Volume 25, Supplement, January 1999

Should Eliminating Unmet Need for Contraception Continue to Be a Program Priority?

By Anrudh Jain

Context: Estimates of unmet need for contraception derived from cross-sectional surveys are often taken as measures of a family planning program's success, and programs typically focus on eliminating unmet need. But that strategy may not be the most effective way of meeting women's total contraceptive needs.

Methods: Data on 1,093 women in Nor-Oriental del Marañón and Lima, Peru, who participated in the 1991–1992 Demographic and Health Survey and a 1994 follow-up survey are used to examine the family planning program's effectiveness in satisfying unmet need and averting unintended pregnancies.

Results: Although aggregate-level data suggest little effect of the program between surveys, individual-level data show that 72% of women who had an unmet need in 1991–1992 no longer had an unmet need in 1994. However, between surveys, 12% of the sample went from not having an unmet need to having an unmet need. Moreover, 20% of respondents had an unintended pregnancy between surveys: 32% of those who initially had an unmet need and 17% of those who did not. Applying the proportions who had an unintended pregnancy to the distribution of women according to their unmet need status in 1991–1992 shows that if the program had focused on eliminating unmet need, the proportion having unintended pregnancies would have been 17%; if the program had emphasized eliminating unintended pregnancies among women who initially had no unmet need, the proportion would have been only 6%.

Conclusions: Family planning programs may be more effective if they emphasize eliminating unintended pregnancies among women who are already practicing contraception than if they focus on persuading nonusers to become users.

International Family Planning Perspectives, 1999, 25(Supplement):S39–S43 & S49

Sample surveys carried out during the last three decades have demonstrated the existence of a significant gap between the apparent need for contraception and actual levels of method use in developing countries. The measurement of use has been fairly direct, but the measurement of need has gone through several modifications. In the 1970s, need was measured by the extent of women's awareness of family planning methods; a gap between women's knowledge, or awareness, and their actual practice of contraception was taken as an indication of unmet need for family planning. Currently, the measurement of need incorporates women's expressed intentions with regard to childbearing and, sometimes, with regard to contraceptive use. The existence of unmet need for family planning is typically indicated by married and fecund women who desire to space or limit their births but are not practicing contraception.

These measures of unmet need at times have been used as the main justification for the investment of public resources in family planning programs. Some experts have argued that in many developing countries, the elimination of unmet need for contraception would achieve--and therefore do away with the need for--national targets for contraceptive prevalence or fertility levels.1 An analysis of Demographic and Health Survey (DHS) data from 27 countries suggests that a realistic estimate of the satisfaction of unmet need could be expected to decrease fertility by anywhere from 7% in Ghana to 26% in Kenya and Madagascar.2

How can the current unmet need for contraception be satisfied? Some researchers maintain that the existence of unmet need is due mainly to a lack of access to service facilities.3 Others argue that the causes of unmet need include "lack of knowledge, fear of side effects, and social and familial disapproval."4 A better understanding of the causes of unmet need is essential to modify service delivery programs and to assess their contribution toward satisfying the existing unmet need.5 At the same time, it is also important to broaden the definition of unmet need--for example, by including unmarried women.6

The existence of an unmet need for contraception is taken as an indicator of deficiencies in accessibility7 and quality8 of family planning programs. There has been an implicit assumption in the literature that unmet need can be greatly reduced through improvements in these program characteristics. It is therefore logical to ask whether unmet need can be used as an indicator of program performance. Specifically, can a decline (or an increase) in the unmet need estimated from cross-sectional surveys in a country be used to assess the success (or failure) of the family planning program?

Westoff and Bankole pointed out that "change in unmet need is a function of changes in contraceptive practice and changes in desire to regulate fertility."9 Thus, unmet need at the aggregate level will decline if the proportion of women using contraceptives increases faster than the proportion who wish to regulate childbearing; unmet need will increase if the need for fertility regulation increases faster than contraceptive prevalence.

Estimates of unmet need are routinely available from cross-sectional surveys such as those conducted under the DHS program. An analysis of eight countries in which two DHS surveys have been conducted found that unmet need has declined by 2–3 percentage points in six countries (Colombia, the Dominican Republic, Egypt, Ghana, Kenya and Morocco) and by 12 points in the other two (Bolivia and Peru).10 These findings might imply that the family planning programs in the first six countries have been somewhat successful and that the programs in Bolivia and Peru have been remarkably successful; but such a conclusion may not be accurate, because we know that the program in Kenya has been more successful than, for example, Ghana's program in terms of an increase in contraceptive prevalence or a decline in fertility.

