Fertility, Biology, and Behavior
Citation: Bongaarts, John, Potter, R. G. Fertility, Biology, and Behavior.
CHAPTER ONE Introduction and Overview Proximate determinants of fertility: the biological and behavioral factors through which social, economic and environmental variables affect fertility. If a proximate determinant changes, fertility necessarily changes. Fertility: actual reproduction Fecundity: the ability to reproduce Infertility: childlessness. Either caused by voluntary decision not to have children or by (biological) infecundity Infecundity: sterility Fecundable: able to conceive. A woman may be fecundable, but infecund because they are physically incapable of carrying a pregnancy to term. Fecundability: the probability of conceiving per month (among cohabiting women who are not pregnant, sterile, or temporarily infecundable) Natural fertility: no contraception or induced abortion is practiced Controlled (Regulated) fertility: fertility control practices are practiced Potential reproductive years start at menarche, though socially sanctioned childbearing is generally limited to women in relatively stable sexual unions. In practice, marriage or cohabitation may be taken as the starting point of the actual reproductive years. Therefore, changes in the age at menarche can only affect fertility by influencing the age at marriage. The length of the birth interval is determined by the postpartum infecundable interval (primarily a function of breastfeeding behavior), the waiting time to conception (which is inversely related to the natural fecundability (which is largely determined by frequency of intercourse) and to the use and effectiveness of contraception), and a full-term pregnancy (which varies little). In the case of spontaneous or induced intrauterine death there is an additional period of a shortened pregnancy, a brief infecundable period, and a conception delay. This article identifies seven proximate determinants: marriage (and marital disruption); onset of permanent sterility; postpartum infecundability; natural fecundability or frequency of intercourse; use and effectiveness of contraception; spontaneous intrauterine mortality; induced abortion. In countries in which contraception and/or abortion is widely available, we can observe large differences in the number of children that women have because of differences in desired family size and non-voluntary reasons such as the premature onset of sterility. In natural fertility populations, contraception and abortion are virtually absent, but the length of birth intervals varies widely due to differences in age at marriage, breastfeeding practices, frequency of intercourse (i.e. taboos against postpartum sexual relations), the risk of spontaneous intrauterine mortality, and the onset of permanent sterility. There is also a random component. Rate of childbearing is lowest at youngest and oldest ages and highest in the central childbearing years. Populations differ little in prevalence of natural sterility. Any age-related increases in the risk of intrauterine mortality, decline in frequency of intercourse, and rise in postpartum infecundability have a modest impact on the age pattern of fertility. Health and nutrition are unimportant determinants of fertility. Chronic malnutrition has a minor influence on fertility, but acute starvation causes a substantial reduction in fertility. The weak effect of malnutrition operates through the age at marriage (through timing of menarche) and duration of postpartum infecundability. The strong effect of famine is because substantial numbers of women become amenorrheic. It also may be due to separation of spouses in the search for food or work, decline in libido, and voluntary birth control, through contraception, induced abortion, or abstinence. The exceptional case in which health plays an important role in fertility is where there are high levels of sexually transmitted diseases that cause sterility. There are two indirect effects of improved health on fertility: declines in infant mortality, which lead to lengthening of breastfeeding period and postpartum infecundity; and declines in risk of widowhood during the reproductive years. But these two effects tend to compensate one another. CHAPTER TWO Natural Fertility and Its Proximate Determinants Natural fertility: fertility in the absence of deliberate birth control that is "bound to the number of children already born and is modified when the number exceeds the maximum which the couple does not wish to exceed" (from Henry, 1961). Prolonged breastfeeding and postpartum abstinence lower fertility, but are considered to be natural when they are not modified deliberately as the number of children already born rises. Natural fertility is also reduced through delays in marriage and marital disruption. A society's level of natural fertility is determined by its marriage pattern and it level of natural marital fertility. The age-pattern of marital fertility is relatively invariant. Age-specific natural fertility rates show a convex shape, declining slowly until the late-30s then dropping sharply after age 40. This contrasts with the concave pattern of controlled fertility, in which fertility declines much more quickly at younger ages due to the use of contraception and abortion. There is a great deal of variation among societies in their levels of natural fertility. Total fertility has been found to range from 3.7 in one French village in the eighteenth century to 9.5 among the Hutterite population in the early-twentieth century. It is estimated that natural fertility can reach as high as 15 births per woman, though no society has approached this level. There are five proximate determinants of natural fertility: proportion of women married; duration of postpartum infecundability; (natural) fecundability; the risk of intrauterine mortality; and the onset of permanent sterility. The last four are the proximate determinants of natural marital fertility. Postpartum infecundability: very prolonged periods have been observed, ranging up to two years. In most societies the duration so postpartum infecundability equals the duration of amenorrhea (or anovulation) because little or no postpartum abstinence is practiced. However, in some societies the period of postpartum abstinence exceeds the amenorrheic intervals of many women. Without breastfeeding the average amenorrheic interval is usually 1.5-2.0 months. An additional month of