1This article addresses the sensitive, complex and largely unexplored question of abortion in Africa, drawing on original survey data to complete data from the Demographic and Health Surveys. Agnès Guillaumefocuses specifically on the determinants of fertility decline in an African metropolis, Abidjan. For a country where abortion is illegal and access to family planning limited, she explores the interaction between the recourse to abortion and contraceptive practices (natural or traditional versus modern). The author shows the role of abortion in fertility regulating strategies, which has potentially serious consequences both for women and for society in general.

2Fertility in most African countries has entered a phase of transition whose rhythm, intensity and determinants are variable. Of the 31 sub-Saharan African countries where Demographic and Health Surveys (DHS) were conducted in 1990-2000, fertility stood at between 4 and 4.9 children per woman in 10 countries, between 5 and 5.9 children in 11 countries, and over 6 children in the remaining 10 countries, the highest level (7.2 children) being observed in Niger. Fertility is lower in North Africa, where it is below 4 children.

3Fertility has declined more among populations that have benefited from an improvement and modernization of their living conditions. Urbanization, education, access to the media, and health improvements are all powerful factors influencing these changes (Vimard et al., 2001). An “increasing diversification of African reproductive regimes” is appearing across and even within countries, and is manifested in large disparities between regions or socio-economic groups (Tabutin and Schoumaker, 2001), with fertility decline being greatest among urban and educated groups (Cosio-Zavala, 2000).

4The divergence of trends even within the various populations raises questions about the role of each determinant of the demographic transition. Traditional birth spacing behaviour based on postpartum abstinence and breastfeeding is important in Kenya, Ghana, Senegal and Sudan, where it accounts for between 37% and over 44% of the fertility reduction (Vimard et al, 2001; Jolly and Gribble, 1996). Delay in first marriage plays a preponderant role in some countries, while in others modern contraception is the main factor behind the change, notably in the English-speaking African countries where governments have actively supported family planning programmes (Mboup, 2000).

5Studies of fertility decline, particularly in countries where contraceptive use remains low but where women express a demand for regulation of their fertility, suggest that other factors are present, prominent among which is abortion (United Nations, 2001). In some countries and more specifically in some cities of Africa, “fertility has declined rapidly, but without widespread use of contraception” (Tabutin and Schoumaker, 2001). Although abortion is one of the components of the Bongaarts model, which is frequently used to measure the contribution of the different variables in fertility reduction, few studies have incorporated it fully. The role of abortion in this trend is in effect assumed, but it is rarely quantified due to the lack of available data on the subject (Desgrées du Loû et al., 1999; Locoh, 1994). Abortion is illegal in the majority of African countries and since few studies have been conducted on the phenomenon it remains poorly understood. However, studies on the complications arising from abortion and on the maternal mortality that is attributable to it have highlighted the extent of the practice (Thonneau et al., 2002; World Health Organization et al., 1998; Benson et al., 1996). Other research has examined abortion from the angle of the growing burden on hospitals of complications arising from illicit abortions (Huntington et al., 1998). The frequent recourse to abortion in some African capital cities, particularly among young women, has already been established (Desgrées du Loû et al., 1999; Guillaume, 2003; Konate et al., 1999; Locoh, 1994).

6In an African context where the modalities of the demographic transition are varied, this article studies the extent of abortion practices in the particular case of the population of the city of Abidjan. Although a process of fertility decline has recently got under way in Côte d’Ivoire, the phenomenon is much more marked in urban zones, and especially in Abidjan, than in the rest of the country. Between the Demographic and Health Surveys (DHS) of 1994 and 1998-1999, the total fertility rate fell by 0.5 children per woman for the country as a whole, from 5.7 to 5.2 children per woman (Table 1). During this period the gap widened between Abidjan, where the rate fell to 3.4 children per woman at the end of the 1990s (a decline of 0.7 children), and the rural zones, where it fell only slightly to stand at 6 children.

Table 1

Fertility, postpartum variables, contraceptive use and age at marriage in Côte d’Ivoire, in the 1990s

DHS 1994 DHS 1998-1999 Abidjan Urban total Rural total Country total Abidjan Urban total Rural total Country total TFR 15-49 years 4.1 4.7 6.4 5.7 3.4 4.0 6.0 5.2 % of sterile women (40-49 years) 4.6 3.4 2.1 2.6 3.3 2.4 0.8 1.8 Median duration (months) Postpartum amenorrhoea 8.3 10.6 12.9 12.3 7.9 9.3 12.5 10.6 Postpartum abstinence 8.5 9.3 12.2 11.8 9.1 10.1 18.9 11.8 Postpartum insusceptibility 15.0 16.0 17.2 16.6 12.4 13.9 20.2 18.9 Breastfeeding 17.9 18.7 21.3 20.3 15.9 17.6 21.6 20.5 Current use of contraception* (%) None 71.4 74.7 89.9 83.5 65.0 69.0 86.7 79.3 Rhythm method, withdrawal 17.2 15.0 5.8 9.8 17.7 14.7 6.2 9.7 Pill 5.4 4.7 1.0 2.5 5.8 6.0 2.1 3.7 Condom 2.4 2.3 1.7 1.9 7.8 7.0 2.5 4.4 IUD 0.7 0.5 0.1 0.3 0.6 0.4 0.2 0.3 Injectable 1.0 1.0 0.4 0.7 1.2 1.3 1.0 1.1 Other (incl. traditional methods) 1.9 1.8 1.0 1.4 1.9 1.7 1.3 1.5 Median age at first marriage (women aged 25-49) 19.4 18.7 17.9 18.1 20.2 19.2 18.2 18.7 * Among all women. Source: Côte d’Ivoire National Statistical Institute, DHS 1994 and 1998-1999.

