Menstrual patterns and fecundity among non-lactating and lactating cycling women in rural highland Bolivia: implications for contraceptive choice

A predictive model for next cycle start date that accounts for adherence in menstrual self-tracking

OBJECTIVE

The study sought to build predictive models of next menstrual cycle start date based on mobile health self-tracked cycle data. Because app users may skip tracking, disentangling physiological patterns of menstruation from tracking behaviors is necessary for the development of predictive models.

MATERIALS AND METHODS

We use data from a popular menstrual tracker (186 000 menstruators with over 2 million tracked cycles) to learn a predictive model, which (1) accounts explicitly for self-tracking adherence; (2) updates predictions as a given cycle evolves, allowing for interpretable insight into how these predictions change over time; and (3) enables modeling of an individual's cycle length history while incorporating population-level information.

RESULTS

Compared with 5 baselines (mean, median, convolutional neural network, recurrent neural network, and long short-term memory network), the model yields better predictions and consistently outperforms them as the cycle evolves. The model also provides predictions of skipped tracking probabilities.

DISCUSSION

Mobile health apps such as menstrual trackers provide a rich source of self-tracked observations, but these data have questionable reliability, as they hinge on user adherence to the app. By taking a machine learning approach to modeling self-tracked cycle lengths, we can separate true cycle behavior from user adherence, allowing for more informed predictions and insights into the underlying observed data structure.

CONCLUSIONS

Disentangling physiological patterns of menstruation from adherence allows for accurate and informative predictions of menstrual cycle start date and is necessary for mobile tracking apps. The proposed predictive model can support app users in being more aware of their self-tracking behavior and in better understanding their cycle dynamics.

Recognizing normal reproductive biology: A comparative analysis of variability in menstrual cycle biomarkers in German and Bolivian women

The idealized "normal" menstrual cycle typically comprises a coordinated ebb and flow of hormones over a 28-day span with ovulation invariably shown at the midpoint. It's a pretty picture-but rare. Systematic studies have debunked the myth that cycles occur regularly about every 28 days. However, assumptions persist regarding the extent and normalcy of variation in other cycle biomarkers. The processes of judging which phenotypic variants are "normal" is context dependent. In everyday life, normal is that which is most commonly seen. In biomedicine normal is often defined as an arbitrarily bounded portion of the phenotype's distribution about its statistical mean. Standards thus defined in one population are problematic when applied to other populations; population specific standards may also be suspect. Rather, recognizing normal female reproductive biology in diverse human populations requires specific knowledge of proximate mechanisms and functional context. Such efforts should be grounded in an empirical assessment of phenotypic variability. We tested hypotheses regarding cycle biomarker variability in women from a wealthy industrialized population (Germany) and a resource-limited rural agropastoral population (Bolivia). Ovulatory cycles in both samples displayed marked but nonetheless comparable variability in all cycle biomarkers and similar means/medians for cycle and phase lengths. Notably, cycle and phase lengths are poor predictors of mid-luteal progesterone concentrations. These patterns suggest that global and local statistical criteria for "normal" cycles would be difficult to define. A more productive approach involves elucidating the causes of natural variation in ovarian cycling and its consequences for reproductive success and women's health.

Field methods and strategies for assessing female reproductive functioning

A detailed understanding of female reproductive functioning is important to many disciplines including anthropology, evolutionary theory, demography, psychology, and biomedicine. In this article, I describe strategies and methods that have been used successfully in community-based studies of human reproduction, many in remote locales, to produce high quality biomarker data. These techniques are applicable to a wide range of research questions and populations, and to persons from adolescence through senescence. I give particular attention to the inherent challenges imposed by the cyclical and somewhat unpredictable nature of the hypothalamic-pituitary-ovarian axis including the necessity and difficulty of ascertaining the timing and occurrence of ovulation, the limits of different sampling regimes for capturing fluctuations in reproductive hormones, and the critical importance of recognizing and, when possible, reducing selection bias. I discuss the relative advantages and disadvantages of collecting saliva, urine, and dried blood spots, and describe some of the subtleties involved in collecting contamination-free samples. Once samples are collected, they must be stored in a manner that minimizes degradation; I describe techniques to keep samples cold even without access to electricity or dry ice. I also discuss various issues that should be considered during initial discussions with a laboratory and when samples are assayed by the laboratory. I include examples of techniques that have worked well in actual field studies, and examples of flawed analytical approaches that should be avoided. With these and other tools, even under technology-sparse conditions, researchers can investigate variability in human physiology across the breadth of human habitats.