Using cross-sectional surveys to estimate unmet need is problematic for several reasons. The magnitude of an increase in unmet need suggested by such an approach would indicate failure of a country's family planning program only if it is assumed that program activities (e.g., information, education and communication) have no effect on women's stated desire to regulate their fertility. Another problem is that a lack of decline in unmet need at the aggregate level does not necessarily mean that unmet need has not been satisfied at the individual level. Moreover, satisfaction of unmet need for contraception at one point in time does not necessarily imply that women subsequently do not have unintended pregnancies.

Finally, an emphasis on meeting unmet need implies that a program's focus is on encouraging women who are not practicing contraception to begin doing so. In many programs, though, method discontinuation is high, and many women who stop using one method have an unintended pregnancy before they get a chance to switch to another. Moreover, some women who are practicing contraception may be at risk of pregnancy because they use their method inconsistently or incorrectly. Therefore, a program too closely focused on unmet need might miss the opportunity to assist women who have already started to practice contraception.

Thus, this article addresses two important questions related to unmet need. First, can the success (or failure) of family planning programs be evaluated in terms of a decrease (or an increase) in unmet need estimated from cross-sectional surveys? Second, is a sole emphasis on the satisfaction of unmet need an appropriate strategy for programs?

To examine these issues, I analyze data on 1,093 married women who were interviewed in both the 1991–1992 DHS in Peru and a 1994 follow-up survey conducted in two regions, Nor-Oriental del Marañón and Lima. The original sample from these two regions in the DHS included 1,850 women, and interviewers conducting the follow-up survey were able to locate 1,370 women (74% of the original sample), of whom 63 were excluded because of coding errors. Of the remaining 1,307 women, 1,093 (84%) were believed to have been interviewed in 1991–1992 because the year of birth they provided in the two surveys was the same or differed by no more than one year, and the number of children ever born reported in the follow-up survey was equal to or (allowing for mortality) one fewer than the number given in the original survey plus the number of births reported for the interval between surveys.*

Unmet Need and Program Performance
Aggregate-level data from Peru reveal that the proportion of women who wanted to have no more children changed very little between the original survey (68%) and the 1994 follow-up (71%), and the proportion using a contraceptive method remained stable (70%), although there was a slight shift toward modern methods (Table 1). Small declines occurred in the proportion of women with an unmet need for any method (from 18% to 17%) and for a modern method (39% to 36%). These findings suggest that the family planning program in Peru had very little effect on contraceptive prevalence or unmet need between 1991 and 1994.

Table 1. Percentage of married women, by contraceptive-related measures, 1991–1992 Demographic and Health Survey and 1994 follow-up survey, Nor-Oriental and Lima, Peru (N=1,062)
Measure 1991–1992 1994
Want no more children 68.2 71.0
Use a method
69.7 69.7
40.2 44.5
29.5 25.2
Have an unmet need
For a modern method*
38.9 35.8
For any method
18.1 16.9
*Based on women who wanted no more children but were not using a modern method of contraception, some of whom may have been using a traditional method.†Based on women who wanted no more children but were not using any contraceptive method (modern or traditional). Note: Thirty-one women whose reproductive preference or contraceptive status in 1994 was not known are excluded.

However, such a conclusion would be misleading, as may be seen by comparing changes in individual women's need over time. For example, 18% of the original sample had an unmet need for any method in 1991–1992, but only 5% of these women still had an unmet need at the time of the follow-up survey, while 13% no longer did (Table 2). In other words, among those with an unmet need at the earlier date, 72% had had their need satisfied by the time they were reinterviewed: Twenty-three percent had changed their reproductive intentions and decided to have more children, and 49% had begun practicing contraception (not shown). The shift in contraceptive behavior may be due in part to improvements in contraceptive services, and therefore would imply that the family planning program has been quite successful in satisfying the unmet need of individual women.

