(introduction...)

National family planning programs have been associated with notable increases in contraceptive use and consequent declines in fertility. Family planning programs predominantly emphasize voluntary acceptance of contraception, usually include an important informational and educational component, and focus especially on the needs of married women but often also reach others The "cafeteria" approach, meaning the provision of many types of contraceptives so that each person can choose the most appropriate, is often favored but often also honored in the breach. Oral contraceptives and condoms are the base of many programs' inventories, but some earlier programs relied and continue to rely heavily on methods, such as the intrauterine device (IUD), that are less prominent in programs that started later. Over time, such newer methods as injectables and implants have found their own niches. Overall, however, the main trend has been toward permanent methods: sterilization has become simpler and more demanded and now accounts for half of all contraceptive use.

How contraceptives are delivered to potential users varies. Building on efforts by private doctors and voluntary organizations, early programs (and most in their earlier phases) depended on clinics, which can offer a wide range of methods and ensure appropriate medical facilities. The limited number of clinics quickly poses a problem in a developing country; over time, various ways to bring contraceptives to people in their villages, homes, and work places have been devised. Many programs have diversified, adding such approaches as community-based distribution, mobile teams, home visiting, subsidized commercial sales, or employee programs. The clinical component in many programs receives relatively less emphasis, and program personnel have become more varied as nurses, midwives, pharmacists, and lay distributors have taken on larger roles.

Focusing primarily on contraceptive use, programs target the primary behavior change that causes fertility to decline in the transition from high, traditional, relatively stable fertility to low, modern levels at which couples are only replacing themselves. Fertility is also affected by other important behaviors (called "proximate determinants" by demographers). Of these, marriage delay in the course of modernization contributes to lower fertility, although to a lesser extent; breastfeeding tends to decline, raising fertility somewhat, because the menses return more quickly; and abortion has an important though highly variable effect. Individual programs may include interventions to affect marriage, breastfeeding, and abortion, but such interventions have tended to be limited. Nevertheless, programs have affected these other proximate determinants at least indirectly. In the long run, readily available contraception reduces resort to abortion. Trend data over 30 years for Hungary, for example, show abortion declining with increased contraceptive use, as do data from other settings, such as South Korea. Recent reports from Russia and Kazakhstan show declines in abortion in the 1990s as contraceptive services expanded.¹ However, in the early program, abortion may become more common if services cannot fill the demand to limit fertility.

¹But even at high levels of contraceptive prevalence, abortion does not entirely disappear, partly because of contraceptive failures. The legality or illegality of abortion has little influence on such trends (Cohen, 1997).

Experimental, country-specific, and cross-national evidence indicates that, by increasing contraceptive use, programs do contribute substantially to reducing fertility. This effect, however, is not universal. It depends on a program adopting a strategy responsive to whatever specific demand for contraception exists in the sociocultural setting, as well as on the staff mastering the basics of service delivery.

The Effect of Family Planning Programs

National programs were established beginning in the 1950s and multiplied most rapidly from the mid-1960s to the late 1970s when on average five new programs were established each year. In a number of instances, national programs expanded the work done by earlier, pioneering private programs, such as those supported by the International Planned Parenthood Federation. National programs naturally required government sanction and often also enjoyed substantial support and technical assistance in many phases of their operation from international donors.

Growth in contraceptive use after programs were established is illustrated in Figure 12, showing a range of experience and some notable spurts in contraceptive use that led to declines in fertility. These examples could be multiplied: Increases in contraceptive prevalence (the proportion of married women of reproductive age using contraception) and consequent fertility declines have occurred in every region of the developing world, from Peru to Mongolia, and from Iran to South Africa. But as Figure 12 illustrates, simply starting a program does not guarantee immediate increased contraceptive use. Contraceptive use did rise quickly in South Korea, where prevalence grew close to 3 percentage points a year - sufficient to reach low, replacement fertility within 25 years, essentially in a generation. But, in contrast, growth in contraceptive prevalence began late and proceeded haltingly in Ghana, where the increase was only 0.7 points a year. Even in such cases as South Korea, the credit that programs can fairly claim for fertility reduction has been a topic of some debate. Perhaps, some have argued, economic growth in Korea was so fast that fertility would have declined even in the absence of a program. Indeed, analysis suggests that programs cannot claim all the credit for fertility decline, but they can clearly claim some credit.

Figure 12 - Starting Date for Family Planning Program and Contraceptive Prevalence, Selected Countries

SOURCE: Based on World Bank (1993a).
^aDates indicate year of program start.

