Fertility Regulation and Reproductive Health in the Millennium Development Goals: The Search for a Perfect Indicator

Abstract

Although several key elements of sexual and reproductive health are included in the United Nations Millennium Development Goals, a measure of women’s capacity to regulate their fertility safely and effectively is missing. We considered the usefulness of 3 pairs of indicators in monitoring this component of reproductive health: contraceptive prevalence and total fertility; unmet need for contraception and unplanned births; and unsafe abortion and abortion mortality.

A single measure of contraceptive use is insufficient. The risks women face from unplanned births and unsafe abortion should also be incorporated into the monitoring process, either directly within the Millennium Development Goals framework or as a parallel effort by reporting governments and other agencies.

The Program of Action adopted at the 1994 International Conference on Population and Development (ICPD) in Cairo stresses the importance to social and economic development and to individual and family well-being of achieving reproductive health for all.1 The United Nations Millennium Development Goals (MDGs) adopted by heads of government in 2000 embrace much of the ICPD Program of Action in their emphasis on eradicating extreme poverty, achieving universal primary education, promoting gender equality and empowering women, reducing child mortality, improving maternal health, combating HIV/AIDS, malaria, and other diseases, ensuring environmental sustainability, and forging a global partnership for development.2 Although reproductive health is not specifically named (the narrower term “maternal health” is used instead), it is widely recognized that ensuring universal access to reproductive health care, including family planning and sexual health, is essential for the achievement of all of the MDGs, and vice versa.3

Two benchmark indicators of reproductive health adopted by the UN General Assembly at its 1999 review of progress in implementing ICPD agreements (ICPD+5) are included in the MDG framework: (1) the percentage of births that are attended by skilled health personnel (together with the maternal mortality ratio [MMR]); and (2) the knowledge of how to prevent HIV/AIDS among young people aged 15–24 years (or HIV prevalence among young pregnant women).4 However, the fertility regulation indicator from ICPD+5—the percentage of women with an unmet need for contraception—is missing, despite its clear relevance to the goal of improving maternal health. Also missing from the MDGs is the consensus from ICPD+5 that all primary health care and family planning facilities should be able to provide a full range of reproductive health services. In response, the Millennium Project Task Force on Child Health and Maternal Health has recommended that a target of “universal access to reproductive health services by 2015” be added to the maternal health goal, with appropriate indicators.5

The task force has also recommended the addition of a fertility regulation indicator of maternal health: the percentage of the total demand (or desire) for contraception that is satisfied by current use, which is a variant of unmet need. In considering this and other recommendations, we ask: is there a perfect indicator? Is any single measure of contraceptive use or fertility outcomes sufficient to capture the health- and rights-based concept of fertility regulation articulated at Cairo? If not, what additional measures are needed to create a more inclusive set of monitoring tools for this purpose?

To answer these questions, we assessed 3 pairs of process and outcome indicators with respect to their utility in tracking women’s capacity (i.e., the exercise of their right) to regulate their fertility safely and effectively in accordance with their reproductive intentions: contraceptive prevalence and total fertility rates (TFRs); unmet need for contraception (or proportion of demand or desire satisfied) and unplanned births; and unsafe abortion and abortion mortality. Reported here as aggregates for countries (a data table is available on request), each indicator is amenable to analysis by age, marital status, socioeconomic status, rural-urban residence, and other characteristics of interest, because each is (or can be) derived from population-based surveys. Indeed, disaggregating the data is essential for identifying the nature and degree of within-country inequalities and for designing approaches to overcome them, especially for adolescents, marginalized ethnic groups, and the poor.

CONTRACEPTIVE PREVALENCE AND TOTAL FERTILITY

In 2000, a WHO/UNICEF/United Nations Population Fund (UNFPA) working group selected the contraceptive prevalence rate (CPR) as one of 17 core indicators of reproductive health for global monitoring.6 The CPR enumerates current contraceptive users (of modern and traditional methods, separately or combined) as a percentage of all women of reproductive age who are in a marital or consensual union and, in the Caribbean, in a regular visiting union, whether or not they are currently sexually active and at risk of pregnancy (i.e., fecund and not currently pregnant). Some demographic and health surveys (DHSs) and related studies also solicit information on method use from sexually active adolescents and unmarried adults.

