THE DEMOGRAPHY OF FERTILITY AND INFERTILITY
Global Health Situation Assessment and Projections Unit
Division of Epidemiological Surveillance and Health Situation and Trend Assessment
World Health Organization
1211 Geneva 27, Switzerland
To approach the topic of human reproduction at the demographic level—at the level that aggregates individual experience in procreation—requires a shift in concepts and definitions, as well as in scale.
Fertility, fecundity and fecundability
In demography, fertility indicates the product or output of reproduction, rather than the ability to have children. The physiological ability to have children—that is manifest roughly in the period between menarche and menopause in women—is termed fecundity. Demographers define a third, further aspect of reproduction—fecundability—which is the probability of becoming pregnant, or the likelihood of exposure to the possibility, that depends on the pattern of sexual and pregnancy preventive behaviours.
Similarly, demographers use the term infertility to refer to the absence of liveborn children, or the presence of few children, rather than the term sterility that refers to the possible physiological status underlying childlessness.
Conception and pregnancy that is followed by fetal loss—whether due to spontaneous or induced abortion, or at term (stillbirths)—is not considered to contribute to fertility by demographers, as its occurrence is not discernible demographically. Accordingly, fertility refers to live births only, and infertility refers only to shortfalls in liveborn children, whether or not pregnancy(ies) occurred.
The proximate determinants of fertility
Demographically observed fertility or infertility is the result of a well-defined number of both biological and behavioural factors, which serve to mediate the influence of culture, society, economic conditions, living standards, and other similar background determinants on individual reproductive behaviour. Together, these biological and behavioural factors are called the proximate determinants of fertility (see Bongaarts and Potter (1), and Frank and Bongaarts (5)), and they are the factors through which, and only through which, the social and economic environment can influence individual procreation. Together, these factors constrain fertility, and explain why women do not have the maximum possible number of children, which would be about 35 if they reproduced continuously from, say, the age of 18, to about the age of 45, having a birth every nine months.
Biological constraints on fertility
The biological constraints on the number of liveborn children include not only the time actually lost during pregnancy(ies), but also the time lost after delivery before fecundity resumes (postpartum infecundability), the waiting time to conception, the time lost because of naturally occurring intra-uterine mortality, and time lost because of sterility arising naturally with age or induced by a pathological condition, which varies the most widely of all biological determinants because of the variability of associated sexual behaviours. On average, assuming the risks spread out over all women, in addition to the 9 months of pregnancy, postpartum infecundability adds 1.5 months, the waiting time 7.5 months, intra-uterine mortality 2 months, and sterility a further, variable period to the interval between two births. Together, the biological constraints lower the maximum feasible fertility from 35 births to about 15, which is called total fecundity, or the limit in physiological capability of childbearing.
Behavioural constraints on fertility
The behavioural constraints on the number of children borne by a women include the extent of exposure to the possibility of conception, i.e. the time spent married and/or having sexual relations, as well as the extent of practice of breastfeeding, which prolongs postpartum infecundability, of a contraceptive method, whether traditional or modern, and of induced abortion. Together, behavioural constraints lower fertility to the levels of total fertility we currently observe—about two children per woman in industrialized countries, or that we have heard of—for example, a family size range of five to seven or eight children in Africa. The most important single determinant of the difference overall has come to be the practice of contraception: consequently, in contemporary societies, fertility has become largely determined by behaviour, and by choice, through voluntary regulation. Biological factors nevertheless still play a considerable role in the variability of individual fertility that will be illustrated below.
This brief note will first review the distribution of fertility worldwide, then review the worldwide distribution and etiology of infertility, and finally illustrate the contribution of both biology and behaviour to demographically observed fertility, by focussing on the role of women’s age.