Field experiments usually indicate that well-done family planning programs are welcomed and make a difference. Researchers have looked at various elements of programs - the use of home visits, field-worker incentives, consistent supervision of field workers, pill prescription practices, postpartum education, etc. - to confirm a small but significant program effect on contraceptive acceptance or continuation (Bauman, 1994). Focusing on specific features of small-area programs, experiments can clearly demonstrate the contribution of program activities and exclude other explanations for changes in contraceptive use and fertility, in the process illustrating useful options for programs. However, such experiments do not address whether programs can have broader, nationwide effects.

This effect must be assessed against that of socioeconomic development. Improved standards of living should make contraceptives more affordable, as well as convincing many couples to opt for smaller families. Do programs then merely substitute for private provision of contraceptives - are they a natural but superfluous government response to consumer demand and consequent public pressure? Or do programs, through their promotional activities and the services they provide (much more promptly than market mechanisms would), play a leading role in reducing fertility?

Answering such questions and determining the historical contribution of large national programs is difficult. Comparisons are necessary across countries with different types of programs or with no programs. Analysts have generally relied on a reputational measure of "program effort," produced by rating 90 or so national programs with regard to the access they provide to contraceptive methods, their management effectiveness, their efforts at informing and educating people, and a number of other such dimensions, 30 in all. Analysis of program effort scores confirms that programs have been more active in more advanced developing countries. Socioeconomic development allows programs to operate more efficiently and, in leading to lower fertility preferences, to be more effective at providing contraceptive services. But relatively high levels of development, although helpful, do not appear to be essential. Bangladesh, one of the world's 20 poorest countries, has a program rated among the 10 best in the developing world and has seen a substantial decline in fertility over the last decade.

Multivariate statistical analysis suggests, in fact, that family planning programs do contribute independently to reduced fertility, even when the effects of socioeconomic development are accounted for. Cross-sectional regressions indicate that effective programs have a smaller effect than such factors as rising levels of female education but still reduce fertility - net of such other factors - by perhaps one-and-a-half births per woman. Skeptics put the reduction at only one birth instead but do not deny it exists (Pritchett, 1994a, 1994b). The reduction is in the region of 40 percent of the fertility decline in developing countries from the 1960s to the end of the 1980s (Bongaarts, forthcoming).

Cross-national statistical analysis, like experimental work, therefore indicates that family planning programs contribute to contraceptive use and lower fertility. But both types of studies agree that the contribution is contingent. Programs are not uniformly successful; their effects are influenced by the social context (level of development, cultural factors, political support) and depend on program performance (on the quality of program effort, the design of interventions, the adequacy of supervision of service providers, etc.). A brief overview of these factors affecting program effectiveness helps clarify when and why programs work.

Socioeconomic and Cultural Factors

Socioeconomic factors are not always favorable when national programs are launched. Does it follow that programs have to wait for incomes to grow, education levels to rise, and cities to come to dominate the countryside before they have any effect? Apparently not. No socioeconomic thresholds are evident in such data as those in Table 2. Fertility decline started in East Asia, in the 1960s and1970s, at income levels typical of low-income economies today. In Indonesia, in particular, GNP per capita was about US$250, which is below the mean level for low-income economies. Even lower income levels were typical of South Asian countries at the start of their transitions. Latin America was different, with a number of countries not starting transitions until reaching much higher income levels, up to US$1,500 per capita. Where transitions have started most recently, in sub-Saharan Africa, income levels at the start were intermediate between Asian and Latin American levels.

Table 2 - Socioeconomic Indicators at the Start of Fertility Transition, Selected Countries, and Comparative Aggregate Data

Region and Country		Start of Fertility Transitiona	GNP per Capitab	Infant Mortality Rate	Female Secondary Enrollment	Percent Urban	Start of Population Program
East Asia
	Indonesia	1975	253	109	15	19	1968
	Korea, Rep. of	1960	550	70	14	28	1961
	Philippines	1970	488	66	50	33	1970
	Thailand	1970	471	73	15	13	1970
South Asia
	Bangladesh	1975	138	138	11	9	1971
	India	1965	218	150	13	19	1965
	Pakistan	1985	304	113	8	30	1960
	Sri Lanka	1965	216	63	35	20	1965
Latin America
	Brazil	1965	889	104	16	50	1974
	Colombia	1965	676	86	16	54	1970
	Costa Rica	1965	1,109	72	25	38	1968
	Mexico	1975	1,504	64	28	63	1974
Sub-Saharan Africa
	Botswana	1980	721	63	22	15	1971
	Kenya	1980	358	83	16	16	1967
	Zimbabwe	1970	544	96	6	17	1968
Comparative data (1995)c
	Low-income economies (except China and India)		290	89	21	28
	Lower-middle income economies		1,670	41	62	56
	Upper-middle income economies		4,260	35	75	73

SOURCES: World Bank (1993, pp. 20-21). Comparative data from World Bank (1996, 1997b).