breastfeeding increases amenorrhea by less than one month. Amenorrhea is affected by the duration, type, and intensity of breastfeeding. Waiting time to conception and fecundability: the conception wait in historical population is calculated by subtracting nine months for pregnancy and two months for time lost due to intrauterine mortality from the interval from marriage to first birth. Average conception waits range from about 5 to 10 months with an average of about 7 months for women in their twenties. Waiting times tend to be somewhat longer for younger and older women and women who experience prolonged periods of separation from their spouses. The most widely used measure of waiting time is fecundability (see definition above). A typical value is 0.2, which means that in a group of fecundable women, 20% can be expected to conceive in the first month of exposure. Fecundability varies because women have difference frequencies of intercourse and difference biological characteristics. Conception is only possible during the fertile period around the time of ovulation (approximately 48 hours), but even if intercourse occurs during this period, fertilization does not always occur (p=0.95) and even when it does, only half of all fertilized ova yield a recognizable conception. The estimate for fecundability for women in the central childbearing years is then: f = 0.45 (1 (M n)(M n 1)/(M2 M)) where n is the number of coital acts occurring during an interval of M days. This implies that in a given menstrual cycle, as the frequency of intercourse rises from 1 to 20 per cycle (assumes sex once per day), fecundability rises from 0.035 to 0.429. Spontaneous intrauterine mortality: usually include both spontaneous abortions (fetal deaths before the 28th week of gestation) and stillbirths (fetal deaths after fetus becomes viable (28 weeks)), but exclude embryonic deaths before the first missed menses. Intrauterine mortality is difficult to measure because of incomplete reporting, unrepresentativeness of study populations, overstatement due to reporting of induced abortions or misreporting of delayed menses, and errors in sampling or statistical estimation procedures. The risk declines steeply from a high of 8.1% in the second month of gestation to a low of 0.2% in the eighth month, rising thereafter. The risk of fetal death after six weeks of pregnancy should be near 15%. Differences between countries appear to be quite small. Most early mortality is due to genetic defects or ectopic pregnancies which are not generally affected by socioeconomic characteristics. Prevalence of Permanent Sterility: Mean age at menopause tends to range from 44 to 50 years. Sterility can also occur because of abnormalities in the reproductive system; high incidence of anovulatory cycles in years preceding menopause; rapid rise in intrauterine mortality and undetected embryonic mortality among women in their forties; and high prevalence of specific diseases. The first three do not vary much across populations, but the last can vary dramatically from being highly significant in some African societies to small or absent in most other populations. The mean age at the onset of sterility can be estimated at 41.7 years. Generally, the end of childbearing or onset of infertility is measured by the mean age at last birth. This age is generally invariant, falling between the ages of 39 and 41 years old when sexually transmitted diseases are not prevalent. The largest variations in the total fertility rate are caused by changes in age at marriage and in the duration of postpartum infecundability, and the smallest fertility variations are found for the end of the childbearing years and spontaneous intrauterine mortality. Postpartum amenorrhea is the only one for which there is clear evidence of an important change with modernization. CHAPTER THREE Regulated Fertility and Its Proximate Determinants Theoretical effectiveness: the effect of a contraceptive under ideal laboratory conditions. Use effectiveness: the protection from unintended pregnancy under real-life conditions. Extended use-effectiveness: takes into account unintended conceptions following discontinuation of a contraceptive in addition to failures during periods of use. Pearl pregnancy rate (P): estimates the number of contraceptive failures per unit of exposure to risk of conception. The formula is P = (number of undesired pregnancies/number of months of exposure) * 1200 Cumulative failure rate (F): proportion of women who become unintentionally pregnant within a given time from the beginning of a segment of contraceptive use. Requires the application of life table methodology. Total abortion rate: the average number of induced abortions per woman at the end of the reproductive years if current age-specific abortion rates prevail throughout the childbearing years. Populations can control fertility in two ways: by reducing the exposure time to childbearing (through delayed marriage and marital disruption) and by deliberate control of marital fertility (through contraception and abortion). During the third quarter of the twentieth century, Asia experienced the largest decline in fertility, while Africa experienced very little change. Bongaarts selects 8 countries (Colombia, France, Indonesia, Kenya, Korea, Mexico, Pakistan, and the United States) that represent a wide range of fertility regimes to show the differences in fertility throughout the world. In 1975, the total fertility rate ranged from a high of 8.0 in Kenya to a low of 1.8 in the United States. Kenya and Pakistan have essentially natural fertility. The remaining six countries have regulated fertility. Bongaarts identifies a number of variables that lead to marital fertility differentials. Those with urban residence, better education, and higher occupation groups tend to have lower overall fertility, but there are exceptions. Marital fertility is lower in rural areas of Pakistan and Indonesia and among women with lower education in Kenya and Indonesia. The mean age at marriage in the sample of countries is 21.4 (up from 20.4 in the 1960s). Fertility tends to decline as the mean age at marriage rises. Most women marry at some point in their life in all 8 societies, though rates drop as low as 86.6% of all women is some countries. The large majority of women who ever marry do so between the ages of 15 and 25. The lowest proportions married at these ages are found in France, Korea, and the United