The demography of fertility
The level of fertility in the world varies broadly by country and culture, social and economic conditions, as well as by individual characteristics such as age. Generally, more industrialized and economically developed societies have lower fertility than agricultural, less developed societies. Also, within countries, generally, more educated groups with higher incomes have lower fertility than less educated groups with lower incomes. Historically, as groups within countries have improved their living standards, and nations have become more economically developed, health conditions have improved, morbidity and mortality have declined, and fertility has declined due to the adoption of fertility-constraining behaviours, such as the limitation of sexual relations or marriage, practice of contraception, and resort to induced abortion. This sequence of events has been observed in western industrialized societies over the last two centuries, and in developing regions in the last half century. The process is often referred to as the demographic transition, and it comprises the principal theoretical base for research conducted by demographers, sociologists, anthropologists, epidemiologists, economists, and others, on the determinants and consequences of the levels of mortality and fertility of national and regional populations.
Figure 1 shows the distribution of estimated current levels of total fertility worldwide. It can be seen that, as noted earlier, total fertility ranges from a high of about six children per woman on average in Africa to a low of under two children per woman in Europe. North America, the newly independent states of the former USSR (less the three Baltic states), Oceania, Latin America, and Asia are intermediate, in order of increasing levels of total fertility. These regional averages, although highly different, nevertheless subsume substantial intra-regional and, of course, country variation. Accordingly, for example, the fertility of European subregions ranges from 1.5 in Southern Europe to just over 2.0 in Eastern Europe, of Oceania from under 2.0 in Australia and New Zealand to over 5.0 in some Pacific islands, of Asia from just over 2.0 in Eastern Asia to about 4.7 in Western Asia, and of Africa from about 4.2 in Southern Africa to about 6.8 in Eastern Africa.
The demography of infertility
In view of the much higher level of fertility in Africa than in other regions, it is surprising to observe that, contrary to expectation, the level of infertility is higher in Africa than elsewhere. The levels of infertility are illustrated in Figure 2. The measure of infertility used is here is the proportion of women who are childless by the age of 40 to 44 years—or 45 to 49 years—i.e., at the end of their fecund period of life. Again, the regional averages subsume substantial variation between countries and within regions. For example, the proportion of women in their forties who are childless in Africa ranges from a level of a few percent in western Africa, to fully a fifth or a third of women in central Africa.
As the level of infertility one could expect due to inborn errors, congenital factors, and lifetime celibacy in women is on the order of about 3 percent worldwide, it can be seen that childlessness in most major regions -except the Middle-East and South America, taken as a whole—is excessive.
A major cause of excess infertility in women is acquired sterility due to pathological conditions. This is especially notable in Africa. For example, whereas in developed countries, although no demonstrable cause could be found in over a third of infertile women who were examined (39 percent), and the two major diagnoses associated with infertility were endocrine factors, including menstrual and ovulatory disturbances, (35 percent) and tubal factors, including occlusion of one or both fallopian tubes and/or pelvic adhesions (27 percent), of infertile women examined in Africa, the proportion with no demonstrable cause was less than half as much (16 percent), but the proportion who had a diagnosis of tubal infertility was nearly two-thirds (64 percent), and only a quarter (24 percent) were diagnosed to have ovarian or endocrine disturbances (Farley and Belsey (2)). Figure 3 illustrates the differences between more developed countries and sub-Saharan Africa in the distribution of diagnoses (Figure 3a and Figure 3b), as well as the prevalence of a diagnosis of tubal factors as compared to the finding of no demonstrable cause by major region (Figure 3c and Figure 3d).