Table 2. Percentage distribution of married women, by unmet need status at initial and follow-up surveys, according to type of method
Type of method and status in 1991–1992 Status in 1994
Total Unmet need No need
Any method 100.0 16.9 83.1
Unmet need 18.1 5.0 13.1
No need 81.9 11.9 70.0
Modern method 100.0 35.8 64.2
Unmet need 38.9 22.3 16.6
No need 61.1 13.5 47.6
Note: See Table 1 for definitions of unmet need for a modern method and for any method of contraception.

Similarly, 39% of women interviewed in 1991–1992 had an unmet need for a modern method, but 17% had had their need met by 1994. Stated differently, 43% of those with an unmet need in 1991–1992 no longer had an unmet need in 1994--21% because they had changed their reproductive desires, and 22% because they had begun using a modern method. However, 57% of unmet need remained unsatisfied between the two surveys. Thus, the satisfaction of unmet need for any method or a modern method at the individual level could be used as an indicator of program performance.

Unmet Need as a Program Focus
In the Peru survey, 82% of women did not have an unmet need for any method of contraception in 1991–1992, 32% because they wanted additional children and 50% because they were using a method. However, 12% of all women had no unmet when originally interviewed but had an unmet need at the time of the follow-up survey (Table 2). Some of these women shifted from wanting more children to wanting no more (a logical progression), and some discontinued contraceptive use. In terms of modern methods, 61% did not have an unmet need in 1991–1992, including 14% who had an unmet need for a modern contraceptive method in 1994.

Thus, women switch between categories of unmet need, and a program strategy in 1991–1992 focused only on meeting the current unmet need would have ignored women who, for example, did not have an unmet need at that time but who subsequently experienced unintended childbearing. The data show that 17% of women who had no unmet need when originally interviewed had an unintended pregnancy between surveys (Table 3), as did 32% of those who had an unmet need in 1991–1992.

Table 3. Percentage distribution of all married women and percentage who had at least one unintended pregnancy or unwanted birth between the 1991–1992 DHS and the 1994 follow-up survey--all by their unmet need status when originally interviewed
Need status N % distribution Unintended pregnancy Unwanted birth
All 1,093 100.0 19.6 10.2
Have unmet need 199 18.2 31.6 26.1
Pregnant* 54 4.9 33.3 25.9
Use no method 145 13.3 31.0 26.2
Have no unmet need 894 81.8 16.9 6.7
Use a method 545 49.9 16.2 11.0
228 20.9 24.6 17.1
317 29.0 10.1 6.6
Want more children 349 31.9 18.1 0.0
*Women who were pregnant in 1991–1992 are included if they said that they want no more children after the current pregnancy. Note: These data are based on the questions about fertility intentions from the original survey and reports of fertility experience as recorded in the follow-up survey. The minimum interval between the surveys was 29 months; thus, pregnancies that began within 29 months after the original interview are included in the computation, and pregnancies that began later are excluded. An unintended pregnancy was one that was mistimed (i.e., the woman wanted to conceive, but not at that time) or unwanted (i.e., the woman did not wish to have any more children); the category includes both pregnancies that ended in live births and those that did not. An unwanted birth was a live birth following an unwanted pregnancy.

Among contraceptive users, the likelihood of having an unintended pregnancy between surveys depended upon the method women used at the beginning of the period. Some 10% of women who wanted no more children and were using a modern method in 1991–1992 had at least one unintended pregnancy by 1994, compared with 25% of those who wanted no more children and were using a traditional method and 31% of those who wanted no more children and were not practicing contraception.

Data on unwanted births reflect a similar pattern: Seven percent of women who did not have an unmet need at the time of the first survey and 26% of those who did have an unmet need had had an unwanted birth by the time they were reinterviewed. And women who were using a modern method at the beginning of the period were less likely to have an unwanted birth (7%) than were those who were using either a traditional method (17%) or no method (26%).

Using a simulation model that contrasted results of a strategy involving high levels of contraceptive acceptance and low rates of continuation with one involving low acceptance and high continuation, an earlier study demonstrated that recruiting a small number of acceptors per year and taking good care of them is a better strategy for family planning programs than trying to recruit a large number of acceptors whose needs cannot be met.11 The data from Peru illustrate a similar point.