Fertility, postpartum variables, contraceptive use and age at marriage in Côte d’Ivoire, in the 1990s

7Women wish to have fewer children than they currently have. In 1998 the desired number of children was 4.5 in the country as a whole and 3 in Abidjan. The demand for family planning confirms this wish among women living in union. According to the 1998-1999 DHS slightly more than half of women in Abidjan want to plan their childbearing, 17% of them for birth-limiting and 34% for birth-spacing. These shifts in women’s reproductive ideals probably result partly from socio-economic changes related to the economic crisis affecting the country, and from the impact of structural adjustment programmes that have contributed to raising the cost of access to social services (health, education, etc.).

8Contraceptive practice increased between the two DHS but remains very low. In 1998, 21% of women in Côte d’Ivoire and 35% of women in Abidjan were using contraception, representing an increase of 4 and 6 points respectively since 1994. Use of modern contraceptive methods, however, is barely progressing. In Abidjan, 17% of women used natural [1] contraceptive methods and this prevalence remained practically stable between the two surveys at 10% for the country as a whole. Natural methods remain more widely used than modern methods even though they are of limited effectiveness, since only 36% of women know the exact timing of their fecund period (Institut National de la Statistique and ORC Macro, 2001).

9In 1998-1999, medicalised methods (the pill, the IUD, injectables) were used by only 7.6% of women in Abidjan and 5.1% in the country as a whole. The spread of modern methods is due mainly to the condom, use of which increased threefold over five years in Abidjan (2.4% in 1994, 7.8% in 1998) and more than twofold in the country as a whole (1.9% in 1994 and 4.4% in 1998). The health-education campaigns initiated in response to the high prevalence of HIV in Côte d’Ivoire, estimated at 15% among women in Abidjan (Welffens-Ekra et al., 1996), partly explain this increase in use of the condom, which is employed to prevent sexually transmitted infections but has certainly also contributed to reducing conceptions.

10Although data on contraception exists for Côte d’Ivoire, the information available on abortion is limited to a few ad hoc studies on specific populations or derives from analysis of abortion-related complications. Some reproductive health studies have included a question about the number of abortions women have undergone in the course of their reproductive life. In four regions of Côte d’Ivoire (Aboisso, Tanda, Centre-North and Niakaramandougou), between 7% and 16% of women reported having recourse to at least one abortion (Table 2). This percentage rises with the women’s educational level. In the Aboisso region (in the south east of Côte d’Ivoire), the proportion of women with secondary level education who reported at least one abortion was four times that of women with no education (Guillaume et al., 1999). The practice is also more frequent in urban zones than in rural zones. In the city of Aboisso, 18% of the women mentioned at least one abortion; in Abidjan the proportion rises to 30%, both in the results of a survey on women during prenatal consultations (Desgrées du Loû et al., 1999) and in those from a survey on the wives of military based in the city (Koffi and Fassassi, 1997). Desgrées du Loû et al. (1999) have drawn attention to the importance of this practice among young women from the start of their reproductive life.

Table 2

Women who have aborted, by number of abortions and region of residence (%)

Abidjan, wives of military Abidjan, prenatal consultation* Aboisso Tanda Centre-North Niakaraman-dougou Survey year % of women who have undergone: 1996 1997 1993 1994 1995 1998 1 abortion 18.7 17.8 10.2 8.1 4.6 6.0 2 or more abortions 13.5 12.7 5.8 2.0 1.9 1.8 At least 1 abortion 32.2 30.5 16.0 10.1 6.5 7.8 Of which: Rural zone – – 13.6 – 4.9 5.5 Urban zone – – 17.8 – 9.8 9.0 Total 2,183 1,200 1,900 1,578 1,055 1,684 * Among women who have been pregnant at least once. Sources: Guillaume et al., 1999; Desgrées du Loû et al., 1999.

Women who have aborted, by number of abortions and region of residence (%)

11The present article is based on a study of individuals consulting in Abidjan’s health centres, which supplied elements for understanding the fertility changes. We begin by examining the extent of contraception and abortion practices, then the relationship between abortion and contraception, and conclude with the link between abortion and the decline in fertility.

I – Population and data

2 – Profile of the respondents

16The surveyed population differs depending on the health centres and reflects the social structure of the neighbourhoods (Table 3). The sample consists mainly of young women (52% are under 25) of low educational level: 39% of the women had received no schooling and around a third (32%) had only reached primary level. More than two-thirds of the women (69%) were living in union at the time of the survey. The majority were Christian (41%) or Muslim (36%).

Table 3

Socio-demographic characteristics of respondents in Abidjan In the FSU Health Centres and the DHS, 1998-1999 (%)

FSU Health Centres DHS Abobo Abobo Abidjan Toit rouge Ouassakara Total Sagbe Avocatier Age 15-24 years 39 49 62 57 52 51 25- 34 years 43 39 32 36 37 29 35 years and over 18 12 6 7 11 20 Educational level No schooling 24 56 34 41 39 34 Primary 29 28 37 33 32 33 Secondary or above 47 16 29 26 29 33 Religion Christian 53 25 50 36 41 51 Muslim 22 56 32 34 36 34 Other 25 19 18 30 23 15 Marital status Not in union 36 18 42 27 31 56 In union 64 82 58 73 69 44 Ethnic group Akan 41 22 44 34 35 33 Krou 25 9 14 16 16 16 Mandé Nord 8 6 5 10 7 13 Mandé Sud 6 19 14 16 14 7 Voltaic 7 13 10 8 10 10 Other 13 31 13 16 18 21 Number 600 600 600 600 2,400 766 Sources: FSU 1998 and DHS 1998-1999 surveys.