The most likely origin of acquired sterility that manifests itself at least in tubal abnormalities is infection with a sexually transmitted disease, particularly with the organisms of N. gonorrhoeae or C. trachomatis, referred to commonly as gonorrhoea or chlamydia infection. Several findings support this probability. Depending on the age and prior fertility of women, between 10 and 20 percent of gonorrhoea infections and 8 to 10 percent of chlamydia infections lead to pelvic inflammatory disease (p.i.d.), an ascending infection of the reproductive organs. Fully three-quarters of cases of p.i.d. (75 percent) have been found to be due to sexually transmitted diseases, at least in developed countries, although the relative contribution of the two organisms to the overall level of p.i.d. does vary by region (chlamydia is the more recent disease of the two; see Frank (3)). Tubal occlusion has been associated with p.i.d. due to both organisms, particularly chlamydia, which might be responsible for up to two thirds of cases of tubal occlusion. The incidence of tubal occlusion can be tied to the number of episodes of p.i.d.: in a Swedish study, one episode was associated with 11 percent subsequent sterility, whereas three episodes were associated with a probability of sterility of over half (54 percent). Finally, the likelihood of sterility to result from p.i.d. depends on the general use of antibiotics; in view of their low availability in Africa, it has been estimated that the probability of sterility after p.i.d could be as high as 75 percent, a level comparable to the risk in Europe before the advent of antibiotics (see the review in Farley and Belsey (2)).
The prominent role of sexually transmitted diseases in the etiology of underlying sterility and manifest, demographic infertility points to the important interaction of biology and behaviour in the determination of fertility, because the transmission and distribution of sexually transmitted diseases such as gonorrhoea and chlamydia are largely determined by patterns of sexual behaviour that are part of the social and cultural fabric of human societies. In addition to differences in the availability of medical and therapeutic interventions, the variation in these patterns probably goes far to explain regional variability in the prevalence of infertility. Another interesting interaction between biology and behaviour in the determination of fertility centers on the role of women’s age.
Age, fertility, fecundity, and fecundability
The relationship of age to fertility, fecundity and fecundability illustrates the complex interplay of behavioural and biological determinants of fertility, as observed at the demographic level.
Whereas the onset of fecundity is determined by first ovulation and menarche, the beginning of fertility is determined by social and cultural factors: early fertility will either be desired in order for a society to maximize fertility, which may grant social and economic advantages, or not desired in those societies that encourage later childbearing for demographic, economic and/or social reasons. Consequently, actual onset of fecundity is delimited by biological events—the true capacity for childbearing—and behavioural constraints. Behavioural constraints on early childbearing, which are generally increasing in all human societies, influence the materialization of fecundity by lowering fecundability, or the probability of a conception: social norms dictate behavioural constraints that prevent individual women from realizing their biological capacity to conceive.
Similarly, an interplay of biological and social factors serve to determine when fertility ceases. In this case, however, the actual timing, or onset, of natural sterility, is less evident, and it is likely that behavioural factors obscure the biological determination of childbearing cessation even more than its initiation.
Evidence of fecundity cessation
Until recently, there were three sources of information on the end of childbearing: data on menopausal status by age, data on age of women at last birth in societies that do not limit fertility (i.e. that have a pattern of natural fertility), and data on pregnancy rates from artificial insemination of women of different ages.
Menopausal status data show that in most populations a proportion of women cease menstruating by the late thirties that increases to about 20 percent in the mid-forties to reach 100 percent by the mid-fifties. This observation informs us little on the end of fecundity, however, because menopause is not well-circumscribed enough to serve as a marker for the end of menstrual cycles, and presumed ovulation.
Data on age of women at last birth in natural fertility societies suggest, furthermore, that even if the timing of menopause were better determined, it might not be an accurate indicator of the end of fecundity, because whereas the average age at menopause is about 48-50 years, women appear to have their last birth on average many years earlier, at about 39-41 years. Other sources serve to support the finding that infecundity occurs on average about eight years before manifest menopause (Gray (6)).
According to data on artificial insemination, infecundity onset would occur even earlier, as noticeable declines in pregnancy rates are observed by the mid-thirties. Several biases, however, serve to reduce the usefulness of data from artificial insemination studies to time the natural decline of fecundity with age (see Frank et al. (4)).