Characterizing physiological and symptomatic variation in menstrual cycles using self-tracked mobile-health data

The menstrual cycle is a key indicator of overall health for women of reproductive age. Previously, menstruation was primarily studied through survey results; however, as menstrual tracking mobile apps become more widely adopted, they provide an increasingly large, content-rich source of menstrual health experiences and behaviors over time. By exploring a database of user-tracked observations from the Clue app by BioWink GmbH of over 378,000 users and 4.9 million natural cycles, we show that self-reported menstrual tracker data can reveal statistically significant relationships between per-person cycle length variability and self-reported qualitative symptoms. A concern for self-tracked data is that they reflect not only physiological behaviors, but also the engagement dynamics of app users. To mitigate such potential artifacts, we develop a procedure to exclude cycles lacking user engagement, thereby allowing us to better distinguish true menstrual patterns from tracking anomalies. We uncover that women located at different ends of the menstrual variability spectrum, based on the consistency of their cycle length statistics, exhibit statistically significant differences in their cycle characteristics and symptom tracking patterns. We also find that cycle and period length statistics are stationary over the app usage timeline across the variability spectrum. The symptoms that we identify as showing statistically significant association with timing data can be useful to clinicians and users for predicting cycle variability from symptoms, or as potential health indicators for conditions like endometriosis. Our findings showcase the potential of longitudinal, high-resolution self-tracked data to improve understanding of menstruation and women's health as a whole.

Fifty fertile years: anthropologists' studies of reproduction in high altitude natives

Early European colonists of the Andes had difficulties in reproducing, a fact that underpins the hypothesis that reproduction is impaired amongst all humans at high altitudes. Yet a 16th century missionary wrote, "… the Indians are healthiest and where they multiply the most prolifically is in these same cold air-tempers, … [yet most children of the Spaniards] when born in such regions do not survive." These observations suggest that humans at high altitudes are subjected to strong natural selection from hypoxia, cold and limited food sources and, furthermore, that human populations can and have adapted, and continue to adapt, to these conditions. Informed by multiple approaches and theoretical frameworks, anthropologists have investigated to what extent and precisely how high altitude environments impact human reproductive functioning and fertility. Analyses of the proximate determinants of natural fertility suggest that behaviors (breast/infant feeding practices in the Andes, and marriage practices and religious celibacy in the Himalaya) are major determinants of fertility in high altitude populations. Furthermore, data from Project REPA (Reproduction and Ecology in Provincía Aroma), a longitudinal study in rural Bolivia, demonstrate that fecundity is not impaired in this indigenous altiplano population, and that the risk for early pregnancy loss (EPL) is not elevated by environmental hypoxia but does vary seasonally with the agricultural cycle (contra to the assumption that EPLs are due almost entirely to genetically flawed concepti). This review discusses these and other findings that reveal the complex and dynamic adaptations of human reproductive functioning in high altitude environments.