^aFertility transitions are dated from initial declines of at least 0.7 points in total fertility over a five-year period, following Bulatao and Elwan (1985).

^bConstant 1987 U.S. dollars. These and the other indicators are as of the transition date, except for the comparative data, which are given as of 1995.

^cFemale enrollment is as of 1993. The enrollment figure for upper-middle-income economies is the median across 13 countries.

The changes in personal aspirations and in the acceptability of family planning that trigger fertility transition appear to have occurred at many different socioeconomic levels. For the countries in Table 2, fertility transition started at infant mortality levels as high as 150 deaths per thousand or as low as 63²; where female secondary enrollment was only 6 percent or already 50 percent; and where anywhere from 9 percent to 63 percent of the population was urban. Fertility transition can therefore start with social indicators practically anywhere in the range typical of low-income or even lower-middle-income economies, with no specific levels triggering decline.

²But see Bulatao and Elwan (1985) for a possible mortality threshold (see also Bongaarts and Watkins, 1996).

Nevertheless, socioeconomic development does contribute to lower fertility. Cross-national comparisons suggest that the pace of decline, once it has started, is faster in more advanced developing countries (Bongaarts and Watkins, 1996), either because higher levels of development condition people to be more favorable to smaller families or because they allow programs to operate more efficiently. And once the transition is under way, contraceptive use appears to spread in a diffusion process that takes on a life of its own.

If low levels of socioeconomic development do not deter family planning programs, neither do cultural obstacles. Skeptics have cited one obstacle or another practically everywhere programs were launched. None appears to be an effective impediment. In East Asia the cultural barriers included Confucian traditions (in Taiwan and South Korea) that made the family central and gave household heads control over extended families, including the childbearing of their children. A strong preference for sons to carry on the family line has also often been cited. Political opposition of various sorts has existed to family planning, for instance among Islamic fundamentalists in Indonesia and the Catholic hierarchy in the Philippines. Such opposition was even more important in Latin America. Although the predominant Catholicism did not deter couples from voluntarily adopting contraception, the opposition of the church legitimized intellectual opposition to limiting population growth. In addition, the split in many Latin American countries between the dominant elite and the peasant masses slowed the spread of contraception. In South Asia, the barriers included such factors as traditional family structures, the subordinate social position of women, and continuing dependence on child labor. Similar barriers have been adduced for sub-Saharan Africa, where decisions on childbearing have been assumed to be controlled by older generations who consider children and grandchildren essential assets.

Yet in each of these settings, despite the cultural obstacles and despite varying and often low socioeconomic levels, some degree of interest in smaller families or demand for contraception appears to have existed. Near the start of fertility transition in South Korea, Thailand, and Indonesia, ideal family size was recorded at around four children, already below existing fertility. For a number of Latin American countries, substantial demand for contraception probably exceeded the capacity of early programs. Even in Bangladesh, substantial "latent demand" appears to have existed, with fertility preferences similar to those in other Asian countries at a similar stage of fertility transition (Cleland et al., 1994, p. 48). Sub-Saharan Africa initially presents a different picture, with large families still often highly prized, but in this region considerable interest exists in spacing births.

Program Strategies and Approaches

Though neither socioeconomic development nor cultural and institutional factors determine program success, sensitivity to such factors does seem to be a hallmark of successful programs. Where they have succeeded - and at least some successes have been recorded in every region of the developing world - programs have built on existing demand for contraception and have experimented to develop ways to address socioeconomic and cultural obstacles.

In East Asia, existing demand was exploited in pilot projects that demonstrated widespread desire for contraception, such as the large-scale Taichung project in Taiwan and the Jakarta pilot project in Indonesia. Close attention to the demand was evident in recurrent surveys in Taiwan to learn about attitudes toward family planning. Political support was a benefit of such demonstrations, although support also developed for different reasons, not always within the control of programs. Delivery systems for family planning eventually became massive and well organized in each country but not before extensive experimentation with various means of delivery. Thailand tested optimal delivery systems in the Potharam project; Korea experimented with urban programs in Sundong Gu and rural programs in Koyang. Eventually, "Mothers' Clubs" became an important element of the Korean program, used in rural areas to mobilize support for smaller families and to motivate women to adopt family planning while assisting with such concerns as income generation (Cho et al., 1982, pp. 129-131). An existing institutional capacity to mobilize peasants was also critical in Indonesia, where an aggressive program drew on a local government structure strengthened after the 1965 coup attempt (McNicoll, 1983, p. 86) and worked intensively with community groups.