Socio-demographic characteristics of respondents in Abidjan In the FSU Health Centres and the DHS, 1998-1999 (%)

17Our sample cannot be taken as representative of the population of Abidjan since it does not include women who do not attend the health centres. Health care coverage in Abidjan is relatively high, as was shown by the results of the 1998-1999 DHS according to which 82% of Abidjan’s women gave birth in a health facility and only 2.8% received no prenatal supervision during their pregnancy. The health centres selected for this survey are situated in the two most populous working-class neighbour hoods of Abidjan. The health services and treatment provided in these centres are inexpensive and therefore accessible to a large population.

18To situate the survey population, its characteristics have been compared with those of the population surveyed for the 1998-1999 DHS in Abidjan. While the proportions of women aged 15-24 are identical (52%), our population contains more women aged 25-34 than in the DHS. The health centre respondents are also less educated: 39% have received no education (34% in the DHS) and 30% have completed secondary or higher education (33% in the DHS). The proportions of Muslims are similar in both populations (34% against 36%). On the other hand, fewer Christian women attended the health centres (41% against 51%). Regarding the ethnic structure of the two samples, the proportions of Krou (16%), Voltaic (10%) and Akan (35% and 33%) women were equivalent. Women from the Mandé group were distributed in opposite proportions in the two samples: among the health centre patients 7% belonged to the Mandé Nord and 14% to the Mandé Sud, against 13% and 7% respectively in the DHS. Finally, more foreign women were interviewed in the DHS.

19A major difference between the two populations is the predominance among the health centre respondents of women living in union (69% against 44% in the DHS). This difference is explained by the difficulties young single women face in visiting family planning centres, as much for economic reasons as over confidentiality, and by the decision to interview the women regardless of their reason for attending the centres, which tends to result in an over-representation of mothers and hence of married women.

II – Very low levels of contraceptive use

20In Abidjan, the level of contraceptive knowledge among the women interviewed in the health centres is high, since nearly all of them reported knowing at least one method: 96% for all methods and 95% for modern methods (Table 4). These results are close to those of the 1998-1999 DHS.

Table 4

Knowledge and current use of contraception in Abidjan, 1998 (%)

Knowledge Current use Any method 96 28 Natural method 55 17 Modern method 95 12 Pill 89 4 IUD 29 1 Injectable 6 1 Condom 57 5 Note: Total of women (n = 2,400). Source: FSU survey, 1998.

Knowledge and current use of contraception in Abidjan, 1998 (%)

21The most widely known modern contraceptive methods are the pill (89%), the condom (57%), the IUD (29%), and injectables (6%). The proportion of women aware of other modern methods such as Norplant, the female condom, and spermicides was under 5%. In addition to numerous education campaigns on condoms as part of AIDS prevention, Abidjan women have also benefited from media promotion of the contraceptive pill “Confiance [2]”. Information about other methods depends mainly on health centres or informal information networks, which explains the lower levels of awareness of these methods. Natural methods seem to be less well known since they were mentioned by only just over half of the women.

22Knowing about the existence of contraception does not necessarily mean having a good understanding of how the different methods work or of their possible side-effects. Nearly half of the women reported having obtained information on contraception through informal networks (parents, friends). Although these networks contribute to diffusing knowledge, there is also a danger that they transmit incorrect information or spread rumours that can create a resistance to family planning (Ngom, 2000).

23Being informed, to whatever degree, is therefore not enough, as is shown by the difference between the number of women who reported knowing about contraceptive methods and the number actually using them. In Abidjan, contraceptive use remains uncommon among the health centre patients: 17% use a natural method (mainly the rhythm method) and 12% use a modern method, with the condom being the most important (5%), followed by the pill (4%). Natural methods are therefore more often used than modern methods, despite being less effective, especially when women do not know the correct timing of their fecund period. These results are close to those from the 1998-1999 DHS which showed 18% of women used a natural method, 8% the condom, and 6% the pill, while very few used injectables (1.2%) and the IUD (0.6%). We can therefore conclude that use of all types of medical contraceptive methods is low.

24Multivariate analysis shows that contraceptive use differs according to the women’s characteristics (Table 5). Educational level is a powerful determinant of use of modern and natural contraception. Recourse to these methods is higher among women who have been educated than among those who have not, and the odds ratio is twice as high for women with primary education as for women with no education.

Table 5

Logistic regression on current contraceptive use and recourse to abortion, by characteristics of women in Abidjan, 1998 (odds ratios)

