Just before Christmas, Taranaki District Health Board in New Zealand announced their plans to make the emergency contraceptive pill freely available through pharmacies for youth aged 12 to 24. The pill was already available through agencies like Family Planning NZ, but the cost without a prescription is non-negligible at about USD40. The announcement produced the rather predictable and boring reactions from either side – with Family Planning welcoming the news while Family First NZ made the following statement: “Sexually active teens need parental involvement - not emergency contraception - and the Health Board should not be handing out contraception like lollies.” (Note: if I can tell your position on a public policy question from your organization’s name, then I am unlikely to pay much attention to said position.)
Nonetheless, given the recent debates in the U.S. (see my earlier post here) and now here in NZ, it might be a good time to provide policymakers (a) a framework within which to ponder the question; and (b) cite some evidence on the most pertinent questions that come up while pondering. The question is: should you provide free emergency contraception (EC) to young people? Here is the executive summary of my answer in three bullets:
· The Taranaki District Health Board was right to make emergency contraception available to youth as young as 12. There are no good public health reasons not to do so. If you are also worried about inequitable access of the poor, then you may also fully or partly subsidize the product: there seems no good reason to turn away a teenager who made the effort to get to the clinic or the pharmacy because she cannot afford the pill.
· On the other hand, do not expect the EC to do public health miracles for you. It is unlikely to make a dent in pregnancy or abortion rates in your district. It is similarly unlikely to increase STI rates, although there is a chance it might.
· Given that EC does only one thing, prevents pregnancies, you should think of other interventions for young people that might not only increase their human capital and reduce their exposure to risk, but also may reduce their likelihood of becoming pregnant or contract STIs. Social protection programs, unconditional or conditional cash transfer programs (see here and here), or other seemingly unrelated interventions targeted at youth or families with young children can have similar effects.
Now, if you’d like to understand the reasoning and research behind these conclusions, you may read on (Warning: it’s a bit longer than our normal posts). There are three parts of the question at hand that are pertinent: (i) access free/subsidizedor at market price; (ii) unrestricted access at pharmacies, supermarkets, etc. without the need for a prescription; (iii) and access at what age?
Let’s start with unrestricted access. To me, this is the simplest question: yes, you should absolutely make this product available anywhere willing to sell it and over the counter. Young women get raped, molested, and make mistakes every day. The former two are sad facts of life and the last one is a consequence of, well, being young. No amount of parental involvement will reduce the incidence of these outcomes to zero. So, when they happen, you should be able to go to the closest pharmacy or supermarket and grab a morning after pill. If the government will put any restrictions on the purchase of a legal drug and reduce individuals’ choice sets, it needs a good reason to do so. Even if there was a good public health reason (I will argue below that the evidence is not there to support that view), it probably would not be enough reason to make getting EC more difficult for some individuals. There is a drug that is safe, legal, and easy to use: the bar to restrict individual access to that drug should be sufficiently high.
The ease of use is also the answer to the question of age of access. FDA in the US has established that the drug is safe and children 11 or older can use it safely. So, yes, if it will be available, it should be available to 12 year-olds.
Should it be subsidized? That question is more difficult. It depends on the future costs of teen pregnancy (and abortions) – both to the individual and to society. If children from single parent families, poorer families, etc. are more likely to need EC but less likely to be able to afford it, it may make sense for the government to subsidize the cost. Even if the costs are solely to the future adult herself and not to the society in general, they can be justified under the principle of ‘second chances’ – after all we’re talking about children and young people here. If there are positive externalities, the argument for subsidies is even simpler. So, I’ll return to the issue of subsidies after reviewing the literature on the individual consequences and public health externalities of making EC freely available to young people.
Now, to think about this, consider EC as a health intervention that affects only one thing: it prevents pregnancies. Other than its effect on other outcomes through the prevented pregnancy, it is expected to change nothing else. Without it, the girl would have had to have an abortion, induce a miscarriage, or carry the pregnancy to term, but using it will not make her better or worse off other than through its effect on the averted conception.
If the intervention we’re considering will prevent pregnancies, then we need to first ask the following question: what is the effect of teen pregnancies on outcomes we care about? This is a two-part question: (a) what is the effect on the baby that is born; (b) what is the effect on the future outcomes of the mother? For the WDR 2012, titled Gender Equality and Development, I did a small background note on the consequences of teen pregnancies and marriages. Below, I plagiarize from my own work.
Age at birth and birth outcomes: In the biomedical literature, there have been investigations of the effects of age at birth on birth outcomes, all of which are non-experimental (it occurs to me that randomized access or free provision of EC can help with identification for this question – provided that EC actually reduces pregnancy rates in the intervention group, unfortunately see below on the chances of this type of ‘encouragement’ design). Outcomes such as very preterm and preterm deliveries, low birth weight, being small for gestational age, infant mortality, late fetal death are all correlated with teenage births. Earlier the age at birth, the worse these outcomes usually are. Biomedical literature suggests two main pathways for the inherent effect of age at birth, rather than the effect of factors correlated with young age: gynecological immaturity (poor intrauterine growth, immature cervical blood supply, etc.) and nutritional competition for growth. However, these pathways have not been empirically confirmed (contact me for citations, which are too many to individually link here and most are gated).