Facing even stronger demand in Latin America coupled with barriers to government involvement, programs evolved in a different way. Demand for contraception was initially met by physicians; by commercial sales; and, increasingly, by private voluntary organizations, such as Profamilia in Colombia. Discovering substantial latent demand (Tamayo, 1989), Profamilia began with urban clinics and expanded after a few years into community-based rural services and also began offering sterilization. The government tolerated the provision of services, probably recognizing the widespread demand for them. Demonstration of this demand was critical in the eventual institution of government services, which eclipsed private services in Mexico. In Colombia, private services continue to be more important; in Brazil, the private organization Sociedade Bem-Estar Familiar no Brasil (BEMFAM) remains a major provider, partly as a contractor to state governments. Experimentation with delivery options has often involved the private sector. With large nationally coordinated programs coming late to the region, contraceptive growth has been slower in Latin America than in East Asia.

South Asia illustrates both initial obliviousness to demand and eventual responsiveness to it. The early spread of contraception was retarded by programs that were heavily bureaucratic and largely administered from the top, especially in Pakistan, where little attempt was made to gauge client demand, and little attention was paid to the needs of front-line staff. The Indian program is harder to characterize because of its diversity and some degree of control by individual states. In general, however, the Indian program also failed to capitalize on all the existing and potential demand for contraception and focused narrowly on sterilization. Targets were set high in the hierarchy, and officials at the top were largely out of touch with village life (Freedman, 1990, p. 39).

In contrast, the program in Sri Lanka, the only one in the region that showed some early success, was much more sensitive to client needs. The Sri Lanka program provided a wide range of methods, including such temporary methods as pills and IUDs, and used community-based distributors in rural areas. The Indian program, on the other hand, focused initially on the IUD and later largely on sterilization, avoiding community distribution of other methods, and experimented instead (to a much greater extent than in Sri Lanka) with incentives and controls on age at marriage, measures with theoretical appeal but limited practical attraction to individuals. The Indian program had more striking political support and better advertising, but the Sri Lankan program delivered the contraceptives. Much more similar to the Sri Lankan approach was the program in Bangladesh as it evolved after 1975. Assuming the existence of latent demand, it focused on mitigating the costs to individuals, both practical and psychic, of using contraception. Frequent contact by trained and caring workers with clients was emphasized, often in the clients' own homes to overcome cultural restrictions on women's mobility. As with earlier programs in East Asia, substantial experimentation with delivery alternatives has gone on in the Matlab area and through the succeeding Extension project (Cleland et al., 1994), with coordinated funding from many international donors.

In the sub-Saharan countries that have made the most progress in reducing fertility - Botswana, Zimbabwe, and Kenya - programs address existing demand for child spacing by providing temporary methods, mainly the pill. Delivery systems have been quite different, however. Botswana, with presumably stronger demand in a relatively more developed setting with a public health system that covers the country fairly extensively, has relied largely on health posts and health centers to provide contraceptives. Zimbabwe placed primary emphasis on community-based distribution to reach out to the rural population. Kenya has also emphasized outreach but has relied to a much greater extent than Zimbabwe on private voluntary organizations to complement public services.

The key to a successful program therefore appears to lie less in a favorable environment than in what the program does with the material it has. No socioeconomic setting, however impoverished, appears devoid of some demand for controlling fertility; even in the most favorable environments, some groups will have unmet need for contraception. Government programs, properly run and complemented where appropriate with private efforts, do appear capable of identifying and satisfying demand, although encouragement and substantial support from international donors has been virtually continual in these cases. Demand for controlling fertility can be fragile and variable; given that family planning is often a sensitive topic, initial approaches need to be tailored to each setting.

(introduction...)

Beyond the general willingness and ability to unearth and respond to existing demand - whatever the form - are there specific structures, activities, or approaches essential for program success? In 1964, Bernard Berelson, a major figure in the population movement, proposed a set of requirements that have been expanded over time into the 30 items that give the family planning program effort score previously mentioned. The 30 items - shown grouped into sets in Table 3 in accordance with factor analysis results - are widely recognized as indicators of serious organizational effort.

The first two sets of items are central and essential. First, adequate program effort means providing access to a variety of contraceptive methods - pills, condoms, IUDs, sterilization - to meet the varying needs and preferences of couples. To provide such access, a good logistical system is needed. In addition - as the second set of items indicates - effective contraceptive provision requires managing the front-line providers: They have to be trained, made to focus on their assigned tasks, motivated to keep essential records on their clients, and properly supervised in their duties. A good management system that accomplishes these things will also periodically assess progress and use evaluation results to improve operations.