Modern Natural At least Socio-demographic characteristics Any method method method 1 abortion Age 15-24 years 0.92 0.85 0.85 13.31*** 25-34 years 0.97 0.96 0.75 8.33*** 35 years and over (Ref.) 1.00 1.00 1.00 1.00 Educational level No schooling 0.40*** 0.38*** 0.46*** 0.34*** Primary 0.69*** 0.65*** 0.79 0.49*** Secondary or above (Ref.) 1.00 1.00 1.00 1.00 Religion Christian (Ref.) 1.00 1.00 1.00 1.00 Muslim 0.81* 0.95 0.63** 0.31*** Other 0.71*** 0.74* 0.68 0.98 Activity Student 1.09 1.13 0.85 0.44*** Housewife 0.90 0.72 0.98 0.60*** Trade 0.90** 0.75 1.00 0.64*** Service sector (except trading) 0.95 0.84 0.92 1.18 Craft worker (Ref.) 1.00 1.00 1.00 1.00 Marital status Not in union 2.20*** 1.66*** 2.00*** 2.36*** In union (Ref.) 1.00 1.00 1.00 1.00 Total number of pregnancies 1.05* 0.99 1.03 1.75*** Survey site Toit rouge 1.48*** 1.25 1.68** 1.83*** Abobo Sagbe 0.29*** 0.34*** 0.37*** 0.73* Abobo Avocatier 1.51*** 0.78 1.99*** 1.13 Ouassakara (Ref.) 1.00 1.00 1.00 1.00 Past contraceptive use No (Ref.) 1.00 Yes 3.33*** Note: n = 2,400; significant: *** at 1% level, ** at 5% level, * at 10% level. Reading: For women not living in union, the ratio between the proportion using any contraceptive method and those using none is 2.2 times higher than for those living in union (Ref.), all other things being equal. Source: FSU survey, 1998.

Logistic regression on current contraceptive use and recourse to abortion, by characteristics of women in Abidjan, 1998 (odds ratios)

25The influence of education on the use of modern contraception has been highlighted in numerous studies which have shown educated women in urban areas to be pioneers in the regulation of their fertility (United Nations, 2001; Vimard et al., 2001).

26Contraceptive use is also much more widespread among women not living in union than among those in union, with the odds ratio reaching 2.2. Muslim women are less likely to use contraception than Christian women, a result consistent with those from the DHS. Variables such as age, number of pregnancies, and women’s economic activity are less closely associated with family planning behaviour in our study.

27The differences in contraceptive practice by survey site are also large and can be explained by the women’s characteristics but also by the family planning services provided in these centres.

III – Abortion practice in Abidjan

32In all the countries of West Africa where induced abortion is prohibited, research on its illicit practice runs up against the classic phenomenon of under-reporting. This risk increases as the time since the event becomes longer, particularly for women at the end of their reproductive life and who had large families. There is no complementary data source (health statistics, for example) with which to verify the completeness of these data because abortions as such are not counted in health statistics. The only source of information is hospitalizations for abortion-related complications, but it introduces a strong bias by giving an extremely partial view of the phenomenon. Not all women who have recourse to abortion suffer from complications requiring hospitalization, and the nature and seriousness of the complications depend largely on the method used to abort. While not purporting to provide a good evaluation of abortive practices, the FSU survey does at least, through the women’s reports, allow a low hypothesis to be formulated for the prevalence of the phenomenon. Furthermore, it gives information about the abortion methods used and about the connections that the women make between contraception and abortion. The measurement of abortion thus obtained is not representative of the population as a whole, but provides valuable information on the circumstances of and trends in recourse to abortion, which is especially useful given that this phenomenon seems to be increasing and remains largely undocumented (Desgrées du Loû et al., 1999).

2 – … with varied methods of abortion

41From time immemorial, women have used a wide range of methods to regulate their fertility, even if the efficacy of these techniques is unproven and if by today’s standards they do not always have scientific foundations (McLaren, 1990). The methods known and used vary by country, as does the availability of contraceptive and abortive products on both official and unofficial markets. It sometimes happens that abortion procedures are unsuccessful; the women can then accept to continue with the pregnancy whereas others seek a more effective technique. Equally complications (haemorrhage, infection, perforation of the uterus, etc.) can occur and lead to health problems, sterility, and, in extreme cases, to the woman’s death (Strickler et al., 2001).

42The traditional pharmacopoeia contains a number of plants reputed for their abortive or contraceptive properties. In the case of abortion, these plants, which are sold in the markets, generally take the form of enemas (18%), herbal teas (5%) or plant pessaries (9%) (Table 6). Only 4% of the women had practised abortion through the insertion of sharp objects, while 12% had tried other methods such as chemicals (bleach, blueing products, detergent, etc.) or manufactured products (alcoholic drinks, coca-cola, sugar or acidic products taken to excess) reputed to have abortive properties (The Alan Guttmacher Institute, 1999).

Table 6

Abortive method used in last abortion, Abidjan, 1998 (%)

Method used* Herbal teas 5 Enemas 18 Plant stem 4 Plant pessary 9 “Pharmaceutical product” 10 Surgical method 62 Other 12 * Other methods may be mentioned. n = 805 women. Source: FSU survey, 1998.

Abortive method used in last abortion, Abidjan, 1998 (%)

43In 10% of cases, the women reported using pharmaceutical products. Often this involved taking overdoses of medicines that are counter-indicated in pregnancy, including anti-malarial drugs (nivaquine, quinine), aspirin, paracetamol, antibiotics, but also some hormonal products such as crynex and synergon. Note that the abortion pill is not available in Côte d’Ivoire.

44However, the majority of women, nearly two thirds, had recourse to a surgical method. The most important of these is abortion by curettage, which is usually administered by medical personnel of varying qualification (gynaecologists, general practitioners, but also nurses and hospital staff, etc.) and in sanitary conditions that vary with location (hospital, private clinic or surgery, local infirmary). In 3% of cases these operations are performed after a failed abortion attempt using traditional methods.

45Some of these methods, particularly traditional methods, carry health risks for women. A study carried out in the Abidjan gynaecological services on complications during the first trimester of pregnancy (Goyaux et al., 1999), showed complications from abortions using “local” methods (herbal teas and insertion of plant stems) to be the cause of large numbers of maternal deaths: 8% of admissions and 3.6% of deaths were due to abortions.

