Sixth in a series.
I wish to challenge the most compelling line of evidence that Plan B works at least some if the time via a post-fertilization mechanism. In this post, I believe those who are very pro-emergency contraception will be quite disappointed.
The most common evidence used to support a post-fertilization effect of Plan B is its proposed effectiveness. Quoting from this post:
1. The reported effectiveness of EC cannot be fully explained by a suppression of ovulation. In other words, if EC worked through non-fertilization events, we would not expect it to be as effective as it appears to be. Therefore there must be some post-fertilization event that is responsible for at least part of its effectiveness.
The effectiveness of Plan B is reported to be somewhere between 90 and 75% in stopping an unintended pregnancy. If this is true, it seems impossible that it could work via a purely anti-ovulatory action. As Steve states here, the most important question to ask would be when did a woman ovulate. If Plan B is taken after a woman ovulates, and it contributes to the 75% effectiveness, then it must work at least part of the time via a post-fertilization event.
If, on the other hand, Plan B works through predominantly anti-ovulatory actions, its effectiveness would be expected to be less than 75%. What does the data support? If one looks at some of the more recent studies, it is reasonable to conclude that the effectiveness of EC is significantly less than 75%.
First, it is a known fact that studying the effectiveness of EC is quite challenging. The main reason is that in order to have a comparison group, an accurate determination of the date of ovulation is essential. An ideal study would use a placebo control group - but that would entail giving a woman who is seeking to prevent a pregnancy a placebo, which is considered unethical. For that reason, estimates of ovulation are used. It has been shown that these estimations are often very inaccurate, as shown by this article: (Contraception 2003 Apr;67(4):259-65)
Calculations of the efficacy of EC depend on knowing the timing of intercourse in relation to the estimated day of ovulation. The results of this study suggest that these calculations are likely to be inaccurate for a significant minority of women.
Is there evidence that EC may not be effective as advertised? Absolutely. I will cite two very recent studies. In this article about EC in JAMA (JAMA 2005;293:54-62), the authors specifically choose a number of women that would have direct access to EC vs pharmacy access in order to show a difference in pregnancy rate between the two groups (they choose about 890 women in each group). They say so in the article, and this was not a small study. As it was, the group who had to go to the pharmacy to get EC used it 197 times, while the group who had direct access used it 309 times. The result on pregnancy: absolutely nothing! The pregnancy rate for the first group was identical despite the fact that they used EC one third more often. This caused the authors of the study to state:
While we set out to demonstrate a large reduction in pregnancy rates, even a10% or 20% reduction in unintended pregnancy rates would be a significant and desirable public health achievement.
What ever happened to 75%? They set out to demonstrate a large reduction in pregnancies, and got zero despite the fact that so many more women took EC. Now they would be happy to see a 10-20% reduction. This is solid evidence that in the "real world", EC doesn't work nearly as often as stated.
Here's one more (Contraception 69 (2004) 361-366). These researchers showed that having EC at home did not reduce pregnancy or abortion rates in that population. Lest anyone think that my reasoning is novel here, the authors seem to agree with me:
Finally, it is possible that EC may be less effective than we belief. Estimates of efficacy are unsubstantiated by randomized trials. Efficacy is based on rather unreliable data and a great many assumptions  and have been questioned both in the past  and more recently .
In conclusion, I believe there is an increasing amount of evidence that the proposed efficacy of EC is not nearly what was originally thought. There is no doubt in my mind that those who support increased access of EC as a public policy issue are not in any hurry to publicize this data. On the other hand, the decreased efficacy of EC is also key evidence that Plan B works predominately prior to ovulation.