The Impact of Contraceptive Use and Abortion on Fertility in sub-Saharan Africa: Estimates for 2003–2014

With the exception of sub-Saharan Africa, fertility has declined substantially in all regions of the developing world. The fertility decline, which started in Latin America and Asia in the 1960s, progressed rapidly through the decades that followed, reaching close to replacement level in these two regions (2.3 births per woman) in 2005–2010 (Bongaarts and Casterline 2013). An incipient decline in fertility in sub-Saharan Africa was first observed in the 1980s, but evidence has shown that the trajectory of decline in the region has not followed the pattern observed for other developing regions. Apart from the fact that the decline took longer to begin, its pace is also slower than in other regions, reaching an average of 5.1 births per woman in 2005–2010. Based on the annual change in the TFR from peak years for each region—1970–75 in Asia, 1975–80 in Latin America, and 1980s in sub-Saharan Africa—the pace of decline in the TFR from the onset of decline to 2005–2010 has been much lower in Africa (0.03 births per year) than in Asia (0.12 per year) and Latin America (0.13 per year) (ibid.). In addition, recent evidence shows that the decline has stalled in several countries (Bongaarts 2008; Ezeh, Mberu, and Emina 2009; Shapiro and Gebreselassie 2009). While a stalled fertility decline is not unique to Africa, the stall is occurring at much higher levels of fertility than is the case in countries with stalled fertility declines in other regions. For example, fertility in Bangladesh, Dominican Republic, Egypt, and Indonesia stalled at about 3 births per woman compared to more than 4 births in Ghana and Kenya and about 4 in Zimbabwe (Bongaarts 2008).

Previous research points to some important factors that help to explain the differences between the pattern of fertility decline in sub-Saharan Africa and the pattern in other developing regions. One factor is differential fertility preferences. Fertility surveys have consistently demonstrated preferences for larger families among women in sub-Saharan Africa than among their counterparts in other regions. For example, pre-transition fertility preferences were higher in sub-Saharan Africa than in Asia and Latin America (Bankole and Westoff 1995;Westoff 1991). Another factor is lower contraceptive use and higher unmet need for contraception in sub-Saharan Africa relative to the other two regions, a situation that may be associated with lower knowledge of and limited access to contraception and weaker family planning program efforts in sub-Saharan Africa (FP2020 2015; Ross and Smith 2011; Sedgh and Hussain 2014).

Fertility transition in sub-Saharan Africa has been relatively slow, with the TFR declining from 6.5 births per woman in 1950–1955 to 5.4 in 2005–2010 (United Nations 2013b). The proportion of all women aged 15–49 in the region who want to avoid pregnancy increased from 39 percent to 42 percent over the years 2003–2012, and the proportion using modern contraceptive methods among women who want to avoid pregnancy rose from 32 percent to 40 percent over the same period (Darroch and Singh 2013). Increased use of contraception is even larger when a longer time period is considered. The proportion of married women using any method of contraception rose from 17 percent in 1990 to 31 percent in 2010 (Alkema et al. 2013).

In this chapter we focus on two proximate determinants of fertility—contraception and induced abortion—and their effect on fertility levels. We focus on use of modern methods because of the high failure rates of traditional methods (Ali, Cleland, and Shah 2012; Trussell 2011)and because, while some women using traditional methods may choose to use these methods, such choices often imply that women perceive other options to be unavailable or are not fully informed of contraceptive options (Wang et al. 2012). Accessible, quality information and services would likely alter the choices available, enabling women to choose from the full range of methods, modern and traditional, to most effectively avoid pregnancy. It is important to note, however, that some of our estimates incorporate the contribution of traditional method use in preventing unintended pregnancy.

Framework, data sources, and estimates

Our estimates are for all women of reproductive age (15–49) in sub-Saharan Africa for two points in time—2003 and 2014—using comparable methodology and data sources. These estimates build on a model that underlies earlier studies on the impact of contraceptive use on averting unintended pregnancies and the costs and benefits of meeting the need for contraception (Darroch and Singh 2011; Darroch et al. 2016; Singh et al. 2004; Singh et al. 2009; Singh and Darroch 2013; Singh, Darroch, and Ashford 2014; Vlassoff et al. 2004).

Framework

Our framework examines the impact of contraception on unintended pregnancy and estimates the distribution of unintended pregnancies by outcome—unplanned births, abortions, and miscarriages that result from unplanned pregnancies. This allows us to estimate the impact of contraception on the fertility rate (through preventing unplanned births) and the impact on fertility of induced abortions that would otherwise have resulted in births. Use of contraception would reduce the unintended pregnancy rate from the level it would have attained if women wanting to avoid pregnancy used no contraception; and changes in contraceptive use would affect the unintended pregnancy rate and, through this, unplanned births, abortion, and overall fertility. Lacking other information that a change in the level of modern contraceptive use would differentially affect pregnancy outcomes, we assume that any increase or decrease in unintended pregnancies would have the same proportional distribution by pregnancy outcome as do current unintended pregnancies.

First, we present two sets of estimates that relate to the impact of levels and trends in contraceptive use and abortion: (a) the impact of current levels of modern contraceptive use on unintended pregnancy, unplanned births, and abortion for the years 2003 and 2014; and (b) the impact of change in modern contraceptive use over the period 2003–2014 on change in the unintended pregnancy rate and its component outcomes. In both cases, the impacts are estimated in terms of absolute numbers of unintended pregnancies and their outcomes and as rates per 1,000 women aged 15–49. Second, we present results from two hypothetical scenarios, namely, meeting the existing need for effective modern contraception and assuming that current users adopt a more effective mix of modern methods.