While these studies do not have experimental identification, and self-admittedly may suffer from confounding factors, I don’t think they should be dismissed. All studies cited here are limited to a fairly homogenous group of young, white females – be it in Utah, Scotland, or Sweden – and control for a variety of maternal background characteristics, such as marital status, education, prenatal care, smoking, height, poverty status, etc. As such, they are not outrageous comparisons of obviously incomparable treatment and control groups – the unobservable confounders would have to be more subtle than that.
Summary: there is some reason to worry about children’s outcomes in teen pregnancies.
Age at birth and mother’s outcomes: The subject of the effects of teen pregnancy on future outcomes of the mother has been examined by economists(“Kids Having Kids” is a good book to consult on this topic. The literature review in this paragraph borrows heavily from Dahl, 2010, from which citations excluded from this post can be obtained). A line of research has attempted to disentangle the effects of teen childbearing on education and wages from preexisting differences between those who parent early and those who delay childbearing. Early research using OLS revealed large and significant consequences associated with teenage childbearing. Subsequent approaches attempting to deal with selection bias have reached disparate conclusions. For example, studies that used a variety of instrumental variables concluded that teenage childbearing has negative consequences. However, others, using sister fixed effects, or random miscarriages as an instrument, found little evidence of a negative effect. In particular, Hotz, McElroy, and Sanders (2005) argued, “much of the ‘concern’ that has been registered regarding teenage childbearing is misplaced.” The debate is ongoing, with recent work by Ashcraft and Lang (2006) and Fletcher and Wolfe (2008) using variations on the miscarriage instrument and finding negative effects (It’s important to note that the use of miscarriage as an instrument, the exogeneity of which is also under debate, answers the question of “what would have been the adolescent mother’s outcomes had she not had a child as a teen?” This question of the effects of teen childbearing is different than the question of teen pregnancy and future outcomes.).
Summary: The evidence on the effects of teen pregnancies on future outcomes of the mother are mixed. There is not enough here to justify a strong stance for a policy decision.
That’s the private side of the equation: some evidence to suggest that it would be desirable to discourage/reduce teen pregnancies, but not particularly strong. There is also not much to inform us about the level of subsidy that would be needed, if any. What about the public health externalities? If we decided that reducing teen pregnancies is desirable, will EC do the job effectively? And, are there any negative externalities associated with making EC available, such as increases in the incidence of sexually transmitted infections (STI) due to potential disinhibition effects?
On the first question, the evidence is surprising: making EC available (or even advance free provision of it) has not been shown to have any discernible effects on reduced pregnancy rates or decreased rates of abortion.This RCT by Raine et al. (2005) found no decreases in pregnancy rate in either the pharmacy access to EC group or the advance provision of EC group, compared to the clinic access group (control). It also notes that previous studies also failed to show significant reductions in pregnancy or abortion rates with advance provision of EC. This RCT in China arrives at the same conclusion. The authors of these studies that I cite here discuss a bunch of potential reasons why pregnancy rates among women, many of whom reported a desire to avoid pregnancies during the study period and all of whom were provided with take-home provisions, were not lower than the control group. In the end, they seem to settle on two possible reasons: first, EC may actually not be as effective as found in previous studies, and, second,a failure to recognize (or acknowledge) a risk of conception on part of the prospective users. The latterreason is not so unlike a finding from a CDC study I cited in my earlier post on this topic, from a Washington Post article: “In a survey of thousands of teenage mothers who had unintended pregnancies, about a third who didn’t use birth control said the reason was they didn’t believe they could get pregnant. Why they thought that isn’t clear.” The evidence is suggesting that (a) unprotected sexual activity is high among people who don’t report not wanting to have children; and (b) they don’t use the free EC that is in their drawer and can be used for up to 5 days after the unprotected sexual activity or a contraceptive incident. You can take a horse to the water, but you cannot make him drink…
What about an unwanted increase in STI risks as a result of greater access to EC? The same study by Raine et al. (2005) also shows no changes in STI risk (although they do find a lower rate of self-reported condom use at last intercourse in the advance provision group, which is consistent with the worry about more risk-taking. Given that there is biomarker data available on pregnancies and STIs, it’s best to put less weight on the self-reported proxy measures). In my earlier post, I cited this paper, which suggests that the introduction of EC over the counter and free of charge in England to adolescents younger than 16 may have had no effect on the conception rates of teenagers while leading to an increase in the prevalence of STIs. Unlike the other studies cited above, this was not an RCT, but nonetheless the evidence is consistent with how an economist would think free provision/availability of EC might affect adolescent risky behavior.
Summary: I think that the authors of the studies above independently stated the summary of their work better than I can do myself: “This study adds to the growing literature casting doubt on the increased use of EC as a quick fix for rising abortion rates. That is not to say that EC will not prevent pregnancy for some women, sometimes, but rather that it may not make much difference to public health.” (Hu et al. 2005) “While removing the requirement to go through pharmacists or clinics to obtain EC increases use, the public health impact may be negligible because of high rates of unprotected intercourse and relative underutilization of the method. Given that there is clear evidence that neither pharmacy access nor advance provision compromises contraceptive or sexual behavior, it seems unreasonable to restrict access to EC through clinics.” (Raine et al. 2005)
So, there you have it. While the topic is complex, the evidence is mixed, and there is definitely bound to be more literature that one could scour, I think there is enough here for policymakers: please make it available to young people (and free if you can afford it), but don’t expect it to solve your public health problems.