IV – The interaction between contraception and abortion

46In a context where abortion is illegal and access to contraception is limited, the question arises of the interaction between these two modes of fertility regulation. Some women explained their abortion by a need to limit (3%) or space (9%) births, expressed in the view that “the last child was too young”. Unwanted pregnancies result from either the absence of contraception or its failure. This situation is characteristic of the postpartum period during which women believe themselves not at risk of another pregnancy because they feel protected by breastfeeding. The interviews with health personnel, in particular midwives, draw attention to the level of risk involved. In our study, among the women who have been pregnant at least twice, 9% of the pregnancies resulted successively in a live birth followed by an induced abortion, and in over 20% of cases this abortion occurred less than a year after the delivery. Yet in Abidjan, the period of postpartum insusceptibility [5] and the median duration of breastfeeding have both declined, by 2.6 and 2 months respectively, between the two Demographic and Health Surveys (see Table 1), which increases the risk of pregnancy for young mothers since in Abidjan contraception is rarely prescribed to women who are breastfeeding.

47The relationship between contraceptive and abortive practices is clearly illustrated by the results of the multivariate analysis (Table 5). The odds ratio of undergoing abortion at least once is three times higher for women who have already used contraception than for those who have never used it. Clearly committed to a strategy for regulating their fertility, they use contraception and, if it fails, resort to abortion. Conversely, those who have never used contraception may start to use it following experience of an abortion, which they do not wish to repeat. Hence, among women who have used contraception during their life, 24% have undergone one abortion and 22% have had two or more, as against 8% and 4% respectively among non-users.

48To determine the links between contraception and abortion more clearly, the women were asked about their contraceptive practices before and after the last abortion (Table 7).

Table 7

Contraceptive use after the last abortion according To the method used before the last abortion, Abidjan, 1998 (for 100 users of each method before the abortion)

Use before the abortion Use after the abortion None Natural or traditional method Modern method None (61 %) 43% 23% 34% 100% Natural or traditional method (25 %) 20% 56% 24% 100% Modern method (14 %) 33% 20% 47% 100% Total 35% 31% 34% 100% n = 805 women. Source: FSU survey, 1998.

Contraceptive use after the last abortion according To the method used before the last abortion, Abidjan, 1998 (for 100 users of each method before the abortion)

49The majority of women (61%) did not use contraception prior to the abortion. They explained this by either a lack of information about the methods or a refusal to use them. 43% of them did not modify their practice after having the abortion, 23% adopted a natural or traditional method and 34% a modern method.

50Of the women who used a natural or traditional contraceptive method before they aborted (25%), more than half (56%) continued using this type of method despite its failure, while 20% stopped using any contraception and 24% adopted a more effective method.

51Among the women who used a modern method before the abortion (only 14%), one-third stopped using any contraception following its failure, 20% changed to a natural or traditional method, and 47% continued using the modern method.

52The experience of an abortion therefore produces changes in the modes of fertility regulation. The proportion of women using no contraception is practically halved, dropping from 61% to 35%. But this experience does not produce a large shift to modern contraception, since nearly identical proportions of women use this type of method (34%) as use none at all (35%) or choose natural or traditional contraception (31%).

53The absence of contraception use following an abortion is explained largely by the deficiencies in post-abortum counselling. Two-thirds of the women received no advice on how to avoid another pregnancy: in 13% of cases, the women were advised to use a modern method and in only 11% of cases was this prescribed. Abortion is practised illicitly in Côte d’Ivoire and although it is often performed by medical personnel, in the majority of cases these practitioners provide no post-abortum care.

54Nevertheless, the recourse to abortion following contraceptive failure is evidence of a real desire among some women to control their fertility. They at first use contraception to avoid an unwanted pregnancy and, in good logic, if they do become pregnant they decide to terminate before going back to their contraceptive practices. These women employ abortion and contraception as two complementary instruments of fertility regulation.

55The situation of women using no form of contraception is more complex. It can correspond to a period when the woman feels no need for contraception (breastfeeding, absence of partner or very infrequent intercourse). But it can also be the situation of young women who are not in a stable union. In Abidjan, and in West Africa more generally, contraception is delivered mainly through the services for the protection of mothers and infants, a source not adapted to the demand from these young women. Irregular sexual intercourse can also explain the low contraceptive use among young women, notably of medical methods since they are more likely to favour intercourse-related methods such as the condom. This method is also strongly promoted by AIDS prevention campaigns, but it makes the young woman dependent on her partner who can refuse to use a condom. Lastly, some women do not use contraception because they judge these methods inconvenient and say that they “prefer” abortion which is “effective” and final. This choice is not without risk, however, since access to medically supervised abortion is difficult in Côte d’Ivoire where abortion remains illegal.

V – Fertility and abortion rates: opposing trends

56The FSU survey confirms the decline in the total fertility rate observed during the 1998-1999 DHS (Table 8). For women aged 15-34 it fell from 3.8 children per woman during the period 10-14 years preceding the survey (1983-1987), to 3.5 during the intermediate period, then to 3.2 during the five years (1993-1997) immediately preceding the survey. For women aged 15-39 the TFR fell slightly more, declining from 4.2 children in 1988-1992 to 3.7 in 1993-1997.

Table 8

Fertility and abortion rates in Abidjan in different periods preceding the survey, by age of mother

FSU survey – 1998 DHS 1998-1999 Age group Period preceding the survey 0-4 years 5-9 years 10-14 years 0-2 years Fertility Abortion Fertility Abortion Fertility Abortion Fertility Age-specific rates (per 1,000) 15-19 125 89 114 51 135 47 50 20-24 160 73 193 58 190 35 157 25-29 194 45 197 33 220 38 152 30-34 150 37 199 23 219 14 157 35-39 113 28 144 23 125 40-44 65 5 43 Total (per woman) 15-44 4.0 1.4 3.4 15-39 3.7 1.4 4.2 0.9 3.2 15-34 3.2 1.2 3.5 0.8 3.8 0.7 2.6 Sources: FSU 1998 and DHS 1998-1999 surveys.