Although classifying women or couples as contraceptive users or nonusers involves widely recognized measurement problems, the CPR and its subcategories of particular method use offer a number of advantages for global monitoring. Familiar and widely reported, CPRs at the aggregate level are a good predictor of the TFR across countries, that is, the average number of children a woman would have in her lifetime on the basis of current age-specific birth rates. Among 102 countries in less-developed regions for which data are available, the correlation between contraceptive prevalence and total fertility is strongly negative (−0.875; P = .00), with fairly consistent patterns across regions.

An additional advantage is that the distribution of users across categories of modern and traditional contraceptive practices can serve as a proxy index of method “choice,” which has been shown to contribute to higher contraceptive uptake, continuation, and safety for the user.7 When combined with data on the user-effectiveness of each method, such distributions can be used to estimate rates of contraceptive efficacy within populations and subgroups.8 Related measures, such as contraceptive knowledge, past use of a method, and awareness of where to obtain modern contraceptives, are also useful markers of the capacity of nonusers to regulate their fertility if and when they wish to do so.

Limitations of the CPR

The CPR has limitations as a monitoring tool, however. The first relates to the assumption that women’s contraceptive needs are somehow “met” when they are classified as users. Yet women may not be choosing the safest method with respect to their individual health status,9 or using the most effective method with respect to the nature and intensity of their fertility intentions (e.g., avoiding conception altogether vs delaying the next pregnancy), or practicing the method correctly or consistently. An estimated 34% of all mistimed and unwanted births worldwide are to women who claim that they or their partners were using some preventive method(s) during the month that conception occurred, and self-reported contraceptive users also account for a substantial proportion of abortion clients in most settings.10 Thus, despite the high cross-country correlation, contraceptive prevalence in the aggregate is a rather poor gauge of women’s ability, at the individual level, to match their reproductive outcomes with their intentions.

A second problem with the CPR is the setting of goals and targets. How should progress be measured, and against what endpoint? Given that contraceptive use or nonuse depends so heavily on individuals’ or couples’ reproductive preferences, it makes little sense to set a target rate in the absence of a specific demographic rationale for doing so. Yet it is exactly this linkage with the “demographic imperative” and its associated policies of population control that made contraceptive prevalence and related measures such as “couple years of protection” and “births averted” so vulnerable to criticism in the past.11 Contraceptive prevalence, with its assumption that “higher is better,” does not adequately reflect either women’s reproductive preferences or their difficulties in achieving them. This is especially problematic in countries with low levels of modern or traditional contraceptive use that are difficult to interpret without additional information.

Total Fertility as an Indicator of Reproductive Health

The WHO/UNICEF/UNFPA Working Group in 2000 also selected the TFR as a core indicator for global monitoring. Like the CPR, the TFR by itself tells us little about women’s reproductive preferences (especially at high fertility levels) or their effectiveness in achieving them. Instead, it serves, for our purposes, primarily as a marker of the frequency of women’s exposure to the prevailing risks of pregnancy and childbirth. In this context, it is not the TFR per se, but rather, its interaction with the MMR that tells the relevant story.

In the abstract, high fertility would not represent a significant threat to women’s health if the ratio of pregnancy-related deaths to live births were very low. But because high fertility is typically found in the same resource-poor settings as is high maternal mortality, their combined effects are explosive. A woman’s lifetime risk of dying during pregnancy, abortion, childbirth, or its aftermath is perhaps the single best indicator of the profound inequities in maternal health across and within populations. According to estimates from the WHO, women living in developed regions have 1 chance in 4085 during their lifetime of dying from pregnancy-related causes compared with 1 chance in 61 in developing regions as a whole, and 1 in 16 in sub-Saharan Africa.12 At the extremes, the risks for Swedish women are 1 in 29 800 compared with 1 in 7 for women in Angola, Malawi, and Niger.

The age- and parity-specific fertility rates from which TFRs are derived are indispensable for identifying potentially high-risk population profiles, such as first births to young adolescents, high-parity births to older women, and closely spaced pregnancies, many of which might otherwise be prevented with effective contraceptive use.13 But the extent to which these potential risks translate into poor maternal mortality and morbidity outcomes depends on the accessibility and quality of antenatal, delivery, and postpartum care. Where quality and accessibility are favorable, age- and parity-specific differentials tend to diminish or even disappear. In addition, the majority of maternal deaths in most populations occur in women who would not be identified as high risk according to these age-parity criteria.14 For this reason, some investigators have concluded that family planning strategies that try to persuade women to avoid potentially high-risk categories of pregnancies are likely to be less effective in reducing MMRs than are health system strategies designed to provide essential obstetric care and emergency referral services for all women during childbirth and the immediate postpartum period. (An indicator of access to emergency obstetric care is proposed in the recommended “universal access to reproductive health services” target for the MDGs.) What family planning programs can do better is to enable women who are already motivated to postpone, avoid, or stop childbearing to do so more effectively, thus circumventing the avoidable risks of unsafe abortion or delivery posed by conceptions or births that they did not intend to have.15