Data sources

Estimates of the numbers of women aged 15–49 in 2003 and 2014 are taken from the United Nations (UN) World Population Prospects 2012 estimates and projections (medium fertility variant) (See Appendix Table 1)¹ (United Nations n.d.a.). The proportions of women of reproductive age who were currently married or in union in each year are based on annual estimates by the UN Population Division (United Nations n.d.b.). The distribution of unmarried (not in union) women into formerly married and never-married groups is based on the most recent UN estimates or national surveys prior to the years 2003 and 2014 (See Appendix Table 2) (United Nations 2013a). These and other changes in population composition may influence the pregnancy rate and pregnancy outcomes. For example, if the proportion of women aged 15–49 who are younger (with higher fecundity) increased between 2003 and 2014, this compositional change may increase the pregnancy rate, all other things remaining equal. The estimated proportions of women unmarried increased in each sub-region of sub-Saharan Africa between 2003 and 2014; this increase will likely reduce pregnancy and fertility rates, but it may be accompanied by an increase in the abortion rate, particularly if sexual activity among unmarried women remains the same or increases. While we present absolute numbers of pregnancies and pregnancy outcomes, these are influenced by changes in the number of women of reproductive age, and the rate per 1,000 women of reproductive age is a better measure of the size of the impact of contraception and abortion on fertility. The rates presented are expressed as the number of events per 1,000 women aged 15–49. We refer to these rates as the General Pregnancy Rate (GPR), General Fertility Rate (GFR), General Abortion Rate (GAR), and General Unintended Pregnancy Rate (GUPR).

Estimates of pregnancies by outcome

Pregnancy estimates are calculated using multiple sources. Birth data were taken from United Nations estimates (United Nations n.d.c.). We used the most recent estimates of unintended pregnancies and their composition (by birth, abortion, and miscarriages) available by sub-region for the years 2003 and 2014 (Sedgh, Singh, and Hussain 2014; Singh et al. 2004). The estimates of unintended pregnancy used here incorporate the most recent estimates of abortion for sub-Saharan Africa available at the time analysis was conducted. The abortion rate for Southern Africa is substantially lower than that for other subregions, affecting results presented. It appears from subsequent research that the abortion rate for Southern Africa in particular was previously underestimated (Sedgh et al. 2016), and this affects the estimated impact of contraceptive use on unintended pregnancy and abortion, a limitation that should be borne in mind in interpreting findings. For both 2003 and 2014, we assumed that the sub-regional estimates of the proportions of births that are unintended apply to country-level births in each sub-region, as was the case for the sub-regional ratios of induced abortions to unintended births. Miscarriages are estimated to include pregnancies that end in spontaneous abortion or stillbirth after lasting long enough to be noted by the woman (from 6 to 7 weeks after the last menstrual period based on existing syntheses of clinical studies) (Bongaarts and Potter 1983; Harlap, Shiono, and Ramcharan 1980).

Estimates of use of contraception and unmet need

Nationally representative surveys are the principal source of data on women’s use of and need for contraception. These include Demographic and Health Surveys (DHS), US Centers for Disease Control and Prevention Reproductive Health Surveys (RHS), United Nations Children’s Fund (UNICEF) Multiple Indicator Cluster Surveys (MICS), and other independent surveys.

For 2003, some survey data on contraceptive need and use were available for countries covering 81 percent of currently married women aged 15–49 in sub-Saharan Africa, including 91–96 percent in Eastern, Southern, and Western Africa, but only 32 percent in Middle Africa. Coverage in 2014 was almost complete. Full contraceptive need and use data were available for sub-Saharan countries covering 98 percent of currently married women aged 15–49 and partial data for the remaining countries. Full data covered 96–100 percent of married women in Eastern, Southern, and Western Africa, and 86 percent in Middle Africa. For 2003 and 2014 there is essentially no overlap in data sources used for the estimates.

The same methodology is used to develop estimates of use of and need for contraception among women aged 15–49 for both years. Separate estimates of the distribution of women aged 15–49 by need and method use were made for each country according to women’s marital status (currently, never, or formerly married/in union). These proportions were applied to numbers of women in each marital status group in the reference year. Country estimates are summed to obtain results by region and sub-region. Where data for a country are missing, we use weighted sub-regional averages, or data from a demographically or socioeconomically comparable country, or information from earlier surveys. In 2003, all of Sudan was included in the sub-Saharan Africa region. To be consistent, the 2014 tabulations presented here include both South Sudan and Sudan.

We used the revised method for calculating unmet need (Bradley et al. 2012) in 2014 for most countries (those with DHS and MICS surveys), and also used this method to update earlier published estimates for 2003. Women were classified into the following groups:

• —Those using modern methods of contraception.²
• —Those with an unmet need for modern contraception. This group comprises women not using a modern method who report (a) using a traditional method,³ or (b) not using any contraceptive method who are married (or unmarried and had sex in the three months before interview), fecund, and want to wait two or more years to give birth or want no more children, or (c) being pregnant with a pregnancy they had wanted later or not at all, or in postpartum amenorrhea after such a birth that they had wanted at least two years later or not at all. These two groups—users of modern methods and those with unmet need for modern contraception—comprise women who want to avoid pregnancy.

For sexual activity among unmarried women, the DHS definition includes those who were sexually active in the 30 days before the survey, while we include those who were sexually active in the last three months, to take account of the generally sporadic pattern of sexual activity among unmarried women. Our classification of group (c) above as having unmet need is consistent with the standard DHS approach to calculating unmet need, and is based on the recognition that need for contraception is best measured over a recent period of time rather than as a point-in-time estimate; the women in group (c) in fact had unmet need for effective contraception at the time of conception of their recent pregnancy. Those not in need of contraception include women who want a birth within the next two years, as well as women who are currently pregnant with an intended pregnancy or postpartum amenorrheic from an intended pregnancy or infecund or unmarried and not sexually active in the three months before the interview.