Table 3 - Family Planning Program Effort Items

Item		Description
Method access
	Pills access	Proportion with easy access to pills or injectables
	Condom access	Proportion with easy access to condoms, diaphragms, or spermicides
	IUD access	Proportion with easy access to IUDs
	Female sterilization	Availability of female sterilization and proportion with easy access
	Male sterilization	Availability of male sterilization and proportion with easy access
	Logistics	How frequently stocks of supplies and equipment are adequate
Management
	Training programs	Adequacy of staff training programs
	Tasks execution	How well all staff categories carry out assigned tasks
	Recordkeeping	Whether client records are kept, summarized, and fed back to clinics
	Managers use evaluation	Extent of use by program managers of evaluation and research results to improve program
	Evaluation	Whether program-related demographic and operations research are conducted and whether staff or other institutions exist for this purpose
	Supervision	Adequacy of supervision at all levels
Mobilization
	Government policy	Whether government officially supports family planning and population control
	Public statements	Whether high officials publicly state support for family planning
	Rank of leader	Rank in the bureaucracy of the family planning leadership
	Other ministries	Involvement in population activities of ministries and government agencies without primary responsibility for service delivery
	Civil bureaucracy used	Use of civil bureaucracy (including central, provincial, district, and county administrators) to ensure program directives are carried out
	Incentives	Whether incentives and disincentives are provided to clients or staff
Marketing
	Advertising restrictions	Freedom from restriction of contraceptive advertising in mass media
	Social marketing	Extent of coverage by subsidized commercial contraceptive sales
	Media coverage	Frequency and coverage of mass media messages
Medical approaches
	Abortion access	Proportion with easy access to abortion under good conditions
	Postpartum programs	Extent of coverage by postpartum programs
	In-country budget	Proportion of family planning budget provided from country sources
Missionary approaches
	Community distribution	Extent of coverage by community-based contraceptive distribution programs
	Private agencies	Involvement of private-sector agencies and groups
	Home visitors	Extent of coverage by family planning workers who visit women's homes
Other
	Administrative structure	Adequacy of national, provincial, and county administrative structure and staff
	Age-at-marriage policy	How high the minimum legal marriage age for women is and how strongly it is enforced
	Import laws	Whether laws facilitate contraceptive imports or local manufacture

SOURCES: Descriptions abridged from Mauldin and Lapham (1985, pp. 8-10). See Bulatao (1996) for grouping of items.

³These could include the health ministry, if it is not the primary service provider; the prime minister's office or its equivalent; and a host of other ministries in the areas of social security, planning and development, finance, interior, education, environment, youth and sports, transport, defense, women's affairs, mass media and information, foreign affairs, rural development, urban affairs, religious affairs, and social affairs.

The third set of items has to do with mobilization of government resources to support family planning. Six items reflect this: the adoption of a national population policy; supportive public statements from political leaders; the rank accorded to the program leadership within the government bureaucracy; the involvement of other ministries besides the service delivery agency itself; the involvement of the civil bureaucracy at regional, provincial, and local levels; and the provision of incentives or disincentives to family planning clients or staff. This set of items is somewhat problematic; official support is essential in obtaining program resources but can easily shade into official pressure. In fact, the last item, the provision of incentives, needs to be carefully handled so as not to lead to undue influence on potential clients (Isaacs, 1995).

More consistently important than mobilization is program activity to reach the population with appropriate messages (the fourth set of items), especially through mass media, advertising, and commercial sales. The last two sets of items reflect delivery-system alternatives chosen in some but not other cases: clinic-based or largely "medical" alternatives, such as postpartum programs and abortion, and outreach systems, often with a strong volunteer component, such as community-based distribution, private agencies, and home visitation.

Rather than strict program requirements, a number of these items are alternatives that may be more or less important in particular settings. Few programs receive high scores across all these areas, and even some quite successful programs do little in some areas. In Bangladesh, for instance, where an extensive review of experience concludes that program-promoted changes in acceptability and availability of contraception have been much more responsible for fertility decline than any changes in preferred family size, a few of the items just discussed were identified as critical (Cleland et al., 1994, pp. 97, 122):

· an organizational culture of excellence, stressing dedication to realistic goals and "insulating the workforce from dysfunctional social pressures"

· frequent contact with clients by outreach workers with basic technical skills and strong supervision

· reliable supply of multiple methods and available follow-up and ancillary services.

The essential elements of organizational effort may vary not only by setting but also by stage of development of the program. Table 4 indicates the stages a program goes through, drawing on several schemes suggested by researchers (Bulatao, 1993; Vriesendorp et al., 1989; Keller et al., 1989; Townsend, 1991; Bernhart, 1991).

