The original article - written for health care professionals and social scientists who deal with contraception and family planning - presented data showing that withdrawal, widely viewed as useless in preventing pregnancy, is only slightly less effective than condoms.  Used consistently and correctly, condoms and withdrawal are both very effective; even with typical use rather than perfect use, both are comparable, at 17% and 18%, respectively. I also noted that while many women use withdrawal at some point in their lives, often as a back-up or secondary method (most commonly in alternation or in conjunction with condoms), there is persistent reluctance among professionals and individuals alike to consider withdrawal as a viable method of contraception.  

It's great to see such lively discussion about withdrawal and intriguing to see the comments, the most interesting of which fall into three broad categories: personal accounts of using withdrawal; skepticism, or outright disbelief, that withdrawal reduces the risk of pregnancy; and claims that withdrawal places responsibility on the male, who may be less than trustworthy given the self-control required. 

The many personal accounts of withdrawal use - in keeping with data showing that 56% of women who have ever used a contraceptive method report using withdrawal at some point in their lives - make clear that the method is part of the contraceptive repertoire for many women, if only occasionally for most. It suggests that sex education curricula and health care providers should present withdrawal in its proper context, with both pros and cons, rather than simply disparaging the method or erroneously presenting it as comparable to using no method at all.  

This general view of withdrawal informs another response - sheer disbelief. In my work I've grown used to promoters of abstinence-only-until-marriage programs dismissing facts about the effectiveness of contraception. However, I'm surprised to see such disparagement of withdrawal among a crowd that is presumably younger, more diverse and more open-minded. Perhaps because most of us have been told for so long that withdrawal doesn't work, we are unable or unwilling to embrace scientific evidence that counters what we "know." 

Most interesting, perhaps, is the response expressed by many that men can't be trusted to withdraw in the heat of the moment. Sometimes it seems men can't win for losing when it comes to sexual and reproductive health. We argue that contraception is a couples' issue and that women should not be solely responsible, yet we don't think men can handle the responsibility. While some women may not be comfortable depending on their partners to pull out before ejaculating, and some men may not be able to do it, that does not mean we should promote a false view of the method's effectiveness. Withdrawal may not be appropriate for some couples and individuals, but it is being used at times by many, so why not ensure they have accurate information about it?

Withdrawal will not prevent exposure STIs and is not appropriate for those at high risk of exposure. But I don't buy the argument that accurate information about withdrawal will discourage folks from using condoms and more effective hormonal methods. The same argument was made about emergency contraception, and numerous studies have demonstrated that the concern was groundless. Disparagement of condoms by conservatives has led to more people putting themselves at risk for STIs, including HIV, because they've been told "condoms don't work." Although no method can guarantee 100% protection, condoms are an important option for both STI and pregnancy prevention, and withdrawal should be considered as an option for preventing pregnancy.

For me, most heartening were comments like the one offered by Flackette Goes Retro on Jezebel: "I think the answer here is really just to give people lots and lots of facts and education on their options. That includes failure rates for all the various types of BC (pills, IUDs, condoms, withdrawal, NFP, whatever-all of it, both perfect and "typical" use), information about the pros and cons, information about protection from STDs (even if condoms aren't perfect they are MUCH MUCH better than nothing) and information about how their bodies work. Then hopefully people will be able to make informed decisions." Well said.

"Better Than Nothing or Savvy Risk-Reduction Practice? The Importance of Withdrawal," by Rachel K. Jones of the Guttmacher Institute, Julie Fennell of Central Connecticut State University, Jenny A. Higgins of the Office of Population Research at Princeton University, and Kelly Blanchard of Ibis Reproductive Health, was published in the June 2009 issue of Contraception.

In fact, according to the review by American Psychological Association (APA), methodological flaws are "pervasive in the literature on abortion and mental health," and "the best scientific evidence indicates that the relative risk of mental health problems among adult women who have an unplanned pregnancy is no greater if they have an elective first-trimester abortion than if they did deliver the pregnancy." 