Darwin's legacy: an evolutionary view of women's reproductive and sexual functioning

On the Origin of species, published just over 150 years ago, has deeply influenced thinking in both scientific and wider communities. Darwin's legacy includes recognition of the fact that all organisms evolve; that variation within and between species is natural and normal; and that an evolutionary approach to understanding the sources and consequences of this variation comprises theoretical frameworks, testable hypotheses, and rigorously collected evidence. With an eye toward facilitating communication and productive collaboration among researchers from different intellectual traditions who nonetheless share a common interest in women's reproductive and sexual functioning, we discuss evolutionary concepts and models, summarize the known variability in ovarian functioning and consider the implications of this variability for conducting sex research, and evaluate the relative merits of various biomarkers that serve as proxy measurements of a woman's reproductive and hormonal status. With these perspectives and methods from reproductive ecology at hand, we examine several contentious issues: the links between hormones and sexuality in premenopausal and perimenopausal women, the causes of premenstrual syndrome, and the existence (or not) of menstrual synchrony. In none of these cases is as much known as is often claimed. In each, there are abundant opportunities for innovative, albeit challenging, research.

Evolutionary Models of Women's Reproductive Functioning

Life history theory posits that natural selection leads to the evolution of mechanisms that tend to allocate resources to the competing demands of growth, reproduction, and survival such that fitness is locally maximized. (That is, among alternative allocation patterns exhibited in a population, those having the highest inclusive fitness will become more common over generational time.) Strategic modulation of reproductive effort is potentially adaptive because investment in a new conception may risk one's own survival, future reproductive opportunities, and/or current offspring survival. Several physiological and behavioral mechanisms modulate reproductive effort in human females. This review focuses on the hormonal changes that vary the probability of ovulation, conception, and/or continuing pregnancy and discusses evolutionary models that predict how and why these hormonal changes occur. Anthropological field studies have yielded important insights into the environmental correlates of variation in ovarian steroids, but much remains to be learned about the evolutionary determinants, proximate mechanisms, and demographic significance of variation in women's reproductive functioning.

The proximate determinants of fertility in populations exposed to chronic hypoxia

It has been hypothesized that hypoxia reduces fertility, but comparative studies of high and low altitude populations have been unable to verify or refute this proposal because it is difficult to control for the behavioral and sociocultural factors that may also either underlie fertility differentials or compensate for physiological changes caused by varying partial pressure of oxygen. Taken collectively, estimates of fertility in populations exposed to chronic hypoxia range widely and do not suggest any reproductive patterns specific to high altitude. Here we review the available data from the Andes and the Himalaya on the proximate determinants of fertility, that set of factors through which any and all environmental, behavioral, and sociocultural factors must act to influence fertility levels. Although hypoxia could potentially affect some of these factors, there is no unequivocal evidence that this occurs in human populations indigenous to high altitude. At this time, it appears that local variations in infant feeding beliefs and practices, often coupled with prolonged breast-feeding, play a major role in determining fertility variation in Andean populations. In the Himalaya, large numbers of adults are not in sanctioned sexual unions as a consequence of polyandrous marriage practices and religious celibacy. The absence of a clear negative effect of hypoxia on fertility in populations indigenous to high altitude, even though migrants report reproductive difficulties, argues that these populations have adapted to the conditions at high altitude. The experimental and clinical evidence presented in this issue suggests proximate mechanisms by which such adaptation, shaped by natural selection and developmental processes, is possible.

Vaginal bleeding patterns among rural highland Bolivian women: relationship to fecundity and fetal loss