Table 4 Stages in Program Development

Initial Level of Family Planning Effort	Dominant Concern	Client Focus	Main Source of Support	Extensiveness of Program Functioning
Very weak (0-24)	Promotion	Highly motivated couples	Donors and voluntary organizations	Very few sites
Weak (25-54)	Management	Couples with unmet demand	Government	Limited coverage
Moderate (55-79)	Outreach	Broad populations	Government	Extensive coverage
Strong (80+)	Efficiency	Least accessible populations	Clients	Increasingly selective coverage

SOURCE: World Bank (1993).

The evidence that program actions in each of these areas contribute to success is varied and complex (and not without occasional contradictions). Some evidence will be briefly reviewed, covering the provision of access to methods and satisfaction of other client needs, management issues, promotion through the media, delivery systems, and political and financial support.

Responding to Client Needs

Consider first some of the evidence that a program that responds to client needs increases contraceptive use:

· Improved access to services helps clients. Access can vary greatly: Some countries may have only one service point for 15,000 women; in others, every community may have at least one such facility. Such variation in proximity to services, according to a review of 16 studies that differ widely in methodology, does affect contraceptive use, even controlling for other factors (Tsui and Ochoa, 1989; Angeles et al., 1996). Expanding access when contraceptive use is still limited is especially important (Phillips et al., 1994).

· Facilities or field-workers must, of course, have contraceptive supplies available. This is often an obstacle in newer programs, such as several in sub-Saharan Africa that suffer from supply shortages and inadequate logistical planning (UNFPA, 1990). Providing not only reliable supplies but also a wider range of appropriate methods increases use. In the Philippines, for instance, providing an additional method through clinics raises prevalence by 5 percentage points (Samara et al., 1996, p. 48). Increasing method choice need not be costly; data from Bangladesh suggest that adding reversible methods to a program that relies heavily on sterilization need not raise the cost per birth prevented (Simmons et al., 1991).

· Measures that promote continued use are also helpful. Improved counseling and better client information, for instance, lead to better method choices, fewer complaints about side effects, less-frequent discontinuation (W H O, 1980), fewer method failures, and less need for abortion. But the possible choices still have drawbacks that discourage some women, although others persevere despite them. Focus-group researchers quote one woman in Karachi as saying: "There is pain in these methods, but at least there is no danger that the woman will conceive" (Snow et al., n.d.). Research to improve contraceptive methods and develop new ones is still critical and requires continued public-sector funding (Harrison and Rosenfield, 1996).

· Whether ancillary services should also be provided is a complicated issue involving difficult trade-offs. Clients have other needs and demands besides contraception, and a program that can address these may be more effective and may also make critical contributions in such related areas as the attempt to contain the human immunodeficiency virus (HIV) and the acquired immun-odeficiency syndrome (AIDS).⁴ Some added services, including pregnancy tests, Pap smears, and screening for sexually transmitted diseases, are sufficiently integral to providing contraception that their role in a clinical program is clear. Other services range farther afield, from emergency obstetric services to income-generating activities for women and female education. Certainly desirable themselves, such activities may also promote lower fertility preferences, and combining them with family planning may have synergistic effects. Recognizing such arguments, the 1994 International Conference on Population and Development in Cairo supported the idea of family planning as one among various reproductive health interventions, all to be pursued together. However, ancillary interventions increase program costs, in some cases (as with emergency obstetric services) by quite substantial amounts, and additional financing is seldom easily available. Family planning program staff are not necessarily the proper people to provide ancillary services. Integrating such staff into larger health organizations risks submerging concerns about family planning, which seldom receive adequate attention within health ministries (Finkle and Ness, 1985). For these reasons, the debate about "vertical" (stand-alone) family planning programs versus integrated programs run as part of health ministries has continued unresolved for decades. A series of careful quasi-experiments in Bangladesh concludes that

minimal health provision indeed may indirectly benefit contraceptive acceptance. However, the addition of a broader range of health skills and supplies made no further impact on the success of family planning. Rather there was evidence of a drop in contraceptive use as the attention and energies of workers were diverted. ... Integration must be justified on grounds other than enhanced family planning effectiveness (Cleland et al., 1994, p. 146).

⁴Attempts at integrating services for sexually transmitted infections (STIs) with family planning, as in Mombasa, Kenya (Twahir et al., 1996), and eastern Uganda (Mukaire et al., 1997), suggest that there are advantages for family planning programs in offering more services, but also difficulties posed by the increased responsibilities. Serious questions also remain about the accuracy of STI diagnoses and the cost-effectiveness of integrated services (Maggwa and Askew, 1997).

