Where Fertility Data Comes From: Making the Modern Fertility Timeline Tool

You may have heard the age-old question (pun definitely intended) “how and when does female fertility decline?” Maybe you've even heard some hard numbers ("falls off a cliff at 35" is something I hear at my family's Thanksgiving, for instance). Using Modern Fertility's Timeline Tool you can now move AWAY from adages, old wives tales, and plain old misinformation and see how female fertility changes with age using some of the best data out there. So now, speaking of data - let's get into how we actually built this thing.

Background

This post is definitely not the first article to tackle explaining how fertility changes with age and how it is investigated. Demographers, anthropologists, doctors, and historians have been grappling with this question in a scholarly way for decades, but often their work is misinterpreted by the popular press. This question is especially relevant as women (and men) are putting off starting families until later ages (1). With the launch of Modern Fertility’s “Timeline Tool,” which includes estimates of the risk of not being able to have a baby naturally, we thought it high time to break down how this number was calculated.

First, let’s lay out why calculating fertility rates is so darn hard. The number one issue is the pill (and condoms, and the IUD, and...you get the idea). Few people will argue that contraception hasn’t been helpful as men and women plan their families. But something we don't often hear about is what other effects, effective contraception can have (say that five times fast!). When you introduce effective contraception, you introduce an element of choice to reproduction (which, don’t get me wrong, is important!). If a sexually active woman does not get pregnant today at age 25, it doesn’t necessarily mean she can’t but that she (and/or her partner) have taken measures to help ensure that she won’t. This makes looking at large-scale demographic data from populations with effective contraception virtually useless when calculating fertility rates.

Additionally, developed nations have gone through what is called the “demographic transition,” a phenomenon where birth rates eventually become low after industrialization because of decreases in child mortality and increases in women's access to education and contraception (2). These advances allow women to choose to have fewer children. This element of choice is particularly relevant when we try to calculate fertility rates at later ages. When the pill initially became available, the first change demographers noticed was that women used it to stop having children as they got older, generally after they already had a couple of kids. They termed it, understandably so, “stopping behavior.” This has made it hard to calculate fertility at later ages because fertile women can now choose to not have children later in their reproductive lives (again, not that they necessarily can’t). In addition to preventing pregnancy at later ages, effective contraception has also allowed women to delay having children, control the time between children, and decide how many children they’d like to have. As a result, record keeping of birth rates in developed nations reflects “choice” more than it reflects “fertility.”

Contemporary Studies

However, there are women who are actively trying to get pregnant and if you survey enough of them from the time they try to conceive to the time they get pregnant, you can collect a better idea of fertility rates. These studies are commonly referred to as “time-to-pregnancy studies” and are conducted by enrolling women before they start trying to get pregnant. Researchers then follow the same women for months or years to determine how long it takes them to conceive. As you might imagine, this is a laborious and time-consuming task––and one that results in small study numbers. The one very important limitation of contemporary time-to-pregnancy studies is that we lack data on women at older ages, notably over 40. Some of the best studies of the past 20 years have not been able to observe or measure pregnancy rates adequately at these older ages. Again, this harkens back to “stopping behavior,” in that there are few older women who enroll in these studies because they have already completed their families and aren’t trying to get pregnant.

I’ve compiled some quotes from some of the best “time-to-pregnancy” studies that show that even the best ones can’t always overcome this limitation:

  • “This study has some weaknesses. First, although the cohort was designed to only include women of ‘older reproductive age,’ few women enrolled in the cohort who were in their forties.”-Steiner & Jukic, 2016 (3)
  • “Our study enrolled women only up to age 40 years, so the possibility of increases in sterility for couples in their 40s requires further investigation” -Dunson et. al., 2004 (4)
  • “However, the proportion of women falling into this age category [>35] is too small to prove (statistically) a significant decline in fertility.” -Gnoth et. al., 2013 (5)

Natural Fertility

Ok, so now we’ve reached a point where we have the pill, and as a result the demographic studies of developed nations do not reflect the potential for births at older ages, and thus can’t capture the decline in female fertility at later ages. And even those studies that do look at women trying to get pregnant at later ages aren’t adequately powered to give us good data either. So where does one turn? Well, fortunately for researchers, there are populations that don’t use birth control, termed “natural fertility” populations.

