Plan B; dig deeper

The controversy over the recent approval by the FDA of Plan B, an "emergency contraceptive", has taken an interesting twist. An otherwise strongly pro-life Catholic blog, Ales Rarus, has protested that Plan B is emphatically not an abortifacient (and thus not morally objectionable on that ground), despite that being exactly what the controversy is all about. Why such an opinion? Is the blog correct to claim that science demonstrates this?

One of the most widely quoted scientific papers on this subject is Pituitary–ovarian function following the standard levonorgestrel emergency contraceptive dose or a single 0.75-mg dose given on the days preceding ovulation, by Croxatto et al. ( doi:10.1016/j.contraception.2004.05.007 ). The abstract of this paper includes the claim that levonorgestrel (LNG, the active ingredient in Plan B):

can disrupt the ovulatory process in 93% of cycles treated when the diameter of the dominant follicle is between 12 and 17 mm. It is highly probable that this mode of action fully accounts for the contraceptive efficacy as well as the failure rate of this method.

That is a very strong claim ("It is highly probable ..." and "fully accounts").

So, isn't the claim of Ales Rarus thus confirmed? Well, we would have to look more closely at the body of the paper, not just the abstract. After all, at least one survey of the accuracy of the abstracts of articles in medical journals came to the startling conclusion that:

Significant results in abstracts should generally be disbelieved.

And this is important because the vast majority of people only ever read the abstract, and never examine the body of such a paper — as is well known to authors.

So, the first thing to look for inside the Croxatto paper is any indication of the numerical uncertainty of the results. Studies such as Croxatto's always have associated uncertainties, and it is extremely difficult to assess the reliability or probability of any conclusion without some indication of what those uncertainties are. In particular, we are looking for the basis on which the conclusion "highly probable" was reached.

There's nothing.

In fact, not only does the body of the paper not back up the claim made in the abstract, it only makes much weaker claims about the significance of the results. One conclusion in the body says:

Based on all the above, it is therefore plausible that the high proportion of cycles with anovulation or ovulatory dysfunction when the EC is given with follicles between 12 and 17 mm (93%), explains most if not all the contraceptive effectiveness of this EC method.

Somehow, without any given reasoning, what is merely "plausible" in the body, has become "highly probable" in the abstract.

Another part of the body of the paper says:

In conclusion, the results of this study indicate that when LNG is used for EC, it prevents pregnancy primarily by interference with the ovulatory process and that method failures are most likely due to treatment given too late to effect such interference.

Which is again a notably weaker claim than the one given in the abstract, since it only concludes (though again, without any provided numerical basis) that the study has identified a primary method of action, leaving entirely open what the scale of secondary effects might be. The abstract makes the claim that the effect "fully accounts" for things, but the body only claims that it is a primary effect.

So, Ales Rarus (and anyone else pointing to this paper) needs to dig deeper into the science of these issues before making claims, or relying on the claims of others.

Comments

[Funky Dung]

The Croxatto paper you mention is not the only one that supports my claim, nor is it the only one cited in the 2005 Population Council press release I linked to. Also, follow the links I've given (as an addendum to the post to which you've responded) to LTI Blog for more links to papers.

[Paul]

The stated position on Ales Rarus is "Plan B is not abortifacient". Zero is an extreme position. And while that number is certainly consistent with the current state of knowledge, other equally consistent larger numbers simply aren't ruled out.
The paper of Croxatto et al chooses not to address the numerical uncertainties in its conclusions, but a reasonable figure based on what they do provide would suggest that 10-20% of the effect could be post-fertilization, without their study being able to detect this.
I had already looked at the other studies you point at -- and more besides -- and not found anything much better. For obvious reasons, more precise studies are extremely difficult to do. It may well turn out that zero is the right answer. But nothing currently rules out the post-fertilization effect of Plan B being at the 10-20% level.
In which case, some caution is necessary. I would agree that a statement like: "Plan B is a proven abortifacient" is not supported by current studies, and is not a scientific statement. Conversely, the statement "Plan B is not abortifacient" is also non-scientific.
(I have also looked at the other statements of Croxatto et al as it relates to possible abortifacient properties of Plan B. As far as I can tell, the study was performed so as to help remove a political road-block to providing emergency contraception. Which would help understand why the study seems to be promoted by them as far more definitive than it actually is.)

[Funky Dung]

The stated position on Ales Rarus is "Plan B is not abortifacient". Zero is an extreme position.

"There's no evidence that Plan B is abortifacient", "Plan B has not been proven to be abortifacient", and similar statements are a bit long for a post title. ;)

"I had already looked at the other studies you point at -- and more besides"

Then please say so. Your post strongly implies that I relied solely on Croxatto for scientific support in this matter. Your fisk of Croxatto is an interesting read and useful in this debate. However, you've written in such a way that your fisk is set up as a rebuttal to my points. Croxatto's findings are not the only ones cited directly or indirectly on my blog.

In which case, some caution is necessary. I would agree that a statement like: "Plan B is a proven abortifacient" is not supported by current studies, and is not a scientific statement. Conversely, the statement "Plan B is not abortifacient" is also non-scientific.