UNMET NEED FOR CONTRACEPTION AND UNPLANNED BIRTHS

The concept of unmet need for contraception has the considerable advantage of taking into account women’s expressed interest in postponing or avoiding a pregnancy in assessing her “need” for a contraceptive method. In its 5-year review of ICPD, the UN General Assembly at ICPD + 5 adopted a target of eliminating the global unmet need for family planning by 2015—that is, closing “the gap between contraceptive use and the proportion of individuals expressing a desire to space or limit their families.”16

Unmet need is conventionally defined as the percentage of women in marital or consensual unions who (or whose partners) are not using any contraceptive method even though they are at risk of conceiving (that is, are sexually active, fecund, not currently pregnant) and do not want to be pregnant any time soon. Unmet need can be calculated for spacing purposes (to postpone pregnancy by at least 2 years) and limiting (stopping or avoiding childbearing entirely) according to women’s reports of their reproductive intentions at the time of the interview.17 On the basis of the conventional definition, the proportions of women with an unmet need for some form of contraception range from under 10% in countries such as Kazakhstan, Indonesia, Viet Nam, Brazil, and Colombia, up to 35% or more in Rwanda, Uganda, Senegal, Yemen, and Haiti.18

To obtain a more realistic picture of the scope of unmet need, many surveys now interview sexually active adolescent and adult women who are not currently in a marital or consensual union. Some also count users of traditional methods, such as withdrawal, periodic abstinence, or postpartum amenorrhea, as having an unmet need for modern (that is, presumably, “better”) contraception. This second definitional shift raises the estimates of unmet need in the developing world from 113.6 million to 201 million couples, the latter representing 3 in every 10 women at risk of conceiving unintentionally.19

Are high CPRs associated with low levels of unmet need for contraception across countries? Not necessarily. Unmet need is not a mirror image of the CPR, for each taps a slightly different dimension. Whereas the overall correlation between contraceptive prevalence and unmet need among all 56 countries in developing regions for which data are available on both measures is in the expected negative direction (r = −0.71; P = .00), the association is positive and the points are more dispersed among the majority of sub-Saharan African countries, plus Yemen, Haiti and Cambodia. This is a familiar reverse U-shaped pattern (one finds the same result with contraceptive prevalence and unplanned births), in which women’s rising aspirations for postponing the next pregnancy (if not to avoid it altogether) in the early stages of a population’s adoption of contraception are, for various programmatic and personal reasons, not yet accompanied by an equivalent uptake in contraceptive practice.20 It is only at higher levels of prevalence that the expected negative relation appears, and then imperfectly.

What Does “Unmet Need” Reflect?

The concept of unmet need—whether for any contraception or for a modern method—clearly taps into the ICPD concept of fertility regulation as a matter of sexual and reproductive health and rights. It takes into account a woman’s reproductive desires and has—at least hypothetically, if not realistically—a zero end-point.

Despite its relative advantages as a monitoring tool, however, unmet need also has its limitations. Primary among these is the interpretation of its underlying causes. Like contraceptive use, unmet need in a population can be low because most women want to become pregnant, for example, and not because their contraceptive needs have been satisfied. Alternatively, high levels of unmet need do not necessarily reflect a lack of access to information and services. Surveys that ask women why they are not using contraception even though they say they do not want to get pregnant right away do cite such reasons, but more frequently they mention causes such as health concerns and fears of side effects, husband’s or community opposition, and other personal and cultural explanations that are, by and large, not amenable to direct programmatic intervention.21 In addition, a woman’s reproductive intentions may be changeable, unclear, situation-specific (“it depends”), or not shared by her partner.22 Qualitative, in-depth investigations of unmet need usually identify complex mixtures of feelings and behaviors relating to a possible pregnancy, such as ambivalence, miscommunication, denial, uncertainties about the sexual relationship, and lucky (or unlucky) experiences with the use or nonuse of contraception in the past.23