Estimates of contraceptive use-failure and no-method pregnancy rates

As described above, we estimated the total number of unintended pregnancies by outcome, in each country in 2003 and 2014 from external sources that provide sub-regional estimates of the number of unintended pregnancies as well as their distribution by outcome (unintended births, induced abortions, and miscarriages).

We calculated another estimate for each country by applying 12-month use-failure rates (Ali, Cleland, and Shah 2012; Trussell 2011) and an initial pregnancy rate of 40 percent for women wanting to avoid pregnancy who use no method (Darroch and Singh 2011) to the numbers of women estimated to be using each method or no method while wanting to avoid pregnancy (see Appendix Table 3). These initial method-specific use-failure rates and the pregnancy rate for nonusers in need for each country were adjusted by the ratio of the number of unintended pregnancies (from external sources, as described above) to the number estimated from method-use and use-failure rates, so that the two estimates of total unintended pregnancies are equal (See Appendix Table 4). The adjusted use-failure rates were used in estimating the impacts of the various scenarios of contraceptive method use presented below.

These adjustment ratios ranged from 0.828 to 1.068 across sub-regions of sub-Saharan Africa in 2014, except for Western Africa where the adjustment ratio was 0.630. Initial use-failure rates and the pregnancy rate for nonusers wanting to avoid pregnancy were adjusted by the sub-regional and focus year ratios. For example, for sub-Saharan Africa as a whole, the initial nonuser pregnancy rate was 40 percent. An adjustment ratio of 85.4 percent reduced this initial value to 34 percent. It is worth noting that the initial and adjusted pregnancy rates for nonusers differ from the 85 percent value commonly cited (Trussell 2011). The 85 percent level is an estimate for women who stop contraceptive use to become pregnant, not for a general population of women using no method who want to avoid pregnancy. Further, using the DHS definition, nonusers in need include women who are pregnant or postpartum following an unintended conception, whose pregnancy rate over a year’s time would be quite low.

Estimating the impact for different scenarios

We estimate the impact of contraceptive use by women in need on unintended pregnancies, and through this on unplanned births and induced abortions, and ultimately on General Pregnancy, Fertility, and Abortion Rates, by using three different scenarios for 2014. The scenarios are shown in Figure 1. Scenario 1 assumes that all current users of modern contraceptive methods use no method; Scenario 2 assumes that all current users of both modern and traditional methods use no method; and Scenario 3 assumes that all current users use no method and there are no induced abortions. These scenarios are counterfactuals and are used to calculate the fertility-reducing impact of current patterns of contraception and induced abortion. They are not intended to represent potential behavioral changes in the region.

FIGURE 1.

Scenarios used to estimate the impacts of current (2014) contraceptive use and current levels of abortion through counterfactual assumptions, in sub-Saharan Africa

In addition, we also estimate the impact of two other hypothetical scenarios that include meeting 100 percent of the existing need for contraception and increasing the effectiveness of the method mix (while holding abortion incidence constant) (see Figure 2). Scenario 4 assumes that all women with unmet need for modern contraception begin to use contraception and use the same mix of modern methods used by current users in their own country; current users of modern methods continue to use their methods. Scenario 5 assumes that all current users of reversible modern methods and all women with unmet need will use highly effective methods (injectables, implants, and IUDs).

FIGURE 2.

Scenarios used to estimate the potential impact of improvements in contraceptive use levels and method mix, in sub-Saharan Africa

^aInjectables, implants, and IUDs.

The total number of unintended pregnancies in each scenario is calculated by applying the adjusted method-specific failure rates to the method-use distributions. We assumed that the distribution of outcomes of unintended pregnancies in these different scenarios would be the same in each country as the outcome for all unintended pregnancies occurring in the sub-region. We present results for sub-Saharan Africa and for each sub-region. The estimates are for women aged 15–49 for all union statuses combined.

Results

Trends in pregnancy, fertility, and abortion

In 2014, 220 per 1,000 women aged 15–49 in sub-Saharan Africa (SSA) were pregnant, with the General Pregnancy Rate (GPR) ranging from 110 per 1,000 women in Southern Africa to 251 per 1,000 in Middle Africa (Table 1). Over the 11-year period covered by this analysis, the GPR declined in SSA and most of its sub-regions. The rate for SSA declined by 1.2 points per year (or by 0.5 percent per year). The greatest declines were in Middle and Eastern Africa, with a drop of about 2.4–2.6 points in the GPR per year. In West Africa, the changes in the GPR and GFR were minimal and in opposite directions, and essentially there was no trend in these rates over the 11-year period. In Eastern, Middle, and Southern Africa, the GPR fell because of decreases in the intended GPR while the unintended GPR rates rose slightly. In Western Africa, in contrast, the intended GPR rose and the unintended GPR dropped.

TABLE 1.

Numbers (000s) and rates per 1,000 women aged 15–49 of pregnancies, births, abortions, and miscarriages, by intention status, in sub-Saharan Africa and sub-regions, 2003 and 2014