What is a “natural fertility” population? It is one that does not use birth control, and even if it does, it is fairly ineffective. There are contemporary populations that, for cultural, economic or religious reason, do not practice contraception. For example, the Hutterites, an Anabaptist group that live in the plains of the northern United States and western Canada, did not use birth control widely in the last century. Hutterites families were also religiously compelled by church doctrine to have as many children as possible, and the communal setup of their “colonies” helped to ensure that these large family sizes could be supported. In a classic study conducted by Tietze in 1957, he showed that Hutterite women have an average of 10 children, 87% had their last birth at age 45 and there were no births after age 49 (6). Others natural fertility populations that have been studied include ones in Taiwan, Mali, Bangladesh and Pennsylvanian Amish communities (among many others) (7).

Though the Hutterite data is robust, it is still based on a small number of women (209) and might not be reflective of all other natural fertility populations. Again, in order to measure something like the decline of fertility in women and age at last birth, it is important to have a large database. As a reminder: this is especially important because as women (and men) get older, birth rates are lower, so there are fewer people to measure and it is harder to compile this data. So, what natural fertility populations are there that are large enough to statistically measure fertility decline? This is when we crack open the history books and acknowledge the work of the father of modern demography, Louis Henry.

Louis Henry was a French historian who recognized that the parish records of France could be a powerful tool in measuring natural fertility. These parish records track birth dates, marriage dates, and death dates (in the form of baptism, wedding and funeral dates...the old-school version of save-the-date cards stuck to your fridge). These records are extremely robust, and the events can be linked to reconstruct the history of entire families. In total, Henry compiled family histories spanning 159 years (1670 to 1829) (8). That’s several generations of data! From this he calculated the proportion of infertile couples at different ages (among many other things). Unfortunately, the pioneering work of the Louis Henry is often misunderstood and cited in the popular press as “questionable data” because it is over 200 years old and thus people assume that it can’t be applicable to modern women. When in actuality, this ignores the fact that the fertility data from contemporary populations has more restrictions and caveats in terms of how it can be interpreted than Henry’s data does.

If we are to believe that Henry’s data holds for more than just agricultural French society, it would be helpful to compare this across other natural fertility populations. Fortunately, this has been done. Marinus Eijkemans and co-authors selected the highest quality historical databases of natural fertility populations and measured the age of last birth for women at all ages (9). All told these included six populations with records of over 58,000 women. They then compiled this data to create a fertility curve. They found that the curves were very similar and that age at last birth was around 40-41 years average in all populations. However, they note that “for women aged 32-34...the lesson is: do not wait much longer, certainly not if you want more than one child. At the same time, the curve contradicts the occasionally exaggerated pessimism about women who intend (or by circumstances are forced) to have children in their late thirties...For them the lesson is: you still have a good chance to succeed if you don’t wait any longer. For women in their early forties the lesson is: just try, your chances are still far from hopeless” (Eijkemans et. al., 2014, page 1311, emphasis mine). This guidance is more nuanced and grounded in data compared to your great aunt’s well-meaning “try by 35” Thanksgiving dinner advice.

Screen-Shot-2018-11-05-at-4.25.28-PM
Figure 1. Estimated percentage of the population that is infertile from: Stelling, Tamar. ”Want children? Do not be fooled by the IVF industry.” de Correspondent, 19 September, 2018, p. 10. (10) Reprinted with permission.

How does it all compare?

But what about contemporary time-to-pregnancy data from the United States and other developed countries? How does it look in comparison to these larger studies of historical data? The interesting piece here is how remarkably similar the historical data estimates are to the contemporary time-to-pregnancy estimates for natural fertility at later ages (though you can see in the table below that there are few older women enrolled in these studies, as mentioned earlier). This is most likely attributable to the fact that there is little reason to believe that natural selection would have worked to extend the length of female fecundity into later ages in the last 200 years, which is only about four to eight generations.

Rate of infertility by age comparison of historical and time-to-pregnancy studies of contemporary populations.
Screen-Shot-2018-10-24-at-6.21.37-PM
Note, the measure of not being able to become clinically pregnant may be lower than that of live birth because miscarriage and stillbirth are not accounted for.

What about ART?

