In 1972, based on analysis of women's responses to three KAP surveys in Taiwan, Ronald Freedman and colleagues first identified a specific group of women who might be expected to adopt contraception—even without changing their desired family size—because they said that they wanted to have no more children but were not using contraception (74). In 1974 Freedman and Lolagene Coombs for the first time used survey data to identify the size of this group in several countries, and they found it to be substantial. They coined the term "discrepant behavior" to describe the status of such women (73). Similar evidence of "discrepant behavior" came from surveys of young people in the United States in the early 1970s, where Leo Morris found "a significant gap" between the need for family planning and its use (133).

One of the first published uses of the term "unmet need" appeared in 1977, when Bruce Stokes, citing both the evidence from KAP studies in developing countries and from fertility surveys in the US, wrote that "in disparate ways, the number of ill-timed pregnancies and widespread reliance on abortion among all social classes and groups signal an unmet need for contraception" (205, 206).

Several large-scale survey programs, each with support from the United States Agency for International Development (USAID) and other donors, have helped develop an operational definition of unmet need (102, 115, 135, 174).

The World Fertility Survey (WFS). The WFS, conducted from 1972 to 1984, was first to report extensively about unmet need. The concept was so important to researchers that, when data first became available, unmet need was the first topic analyzed (34). In 1978, based on WFS data from five Asian countries, Charles Westoff published the first comparative estimates of unmet need for limiting births (215, 235). The WFS questionnaire did not ask women about their desire to space births. Also, at that time Westoff excluded pregnant and amenorrheic women because they did not currently need contraception (235).

In 1981 Westoff and Anne Pebley, using WFS data from 18 countries, showed that different definitions of unmet need produced widely differing estimates of the size of the unmet need group (239, 240). Also, they recommended that the unmet need concept be extended to cover desire to space births as soon as the data could be collected (239).

Contraceptive Prevalence Surveys (CPS). The CPS, conducted from the mid-1970s to 1984, made possible further refinement and measurement. The CPS added questions about women's interest in postponing, or spacing, next births. Thus it became possible to calculate unmet need for spacing births as well as for limiting births, helping to distinguish potential interest in temporary methods from that for permanent and long-term methods. In 1981 John Anderson and Leo Morris used the new CPS data to measure the percentage of women of reproductive age who are "exposed to the risk of unintended pregnancy and are not using contraception" in five Latin American countries (9).

In 1982 Dorothy Nortman raised a new point about defining and measuring unmet need (142). She argued that women who were pregnant, breastfeeding, or amenorrheic should be included in the definition of unmet need because they would soon need contraception again. Nortman and Gary Lewis developed a model that estimated unmet need for contraception, not just at the moment of the survey, but over the year following the survey (143). Other researchers used the CPS data to propose variations of the standard measure. For example, in 1985 Bryan Boulier reasoned that, if unmet need is to measure the number of women who would benefit from modern contraception, then users of traditional, inefficient methods should be counted as having an unmet need (27).

The Demographic and Health Surveys (DHS). The DHS have further improved measurement of unmet need. The DHS asks pregnant women whether their current pregnancies were intentional, mistimed, or unwanted and also whether they were using contraception at the time of conception. Also, the DHS questionnaire asks women directly about postpartum amenorrhea, thus avoiding the necessity of using breastfeeding as a proxy, as in past surveys (234). This approach made it possible to classify some pregnant women as having an unmet need for family planning and others, not. Thus Westoff revised the standard definition of unmet need to include pregnant or amenorrheic women whose pregnancies were mistimed or unwanted (231, 234).

Family Planning/Reproductive Health Surveys (FP/RHS). Since 1985, on a more limited scale than the DHS, a number of FP/RHS have been conducted, primarily in Latin America and the Caribbean, with technical assistance from the US Centers for Disease Control and Prevention. These surveys provide estimates of unmet need, including among unmarried women (174).

By adopting strategies to address unmet need effectively, many countries that now have demographic goals could replace them with objectives based on "the stated desires of the people served" and at the same time have equal or greater impact on contraceptive use and fertility levels, Steven Sinding and colleagues have observed. They have estimated that, if all married women with an unmet need were to use contraception, in some countries contraceptive prevalence would rise even above program goals set ac-cording to demographic criteria (201).

It is unrealistic to assume that family planning programs could meet all unmet need, however. Among married women in the unmet need group, some appear to have little interest in contraception and do not intend to use it (see Intention to use contraception in Chapter 3.2) (230, 237). Others apparently face little risk of pregnancy because they are older and have sexual relations infrequently (238). Still, meeting much of unmet need, if not all, would help millions of women avoid unintended pregnancies and probably also would prevent many abortions.