Geographic area	Pregnancies (GPR)						Births (GFR)						Induced abortions (GAR)		Miscarriagesc
	Total		Intendeda		Unintendedb		Total		Intendeda		Unintendedb
	2003	2014	2003	2014	2003	2014	2003	2014	2003	2014	2003	2014	2003	2014	2003	2014
Numbers (000s) of events
Sub-Saharan Africa	38,720	49,890	25,210	31,830	13,510	18,060	28,190	36,130	21,010	26,530	7,180	9,600	4,450	5,940	6,080	7,820
Eastern Africa	15,510	20,020	9,940	11,250	5,570	8,780	11,180	14,110	8,280	9,370	2,900	4,740	1,900	2,800	2,430	3,110
Middle Africa	6,380	8,090	4,540	4,970	1,850	3,120	4,770	5,800	3,780	4,140	990	1,660	600	1,020	1,010	1,270
Southern Africa	1,670	1,790	810	800	860	990	1,210	1,290	680	660	530	630	200	220	260	280
Western Africa	13,680	18,220	8,990	13,560	4,690	4,660	9,930	13,630	7,490	11,300	2,440	2,330	1,600	1,700	2,150	2,890
Events per 1,000 women aged 15–49
Sub-Saharan Africa	233	220	151	140	80	80	169	159	126	117	43	42	27	26	37	35
Eastern Africa	247	221	158	124	89	97	178	156	132	104	46	52	30	31	39	34
Middle Africa	280	251	199	154	81	97	209	180	166	129	43	52	26	32	45	39
Southern Africa	115	110	56	49	59	61	83	79	47	41	37	38	14	13	18	18
Western Africa	235	232	155	173	81	59	171	174	129	144	42	30	28	22	36	36

NOTES: Estimates for sub-Saharan Africa include Sudan, which is in Northern Africa. Numbers may not add up to totals because of rounding.

^aIntended pregnancies include intended births and miscarriages of intended conceptions (not shown separately).

^bUnintended pregnancies include unintended births, induced abortions, and miscarriages of unintended conceptions (not shown separately).

^cTotal miscarriages, including those from intended and unintended pregnancies.

Fertility was also high in 2014, with a General Fertility Rate (GFR) of 159 births per 1,000 women aged 15–49 in SSA. There was wide variation in the GFR across sub-regions: the rate was highest in Middle Africa (180 per 1,000 women) and lowest in Southern Africa (79 per 1,000). Fertility declined noticeably between 2003 and 2014, with a drop of 0.9 points in the GFR per year for the whole region. The decline was largest in Eastern and Middle Africa with a drop of about 2.0 and 2.6 rate points per year, respectively, and was much smaller in Southern Africa (a drop of 0.4 points in the GFR per year); and in West Africa, the GFR increased slightly, by 0.3 points per year.

The General Abortion Rate (GAR) was moderately high in 2014, at 26 per 1,000 women aged 15–49 in SSA. The rate varied widely across subregions, ranging from 31 and 32 per 1,000 in Eastern and Middle Africa, respectively, to 13 per 1,000 in Southern Africa. Unlike for pregnancy and birth rates, there was no systematic pattern in the trend in the GAR across sub-regions. The GAR increased in Middle Africa and declined in Western Africa between 2003 and 2014, but changed very little in Eastern and Southern Africa.

Trends in contraceptive use

Contraceptive use remained low in sub-Saharan Africa in 2014 (Table 2) (Darroch 2013). The proportion of women aged 15–49 wanting to avoid pregnancy increased between 2003 and 2014 in Eastern and Middle Africa, by 5 and 7 percentage points, respectively. There was little change in Southern and Western Africa. The proportion of women using any method in SSA was only 21 percent in 2014. Use of any method was highest in Southern Africa (48 percent) and lowest in Western and Middle Africa (16 percent and 17 percent, respectively). Between 2003 and 2014, only Eastern Africa showed a noticeable increase in use of any method, with an increase of 8 percentage points, or a 4.5 percent increase per year. This increase was greater than the increase in the percent of women 15–49 wanting to avoid pregnancy, reflecting progress toward meeting overall demand for contraception.

TABLE 2.

Percent of all women aged 15–49 who want to avoid pregnancy, percent using any contraceptive method, percent using modern methods, percent using traditional methods, and percent using no method, in sub-Saharan Africa and sub-regions, 2003 and 2014

Geographic area	Want to avoid pregnancya		Using modern methods		Unmet need for modern methods
					Using traditional methods		Using no method
	2003	2014	2003	2014	2003	2014	2003	2014
Sub-Saharan Africa	39	41	12	17	6	4	20	20
Eastern Africa	39	44	12	20	4	4	23	20
Middle Africa	35	42	6	9	12	8	17	25
Southern Africa	63	62	47	47	1	0	15	14
Western Africa	35	36	8	11	7	5	21	20

NOTES: Estimates for sub-Saharan Africa include Sudan, which is in Northern Africa. Numbers may not add up to totals because of rounding. Modern methods include female and male sterilization, hormonal implants, IUDs, hormonal pills and injectables, male and female condoms, and other supply methods. Traditional methods include periodic abstinence, the lactational amenorrhea method (which includes exclusive breastfeeding for up to six months postpartum), withdrawal, and folk methods.

^aSee text for definition.

Only 17 percent of women aged 15–49 in SSA were using a modern method in 2014. Across the sub-regions, this proportion varied widely, ranging from 9 percent in Middle Africa to 47 percent in Southern Africa. Trends in use of modern contraception between 2003 and 2014 were similar to those observed for use of any method, although slightly greater increases occurred in levels of modern method use. For example, while use of any method in SSA increased by 3 percentage points between 2003 and 2014, use of modern contraception increased by 5 percentage points.

For SSA as a whole, the proportion of women aged 15–49 using a traditional method was 4 percent. This proportion declined from 12 percent to 8 percent between 2003 and 2014 in Middle Africa, and was 5 percent or lower in the other sub-regions in 2014.

The percent of women aged 15–49 who wanted to avoid pregnancy but were using no contraceptive remained at 20 percent in SSA between 2003 and 2014 and changed little in Southern Africa (14 percent in 2014) and in Western Africa (20 percent in 2014). In contrast, the proportion using no method declined in Eastern Africa from 23 percent in 2003 to 20 percent in 2014, as modern contraceptive use increased at a greater pace than the percent of women wanting to avoid pregnancy. In Middle Africa, the percent using no method rose from 17 percent in 2003 to 25 percent in 2014; while modern method use rose 3 percentage points between 2003 and 2014, this increase was not enough to make up for the 4 percentage point decline in traditional method use or the large increase in desire to avoid pregnancy.