1. Given the lack of strong evidence indicating that Pan B has abortifacient properties, don't you think some pro-lifers are making the movement look foolish by confidently decrying as an abortion drug?
2. You can't prove a negative. The burden of proof is one whoever claims the positive.
3. "Lollipops don't cause mouth cancer" is an unscientific statement, too, but that doesn't mean it isn't wrong.

"So, Ales Rarus (and anyone else pointing to this paper) needs to dig deeper into the science of these issues before making claims, or relying on the claims of others."

Likewise, 153 needs to digger deeper into the scientific claims on other blogs before assuming that fisking a single study will dismantle the blogger's entire argument, when in fact he has cited or linked to other who have cited several other studies.

[Lightwave]

I have to say, I'm with funky on this one, Paul. All you've done is stated there is no conclusive evidence that something "doesn't exist". Indeed there is *almost never* any evidence that something "doesn't exist". Instead, we have to say that there is such little evidence that it actually exists, that we have to generally accept that something "doesn't exist".

What does this mean? Well, I can't prove to you that unicorns don't exist. I can, however, state that there is so little evidence that they do exist that we can generally accept that they don't exist.

The implications for your article here, are that we are presented with the same type of evidence as the unicorn example. We are left in a case where it is far more reasonable to accept that abortifacent effects do not exist.

Now, some folks will begin to say, "but we should err on the side of caution". Indeed, having studied the effects, in this case we have to realize we know more about the abortifacent effects (or lack thereof) for Plan B than most other substances on earth. This means that we have a higher degree of confidence in the saftey of this substance than, say, pennut butter (which has not been studied). This results in a reductio ad absurdum, saying to "err on the side of caution" would mean using Plan B frequently, and never eating pennut butter (or any other substance that has not been studied for abortifacent effects). Thus, the "err on ths side of caution" argument just doesn't hold water.

(By the way, though my argument doen't further my agenda, for the record I find I sit on the pro-life and conservitive catholic side of the fence.)

[Paul]

In reply to Ales Rarus:

Your post strongly implies that I relied solely on Croxatto for scientific support in this matter.

Your post here relies very heavily on summaries of Croxatto et al's studies, with the most important of those studies (so far as I can see) being the one involving humans. I think that the studies of Croxatto et al have generally been summarized wrongly, inaccurately, or misleadingly (and in fact only weakly and indirectly provide some support for your position), and I concentrated on showing this in the study involving humans. If you think that some of the other studies (perhaps by other authors) do in fact provide good support for your position, then detail the actual studies (not the summaries) and say why. I have found nothing in particular, since the other studies seem to be weaker still than even Croxatto et al.

1. Given the lack of strong evidence indicating that Pan B has abortifacient properties, don't you think some pro-lifers are making the movement look foolish by confidently decrying as an abortion drug?

I welcome accuracy everywhere.

2. You can't prove a negative. The burden of proof is one whoever claims the positive.

There simply aren't any fixed and universal rules for assigning the burden of proof. In the current case, since the issue is one of life and death, caution is necessary. If person A says to me: "I don't have proof, but my medical training suggests to me that the substance you are about to ingest may kill you", and person B replies, "See, he says he has no proof, so go ahead and take it", then I know exactly who it is wise to trust.

Another two points relating to statements made on your blog:

• in relation to ovulation, Plan B's method of action is not solely by preventing ovulation (in which case there could be essentially no post-fertilization effect, since fertilization could not occur), but also but by allowing ovulation, but with altered hormone levels which may prevent pregnancy (or so Croxatto et al, claim, but do not demonstrate or measure);


• the AAPLOG association, rather than always leave everything up to each member physician to decide, has recently issued a letter to the FDA pointing out an apparent increased risk of ectopic pregnancy with the use of Plan B.

[Paul]

In reply to Lightwave:

we have to realize we know more about the abortifacent effects (or lack thereof) for Plan B than most other substances on earth.

Based on the studies I've read, I think it is hard to say anything either very firmly positive or negative about the abortifacient effects of Plan B.

This means that we have a higher degree of confidence in the saftey of this substance than, say, pennut butter (which has not been studied)

There are far stronger rational grounds for supposing that Plan B is abortifacient than peanut butter.

[Funky Dung]

I'm curious what you think of the posts on LTI Blog specifically addressing whether or not there are post-fertilization effects at all. One of the interesting things that Serge has pointed out is that if ovulation is estimated by simply counting 14 days from menstruation, the number of expected ovulations may be very far off, since 28-day cycles w/ day 14 ovulation are mythical. Any woman who practices NFP, or the attemtive husband thereof, knows this well.

[Paul]

I'm curious what you think of the posts on LTI Blog specifically addressing whether or not there are post-fertilization effects at all.

I've read those, but they don't provide definite enough reasoning to overcome the central moral problem. The effect of Plan B may turn out to be purely pre-fertilization, but there is nothing currently known that rules out a 10% (or whatever) effect. In which case, it's not moral to use it.

[Funky Dung]

There's another post in the series up now. He specifically addresses the claim that Plan B is 75% effective (which would require some post-fertilization effect(s)).