The issue of what services to combine with family planning therefore remains to be approached in each case as a matter of values and objectives, but with full awareness of the practical consequences.⁵

⁵See Tsui, Wasserheit, and Haaga (1997, pp. 158-163) for a further discussion of the issues from the perspective of reproductive health services - a discussion that does not, however, appear to come to a clear conclusion.

Managing Effectively

Elements of good management also contribute to more contraceptive use. Informal comparisons of national programs easily demonstrate this, but it is difficult to control complicating factors. More precise evidence therefore usually comes from experiments or small studies.

· Training can improve provider performance. This is an article of faith for most programs. Several hundred short-term training courses in family planning and maternal health are conducted each year around the world in more than two dozen different languages. Studies in such countries as Bangladesh, Ecuador, Morocco, and Ghana confirm that trained field-workers and supervisors perform better (Finkle and Ness, 1985; Gallen and Rinehart, 1986, p. 825; Brown et al., 1995), and other studies indicate the need for periodic retraining (Gallen and Rinehart, 1986, p. 826). Training does have to match the specific provider's needs, reflect actual problems and options, emphasize practical skills, and focus on developing competence.

· Good supervision is often cited as a central element in program success, given the many shortcomings of supervisors who must labor under "low salaries, harsh working conditions, and the absence of performance-based rewards" and are distracted by administrative requirements and their own pressing personal affairs (Simmons, 1987, p. 251). Studies in Nigeria, Guatemala, Turkey, and Brazil (Townsend, 1991, p. 55; Foreit and Foreit, 1984) indicate that frequent supervisory visits increase program effectiveness. What supervisory visits cover is critical: Providing training and reinforcement and actually observing worker contacts have much more effect than merely collecting management information.

· Strategic planning, aided by reliable evaluation data and good applied research into program options (usually referred to as "operations research"), contributes to program success. The East Asian experience, in which extensive field experiments and continuing research and evaluation were used to guide program expansion (as well as to generate public support), and the similar experience in Bangladesh suggest the importance of these factors. Reviews of development programs similarly suggest the importance of strategic planning or strategic management (Paul, 1983), which requires sensitivity to the environment a program faces.

Targeting services to those most in need, or collecting some payment from those who can afford it, is a sensible option that can improve financial sustainability. Nominal price increases do not appear to affect contraceptive use substantially, as studies of price elasticities in Jamaica and in several Southeast Asian countries suggest (Lewis, 1996). However, the time and the manner for introducing such changes require careful consideration in each case, especially so they do not affect the poor (Lande and Geller, 1991).

Promoting Family Planning

Promotional activities to reach the public, particularly through mass media, can have substantial effect but have to be properly done. Brief 30-second "spots" in Peru had only a small effect (Westoff et al., 1994). In contrast, continuing publicity about family planning had a greater effect in India. A large national survey established that exposure to a family planning message in the past month increased the number intending to use contraceptives by 6 percentage points, controlling for socioeconomic factors and for general media exposure (Ramesh et al., 1996). Finally, an extensive media campaign in Nigeria, with radio and TV dramas, music videos with popular artists, billboards, bumper stickers, and so on, demonstrated a clear effect on behavior. Among those exposed to media messages in 1990, almost twice as many were using contraception three years later than among those not exposed. The influence of exposure appeared more often among those who then discussed family planning issues with current users and was greater on initial adoption than on continuation. Media exposure also appeared responsible for reducing fertility preferences by about half a child (Bankole et al., 1996; Bankole and Adewuyi, 1994). Advertising in particular can make a difference: Condom sales are more closely linked to advertising than to any other factor (Boone et al., 1985).

Selecting a Delivery System

The major distinctions among delivery systems are between clinic-based delivery, community-based distribution (or outreach systems more generally), and commercial distribution. An overlapping distinction may also be made between public and private distribution, although private systems may involve some degree of public subsidy. These different systems are often complementary:

· Clinics are needed for such methods as sterilization and to provide backup for outreach workers and referral, when needed, for clients' medical problems. Proximity to clinics increases contraceptive use, as noted earlier, up to the point at which clinics and other facilities are easily accessible to everyone.

· Community-based distribution involves bringing contraceptives to women in their own communities, rather than requiring them to travel to clinics. Such efforts are especially needed where clinics are too sparse and too expensive to build and maintain. They often depend on local residents with a few weeks' training, who receive no fixed salary but may have various incentives. Their effectiveness at increasing contraceptive use has been demonstrated in many countries, including Egypt, Mexico, the Democratic Republic of Congo, and Bangladesh (Gallen and Rinehart, 1986; Bertrand, 1991). However, at higher levels of contraceptive prevalence in Brazil and the Eastern Caribbean, community-based distribution had minimal effect. With the large workforce they require, the costs of community-based distribution can be a concern. These costs can vary from US$5 to US$150 per user per year (Huber and Harvey, 1989). For each setting in which outreach is needed, the most cost-effective way to provide outreach needs to be considered.