In summary, over the 11 years covered by this study, use of contraception increased slightly in sub-Saharan Africa and increased moderately in Eastern Africa. In all cases, the increase was due almost entirely to an increase in use of modern methods. In Eastern Africa the increase in modern contraceptive use outpaced the rising proportion of women wanting to avoid pregnancy, but in Middle Africa the increase in modern method use was smaller and not enough to keep up with declining traditional method use or the increased proportion of women wanting to avoid pregnancy.

Impact of modern contraceptive use on unintended pregnancies and births

We attempt to quantify the size of the impact of modern contraceptive use on unplanned pregnancies and births in sub-Saharan Africa in the past decade. We estimate the extent to which use of modern contraception reduces the incidence of unintended pregnancy and its component parts in 2003 and 2014. We also examine potential changes in the number and rates of pregnancies and births based on hypothetical changes in the level of contraceptive use and method mix in 2014.

For SSA in 2003, use of modern methods prevented about 6.3 million unintended pregnancies (Table 3, upper panel). The vast majority of these unintended pregnancies (56 percent in SSA and 52–62 percent across sub-regions) would have resulted in unplanned births. About 30 percent of unintended pregnancies in SSA and 23–34 percent in the sub-regions would have ended in induced abortion, while the remainder would have resulted in miscarriages.

TABLE 3.

Numbers (000s) and rates per 1,000 women aged 15–49 of unintended pregnancies prevented by modern contraceptive use, by pregnancy outcome, in sub-Saharan Africa and sub-regions, 2003 and 2014

Geographic area	Prevented by modern contraceptive use in 2003				Prevented by modern contraceptive use in 2014
	Total unintended pregnancies	Unplanned births	Induced abortions	Miscarriages	Total unintended pregnancies	Unplanned births	Induced abortions	Miscarriages
Numbers (000s) of events
Sub-Saharan Africa	6,340	3,520	1,920	900	12,550	6,920	3,860	1,770
Eastern Africa	2,380	1,240	810	330	7,220	3,910	2,290	1,010
Middle Africa	430	230	140	60	790	420	260	110
Southern Africa	2,110	1,310	490	310	2,460	1,550	540	360
Western Africa	1,330	690	450	180	1,980	990	720	270
Events per 1,000 women aged 15–49
Sub-Saharan Africa	38	21	12	5	55	31	17	8
Eastern Africa	38	20	13	5	80	43	25	11
Middle Africa	19	10	6	3	25	13	8	3
Southern Africa	146	90	34	21	151	95	33	22
Western Africa	23	12	8	3	25	13	9	3

NOTES: Estimates for Sub-Saharan Africa include Sudan, which is in Northern Africa. Numbers may not add up to totals because of rounding.

In 2014, use of modern contraception in SSA was estimated to have prevented about 12.6 million unintended pregnancies. Similar to the findings for 2003, 55 percent of these unintended pregnancies would have resulted in unplanned births and 31 percent in induced abortions. The proportion associated with each of the components at the sub-regional level is similar to that observed for 2003: for example, the proportions that would have ended in unplanned births in 2014 were highest in Southern Africa (63 percent) and lowest in Western Africa (50 percent).

Expressing the number of unintended pregnancies prevented by use of modern contraception in terms of rates allows us to identify where the impact of modern method use is greatest across the sub-regions (Table 3, lower panel). For example, in 2003, use of modern contraception averted 38 unintended pregnancies per 1,000 women aged 15–49 in SSA. The impact was highest in Southern Africa, where contraceptive use is also much higher than in other sub-regions (146 pregnancies are averted per 1,000 women). The lowest impact was observed in Middle Africa, where use of modern contraception averted just 19 unintended pregnancies per 1,000 women. In terms of unplanned births prevented by use of modern methods in 2003, the rate for SSA was 21 per 1,000 women aged 15–49 and, not surprisingly, the size of the impact across sub-regions is similar to that observed for unplanned pregnancies. By 2014, while Southern Africa remained the sub-region with the largest impact of modern contraceptive use on unplanned pregnancies and births prevented, with rates of 151 and 95 per 1,000 women, respectively, Eastern Africa took the second place with rates of 80 unintended pregnancies and 43 unplanned births per 1,000 women averted by use of modern methods.

Impact of changes in contraceptive method mix on unintended pregnancies and births

What would be the impact on unintended pregnancy and its components in 2014 if the contraceptive method mix in 2003 was observed in 2014, all else being equal? If the contraceptive method-use pattern of women wanting to avoid pregnancy in 2014 was the same as observed for 2003 (a less effective method mix), the number of unintended pregnancies would have increased by 2.4 million (Table 4, top panel). As a result, the General Unintended Pregnancy Rate (GUPR) in 2014 would have risen by 11 per 1,000 women aged 15–49 (Table 4, lower panel). Eastern Africa would have recorded the highest increase in this rate (by 26 per 1,000 women). On the other hand, a decline in the GUPR of 6 per 1,000 women would have occurred in Middle Africa. Given this contraceptive use scenario, the number and rate of unplanned births for SSA in 2014 would have increased by 1.3 million and 6 per 1,000 women aged 15–49, respectively. Changes similar to those for pregnancy rates would have occurred for the sub-regions, with the increase in the GFR being highest for Eastern Africa (by 14 per 1,000 women).

TABLE 4.