Another comprehensive review of the scientific literature, by researchers at Johns Hopkins University, likewise found that "the highest-quality research available does not support the hypothesis that abortion leads to long-term mental health problems. Lingering post-abortion feelings of sadness, guilt, regret, and depression appear to occur in only a minority of women." The Johns Hopkins review also found a "clear trend" by which "the highest quality studies had findings that were mostly neutral" in terms of mental health outcomes for women obtaining abortions. By contrast, they found that studies with the most flawed methodologies consistently found negative mental health outcomes to be associated with abortion.  

Assessing the evidence  

The APA identified a number of major flaws in the literature on abortion and mental health: failing to compare mental health outcomes between women who have abortion and women who have unintended pregnancies, failing to take into account preexisting mental health conditions (that could account for both abortion and later mental health problems), using samples that were too small or too restricted to allow for generalizations to the larger population of women, failing to account for the fact that many women do not report abortions, and using faulty measurement of mental health outcomes.  

Take, for example, a recent study authored by Priscilla Coleman and her colleagues, published in the Journal of Psychiatric Research. The authors use data from a nationally representative sample of women aged 15-54 years old in 1990-1992. After taking into account 21 background characteristics of respondents, they assert that abortion is associated with 12 of the 15 mental health outcomes examined in their analysis. But several fundamental flaws detract from the validity of their findings. Previous research has established that only about half of abortions are reported on nationally representative surveys, a shortcoming for which Coleman fails to correct. If half of women who have abortions aren't reporting them, then researchers cannot determine if those abortions are, or are not, associated with subsequent mental health outcomes. The authors are also unable to adequately control for potentially pre-existing mental health problems. For example, women who have bipolar disorder may be more likely to have an unintended pregnancy and subsequent abortion than women who do not have bipolar disorder. The authors' inability to control for these types of pre-existing mental health condition leads them to assert that abortion leads to bipolar disorder - an unlikely causal chain given that genetic factors are a major contributor to this particular disorder.  

Even studies with better research designs still have their limitations. For instance, two studies from New Zealand suggest that abortion may be associated with, or cause, later mental health problems. These studies employed a more rigorous methodology than many prior studies addressing the issue. The principal methodological strength of both studies, which were conducted by David Fergusson and colleagues, is that they followed the same groups of women over an extended period of time. Nonetheless, the APA review of the 2006 study cautions that "several design features limit conclusions that can be drawn from this study," among them failing to control for the wantedness or intentionality of pregnancy, not separating women who had multiple abortions from those who had only one, and not accounting for the underreporting of abortion.  

A more recent Fergusson study (published in 2008) likewise did not separate women who had multiple abortions from those who had only one, and it did not account for underreporting of abortion. Another problem lies in its inadequate comparison groups. Women unable to obtain abortions go on to have unintended, or even unwanted births, and this outcome may also have negative mental health outcomes. The challenge, therefore, is to compare mental health outcomes between these women and women who have abortions. The manner in which Fergusson and his colleagues attempted to address this in their most recent study was inadequate. Instead of directly comparing these two groups of women, the authors compared women who had abortions to all those who had not and women who had unintended births to all those who had not.  

The research design of one study is judged by experts, including at the APA, as close to ideal. The study was conducted in the United Kingdom by the Royal College of General Practitioners and the Royal College of Obstetricians and Gynecologists. According to the APA, this study of 13,000 women in England and Wales "stood out from the rest in terms of its methodological rigor." It was based on a large, representative sample; used established diagnostic categories to measure post-pregnancy/post-abortion mental health; controlled for mental health and other factors that may have existed prior to the pregnancy; and had appropriate comparison groups. The study, according to the APA "provides high-quality evidence that among women faced with an unplanned pregnancy, the relative risks of psychiatric disorder among women who terminate the pregnancy are no greater than the risks among women who pursue alternative courses of action."  

The bottom line is this: A woman's mental health before she faces an unwanted pregnancy is the best indicator as to her likely mental health after an abortion. As the APA report says, "Across studies, prior mental health emerged as the strongest predictor of postabortion mental health."