Fertility and abortion rates in Abidjan in different periods preceding the survey, by age of mother

57The total abortion rate, by contrast, has increased. For ages 15-34, after rising slightly from 0.7 to 0.8 abortions per woman between 1983-1987 and 1988-1992, it has reached 1.2 in the last five years. For women aged 15-39 the increase is even sharper, from 0.9 to 1.4 abortions per woman in the last ten years. For women under 40, fertility has fallen by 0.5 children over ten years while the number of abortions per woman has increased by 0.5.

58The growth in the recourse to abortion is observed at all ages, but with a varying intensity. The increase in the abortion rate is particularly large among young women under 20: during the last two five-year periods the rate rose from 51 per 1,000 to 89 per 1,000, multiplying by a factor of 1.7, while at other ages the multiplying factor is not above 1.4, except for 30-34 years (1.6).

59The younger the women, the higher the abortion rate. In the most recent period the rate was three times higher at 15-19 years than at 35-39 years (Figure 1). Recourse to abortion tends to decrease with age; for the intermediate period the rate peaks slightly at 20-24 years and then falls rapidly. These trends reflect a differential practice of abortion over the reproductive life-span. Abortions are more frequent during first pregnancies: 23.5% of the women terminated their first pregnancy and 21.5% terminated their second, whereas the proportion is not over 20% for third and subsequent pregnancies.

Figure 1

Age-specific fertility and abortion rates at Abidjan in the five years preceding the survey (per 1,000 women)

Age-specific fertility and abortion rates at Abidjan in the five years preceding the survey (per 1,000 women)

60Fertility rates, by contrast, have declined (Table 8). This trend is very distinct in the ten years preceding the survey (by between 10 and around 50 children per 1,000 women depending on the age group). For women aged 20-24, the decline occurred in the five years preceding the survey. For older women the fertility decline was continuous but of variable intensity according to the period.

61These results differ slightly from those of the DHS, which show a lower total fertility rate (3.4 children per woman at 15-44 years for the three years preceding the survey) and lower age-specific fertility rates. The DHS fertility rates are lower at all ages except 30-39 but the differences are particularly marked at 15-19, 25-29, and over 40 years (Table 8). These differences are partly explained by the higher proportion of married women in our study.

62The importance of abortion relative to fertility has thus increased over the last five years (Figure 2). For young women, 70 abortions were performed for 100 births in the recent period, a fact which underscores the role played by abortion in the timing of family-building. For women aged 20-24, the proportion of abortions is 45% for the most recent period, but an increase in the relative importance of abortion in recent years is also observed at these ages.

Figure 2

Ratio between the number of abortions and live births by age group of women, by five year periods preceding the survey (%)

Ratio between the number of abortions and live births by age group of women, by five year periods preceding the survey (%)

63The importance of abortions relative to live births is lower among women aged 25-39 and varies little: these women in the middle of reproductive life are in an active phase of their childbearing. For women at the end of their reproductive life abortions represent only 8% of births.

64The scale of these developments reveals the extent of the changes in behaviour; they show that abortion has an increasing role in the reproductive life of women, notably to delay childbearing. Abortion rates have in fact always been higher in the early reproductive years and this pattern is simply being reinforced. Even if the reporting of abortions can be assumed to have improved in the recent period, the scale of the increase observed reflects the greater recourse to abortion, which has been confirmed by other studies carried out in Abidjan (Desgrées du Loû et al., 1999).

How much of the fertility decline is attributable to abortion?

65What would the potential cumulated fertility of women in Abidjan be in the absence of abortion? One abortion does not avert one birth, taking into account the mean waiting time to conception , the gestation period (g_m) and the postpartum insusceptible period (t_m) (Leridon, 2002). The number of abortions N necessary to avert one birth is equal to:

66

67where E is the effectiveness of contraception, g_aand t_aare the durations, respectively, of gestation and of the postpartum insusceptible period in the case of abortion.

68The waiting time to conception is estimated at 4 months, the gestation period (g_m) is 9 months but is reduced to 3 months in the case of abortion (g_a). Data concerning the postpartum variables is not available in our survey but we assume that the duration for the observed population is close to that measured in Abidjan during the last DHS, i.e. a median postpartum insusceptible period (t_m) of 12.4 months; this duration is equal to 1 month in the case of abortion. The effectiveness of contraception (E) has been calculated from the prevalence of different types of contraceptive methods observed in our survey [6] and is 49%. The variable N therefore has a value of 2.48, which means that 2.48 abortions are necessary to avert one birth. When this model is applied to our data (Table 9) it shows that abortion reduced the total fertility rate of women aged 15-44 by 12%, equivalent to 0.6 children per woman.

Table 9

Impact of abortion on age-specific and total fertility rates, Abidjan, 1998

Age group Abortion (1) “Averted” fertility (2) = (1)/N Fertility (3) Potential fertility without abortion (4) = (2) + (3) Reduction in potential fertility from abortion (%) Age-specific rates (per 1,000) 15-19 89 36 125 161 – 22 20-24 73 29 160 189 – 16 25-29 45 18 194 212 – 9 30-34 37 15 150 165 – 9 35-39 28 11 113 124 – 9 40-44 5 2 65 67 – 3 Total (per woman) 15-44 1.4 0.6 4.0 4.6 – 12 * N = 2.48; 1 – E = 0.51. Source: FSU survey, 1998.