Numbers (000s) and rates per 1,000 women aged 15–49 of unintended pregnancies in 2014, and numbers and rates prevented by 2003–2014 change in method use among women wanting to avoid pregnancy, by pregnancy outcome, in sub-Saharan Africa and sub-regions

Geographic area	Pregnancies among women wanting to avoid pregnancy, 2014				Pregnancies prevented among women wanting to avoid pregnancy in 2014 by change from 2003 method-use distribution
	Total unintended pregnancies	Unplanned births	Induced abortions	Miscarriages	Total unintended pregnancies	Unplanned births	Induced abortions	Miscarriages
Numbers (000s) of events
Sub-Saharan Africa	18,060	9,600	5,940	2,510	2,440	1,330	770	340
Eastern Africa	8,780	4,740	2,800	1,230	2,330	1,280	720	330
Middle Africa	3,120	1,660	1,020	430	−180	−100	−60	−30
Southern Africa	990	630	220	150	10	4	1	1
Western Africa	4,660	2,330	1,700	640	290	140	100	40
Events per 1,000 women aged 15–49
Sub-Saharan Africa	80	42	26	11	11	6	3	2
Eastern Africa	97	52	31	14	26	14	8	4
Middle Africa	97	52	32	14	−6	−3	−2	−1
Southern Africa	61	38	13	9	0	0	0	0
Western Africa	59	30	22	8	4	2	1	0

NOTES: Estimates for sub-Saharan Africa include Sudan, which is in Northern Africa. Numbers may not add up to totals because of rounding. Negative numbers indicate that the 2003–2014 change in method use among women wanting to avoid pregnancy resulted in an increase over the actual 2014 number; all other numbers indicate a decrease.

Impact of current levels of contraceptive use and induced abortion on pregnancies and births

In Table 5, we present three scenarios that estimate the impact of contraceptive use and induced abortion on pregnancy and fertility in 2014. As discussed above, Scenarios 1, 2, and 3 apply hypothetical assumptions to calculate the fertility-reducing impact of the use of contraception and induced abortion in that year, and are not meant to represent potential behavioral changes in the region.

TABLE 5.

Numbers (000s) and rates per 1,000 women aged 15–49 of additional unintended pregnancies and unplanned births that would occur under three hypothetical scenarios regarding modern and traditional contraceptive use and induced abortion in sub-Saharan Africa and sub-regions, 2014

Geographic area	Numbers and rates of additional unintended events that would occur under:
	Scenario 1: Modern method users used no method; abortion, traditional, and no-method use unchanged				Scenario 2: Modern and traditional method users used no method; abortion and no-method use unchanged				Scenario 3: Modern and traditional users used no method; no induced abortion
	Unintended pregnancies	Unplanned births	GPR	GFR	Unintended pregnancies	Unplanned births	GPR	GFR	Unintended pregnancies	Unplanned births	GPR	GFR
Sub-Saharan Africa	12,550	6,920	55	31	14,380	7,890	63	35	14,380	17,430	63	77
Eastern Africa	7,220	3,910	80	43	7,990	4,330	88	48	7,990	9,230	88	102
Middle Africa	790	420	25	13	1,270	380	39	21	1,270	1,990	39	62
Southern Africa	2,460	1,550	151	95	2,460	1,550	152	96	2,460	2,250	152	139
Western Africa	1,980	990	25	13	2,530	1,260	32	16	2,530	3,670	32	47

NOTES: Estimates for sub-Saharan Africa include Sudan, which is in Northern Africa. Numbers may not add up to totals because of rounding. General Pregnancy Rate (GPR) and General Fertility Rate (GFR) are measured as rates per 1,000 women aged 15–49.

Scenario 1—No modern method use in 2014

We first explore the impact of a scenario in which there was no use of modern methods in 2014, instead of the contraceptive use pattern that existed in that year. Modern contraceptive users are assumed to be using no method and only users of traditional methods had some contraceptive protection. We also assume that induced abortion remains at its 2014 level. According to this scenario, the number of pregnancies in SSA in 2014 would have increased by 12.6 million, implying an increase in the General Pregnancy Rate (GPR) of 55 per 1,000 women aged 15–49 (Table 5). This hypothetical change would have resulted in an increase in pregnancy rates for all sub-regions. The increase in the GPR would have been highest in Southern Africa at 151 per 1,000 women and lowest in Middle and Western Africa (25 per 1,000). This scenario would result in 6.9 million unplanned births in SSA. The GFR would also increase by 31 per 1,000 women. The change in the GFR would be positive for all subregions, ranging from an increase of 13 births per 1,000 women in Western and Middle Africa to an increase of 95 births per 1,000 women in Southern Africa.

Scenario 2—No modern or traditional method use in 2014

In the second scenario, all current users of contraception (modern and traditional) are assumed to be using no method, and current non-users remain in the same status, as does abortion incidence. The number of pregnancies in SSA would increase by 14.4 million or 63 per 1,000 women aged 15–49 (Table 5). This increase in the GPR would be evident in all sub-regions, with Southern Africa exhibiting the highest increase (152 per 1,000 women), while the lowest increase (32 per 1,000) would be in Western Africa. The GFR would increase by 35 per 1,000 in SSA, with the increase ranging from 16 per 1,000 women in Western Africa to 96 per 1,000 in Southern Africa.

Scenario 3—No modern or traditional method use and no induced abortion in 2014

According to this scenario, the GPR would remain at the level observed under the second scenario (because the only change between Scenarios 2 and 3 is that there is no induced abortion, which is already included in pregnancies). So, as with Scenario 2, the GPR in SSA would increase by 63 per 1,000 women. However, in the absence of abortion, the GFR would increase by 77 per 1,000 women (Table 5). This scenario would therefore result in an additional increase of about 42 births per 1,000 women compared to Scenario 2, where there was no contraceptive use but there was induced abortion. In effect, in the absence of induced abortion in 2014, the GFR in SSA would increase by an additional 42 per 1,000. This additional impact would be smallest in Western Africa (31 per 1,000) and largest in Eastern Africa (54 per 1,000).