Because the need for contraception is defined by the research design and not by the woman herself, unmet need may or may not represent an active demand or desire for contraception. In this sense, the indicator proposed for the MDGs by the Task Force on Child Health and Maternal Health, the “percentage of demand (or ‘desire for family planning’) satisfied,” is somewhat misleading in its implication.24 Indeed, it is this very elusiveness of the multiple factors underlying unmet need that have led some observers to suggest that a program strategy that focuses on serving current contraceptive users better (e.g., with methods that are more appropriate to their needs and that support correct use and continuity) may be more effective in avoiding unintended pregnancies than one that attempts to reduce or eliminate unmet need, which in some settings may be quite resistant to change.25

The Relationship Between Unmet Need and Unplanned Births

The birth of a child that a woman says was unplanned—either mistimed or unwanted—at the time of its conception is a clear indication of a mismatch between reproductive intentions and outcomes. DHSs ask every woman who had any live births during the past 5 years to think back to the time she became pregnant and to say if she had wanted to become pregnant then, did not want to until later, or did not want to have any (more) children at all. Because interviewers refer to the child by name when asking about each pregnancy, women may redefine pregnancies that were unwanted at the time they occurred as having been wanted (perhaps less so for those that were mistimed).26 Even so, despite this almost certain bias, the proportions of births that women say were unwanted at conception range from 5% or fewer in 13 sub-Saharan African countries and in Kyrgyzstan, Turkmenistan, and Uzbekistan, up to 20% or more in Malawi, Yemen, Nepal, Cambodia, Haiti, Mexico, Bolivia, Brazil, Colombia, and Peru.27 Proportions of births unplanned (unwanted plus mistimed) are far higher, exceeding half of all births in 9 of the 67 countries in developing regions for which data are available.

The limited value of unmet need for contraception as a predictor of proportions of births unplanned among 56 developing countries with data for both variables is revealed in the low correlation between these 2 measures (r = 0.26; P = .06). In addition, the proportions of births that are unplanned almost always exceed the proportions of women with an unmet need for some form of contraception. This occurs most dramatically among countries in which 10% or fewer of women have an unmet need for contraception as conventionally measured. Among these is Turkey, where women terminate 20% of their pregnancies by legal abortion, and define an additional 30% of their births as unplanned. Legal abortion ratios are also high in Kazakhstan (41%), Turkmenistan (23%), and Viet Nam (44%), even though unmet need is below 10%. In Brazil, Colombia, and Peru, where similarly low levels of unmet need are combined with CPRs of 70% or more, women terminate perhaps 30% of their pregnancies with induced clandestine abortion (a regional estimate),28 and report half or more of their births as unplanned. Thus, low levels of unmet need are not necessarily associated with low proportions of births being unplanned.

Unplanned pregnancies and births represent unanticipated and largely avoidable social, economic, and health costs to women, couples, and families that are relevant to all of the MDGs—not just maternal health. A review of surveys of abortion clients in 23 countries found that economic reasons (“cannot afford a baby”) were second only to the nonspecific reasons of “want to postpone childbearing” or “want no (more) children” given by women seeking legal terminations in the 1990s.29 Poverty alleviation—the overriding concern of the MDGs—is a prime motivation. Other reasons include the disruption of education or work, relationship problems, partner does not want the pregnancy, the woman is too young or is unmarried, and risks to the health of the woman or the fetus.

Unplanned births tap into other dimensions of maternal and child health as well. In some but not all settings, they are associated with adverse child health outcomes that begin with delayed antenatal care and continue through a sequence of depressed levels of maternal investment in the newborn, such as breast-feeding and other expressions of physical and emotional care.30 Moreover, unplanned pregnancies and births may be related to unwanted intercourse (within or outside marriage) or difficulties that girls and women face in negotiating protection from unwanted pregnancy, as well as lack of access to information and services for terminating an unwanted pregnancy safely and humanely, especially among unmarried adolescents and the poor.31

UNSAFE ABORTION AND ABORTION MORTALITY

When all else fails (or, for various reasons, is not attempted), the ultimate decision for safe and effective fertility regulation centers on the possible interruption of an unintended or unwanted pregnancy. Although promoting more effective contraceptive use with good information and wider access to high-quality care can go a long way toward reducing the numbers of abortions, it will not eliminate the need or demand for safe services.32

An estimated 26% of all pregnancies worldwide are terminated by induced abortion: 41% in developed regions and 23% in developing regions.33 Of the estimated 76 million unintended pregnancies that occur annually in developing countries, perhaps 34 million result in unplanned births. Among the rest, an estimated 10 million end in miscarriage and 32 million are interrupted by induced abortion. Reliable data from the developing world on the ratio of abortions to pregnancies are limited almost entirely to those Asian countries in which abortion is permitted on broad grounds, ranging from 10% in Uzbekistan to 40% in Armenia, Kazakhstan, and Viet Nam. For other regions, the estimated ratio of abortions to pregnancies is 15% in sub-Saharan Africa (of which 99% are probably illegal); 21% and 30%, respectively, in Central and South America (of which virtually 100% are illegal, except in Cuba); and 18% and 20%, respectively, for Central and Southern Asia and for Southeast and Eastern Asia (of which an estimated 78% and 60% are illegal).34