States; the highest in Kenya, Pakistan, and Indonesia. During their reproductive years, women are married on average for about 2/3 and single for ? of their reproductive lives. This varies from 17% single in Kenya, Pakistan, and Indonesia to over 25% single in the other five countries. On average the mean age at marriage is higher for urban and more educated women. Worldwide in 1975, 1/3 of women were currently using contraception. In all 8 countries contraceptive use increased between 1965 and 1975, but the rise was negligible in Kenya and Pakistan. No relationship between prevalence of contraception and methods used. A little over half use the pill or wear an IUD and 1 in 8 is or has a partner who is sterilized. Prevalence increases with age until ages 30-34 then declines slightly. The only significant difference between countries is found in France and the U.S., where prevalence is high among young women. Contraceptive use tends to be higher among urban (40% vs 30% for rural) and educated women (47% with secondary education vs 34% with primary education). The standard measure of effectiveness of contraception is, e, where e = 1 fr/fn where fn is natural fecundability and fr is residual fecundability. With a perfect contraceptive fr is 0 and effectiveness is complete, e = 1.0. Other measures are the Pearl pregnancy rate and the cumulative failure rate, defined above. The Pearl pregnancy rate has the drawback that in a group of women initiating contraception, the women with the highest frequency of intercourse and lowest effectiveness have the highest risk of conceiving early, leaving increasing proportions of women with low risk in the sample over time. The "improved Pearl index" fixes this problem by observing women for a fixed interval of one year. It is closely related to the 1-year cumulative failure rate. It is usually necessary to derive effectiveness levels from available estimates of Pearl rates or cumulative failure rates (see Appendix of this chapter for formulas). Effectiveness of contraceptive methods in the U.S. ranges from 0.99 for pills to 0.87 for the rhythm method. Estimates for the Philippines range from 0.95 for pills to 0.62 for condoms. In the U.S. failure rates decline with rising age and women who want to prevent a pregnancy have lower failure rates than women who want to delay a pregnancy. After controlling for age and intention, there is little difference by socioeconomic status. Access to abortion varies substantially throughout the world. In general, access can be summed up as follows: 28% - Abortion is prohibited without exception, or only allowed to save the life of the woman (most of Latin America and Africa, and Muslim countries of Asia) 10% - Abortion is authorized on broader medical grounds to avert threat to women's health and on eugenic or juridical indications (Korea) 24% - Abortion allowed on socio-medical grounds such as health, unmarried status, or inadequate income (most of Eastern Europe, Japan, India) 38% - Abortion on request, but generally limited to first trimester (U.S., USSR, China, France) The Total abortion rate (TA) is over 1.0 in five developed countries in Eastern Europe. Intermediate levels of 0.5-1.0 are observed in 8 developed countries mostly in eastern and northern Europe. Low levels (<0.5) are found in the remaining ten developed countries. In the developing world, TA is highest in Cuban and Korea, intermediate in Singapore and lowest in Tunisia, Hong Kong, and India (China is not included). Age-specific abortion rates typically have an inverted U-shaped pattern. In the West often more than half of all abortions occur among single or previously married women, but in the other countries the majority of abortions take place when women are married. CHAPTER 8: Birth Spacing As is evident by its title this chapter focused on spacing control. In the absence of efforts to space births, the time of the last desired birth varies appreciably as the result of random variations from conceptive delays and numbers of spontaneous fetal losses. Lengthy anovulation adds another source of variation. This stochastic variation accumulates rapidly over successive births, meaning that the standard deviation of duration from marriage to last desired birth increases sharply as desired family size increases. There are two types of spacing: master-schedule spacing and serial spacing. Under master-schedule spacing, the couple tries to keep their successive births as close to a chosen schedule (based on age at marriage and the number of children that the couple ultimately wishes to have) as possible, and therefore, abbreviates or prolongs their practice of contraception depending on whether they fall behind or forge ahead of their timetable. The standard deviation of duration to last birth still increases as the number of children desired increases but the cumulation of variation is much slower than in the absence of spacing-control efforts. In a youthful marriage (around age 20) with efficient contraception, a wide- (3 years) rather than a narrow-spacing (<=2 years) goal favors spacing control by affording more opportunity to compensate for intermediate births that are late. However, if contraception is inefficient, the advantage switches to a narrow-spacing goal. When marriage is late (around age 30), a wide-spacing goal pushed childbearing into more advanced ages when the risks of pregnancy wastage and decreased fecundability are heightened and control over the last birth interval is weakened. Under serial spacing the couple takes each birth interval in its own right, which means a minimum and maximum length that they do not wish to violate and a most preferred intermediate length that they aspire to. Optimum strategy is to raise natural fecundability (increase coital frequency) as high as possible and then to interrupt contraception shortly before 9 months prior to the target date for the next birth. There a number of possible decision rules that couples can use to decide when to end contraception. However, when natural fecundability is low, none of these decision rules work very well, whereas if natural fecundability is high, all become efficient and lead to approximately the same prescriptions. An augmented rate of marital intercourse is effective in raising natural fecundability only if properly times within the menstrual cycle and extended over a sufficient number of days to give a high probability of encompassing the fertile period.