Impact of abortion on age-specific and total fertility rates, Abidjan, 1998

69The scale of this reduction varies according to age group: the fertility rate of women aged 15-19 is reduced by 22% and that of women aged 20-24 by 16%, which confirms the role of abortion in postponing childbearing for these young women. This impact is only 9% for women aged 25-39 and 3% at ages 40-44. The practice of abortion is therefore not constant over women’s lives, but occurs primarily in the early reproductive ages. What is the role of abortion compared with the other fertility determinants?

70To show the role of abortion and contraception in fertility changes, we have applied Bongaarts’ model (in Jolly and Gribble’s revised version) to the data from the two DHS surveys for Côte d’Ivoire as a whole and for the city of Abidjan, and to that from our survey in the Abidjan health centres. This model enables us to measure the fertility-inhibiting effects of the various determinants: postpartum variables (abstinence, amenorrhoea, and breastfeeding), nuptiality, contraception, primary sterility, and abortion. It will be recalled that the DHS for Côte d’Ivoire provide no data on abortion.

71According to the DHS data, postpartum infecundity has the largest role in the fertility decline since it accounts for 44% of the fertility decline in the country as a whole (Table 10). Its impact lessened in Abidjan between the two surveys, falling from 42% to 37%: the median duration of the postpartum unsusceptible period declined from 15 to 12.4 months, and the duration of breastfeeding also declined (Table 1). In 1998, based on a theoretical fertility of 15.3 children, the number of live births was reduced by 3.4 children in Abidjan through the effect of these postpartum variables. In the absence of back-up contraception, the shortening of the postpartum unsusceptible period will expose women to the risk of pregnancy and contribute to an increase in fertility.

Table 10

Inhibiting effects of nuptiality, contraception, postpartum insusceptibility, sterility, and abortion in Côte d’Ivoire during the 1990s

DHS 1994 DHS 1998-1999 FSU 1998 Abidjan (a) Whole country(b) Abidjan (a) Whole country(b) Abidjan (a) Adjusted marriage index C’m 0.66 0.74 0.61 0.70 0.75 Nonmarital births M0 1.12 1.13 1.18 1.13 1.15 Marriage index Cm 0.74 (26.4) 0.84 (16.5) 0.73 (27.7) 0.78 (21.9) 0.86 (14.2) Contraception index Cc 0.86 (13.7) 0.93 (6.6) 0.83 (16.6) 0.91 (9.5) 0.85 (15.0) Post-partum index Ci 0.58 (42.1) 0.56 (43.8) 0.63 (37.0) 0.56 (44.2) 0.63 (37.0)* Sterility index Ip 0.93 (7.4) 1.00 (0.5) 0.99 (1.0) 1.05 (– 4.5) 0.98 (2.0) Abortion index Ca n.a. n.a. n.a. n.a. 0.87 (13.2) TFR estimated from a cumulative fertility of 15.3 children 5.2 6.7 5.7 6.0 6.0 Difference: estimated TFR – observed TFR 1.1 1.0 2.3 0.8 1.6 Reading: The numbers in brackets indicate the percentage of inhibition. n.a. : not available. * Ci was calculated by applying data on postpartum insusceptibility from the 1998-1999 DHS in Abidjan since these data were not available in the FSU survey. (a) Author’s calculations. (b) Vimard et al., 2001. Sources: DHS 1994 and 1998-1999; FSU survey, 1998.

Inhibiting effects of nuptiality, contraception, postpartum insusceptibility, sterility, and abortion in Côte d’Ivoire during the 1990s

72The second most important factor in the fertility decline is nuptiality, which explains 22% of the reduction in the country as a whole and 20% in Abidjan in 1998-1999. The median age at first marriage has increased slightly, from 18.1 in 1994 to 18.7 in 1998-1999 for Côte d’Ivoire and from 19.4 to 20.2 for Abidjan, while the effect of nonmarital births on cumulative fertility has increased from 12% to 18%.

73Although its effect remains large compared with the other variables, contraception plays a lesser role in the fertility changes. The contraception index went from 0.86 to 0.83 in Abidjan, and from 0.93 to 0.91 in the country as a whole

74The difference between, on the one hand, the total fertility rate observed and, on the other, that estimated on the basis of a potential fertility of 15.3 children, minus the various fertility-inhibiting effects calculated, is particularly large in Abidjan where it increased from 1.1 to 2.3 children per woman between the two DHS, while it remained at around 1 child for the country as a whole (Table 10). These differences can be explained by the factors not included in this model, notably abortion.

75The FSU survey enables this role to be taken into account. To estimate its impact, we made the assumption that for the postpartum variables the data from the 1998 DHS for Abidjan applied to our results. Postpartum insusceptibility is obviously still the main determinant, reducing theoretical fertility by 37%. Contraception, marriage, and abortion make approximately equal contributions to fertility decline (between – 15% and – 13%). Based on a potential fertility of 15.3 children, the reductions in the number of births would be 1.06 from contraception, 0.99 from marriage and 0.91 from abortion. The reduction attributable to abortion would be larger if the abortion rates for all women were taken into consideration, rather than only for women in union: the number of live births averted would then reach 1.15.

76All in all, the study of fertility transition in Abidjan shows that women regulate their fertility in several ways. The observed population can be divided into four distinct groups depending on their recourse to contraception and/or abortion (Guillaume and Desgrées du Loû, 2002).