Impact of meeting unmet need for contraception and a more effective method mix

In Table 6 we show the potential impact on pregnancies and births of two scenarios in which all existing need for modern contraceptive methods is met. In Scenario 4 current users of modern methods and all other women with unmet need for modern methods use modern methods, and women use the same mix of modern methods as users in their country with similar marital status and intention to space or stop childbearing. Scenario 5 estimates the impact of a hypothetical situation in which all women with unmet need for modern contraception and all users of reversible modern methods use a highly effective mix of long-acting reversible contraceptives (LARCs)—injectables, implants, and IUDs—instead of their country’s mix of modern methods.

TABLE 6.

Numbers (000s) and rates per 1,000 women aged 15–49 of unintended pregnancies and unplanned births that would be prevented under two hypothetical scenarios of meeting all unmet need for modern methods and improving the mix of modern reversible methods, in sub-Saharan Africa and sub-regions, 2014

Geographic area	Scenario 4: All women wanting to avoid pregnancy use modern methods with method mix similar to that of current users in their countrya				Scenario 5: All women wanting to avoid pregnancy use LARC methodsb
	Unintended pregnancies	Unplanned births	GPR	GFR	Unintended pregnancies	Unplanned births	GPR	GFR
Sub-Saharan Africa	15,140	8,040	67	35	16,870	8,960	74	40
Eastern Africa	7,560	4,080	84	45	8,160	4,410	90	49
Middle Africa	2,540	1,360	79	42	2,950	1,580	92	49
Southern Africa	720	450	44	28	870	550	54	34
Western Africa	3,880	1,940	49	25	4,390	2,190	56	28

NOTES: Estimates for sub-Saharan Africa include Sudan, which is in Northern Africa. General Pregnancy Rate (GPR) and General Fertility Rate (GFR) are measured as rates per 1,000 women aged 15–49.

^aAssuming women with unmet need adopt method mix of women of similar marital status and desire to space or limit childbearing.

^bAssuming the initial, unadjusted average failure rate of the mix of long-acting reversible contraceptives (LARCs) (injectables, implants, IUDs) is 1.5 percent per year; users of permanent methods (sterilization) do not switch.

Scenario 4—All women with unmet need for modern methods in 2014 use modern methods (the same method mix as similar current users of modern methods)

Under this scenario, 15.1 million unintended pregnancies, including just over 8 million births, would be prevented in SSA (Table 6). This is in addition to unintended pregnancies already prevented by current users of modern methods in 2014 (12.6 million, Table 3). This would reduce the current GPR by 67 per 1,000 women aged 15–49. The greatest impact would be seen in Eastern and Middle Africa (reductions of 84 and 79 per 1,000 women), given the high level of unmet need for modern methods in these two subregions. The impact would also be large in Southern and Western Africa (reductions of 44 and 49 per 1,000 women). The impact on the GFR would also be substantial: reductions of 42–45 in Middle and Eastern Africa and 25–28 in Western and Southern Africa.

Scenario 5—All women with unmet need for modern methods in 2014 and all users of reversible modern methods in 2014 use a mix of highly effective methods (injectables, implants, and IUDs); women using sterilization continue to do so

In this scenario, we add an improved or more effective mix of modern methods (LARCs) to the assumption in Scenario 4. This would result in a further reduction in unintended pregnancies and births. Under this scenario, 16.9 million unintended pregnancies would be prevented (1.7 million more than under Scenario 4), as would nearly 9 million unplanned births (900,000 more than under Scenario 4). The GPR would decline by 74 per 1,000 women annually in SSA (compared to 67 per 1,000 under Scenario 4), and the GFR would decline by 40 per 1,000 (compared to 35 per 1,000 under Scenario 4)—small but still notable additional impacts. As in Scenario 4, Eastern and Middle Africa would see larger impacts than Southern and Western Africa.

Impact of current contraceptive use combined with meeting all unmet need and improving method mix

The preceding analyses suggest that current contraceptive use has a substantial impact in preventing unintended pregnancies and unplanned births. If there were no use of modern methods in 2014 (first panel of Table 7), the numbers and rates of pregnancies and births would have been substantially higher than they were with the current levels of use shown in panel 2 (first two bars of Figure 3). The GFR in sub-Saharan Africa would be 20 percent higher than it was in 2014 (190 births per 1,000 women compared to 159). In addition, there is a large potential for reductions in fertility if all women who did not want to become pregnant used highly effective contraception. Meeting all existing unmet need for modern methods and attaining a highly effective mix of such methods, as depicted in panel 3, would lead to a 34 percent reduction in the GFR (from 159 to 119 per 1,000 women) and a decline from 15.5 million to one million pregnancies ending in unintended birth or abortion (the last two bars of Figure 3).

TABLE 7.

