Implanon and contraceptive failures

Today a story has been in the news about the "scandalous" contraceptive failures reported with Implanon, a long-term hormonal contraceptive which is implanted under the skin. See here and here for examples from some of our most respected broadcasters. And see here for an example from possibly our least respected "news" source, the Daily Mail.

OK, given that I'm quoting the Daily Mail, you've probably guessed by now that I'm not entirely happy with the way this story has been reported.

We are told that 584 women who used Implanon got pregnant, and this is presented as some sort of terrible scandal. Well, clearly it's pretty rough on those women who thought they were using an effective contraceptive and got pregnant anyway, but the fact is that no method of contraception is 100% effective. Everyone knows that (or should know that).

So the question is whether 584 is more than the number of women using Implanon who might reasonably be expected to get pregnant.

None of the media sources gives us any information on that. And without that information, it's impossible to judge whether there's anything really wrong here, or whether this is just a small and probably unavoidable failure rate that can be expected with any form of contraception.

However, the Channel 4 story gives some numbers that have allowed me to do a quick back-of-an-envelope calculation.

We are told there that 16,000 women used contraceptive implants in 2005 and 82,000 did in 2010. If we assume that the rate of increase was uniform over that period, then that would be 294,000 implants in that 6-year period. We are not told the time period in which the 584 pregnancies were reported, but if it applies to the same 6-year period, then the failure rate is 0.2% per implant. Given that the implants are intended to be used for 3 years, that implies a failure rate of 0.07% per year.

Now, given that the failure rate of the oral contraceptive pill, when used perfectly, is 0.3%, and a whopping 8% in typical use (as not all women use it perfectly, for example they may miss doses), a 0.07% failure rate for Implanon doesn't sound so bad. In fact it also seems to be within the confidence interval for the failure rate calculated from clinical trials of Implanon, which was from 0 to 0.09%.

Much of the media coverage makes the point that failures have generally occurred when doctors have inserted the devices incorrectly. No doubt that's true. And maybe, just maybe, there is some cause for concern here. Clearly inserting the device is a procedure requiring some skill, and if some doctors who are inserting the device have not been adequately trained in its use, then something probably needs to be done about that.

However, doctors are humans, so its inevitable that some of them will sometimes insert the device incorrectly. It's unrealistic to assume that it will be done perfectly every time. Statistics on the failure rate of the device need to take that into account.

It's worth noting that clinical trials of the device may give a misleadingly favourable impression of the failure rate, as doctors participating in clinical trials were probably trained to a higher standard than your average local GP, so you would expect failure rates in real life to be higher than in clinical trials.

So, maybe there is a case for ensuring that doctors who use the device undergo more rigorous training than they currently do. But if my back-of-the-envelope failure rate of 0.07% per year is anything like the real figure, then it seems that Implanon is still a pretty effective form of contraception.

But hey, as they teach them at journalist school, why let the facts get in the way of a good story?