As can be seen in Table 3, 18% of women in the sample had an unmet need in 1991–1992; 82% had no unmet need--50% because they were practicing contraception and 32% because they wanted more children. To eliminate unintended pregnancies, the program strategy in 1991–1992 could have focused either upon helping women in the first category to initiate contraceptive use or upon helping women in the other two groups by providing them with accurate information on the way to use their method, the possibility of switching methods or the sources of supply. What would have happened under each of these scenarios?

The overall proportion who had an unintended pregnancy (20%) is calculated by multiplying the proportion of women in each of the three groups--those who had an unmet need, those who were using contraceptives and those who wanted more children--by the percentage of each group who had an unintended pregnancy (from Table 3) and adding the results.§ When the calculation is repeated, assuming that the program succeeded in eliminating unmet need so that these women were exposed to the same rate of unintended pregnancy as women using a method, the proportion declines from 20% to 17% (Table 4). Alternatively, if the calculation assumes that the program eliminated unintended pregnancies among women with no unmet need so that both those who use a method and those who want more children experience zero unintended pregnancies, the overall proportion of unintended pregnancies declines to 6%.

Table 4. Percentage of women who had an unintended pregnancy or unwanted birth between surveys and percentage who would be expected to have each of these outcomes under two alternative scenarios
Scenario Unintended pregnancy Unwanted birth
Actual 19.6 10.2
Unmet need fully satisfied 16.8 7.5
Unintended pregnancies eliminated* 5.8 4.8
*Among those without unmet need. Note: For definitions of unintended pregnancy and unwanted birth, see Table 3.

The difference between the implied effects of the two strategies diminishes if we look only at unintended pregnancies resulting in an unwanted birth. The first strategy would result in a reduction of 2.7 percentage points in the occurrence of an unwanted birth, while the second strategy would result in a reduction of 5.4 percentage points. Thus, the second strategy turns out to be more effective than the first in Peru.**

The relative effectiveness of these two strategies in a country will depend upon the distribution of women according to their unmet need status. The second strategy, however, turns out to be as effective as the first even under implausible conditions. For example, if the proportion of women in the Peru sample having an unmet need and the proportion using a method in 1991–1992 had been reversed (i.e., 50% had an unmet need, 18% wanted no more children and used contraceptives, and the remaining 32% wanted more children), the overall proportion having an unintended pregnancy would have been 25%. Under a successful strategy to eliminate unmet need, the proportion would have declined to 17%, while if the focus was to eliminate unintended pregnancies among women with no unmet need, it would have fallen to 16%. Ignoring the small difference between the two, it can be said that the second strategy would be as effective as the first strategy.

It is possible to find a situation in which the first strategy may turn out to be superior to the second. However, in almost all situations, the second strategy will significantly reduce the overall rate of unintended pregnancies. Moreover, programmatically, the second strategy may be easier to implement than the first, because women who have started to practice contraception have already overcome some of the hurdles (e.g., lack of availability) associated with the initiation of contraceptive use.

As an indicator of program success, unmet need has an advantage over such indicators as the contraceptive prevalence rate and total fertility rate because it incorporates a woman's expressed desire to regulate her fertility, whereas these other measures do not. As the data from Peru reveal, however, while a decline in unmet need at the aggregate probably reflects improvements in a country's service delivery program, an increase or a lack of decline does not necessarily imply program failure. (Moreover, in cross-country comparisons, the same magnitude of unmet need does not imply the same degree of effectiveness.)

The extent of change in unmet need at the aggregate level therefore is not an appropriate indicator of the success or failure of family planning programs. A better indicator is the satisfaction of unmet need at the individual level. However, since this measure ignores the experiences of women who are reported to have no unmet need, a still better indicator is the reduction in unintended pregnancies (or unwanted births) at the individual level.

The Peruvian data also provide valuable information for addressing the question of program focus. While programs cannot be expected to reduce either the level of unmet need or the level of unintended pregnancies to zero, a sole focus on the reduction of unmet need--i.e., on getting women to initiate contraceptive use--is not the best program strategy. Programs must also pay attention to reducing unintended pregnancies among those who want to space or limit their childbearing and who have already started to practice contraception.

In Peru, the occurrence between surveys of an unintended pregnancy (or an unwanted birth) among women who had no unmet need because they were using a method when initially interviewed reflects one of three possibilities: They were using a method consistently, according to accurate instructions, and they had a method failure; they were using a method inconsistently or according to inaccurate instructions; or they terminated the use of a method and conceived before they had a chance to switch to another method.

