’ve heard some use the word “epidemic”. In terms of chlamydia, there was a 70% rise from 1997 to 2004, from 113.9 to 197.1 per 100,000 population. The current prevalence is164 cases per 100,000 population, or about 0.16% of the total population, assuming a conservatively estimated base population of 35 million.

A 70% rise is a staggering number that necessarily warrants concern. But the prevalence is still only 0.16%. That means that probabilistically, you’d have to have unprotected sex with 625 random people before you were guaranteed to be exposed to someone with chlamydia.

Now let’s break it down further. Let’s say you’re one of these single on-line women, aged 35, currently disease free. The age-specific chlamydia burden among Canadian men aged 25+ is 9374 cases. Using an adult male population of about 20 million, that gives us a prevalence estimate of 0.05%. Thus, one of these single women would then have to have unprotected sex with 2000 random men in order to guarantee being exposed to chlamydia.

So far I’ve been talking about the chances of being exposed to an STI. What about actually contracting one? Well, the transmission rate of chlamydia is between 30% and 40%. In other words, only 30-40% of sexual encounters with an infected person will result in the disease being transmitted. This means that our hypothetical woman, in a worst case scenario, would have to have unprotected sex with 5000 random Canadian men before being guaranteed to contract chlamydia. And, of course, the number is even more outrageous for the less common STIs.

Given these numbers, I think our STI specialist above was being a tad hyperbolic when he said such women were guaranteed to become infected. Don’t you agree? To him I say, “Get a freakin’ grip, buddy.”

The truth is, thousands of people around the country are having unprotected sex all the time, and nothing bad is going to happen to most of them. This is not a reason to advocate unprotected sex– not at all. Do not mistake what I am saying! I am not condoning unprotected sex. Even a 0.16% risk is too much to always take lightly, in my opinion, when you can reduce it to near zero by using very basic controls, like a condom.

Rather, it is never justifiable to exaggerate risk, even if the intentions are pure. I would rather a society that treats its adults like adults and presents the numerical risks reasonably and accurately, instead of one whose public health officials shout out unfounded absolutes, which quickly devolve into medical propaganda.

As I was writing this blog post, I was surprised to find others exploring the same topic. This press release, for example, laments Canadian clinicians’ poor understanding of the true risks associated with unprotected sex and STIs.

I fully expect a public health practitioner to write to me now and tell me what a horrible person am. Here it comes…


Due to the sensitive nature of this post, I thought it best to clarify some things, in response to what was brought up in the comments:

  • It is entirely possible that women who have unprotected sex with men they meet online share common behavioural characteristics that put them in a higher risk group for reasons beyond their “promiscuity”. (I don’t like using such judgment-filled terms, but whatever.) In other words, it’s possible that they’re not sleeping with random Canadian men, but with men who have a higher probability than the average of having an STD; club-goers, for instance, or atypically “promiscuous” men; or single men in general, who have a higher chance of infection than do married men. These women’s risk is therefore possibly orders of magnitude higher than I’ve indicated. But the numbers I’ve presented are based on known data; arguments about a higher risk “promiscuous” population are speculative.
  • I have identified chlamydia as the #1 STD/STI in Canada. This is true, according to the Canadian STI surveillance system. But HPV may have much higher incidence and prevalence.
  • While, according to my analyses above, the risk of contracting an STD is small, the fact remains that there’s a very small –but non-zero– chance of contracting HIV, which can kill you. One can argue that a risk of death is never trivial.
  • The issue boils down to this: knowing how low the actual risk is may end up promoting the risky behaviour. But I’d rather my public health system give me the facts and not rely on fuzzy data and shock tactics to scare me into safe behaviour.