Recent demographic and biological findings
Recent demographic and biological research on fertility at older ages is now beginning to shed more light on the end of fertility, and to underscore the mixed contributions of behaviour and biology. New data from demographic surveys on women’s frequency of intercourse—that indicates their level of exposure to the " risk " of pregnancy—are providing new insights and a new perspective on the issue, giving more weight to the role of infecundability. Whether women are asked their recent frequency of intercourse, or time elapsed since their last sexual intercourse, data for more than a dozen and a half developing countries show that their exposure to the risk of pregnancy declines noticeably after the mid-thirties, and/or after 15 years of marriage. Accordingly, as illustrated in Figure 4, the proportion of women who have not had sexual relations in any elapsed interval increases sharply with age, particularly by forty, but beginning already in the mid to late thirties. Consequently, whereas women biologically need a longer period of exposure to the risk of pregnancy with age to achieve conception, the possibility of conception declines with age on behavioural grounds alone, due to such factors as the decline of exposure with duration of marriage. Actual fecundity is more obscured by factors of fecundability than was apparent earlier.
On the biological side, recent research on medically assisted conception is clarifying the origin of declining fecundity with age, especially in regard to the relative contributions of endometrial and ovarian ageing, as well as providing alternatives to undesired infecundity due to age. Indeed, although decreased fecundity with age is undoubtedly influenced to some degree by decreased sexual activity with age, biological factors still determine the underlying decline in capacity to conceive and carry to term a pregnancy.
The availability of new assisted reproductive technology (ART) procedures has created a new important study model, at the same time that it has provided possible solutions for infecund couples. Recent data accumulating on the question of embryo quality generally support the hypothesis that the reduced quality of oocytes, a problem of ovarian origin, accounts importantly for declining natural fecundity with age, and recent developments in assisted reproductive technology procedures have provided an investigative tool to develop protocols for oocyte and embryo donation in acyclic, prematurely menopausal and postmenopausal women—as well as ovulating, but infecund women—that by extrapolation also permits evaluation of the different factors involved in in vivo decline of fecundity with age (see the review in Frank, et al. (4)).
It is probably no accident that these technologies have been developed and become more accessible at the same time that post-war generations who delayed childbearing created a further demand for technologies to help solve the problems of starting childbearing at older ages. Consequently, in a sense, it is behaviour that has served to determine developments in the biological basis of reproduction.
For developing countries, it is unlikely that the extension of fertility with age will make a change that will be visible demographically. In western societies, however, the possibility of increasing the fertility of older women may have far-reaching implications. Fertility can now be extended to women of age groups for whom childbearing has been biologically unlikely and socially unusual. Extension of life expectancy for older women, the benefits of hormonal replacement therapy, and general improvement of the health and living conditions of older women in many western societies will likely have a favourable influence on the social acceptability of very late childbearing.
- Bongaarts, J., and Potter, R.G. (1983): Fertility, Biology, and Behavior: An Analysis of the Proximate Determinants. Academic Press, New York.
- Farley, T.M.M., and Belsey, E.M. (1988): The Prevalence and Aetiology of Infertility, Proceedings of the African Population Conference, International Union for the Scientific Study of Population (IUSSP), Dakar, 7-12 November.
- Frank, O. (1987): IPPF Med. Bull., 21:6-8.
- Frank, O., Bianchi, P.G., and Campana, A. (1994): J. Biosoc. Sci., (in press).
- Frank, O., and Bongaarts, J. (1991): Stat. in Med., 10:161-175.
- Gray, R.H. (1979): In: Natural Fertility: Patterns and Determinants of Natural Fertility, Proceedings of a Seminar on Natural Fertility, edited by H. Leridon, and J. Menken, pp. 217-250. Liège, Ordina Editions.
- United Nations Population Division, Department of Economic and Social Information and Policy Analysis (1993): World Population Prospects: The 1992 Revision. United Nations, New York.
- World Health Organization Scientific Group (1981): Research on the menopause, Technical Report Series 670. World Health Organization, Geneva.