Just how much would fertility fall if programs effectively addressed unmet need? Estimates range widely, depending on the measures used and the assumptions made. The most recent estimates, prepared by Westoff and Bankole, examine changes in fertility under five different sets of assumptions about how much unmet need family planning programs can meet (236, 237, ). They consider the following most plausible:

Among married women with an unmet need who intend to use contraception, 90% of potential limiters and 80% of potential spacers would follow through on their intention and adopt contraception. Among those who do not intend to use contraception because they believe that they face little risk of becoming pregnant, none would adopt contraception, while half of the other women with unmet need who do not now intend to use contraception nevertheless would do so.

Meeting unmet need to this extent would reduce fertility in the developing world by an average of 18%. This estimate is based on a concept of unmet need developed by John Bongaarts in 1991, in which the level of unmet need in a country equals the amount of additional contraceptive use needed to achieve women's fertility preferences (22). Westoff has acknowledged that this concept is appropriate for estimating changes in fertility due to reduction of unmet need (23, 232), while the standard formulation of unmet need is more appropriate for program planning by family planning managers (237, ).

Such reductions would help many countries approach replacement-level fertility of 2.1 births per woman. In Turkey, for example, meeting unmet need to this extent would cause fertility to drop to the replacement level. In Bangladesh and Indonesia, fertility would fall more than three-quarters of the way to the replacement level. In Colombia, the Dominican Republic, and Peru, the decline would be about half the distance to replacement-level fertility. Even in sub-Saharan Africa, where fertility is high, meeting unmet need to this extent would reduce fertility by an average of about one birth per woman and bring most countries 20% to 30% closer to replacement-level fertility (236).

Thus interest has grown in conducting more small-scale, qualitative studies that use in-depth interviews and focus-group discussions to reveal attitudes, interests, and values that help to explain unmet need (169). In general, public health programs increasingly are using such studies to provide psychosocial data—or "psychographics," a term borrowed from advertising and marketing research (26, 68)—in order to add "color and depth" to data from large-scale surveys (153).

"Main" Reasons Identified by the DHS

Nearly two of every three nonintenders queried in 24 DHS surveys since 1990 give reasons that fit one of three main categories: lack of information; opposition to family planning; or apparent ambivalence about future childbearing. Other reasons include fear of side effects, little exposure to the risk of pregnancy, and unavailability of contraception (237, ). The mix of these reasons differs by whether unmet need is for limiting or spacing (see Limiting or spacing in Chapter 3.2 and Figure 7).

In contrast, the DHS questionnaire used from 1985 to 1990 asked women with an unmet need their main reason for not currently using contraception. Women queried about current use were more likely to cite husband's disapproval and health concerns than the women asked about their intentions. Women asked about current use also were less likely to disapprove of family planning or to give conflicting answers about childbearing desires (25).

Insights from Qualitative Studies

The Philippines. In Manila and several rural areas, women with unmet need were more likely than contraceptive users to think that the health risks of contraception outweigh the risks of pregnancy (37). Together, their fears of side effects and their husbands' fears explained much of the unmet need. Women with unmet need appeared less committed than contraceptive users to avoiding pregnancy. Also, they were less likely to think themselves at risk of pregnancy.

Nepal. In the Chitwan District, where access to services is better than in many other parts of Nepal, many women with unmet need said that they did not use contraception because they received or expected poor treatment at clinics, or they feared side effects of contraceptive use that would cost them time working (204). Many women with unmet need expressed concerns about their health and said that their husbands opposed family planning. Also, they were less likely than contraceptive users to have relatives or friends who were using contraception.

In Dang District, where family planning is not as accessible, some women were not using contraception because they did not know who could provide information and supplies (223). Others believed that they had to bear more children than they ideally would want because they expected some to die before they grew up. Also, women and men alike expressed reluctance to try contraception because they feared that other people would disapprove.

Guatemala. In a peri-urban neighborhood of Guatemala City, women said that uncertainties about the characteristics and safety of contraceptives, fear of side effects, and dissatisfaction with particular methods kept them from using family planning (10). Sometimes faced with sexual violence from their partners, many women deferred to the wishes of their partners despite their own preferences.

India. In two communities of Tamil Nadu, one peri-urban, the other rural, unmet need often resulted from the lack of contraceptive choices; sterilization was the only method available (165). Also, women said that their husbands, often fearing heath risks, discouraged them from using contraception altogether. Another study in Tamil Nadu found that few women were even aware of methods other than sterilization (54). Most wives felt that sterilization entailed too many risks, while most husbands were opposed to sterilization.

Kenya. In rural Nyanza Province women's decisions to use family planning were taken tentatively, following exploratory conversations with friends and then "more strategic conversations" with contraceptive users (184). Most women expressed concerns about side effects of contraceptives. Also, their husbands had considerable power over them. After beginning contraceptive use, women remained ready to discontinue should they change their minds, experience side effects, or face their husbands' opposition. Some used contraception secretly "to test the waters," hoping they might gradually convince their husbands or get some rest before their next pregnancy.