Among the most common reasons given for discontinued use of some contraceptive methods is a disturbance in the menstrual cycle, particularly changes in vaginal bleeding. Work to date suggests marked populational variation in menses duration, but few data have been collected from South America. This longitudinal study of non-contracepting Aymara women (n = 189 providing 837 non-truncated bleeding episodes) identified conceptions and fetal loss via urine tests for human chorionic gonadotropin and classified episodes accordingly to test the hypotheses that (a) vaginal bleeding patterns differ between lactating and non-lactating women, (b) duration of vaginal bleeding accompanying fetal loss differs from that of menstruation, (c) menses preceding a conception are longer than those not followed by a conception. Compared to published values, mean menses duration (3.5 days) in these women was relatively short. Menses duration was not significantly correlated with current age, age at menarche or first birth, parity, time postpartum, or menstrual segment length. Mean menses duration (not preceding a conception) was comparable for lactating and non-lactating women. Mean duration of fetal loss bleeding did not differ from that of menses. Pre-conception episodes were significantly longer than those not followed by conception. Thus, because the rate of conceptions was twice as great among lactating than non-lactating women, the mean duration of all menses (irrespective of conception) was significantly longer in lactating women. Bolivian, and perhaps other South American, women may be particularly disinclined to accept contraceptives (e.g., intrauterine devices) that modify an otherwise relatively brief menses duration. Therefore, a wide variety of contraceptive choices accompanied by population-specific informed counseling is essential. In addition, these findings suggest that studies of fecundability limited to non-lactating women may be biased toward those of relatively lower fecundity and that menses duration may be predictive of risk for some cancers.

The home team advantage: reproduction in women indigenous to high altitude

Although there is substantial evidence that environmental conditions disrupt reproductive function among newcomers to hypoxic settings, it is not certain that low oxygen pressure reduces fertility among those indigenous to high altitude. Even when fertility does appear to be relatively lower, numerous behavioral and sociocultural factors may be responsible. These are best examined within demographic frameworks that delineate a finite list of the proximate determinants of fertility. The findings presented here are based on several studies of indigenous Andean populations (Peruvian Quechua at 4000m, Bolivian Quechua at 3100m, Bolivian Aymara at 4000m). Data on ovarian function suggest that neither progesterone levels nor menstrual cycle length or regularity are significantly different from those of women at lower altitudes. Data on two behavioral factors that determine fertility levels, coital frequency and infant feeding practices, suggest that the former is not likely to be of significance in co-habitating couples, but that variation in breastfeeding patterns has probably made a substantial contribution to differences in fertility among at least some populations at high altitude.

Human genetic adaptation to high altitude

Some 140 million persons live permanently at high altitudes (>2500 m) in North, Central and South America, East Africa, and Asia. Reviewed here are recent studies which address the question as to whether genetic adaptation to high altitude has occurred. Common to these studies are the use of the oxygen transport system and the passage of time as organizing principles, and the recognition of the multifaceted ways in which genetic factors can influence physiological processes. They differ in terms of study approach and sources of evidence for judging duration of high altitude residence. Migrant, family set, and admixture study designs have been used for comparisons within populations. These collectively demonstrate the existence of genetic influences on physiological characteristics of oxygen transport. Differences in oxygen transport-related traits between Tibetan, Andean and European populations have been interpreted as having demonstrated the existence of genetic influences on high altitude adaptation but there is not consensus as to which groups are the best-adapted. Part of the controversy lies in the kinds of evidence used to assess duration of high altitude habitation. More other information is needed for a fuller appreciation of duration of residence and also features of population history (genetic drift, gene flow) but existing data are consistent with Tibetans having lived at high altitude longer than the other groups studied. Another issue surrounds usage of the term "adaptation." The definition should be based on evolutionary biology and physiological traits linked to indices of differential fertility and/or mortality. Two examples are developed to illustrate such linkages; intrauterine growth restriction (IUGR) at high altitude and the prevalence of Chronic Mountain Sickness (CMS). Interpopulational as well as intrapopulational variation exists in these conditions which appear linked to characteristics of oxygen transport. Both adversely influence survival and appear to be less severe (IUGR) or less common (CMS) among Tibetans than other groups. Thus available evidence suggest that Tibetans are better adapted. Needed, however, are studies which are better controlled for population ancestry, especially in South America, to determine the extent to which Tibetans differ from Andean highlanders. More precise information is also needed regarding the genetic factors underlying characteristics of oxygen transport. Such studies in Tibetan, Andean and Europeans as well as other high altitude populations offers a promising avenue for clarifying the adaptive value of physiological components of oxygen transport and the extent to which such factors differ among high altitude populations.