Total numbers (000s) and rates per 1,000 women aged 15–49 of pregnancies and births that would occur under two hypothetical scenarios—no use of contraception and improving the mix of reversible modern methods—in sub-Saharan Africa and sub-regions, 2014

Geographic Area	Hypothetical estimate: Scenario 1- Total numbers and rates in 2014, if there were no modern contraception				Total numbers and rates in 2014, with current levels and patterns of contraceptive use and induced abortion				Hypothetical estimate: Scenario 5- Total numbers and rates in 2014, if all women who want to prevent pregnancy used highly effective methods
	Total pregnancies	Total births	GPR	GFR	Total pregnancies	Total births	GPR	GFR	Total pregnancies	Total births	GPR	GFR
Sub-Saharan Africa	62,440	43,050	275	190	49,890	36,130	220	159	33,020	27,170	146	119
Eastern Africa	27,240	18,020	301	199	20,020	14,110	221	156	11,860	9,700	131	107
Middle Africa	8,880	6,220	276	193	8,090	5,800	251	180	5,140	4,220	159	131
Southern Africa	4,250	2,840	261	174	1,790	1,290	110	79	920	740	56	45
Western Africa	20,200	14,620	257	187	18,220	13,630	232	174	13,830	11,440	176	146

NOTES: Estimates for sub-Saharan Africa include Sudan, which is in Northern Africa. Numbers may not add up to totals because of rounding. General Pregnancy Rate (GPR) and General Fertility Rate (GFR) are measured as rates per 1,000 women aged 15–49.

FIGURE 3.

Impact on number of unintended births and induced abortions of no contraceptive use (Scenario 1) and of fully meeting unmet need for contraception with highly effective methods (Scenario 5), 2014

Discussion

Our results show high rates of pregnancies and births in sub-Saharan Africa in 2003 and 2014, although the rates declined over the 11-year period. A General Fertility Rate of 159 per 1,000 women aged 15–49 in 2014 confirms that fertility is still high in the region and that population is still growing at a fast pace. This high rate is accounted for by a combination of high desired fertility and high levels of unmet need for effective contraception among women who do not want a child soon or want no more children. An earlier analysis of the ideal number of children among married women in the region between 2000 and 2008 showed that this measure ranged from 4.6 to 9.1 among countries in Western and Middle Africa and from 2.7 to 6.3 in Eastern and Southern Africa (Westoff 2010). For most countries in all sub-regions, actual fertility exceeds desired fertility.

According to a recent study, the number of unintended pregnancies that occurred in the region in 2012 ranged from 1 million in Southern Africa to almost 9 million in Eastern Africa (Sedgh, Singh, and Hussain 2014). Of these pregnancies, the proportion that resulted in unplanned births ranged from 50 percent in West Africa to 63 percent in Southern Africa. Non-use of contraception is largely responsible for the unplanned births. As our findings show, among all women in sub-Saharan Africa aged 15–49, only 21 percent were using a method of contraception in 2014, and 17 percent were using modern methods. For the entire region, these proportions increased only slightly between 2003 and 2014. However, evidence from Southern and Eastern Africa shows that, with effective policies and programs, women and couples who want to delay or stop childbearing will use contraception to do so. While use of contraception is already fairly high in Southern Africa (48 percent of women aged 15–49), it increased at a moderate pace in Eastern Africa, rising from 16 percent in 2003 to 24 percent in 2014 (Darroch and Singh 2013; Singh, Darroch, and Ashford 2014).

Our study shows that contraceptive use has a substantial impact on fertility by reducing unintended pregnancies and unplanned births. The level and pattern of contraceptive use in 2014 meant that the GFR was 31 per 1,000 women lower than it would have been in the absence of modern contraception, or 6.9 million fewer births. We also demonstrated that increasing the level of contraceptive use, especially the use of long-acting reversible methods, can have a substantial impact on fertility. For example, if no women used a contraceptive method in 2014 and abortion behavior remained the same, the GFR would have increased by 35 per 1,000 women. In addition, in the absence of both contraceptive use and abortion, the GFR would have increased by 77 per 1,000.

The analysis also shows that meeting the existing need for effective contraception by shifting toward a more effective method mix would have a notable impact on fertility. If all women who want to avoid pregnancy used the same method mix as employed by similar modern method users in their country, the GFR would drop by 35 per 1,000 women. If this same scenario were achieved with one change—that all modern users used long-acting methods—the decline in the GFR would be slightly greater (40 per 1,000 women). The difference between the two scenarios reflects the fact that the majority of the impact of improvement in contraceptive use comes from women who currently have unmet need for modern methods starting to use the mix of modern methods similar to other women in their country, and the added impact of moving to a more effective mix of methods is relatively small.

To effectively promote effective contraceptive use, policies and programs must address both the demand and supply side. Effort should include school-based and community-level education to improve knowledge of pregnancy risk, increase awareness of the importance of contraceptive use for the health of women and the well-being of the family, and dispel misconceptions about the adverse health consequences of contraception. Interventions should also promote communication between partners to ensure proper understanding of each other’s position on fertility intentions and contraception. It is also important to give women and couples access to a wide range of effective methods so that they can choose a method that best suits them and can switch methods when they wish. Ensuring that women and couples are able to use effective methods that they feel comfortable with will reduce contraceptive discontinuation and incorrect and inconsistent method use. Effective and respectful contraceptive services require steady supplies of contraceptive commodities and availability of personnel trained to provide them, including insertion and removal of methods such as implants and IUD. The most widely used methods in sub-Saharan Africa among users of modern methods are injectables and implants, followed by the pill and condoms (Darroch and Singh 2013). Expanded and improved services are needed to offer women and couples choices of the most effective methods—for example, IUDs and implants for those wanting to delay or space births and voluntary sterilization for those who want no more children.

Additionally, contraceptive counseling remains a key component of ensuring contraceptive uptake and continued use. Service providers should make counseling an integral part of their contraceptive service provision. Increased integration of contraceptive counseling into post-abortion care and postpartum services is also needed.

To achieve these measures requires sufficient and sustained funding and political commitment. With renewed interest in reinvigorating family planning and contraceptive use, it is hoped that international and national stakeholders will continue to harness resources and prioritize contraceptive services in the coming years. It is encouraging that the new Sustainable Development Goals include an important target on meeting demand for modern contraceptive use. Sustained effort at the global, regional, and country levels is essential to ensure optimal development and implementation of effective contraceptive policies and programs in sub-Saharan Africa.