We propose two indicators for monitoring trends and variations in abortion-related risks to maternal health: the percentage of induced abortions that are unsafe as defined by the WHO—that is, procedures “that are carried out either by persons lacking the necessary skills or in an environment that does not conform to minimal medical standards, or both,”35 and the ratio of abortion deaths per 100 000 procedures (the abortion mortality ratio). The former is a process indicator of policy environments and health system capacities; the latter an indicator of reproductive health outcomes. Unlike unmet need and unplanned births, both indicators have a feasible zero target where policy and program environments support high-quality and universally accessible services.36

Eliminating Unsafe Abortion and Abortion Mortality

The ICPD Program of Action urges governments and intergovernmental and nongovernmental organizations to strengthen their commitment to women’s health by addressing the impact of unsafe abortion as a major public health concern. Delegates agreed that women with unwanted pregnancies should have access to reliable information and compassionate counseling; to quality services for the management of complications arising from (unsafe) abortion; and—crucially—where abortion is not against the law, that such abortion should be safe.37 At ICPD + 5, governments further agreed that health systems should train and equip health service providers to ensure that such (legal) abortion is safe and accessible.38 Because abortion is permitted by law, on at least some grounds, in 189 of the 193 countries of the world, this mandate to provide safe services within health and family planning facilities for those who are eligible applies virtually everywhere.

Accurate statistics on the incidence of abortion and on related morbidity and mortality are notoriously difficult to collect, however, whether from institutional sources, indirect estimation techniques, or from women themselves.39 The exception is in countries where services are available on request and where most procedures are performed in the public health system or in private clinics—that is, in approved facilities by trained personnel—with good reporting systems. In these countries, one can calculate abortion ratios (per 100 pregnancies or, alternatively, live births) and abortion rates (per 1000 women of the relevant age group or marital status) as well as abortion mortality ratios, recognizing that there may still be some underreporting.

In estimating the percentages of all abortions that are performed under unsafe conditions where data are scarce or unreliable, however, both the numerator and denominator are problematic. On the basis of complex assumptions and a variety of complete and incomplete country-specific data sources, the WHO estimates that 19 million of the approximately 45 million abortions performed annually worldwide are unsafe according to their definition, with 97% of all unsafe procedures being performed in developing countries.40 Perhaps 68 000 women die each year from the complications of unsafe abortions, including those that are self-induced, and untold numbers suffer from infections and other damage to the reproductive tract. Preventable deaths from improperly performed procedures constitute 13% of maternal mortality globally, and 25% or more in some countries where maternal mortality from other causes is relatively low (e.g., Eastern Europe and South America), making unsafe abortion the leading single cause of maternal mortality worldwide.

In developed countries in which safe services are broadly accessible, abortion deaths average about 1 per 100 000 procedures compared with 6 to 25 deaths during pregnancy, delivery, or its aftermath per 100 000 live births.41 Risks are hundreds of times higher in countries where women turn to untrained providers, quacks, or self-induced methods. Abortion deaths average 330 per 100 000 procedures in all developing regions (excluding China) and as high as 680 per 100 000 in sub-Saharan Africa. Ensuring the safety of all voluntary pregnancy terminations and not just those that are acceptable on narrow legal grounds is clearly central to the reduction of maternal mortality as proposed in the MDGs. The persistence of restrictive policies in many countries should be a major cause for alarm, however. These policies impede women’s access to high-quality services, often with disastrous results, in addition to impeding the collection of reliable data for the global monitoring of maternal health.

Can Population Surveys Collect Information About Abortion?

Although the abortion mortality ratio (like the MMR) requires either a complete reporting system or a good indirect method of estimating deaths from this cause, we believe that an approximation of the percentage of all procedures that are unsafe—like the percentage of births attended by skilled health personnel—could be obtained from population surveys. The standard DHS questionnaire does not ask women about abortions, except in the general question “have you ever had a pregnancy that miscarried, was aborted, or ended in a stillbirth?” (respondents are also asked when the pregnancy ended and at how many months). Appropriately worded follow-up questions could distinguish among these outcomes (as in the 2000 DHS of Armenia, for example)42 and ask about the conditions of abortion as they do about childbirth. Abortion methods, providers, and settings could also be routinely identified, as they are in the 2000 Cambodia DHS,43 and the percentages of unsafe procedures reported overall and for subgroups, such as adolescents, ethnic minorities, and the rural and urban poor.