One thing to note about this data is that it does not account for the fact that some women might be able to conceive if they have access to assisted reproductive technology (ART). And that’s definitely true! But remember, this data is called “natural fertility” for a reason, so it’s important to avoid comparing apples to oranges. There are great databases where you can look at the success rates of ART for different infertility diagnoses. One I find particularly useful is the SART.org “Predict My Success” tool (this tool is getting an overhaul in 2019 that will update the success rate of ART at older ages. Stay tuned for that update on the SART page! ). However, one element that ART has not yet been able to overcome is that women’s eggs decline in quality as they get older. Even if a woman utilizes ART to try to conceive, if her eggs are of poor quality, she may not be able to bear a child. This is also reflected in the historical demographic data. But don’t forget: there is the possibility of using donor eggs which helps to avoid the problems of egg quality. And in these modern times, creating a family doesn’t always have to involve bearing a biological child: there are 1,000 and 1 ways to have a family.

Conclusion

So, what is the bottom line? As many other authors have concluded, fertility declines with age. Nonetheless, there is definitely still the possibility of having a child at later ages. You might be wondering what exactly those chances are. Thanks to the painstaking work of researchers such as Henry, Tietze and Eijkemans (among many others) we can now better estimate those chances. And those chances have not changed significantly in the past four to eight generations. We specifically chose to utilize the data from Eijkemans and coauthors in our Timeline Tool because it is the largest cross-cultural study of the timing of last birth and is largely aligned with contemporary time-to-pregnancy studies. All the same, we definitely need more attention and research on fertility and women’s health, especially as our generation puts off having children until later ages.

You might also be wondering about the chances of miscarriage or having a healthy baby, particularly because these odds also change with age. The Modern Fertility Timeline Tool synthesizes the best datasets available for natural fertility, miscarriage and Down Syndrome (one of the most common birth defects that is linked to maternal age) in one place so that you can have a more holistic sense of your chances of becoming a parent at different ages. Don’t forget that these are averages and your individual chances depend on many other factors (like male infertility) that aren’t included in the Tool.

Click here to explore Modern Fertility’s new Timeline Tool. Start the conversation and pass along to the badass women in your life––our collective knowledge is our power!


References

  1. Vespa, Jonathan. "The changing economics and demographics of young adulthood: 1975–2016." Washington, DC: US Department of Commerce, United States Census Bureau 14 (2017).
  2. Coale, Ansley J., and Susan Cotts Watkins. "The decline of fertility in Europe: the revised proceedings of a Conference on the Princeton European Fertility Project." (1986).
  3. Steiner, Anne Z., and Anne Marie Z. Jukic. "Impact of female age and nulligravidity on fecundity in an older reproductive age cohort." Fertility and Sterility 105.6 (2016): 1584-1588.
  4. Dunson, David B., Donna D. Baird, and Bernardo Colombo. "Increased infertility with age in men and women." Obstetrics & Gynecology 103.1 (2004): 51-56.
  5. Gnoth, Christian, et al. "Time to pregnancy: results of the German prospective study and impact on the management of infertility." Human Reproduction 18.9 (2003): 1959-1966.
  6. Tietze, Christopher. "Reproductive span and rate of reproduction among Hutterite women." Obstetrical & Gynecological Survey 12.5 (1957): 727-728.
  7. Baird, Donna Day, and BEVERLY I. Strassmann. "Women’s fecundability and factors affecting it." Women and Health (Goldman MB, ed). New York: Academic Press (2000): 126-137.
  8. Henry, Louis, and Michel Fleury. "Des registres paroissiaux a l'histoire de la population: manuel de dépouillement et d'exploitation de l'état civil ancien." Population (french edition) (1956): 142-144.
  9. Eijkemans, Marinus JC, et al. "Too old to have children? Lessons from natural fertility populations." Human Reproduction 29.6 (2014): 1304-1312.
  10. Stelling, Tamar. "Want children? Do not be fooled by the IVF industry.” de Correspondent, 19 September, 2018, p. 10.

Dr. Erin Burke, PhD

Dr. Erin Burke earned her doctorate at Yale before joining the Modern Fertility team. Her dissertation research focused on the physiological changes that gay men go through after becoming fathers.

Fertility Unfiltered