• Know about family planning methods;
• Find that contraceptives are easy to obtain;
• Think that their husbands and others approve of family planning.

Availability of contraception. Women with unmet need are less likely to think that family planning is readily accessible, according to survey responses. For example, in Morocco 42% of women with unmet need say that it is difficult to gain access to contraceptive services compared with only 24% of contraceptive users.

Attitudes about husbands' and social approval. Women with unmet need are less likely than contraceptive users to report that their husbands, relatives, and friends approve of family planning. For example, in Kenya only 58% of wives with unmet need report that their husbands approve of family planning compared with 86% of those using contraception. Contraceptive users and women with unmet need differ little in their own approval of family planning, however.

Bangladesh and Pakistan. Twenty years ago, Bangladesh and Pakistan were similar in women's stated reproductive preferences, levels of unmet need, and contraceptive prevalence. Before 1971, in fact, the two comprised one country, with a common population policy and a single family planning program. By the early 1990s, however, the level of unmet need in Bangladesh was 18%, while in Pakistan the level was 32%. Contraceptive prevalence in Bangladesh was 45% compared with 12% in Pakistan (see Table 2).

Since 1971, as John Cleland has pointed out, the two countries have followed quite different demographic paths. In Bangladesh successive surveys have reported rapid increases in contraceptive use and declines in unmet need. Surveys in Pakistan repeatedly "have attested to a huge latent demand for fertility regulation" (41).

Differences in economic development do not explain these divergent paths. Pakistan ranks above Bangladesh in most development indicators and thus—other things being equal—might be expected to have more widespread use of family planning. But other things have not been equal. In particular, the national family planning effort has been much stronger in Bangladesh than in Pakistan (122, 181). As Freedman has noted, Pakistan spends much less on family planning than Bangladesh spends, whether measured per capita or as a percentage of gross national product (GNP), and Pakistan also gets little for its money because the program has been poorly run (70). Furthermore, the program has lacked high-level political support.

In contrast, in Bangladesh national and community leaders have supported family planning, attracted substantial external funding because of their commitment (173), and helped contraceptive information and services expand across the country. Family planning increasingly is becoming a community norm.

Indonesia, Thailand, and the Philippines. Comparing Indonesia and Thailand with the Philippines reveals much the same story. In the 1970s the three countries had similar levels of unmet need, while contraceptive prevalence was higher in the Philippines than in either Indonesia or Thailand. By the 1990s contraceptive use had risen substantially in Indonesia and Thailand, far surpassing the level in the Philippines. In Indonesia and Thailand levels of unmet need for family planning are among the lowest in the developing world, at 14% and 11%, respectively (see Table 2). In the Philippines, in contrast, the level of unmet need, at 26%, is among the highest in the world.

"The explanation of this unexpected outcome lies well beyond the realm of statistical evidence," Cleland has observed, "but almost certainly involves the intertwined factors of religion and government policy" (41). In Indonesia and Thailand governments have supported and promoted family planning for the past 20 years. In the Philippines fluctuating government policies have made it difficult until recently to sustain a strong family planning program.

Based on these and other comparisons, Cleland has pointed to the importance of political leaders and other leaders in a country's transition from high to lower fertility. Concerted government action, he has noted, appears to overcome obstacles to contraceptive use even in poor, constrained circumstances, as in Bangladesh (41).

A matrix can be prepared for the national level and, where data are available, also for rural and urban sectors, for geographic regions that differ from national averages, or for the largest cities—wherever program strategies are likely to take different forms. Also, preparing separate matrices for spacing and limiting may prove useful, since these two groups typically have different contraceptive needs and interests in using contraception.

The matrix also identifies program components to reach the various groups. In Table 6, four different aspects of family planning programs are presented: information, education, and communication (IEC); mix of services and methods; delivery channels; and worker types. These are illustrative only and not intended as a complete list of all possible approaches or program components.

While countries differ in their data resources, the approach used in Table 6 can apply to many situations. Many of the data are rough estimates but probably close enough for general program planning. Estimates are shown for three illustrative countries—Kenya, Morocco, and the Philippines. For each country the matrix contains possible program responses for selected population subgroups. Individual women may fit more than one subgroup in the matrix. Also, certain subgroups can add some women to the "basic" unmet need group, since programs must work among the general population and must locate clients in a variety of ways and places.

Data Sources for Unmet Need Subgroups

Unmarried young people. In the developing world as a whole, the 15-19 age group, including both women and men, is close to 10% of the total population, as is the 20-24 age group. In each country the absolute number in the young age groups can be multiplied by the percentage who are unmarried in order to estimate the number of unmarried young people. Most of these people probably are not sexually active, but surveys such as the DHS often estimate the percentage who are. While unmarried women are not included in conventional estimates of unmet need, most unmarried women probably do not want to become pregnant. By pro-viding information and services to those in this group who are sexually active, family planning programs can help avoid many unintended pregnancies. (See Population Reports, Meeting the Needs of Young Adults, J-41, October 1995.)