In addition, in the family planning section of the questionnaire, women could be asked, as they are asked now about delaying or avoiding pregnancy, “have you ever used anything or tried in any way to induce a late menstrual period or terminate a pregnancy? If yes, what have you used or done? Was it successful or not? What else have you tried?”44 They could be asked if they know of a method for doing so (traditional or modern, the latter referring to medical inductions, vacuum aspiration, or dilation and curettage) and of a service provider or facility. There will be underreporting, of course, as there is for questions about other sensitive issues, such as sexual partnerships and exposure to HIV/AIDS.45 Despite this almost certain bias, however, we argue that DHSs should ask routinely about practices of menstrual regulation and pregnancy termination. In our opinion, this is an essential first step in normalizing the political discourse about abortion as a common method of fertility regulation and as a public health issue, and in collecting the necessary data on availability, prevalence, and risks.

CONCLUSIONS

Is there a perfect indicator of fertility regulation for monitoring reproductive (maternal) health in the MDGs? The answer is clearly, no. Our assessment of 3 pairs of process and outcome indicators leads to the following conclusions.

Contraceptive prevalence and TFRs are useful primarily as background information for analyzing levels of unmet need and unplanned births. By themselves, they do not capture women’s capacity to manage their fertility safely and effectively because they do not take reproductive preferences into account. Nor are the safety aspects apparent on the surface; rather, these manifest themselves in relation to women’s capacity to avoid a high-risk abortion or delivery in resource-poor settings with high MMRs. CPRs and TFRs also suffer as indicators from not having a logical target for monitoring purposes.

Indicators of the proportions of women having an unmet need for contraception and proportions of births that are unplanned have the strong advantage of taking women’s reproductive intentions at the time of the interview (unmet need) or at the time of a previous conception (unplanned births) into account. Both are subject to some measurement error, however, and to difficulties in interpreting underlying causes. Their hypothetical “zero” target—that of eliminating unmet need and unplanned births—is unrealistic for these reasons. Perhaps the most that could be accomplished is to reduce their aggregate levels to under 10% and to eliminate within-country inequalities.

Access to safe services for the voluntary termination of a mistimed, unwanted, or otherwise problematic pregnancy represents the ultimate measure of safe and effective fertility regulation in accordance with women’s reproductive intentions. The indicators of unsafe abortion and abortion mortality are intended to monitor progress in this area. Both measures have a feasible zero target that can be reached in supportive policy and program environments. Reliable cross-country data that could otherwise be obtained are critically lacking, however, in large part because of political restrictions by some governments that block not only reproductive health advocacy and safe services, but also the necessary data collection and research. These restrictions need to be removed.

In summation, we believe that no single measure of contraceptive use (or nonuse) or of total fertility is sufficient without addressing both the high proportions of births in the developing world that women (both users and nonusers of contraception) say were unplanned at the time of conception and the percentages of induced abortions that are performed under unsafe conditions. Because the narrative of women’s attempts to manage their fertility safely and effectively—especially in conditions of poverty—inevitably embraces their experiences with unintended pregnancies and potentially life-threatening deliveries or clandestine abortions, these processes and outcomes should be monitored for the maternal health goal of the MDGs.

The major prerequisite for both the data collection and analysis and for the design of responsive policies and programs is the strengthening of political commitment at national and international levels to the ICPD Program of Action in all of its manifestations. Reinforced by the Fourth World Conference on Women held in Beijing in 1995, reaffirmed many times since, and supported in 2004 by a declaration of world leaders and by the adoption of a global reproductive health strategy by the World Health Assembly, the ICPD agenda for sexual and reproductive health and rights remains a vital force.46

At the 2005 World Summit, governments made the strongest commitment yet to reproductive health, recognizing that it is essential for the accomplishment of the MDGs.47 Within this context, there is a crucial need for bilateral and multilateral donors and foundations to fund the data collection and analysis recommended here, as well as the broader policy and programmatic changes and health system capacity–building required to fulfill the Cairo consensus on sexual and reproductive health and rights. The UN needs to play its part by adding a target—universal access to reproductive health—under the maternal health goal, along with a more inclusive set of indicators, so that countries can plan and monitor their progress more effectively.