How can the family planning program help such women to avoid unintended pregnancies? The contraceptive failure rate will likely be quite low among most women using modern methods of contraception, assuming that they receive and follow accurate instructions about how to use their method; if this is the case, the program can do very little more.

The bulk of unintended pregnancies, however, are likely to occur among women in the second and third groups. A program can certainly help these women to reduce the likelihood of an unintended pregnancy, by providing accurate information about how to use the method selected (including traditional methods, like periodic abstinence), informing women about the possibility of switching methods whenever their needs change and informing them about alternative sources of supply. However, data collected from a number of developing countries show the information exchange component of client-provider transactions to be generally very weak.12

Strictly speaking, women in the third group have an unmet need just prior to becoming pregnant. By providing adequate information, both at the time women initiate use and at subsequent contacts, a program might enable women to reduce the period of nonuse between methods and thus reduce the likelihood of unintended pregnancy. Alternatively, a program could devise a strategy to contact women soon after they terminate the use of a method but before they become pregnant. This implies the need for some algorithm to be used by fieldworkers (including community-based distribution workers).

Figure 1 illustrates one such algorithm, which is based on identifying the individual woman's current pregnancy status, desire for an additional birth, desired timing of the next birth, contraceptive use and satisfaction with her current method. This algorithm assumes that the fieldworker is capable of providing minimum information and that his or her basic job is to refer women to appropriate fixed services; it can easily be adapted to situations in which the fieldworker also provides services (supply or resupply) for such methods as the pill, condoms or injectables.

Figure 1. Scheme by which a fieldworker may prioritize services on the basis of a woman's answers to questions about her reproductive health needs
Figure 1

According to this strategy, women who say they are pregnant would receive information about managing their pregnancy safely, including the availability of prenatal care, delivery and abortion services, and (in subsequent visits) immunization, breastfeeding and contraceptive methods. They would also be given an opportunity to involve their partner if they choose to do so. For nonpregnant women, priorities for follow-up visits can be assigned with a view to reducing unintended pregnancies.

Follow-up of women who are using a method to space or limit their childbearing but are not happy with it should be the program's first priority. These women will move to the unmet need category and may become pregnant after terminating the use of their current method, unless they switch to another, more appropriate method. They need information about the possibility of switching to another method.

Women who want to space or limit their childbearing but are not using any method should be the second priority for a subsequent visit by the fieldworker. Most likely, they need information about where to go for services and what to expect once they get to a service facility and begin using a method.

How should fieldworkers handle women who want another child soon? This group of women may need help if they have difficulty conceiving, so fieldworkers should provide information about infertility services. In case these women change their mind, fieldworkers should also provide them with information about contraceptive services. But should fieldworkers try to persuade them to change their reproductive intentions? This is one of the most controversial issues in family planning, and the answer depends upon whether a country is trying to implement the client-oriented perspective of the International Conference on Population and Development, held in Cairo in 1994, and upon whether it has already met the needs of women in the other groups.

I have discussed this algorithm in a number of settings with different types of workers and professionals. People who are close to service delivery and understand the constraints on fieldworkers' time and capability usually like this type of scheme and see no problems in adopting it. However, the further away one gets from the field issues, the more difficult it becomes for people to accept this line of thinking.

Some social scientists, donors and government officials are not convinced that women who want to have their next child soon should be allowed to do so. They see an urgent need to persuade these women to change their reproductive intentions. Some argue that, at the least, women who are at high risk of having a negative pregnancy outcome (because they are very old or very young, they have had many children or their previous birth was very recent) should be counseled against trying to conceive again.13 Advocates of this "high-risk approach" contend that without such counseling, these women cannot make an informed choice about whether to have another child, and it is unethical for programs not to provide counseling.