Newlyweds. While few newlyweds immediately desire family planning, they are an easily identifiable group for whom family planning education is important. Without such education, many soon will have an unmet need. Furthermore, providing information about contraception at this time can help them overcome shyness about reproductive matters, promote spousal discussion of family planning, and encourage them to space their second birth at least two years after the first one. The number of marriages annually can be found in official statistics. An alternative estimate is the number of women arriving each year at the mean age at marriage.

Postpartum women. This group is a key audience for unmet need strategies (see Chapter 3.1). The number of births annually provides a basis for estimating its size. The number of births comes from the best estimates available of the total population size and the crude birthrate. This number can be adjusted by the proportion seen by trained personnel (or in established delivery facilities), an estimate that may come from the Ministry of Health and also from DHS. Some countries have additional information on the proportion seen prenatally and postnatally.

Postabortion cases. The numbers of abortions and menstrual regulation procedures can be estimated to the extent that these women are seen in established facilities either for legal procedures or for treatment of complications after unsafe procedures. Contraceptive services should be offered in either case. Official counts may indicate the number of women who can be reached through health care facilities. National estimates may be derived from an estimate of the abortion rate or ratio, which can be applied to the estimated numbers of women or pregnancies, respectively.

Dissatisfied users. Everywhere, some current users of each contraceptive method would prefer a different method. Dissatisfaction with a particular method leads some women to discontinue contraception even though they do not want to become pregnant. Estimates of the number of dissatisfied users can be based on survey data and other information about preferred methods, side effects, and problems with particular methods.

The Bangladesh study is intended to deepen understanding of unmet need and to help policy makers develop guidelines for allocating field workers' time among various activities (179). Beginning with the national estimate of 4.4 million women with unmet need, the study has made estimates for the rural and urban sectors and, separately, for Chittagong Division, the area where unmet need is highest as a percentage of the population (13).

The study used the CHAID computer software (see Software for analysis in Chapter 5.2) to identify women's characteristics that are most closely associated with unmet need and to identify audience segments. The study found that, among married women of reproductive age with unmet need, about half intended to use contraception within the next 12 months. The same percentage were nonusers who did not currently have unmet need but nonetheless intended to use contraception within 12 months. The absolute numbers were about equal for the two groups of intenders, as well. Contrary to expectations that almost all intenders would already have large families, the researchers found that about 40% of all intenders had no children or just one child (13).

The Gujarat study focuses on the state level to propose recommendations for consideration at the national level. The government of India canceled national acceptor targets as of April 1996 and is looking for new indicators of program performance. Adopting an unmet need approach might help programs establish more client-oriented approaches and choose meaningful indicators, given that about 31 million married women in India have unmet need.

Also, a panel study has been conducted to trace changes in unmet need between 1989 and 1995. In two districts researchers contacted 751 women who had been interviewed in the baseline 1989 study. They reinterviewed these women to compare women who had unmet need for limiting births with women who wanted another child. Preliminary findings indicate that about 60% of women with unmet need in 1989 were unable to avoid unwanted pregnancies during the subsequent 6-year period, and 30% had two births or more (179).

The Vietnam study is concerned particularly with the limited contraceptive mix currently offered: the IUD is virtually the only modern family planning method widely available. Not surprisingly, the level of IUD use is among the highest in the world, at one-third of all couples, but another one-sixth have stopped using the method and have only limited alternatives (179). Largely as a result, unmet need is estimated at about 35%, among the highest levels in the world (149). In addition to using survey and census data and official statistics, the researchers are interviewing 49 couples in depth (98 spouses, interviewed separately).

Broadening the method mix in Vietnam would go far to reducing levels of unmet need, but doing so requires changes in national policies and administrative procedures. A broader method mix also would require training for family planning officials, supervisors, and workers as well as more communication to the public about contraceptive method choices (179).

Other Studies In Nepal a national communication strategy, designed and implemented with technical assistance from Johns Hopkins Population Communication Services, is based on identifying women with unmet need and responding to their reasons for unmet need (85, 170, 207). Results of this project, which began implementation in 1995 and is planned for three years, will guide similar efforts elsewhere.

Other relevant research underway includes a multi-country study of factors underlying unmet need, being conducted in Ghana, Pakistan, and Zambia, following the initial study in the Philippines by Casterline and colleagues, with support from the Rockefeller Foundation. Also, the Demographic and Health Survey Project is conducting an in-depth study of unmet need in Egypt. The International Center for Research on Women (ICRW) is conducting similar research in India. Each of these studies seeks to understand the reasons for unmet need through analysis of focus-group discussions, in-depth interviews, and survey data (35).