· Distributing partly subsidized contraceptives through commercial channels, aided by catchy advertising, has become known as contraceptive social marketing. This provides an alternative source of supply, especially for condoms and pills, that can be more congenial for some users and therefore can expand contraceptive use. To the extent these are new users or switch from government programs, the considerably lower cost of social marketing saves public funds; however, users may also switch from purely commercial outlets.

· Private voluntary organizations operate a range of distribution systems, some involving clinics and others concentrating on outreach. Volunteer programs have been critical in introducing family planning in various settings and are still useful for reaching specific target groups, such as adolescents, and for providing some competition and some savings to government programs.

· Additional delivery systems of many types often coexist with these. Private physicians and midwives often provide some contraceptive services on their own, although they would often benefit from specific training in family planning. Health insurance or other employer-supported programs may cover contraceptive services. Commercial outlets may be better or worse at supplying contraceptives. Since expanding service outlets and generally making family planning more available lead to increased contraceptive use, the multiplication of such distribution systems can be useful, and public programs should be designed to complement rather than to discourage them.⁶

⁶See World Bank (1993a, pp. 68-69) for an example of how commercial services can be disrupted by a public program.

Mobilizing Support

Political support is essential to enable a program in such a sensitive area as family planning to achieve respectability and have sustained effect. Programs depend on politically allocated resources. At least as important, they also depend on a positive, widespread popular response, which is greatly aided by visible political support. Government endorsement of family planning helps legitimize what is initially innovative behavior and can helpfully be supplemented by support from the media, voluntary organizations, and even religious groups. Official support has grown over the decades and is now close to universal, but the depth of this support and the extent to which it is reflected in effective government action still vary considerably.

Government support does have to be tempered to preserve voluntary individual choice. This is important not only to protect human rights but also for pragmatic reasons. Coercive programs cannot succeed or even survive for long in countries that aspire to be democratic. India attempted to force sterilizations after emergency rule was declared in the mid-1970s; partly as a result, the government was soon voted out of office. The most coercive modern program was launched in China by a government intent on controlling every aspect of life, including childbearing. China is the only country that has "penalized people specifically and directly for violating population policy" (Li, 1995, p. 563). Yet the Chinese one-child policy has been remarkably ineffective. An analysis of fertility in Hebei and Tianjin (around the capital) concludes that "The majority of Chinese women ... ignored birth-quota regulations, refused to accept the one-child certificate, and bore the burden of heavy financial penalties" (Li, 1995, p. 582). International donors have generally turned away from such coercive programs and have been increasingly active in improving the quality and service orientation of the programs they assist.

International donors have played an important role in helping to marshal government support for voluntary family planning programs. They have done this not only through financial assistance for services but also through political dialogue, at successive world population conferences for instance (Bucharest in 1974, Mexico City in 1984, Cairo in 1994), which allowed a consensus to build in favor of family planning. Donors have also largely underwritten such activities as the World Fertility Survey and the Demographic and Health Surveys and contributed to various rounds of national censuses, which have demonstrated to governments the specific population-growth scenarios they face. Donor influence in initiating and strengthening programs has been substantial, as has been argued by such observers as Warwick (1982, p. 44), who states that, at least through the 1980s "of all the spheres of national development, population has been the most donor driven."

Some impact of donor funding on ongoing programs appears in statistical analysis, although not consistently. In comparisons across Asian countries up to the early 1980s, Ness and Ando (1984) found that the volume of outside financial aid did not affect program strength. But more recently, Tsui (1997) has shown that, across all recipient countries, population assistance from the United States (through USAID, the agency responsible for bilateral foreign-aid programs) has a small but significant impact on program effort. The inconclusive statistical evidence is understandable: donors naturally are selective in their assistance, sometimes choosing more promising settings, sometimes more impoverished ones, and often ones with which they have some special cultural or political link. This makes interpretation difficult.

Anecdotal evidence and experience suggest several ways in which donor influence has been important to ongoing programs:

· in providing the training and developing the technical and managerial resources on which programs rely

· in encouraging new approaches, from new contraceptive methods, such as implants, to new strategies, such as community-based distribution and social marketing

· in sponsoring research into ways to make programs work better, research that nascent programs seldom can spend their own time and resources on

· in encouraging service standards in such areas as the need for informed consent.

Donor involvement in a program is not a guarantee of success, but it provides resources that, coupled with national commitment and a reasonable strategy, can accelerate the progress of a program.