Indeed, many programs in the past focused on training their fieldworkers to counsel or to motivate women not to have another birth. While focusing on motivational efforts, these programs often neglected the needs of women in the first- and second-priority groups. However, a continuation of such an approach is not consistent with the recommendations adopted at the Cairo conference. For example, the conference's Programme of Action states that "the aim of family planning programmes must be to enable couples and individuals to decide freely and responsibly the number and spacing of their children and to have the information and means to do so."14

Family planning programs that are trying to incorporate a client-oriented perspective and implement the Cairo recommendations cannot continue to focus on persuading these women to change their reproductive intentions. The evidence from Peru suggests that, indeed, the fieldworkers should assign a higher priority to helping women in the first two groups to avoid unintended pregnancies. However, if a choice has to be made, the fieldworkers should give higher preference to those who have already initiated contraceptive use. How many households a fieldworker can visit, how often a particular household should be visited and what type of information and services fieldworkers should provide are some of the issues that need to be resolved through operations research within a particular country context.

1. Sinding S, Ross J and Rosenfield A, Seeking common ground: unmet need and demographic goals, International Family Planning Perspectives, 1994, 20(1):23–27.

2. Westoff CF and Bankole A, The potential demographic significance of unmet need, International Family Planning Perspectives, 1996, 22(1):16–20.

3. Robey B, Rutstein SO and Morris L, The fertility decline in developing countries, Scientific American, 1993, 269(6):60–67; and Ross J, The question of access, Studies in Family Planning, 1995, 26(4):241–242.

4. Bongaarts J and Bruce J, The causes of unmet need for contraception and the social content of services, Studies in Family Planning, 1995, 26(2):57–75.

5. Casterline J, Perez AE and Biddlecom A, Factors underlying unmet need for family planning in the Philippines, Studies in Family Planning, 1997, 28(3):173–191.

6. Dixon-Mueller R and Germain A, Stalking the elusive "unmet need" for family planning, Studies in Family Planning, 1992, 23(5):330–335.

7. Ross J, 1995, op. cit. (see reference 3).

8. Bongaarts J and Bruce J, 1995, op. cit. (see reference 4).

9. Westoff CF and Bankole A, Unmet Need: 1990–1994, DHS Comparative Studies, Calverton, MD, USA: Macro International, 1995, No. 16.

10. Ibid.

11. Jain A, Fertility reduction and the quality of family planning services, Studies in Family Planning, 1989, 20(1):1–16.

12. See, for example, Mensch B, Miller RA and Miller VC, Focusing on quality with "situation analysis,"IPPF Challenges, 1994, No. 2, pp. 17–21.

13. DeGraff DS and de Silva V, A new perspective on the definition and measurement of unmet need for contraception, International Family Planning Perspectives, 1996, 22(4):140–147.

14. Department for Economic and Social Information and Policy Analysis, ICPD Programme of Action, New York: United Nations, 1994, p. 32.

*For further details about the follow-up survey, loss to follow-up and creation of the matched sample, see:
Mensch B et al., Avoiding unintended pregnancy in Peru: does the quality of family planning services matter? International Family Planning Perspectives, 1997, 23(1):21–27.

Strictly speaking, these data do not represent separate cross-sectional estimates, because the surveys involved the same women, interviewed three years apart, and because younger women were more likely than older respondents to be lost to follow-up. However, since unmet need at young ages is close to zero, these comparisons are not far from reality.

These figures are based on the conventional definition of unmet need: the proportion of women who want no more children but are not using any family planning method. While this definition does not provide the most accurate estimate, it is simple to calculate and to understand. Modifications of the definition are unlikely to affect the conclusions.

§For example, the overall proportion is equal to (0.182x31.6)+(0.499x16.2)+(0.319x18.1), or 19.6%.

**A similar point can be made for Morocco, using data from a panel study of 1,682 women interviewed in 1992 and 1995. (Source: Westoff CF and Bankole A, The time dynamics of unmet need: an example from Morocco, International Family Planning Perspectives, 1998, 24(1):12–14 & 24.) Whereas 16% of these women had an unwanted birth between surveys, the proportion would have been 11% if the first strategy were employed and 8% under the second strategy.

Anrudh Jain is senior director of policy and regional programs, International Programs division, Population Council, New York. Funding for data collected through the follow-up survey was provided by the U.S. Agency for International Development through the Evaluation Project, University of North Carolina, Chapel Hill, NC, USA. The author thanks Judith Bruce, John Casterline and Cheikh Mbacke for their comments on an early draft of this article.

© copyright 1999, The Alan Guttmacher Institute.

About AGI | Home | Search | What's New
Feedback | Keep Me Informed | Support