77A first group of women (27%) continue to use strategies that can be described as traditional, using neither contraception nor abortion, and attempting to control their childbearing only by means of postpartum practices (abstinence and breastfeeding). This first group comprises mainly married, poorly educated, Muslim women. But postpartum practices are being eroded in Abidjan, and the period of postpartum insusceptibility has contracted by 2.5 months between 1994 and 1998. In the long term, unless back-up contraception is introduced, these women are at risk of experiencing an increase in their fertility.

78By contrast, a second group of women (30%) use both contraception and abortion to avoid any unplanned or unwanted pregnancies. Most of these women are below 35, relatively highly educated, and Christians. The two methods are used in a complementary manner: abortion is resorted to after the failure of contraception, often a natural method, or the experience of an abortion is what motivates them to use contraception. This pattern of behaviour attests to the women’s desire to regulate their fertility using all available techniques.

79A third group is composed of women who have used only contraception during their reproductive life (39%). Their use of these methods may be effective and it is possible that they are opposed to abortion in the event of contraceptive failure. These women are generally young (the majority are under 25) and are represented at all educational levels.

80The last group (3%) includes women who have recourse only to abortion. They are mainly women at the beginning of their reproductive life, who are currently unable to take on the responsibility of a child. Too young to become mothers, they have great difficulty getting access to contraception and perhaps did not realize that they were at risk of becoming pregnant.

81This simple typology provides evidence of the growing determination of women to regulate their own fertility, but also highlights the force of traditional birth spacing practices.

Conclusion

82The reproductive behaviour of women in Abidjan appears to be evolving rapidly. The sizable decline in fertility results from the slow but continuous increase in contraceptive use and from the growing recourse to abortion. The varied patterns of recourse to contraceptive and abortive techniques reflect women’s changing reproductive ideals and their commitment to the idea of family planning but also their economic, family, and social situations that condition the demand for children. Women do not always enjoy full autonomy in their fertility decisions nor in the adoption of preventive methods. Nevertheless, the current context in Côte d’Ivoire offers a number of prospects for the future, with the adoption by the government of a policy to promote fertility control and the introduction of family planning activities in some health centres.

83As in many other African countries, access to birth control techniques in Côte d’Ivoire remains limited not least by an economic context which is adding to the difficulties of operating the health system. This is reflected in higher health costs for the population as a result of cost-covering policies. The cost of modern contraception can therefore appear as an obstacle to some women, encouraging them instead to use “natural” methods on the grounds that they are “free”.

84Over and above financial cost, the issue of information about contraception and access to the various methods raises the question of counselling, which is certainly insufficient, particularly during the postpartum period. And because information about contraception is distributed to women chiefly during prenatal or postnatal consultations, it cannot reach young single women and those who do not yet have any children. Access to family planning is especially difficult for adolescents, whose sexuality is not always socially accepted and for whom health personnel are sometimes reluctant to prescribe contraceptives.

85Without improvements in information and in the diffusion of family planning programmes, the use of modern methods will remain low, since women are still relatively resistant to methods about which they know little and whose side effects they fear. In general, fertility control in Africa still depends largely on natural and traditional methods (accounting for between 3% of methods used in Southern Africa up to 70% in Central Africa).

86For the fertility transition to be realized fully in East, Central, and West Africa, an increase of at least one percentage point per year in the use of modern contraception is needed, on condition that the other fertility determinants move in the same direction. Over the last ten years, contraceptive use (all methods combined) has increased by only 0.4% and 0.5% per year in Central and West Africa, respectively, and by 1.0% and 1.3% in Southern and East Africa (Guengant and May, 2001).

87In this context of low contraceptive coverage and as a result of the change in women’s aspirations regarding final family size, recourse to abortion is likely to continue and to be used as a substitute for contraception. Its role in the fertility transition, shown here for Abidjan, is increasingly put forward to explain the fertility declines observed in various African capital cities (Amegee et al., 2001; Konate et al., 1999). An in-depth analysis of the increase in abortions is particularly necessary in view of their consequences not only for fertility but also for maternal morbidity and mortality (Thonneau et al., 1996).

88Another factor likely to influence the fertility trend in Côte d’Ivoire is AIDS. The potential consequences of this epidemic are particularly serious in a city such as Abidjan where it is estimated that 15% of pregnant women are HIV-positive (Welffens-Ekra et al., 1996). Foetal mortality is higher in women infected by HIV (Welffens-Ekra et al., 2000; Zaba and Simon, 1998). These women are likely to modify their reproductive goals because of the risk of contaminating the children, whose mortality increases, and also because they fear that these children will be orphaned at an early age. The emergence of AIDS and the need to self-protect has also led to the emergence of new patterns of behaviour: use of preventive practices, reduced remarriage opportunities for widows, changes in postpartum abstinence and breastfeeding practices, changes in sexual behaviour, etc. (Guengant and May, 2001; Zaba and Simon, 1998). The campaigns in the fight against AIDS that have contributed to a more widespread use of the condom have at the same time increased its impact in preventing pregnancies. Condom use in Abidjan has doubled over the last five years. In some African countries, the condom accounts for more than 80% of the modern methods used by unmarried, sexually active adolescents (Guillaume, 2002). All these elements contribute to a change in reproductive ideals and behaviour, albeit one which remains hard to evaluate.

89Contraception, abortion, and HIV will certainly play a major role in fertility changes in Côte d’Ivoire, especially if postpartum practices tend to be eroded over time. Further research into the role of abortion is necessary to improve measurement of its impact on the general population and to understand the interactions between contraception and abortion so as to identify potential barriers to access to family planning.