Plain English summary
Approximately 99% of the global maternal mortality occurred in resource-constrained settings, with Sub-Saharan Africa region reporting about two-third (66%) of maternal deaths. Literature showed that women’s status within the household is a prominent factor for improving the utilization of maternal health services. Nevertheless, the impact of women’s socio-economic class on maternal health care use has not received adequate attention in Sub-Saharan Africa.
We have undertaken nationally representative Demographic and Health Surveys (DHS) data involving the selection of 32 countries based on geographical diversity. We used a 3-level model to explore contextual and compositional factors associated with contraceptive use. Neighbourhood socioeconomic disadvantage was operationalized with a principal component approach using key indicators. More so, Human Development Index (HDI) was used as a measure of a country’s intensity of deprivation, which was the average percentage of deprivation experienced by people in multidimensional poverty. Contextual effects were measured by the intraclass correlation (ICC) and median odds ratio (MOR). There were large disparities in contraceptive use, with about half of the study countries identified with low contraceptive use. This study empirically demonstrated the individual-level, neighbourhood-level and country-level maternal factors associated with contraceptive use among women in Sub-Saharan African countries.
Based on the findings of contraceptive use and its associated women’s empowerment and proximate factors among women of reproductive age in Sub-Saharan Africa countries, there is a need for regional interventions in improving fertility control measures, specifically contraceptive use.
Introduction
The Sustainable Development Goal (SDG-5) targets to achieve gender equality and empowerment of all women and girls [
1]. Reducing gender inequality is a vital policy agenda globally, through access and rights to resources. To enhance the impacts of programs and policies that relate to gender, a crucial step is first to understand how inequality is exhibited throughout the world, specifically, by identifying where it occurs using indicators and approaches that are consistently disaggregated as to find a resolution as possible. Such clear measure of indicators for gender inequality is not known to currently exist. In the past decades, assessments of gender disparity in many parts of the world have failed to deliver on promoting gender equality and empower women [
2‐
4], with evidence suggesting that women in some countries have seen decreased opportunities to improve their welfare [
5]. One reason gender equality is so high on the international policy concern is the growing body of evidence showing that improving the welfare of women and closing the inequality gap can lead to improved childhood nutrition and reduced mortality, increased school enrollment, improved maternal and children’s health and improved natural resource management [
6].
Maternal autonomy in healthcare-seeking behaviour is connected to women’s empowerment and helps to achieve desired health outcomes [
7]. The multifaceted latent nature of women’s empowerment makes it difficult for researchers to directly and accurately measure empowerment. Generally, proxy indicators are commonly used to measure empowerment, including but not limited to decision making power, reasons to justify sexual violence, women’s knowledge level, and labour force participation [
8]. In most communities, specifically in Sub-Saharan Africa region, men have huge control over the women of their social class; particularly within families and households [
6]. The health care system for children and mothers alike in many communities is poorly affected by the women’s subservient status within households, which is consequent upon social and cultural predetermined roles for women that subvert almost every aspect of their lives [
9]. Thus, women’s empowerment is majorly recognized as a vital tool to enable access to reproductive and sexual health care services for improved mother-and-child health [
10].
Debate related to the prominent indicators for measuring women’s empowerment has established that women’s empowerment can be evaluated by their capability to contribute in household decision-making which reflects their economic, domestic and movement autonomies [
7,
11]. The more empowered women are, the more likely to use modern contraception, deliver in a health facility and have a skilled attendant at birth [
12]. In addition, contraceptive use is important in preventing fetal, neonatal, and under-5 deaths; reducing maternal mortality and avoiding high-risk pregnancy including pregnancy among teenage girls and older women. Many women with issues of health care challenges experience gendered power inequalities, especially in their intimate relationships, that prevent them from achieving optimal sexual and reproductive health benefits and exercising their rights [
13]. An increasing body of evidence demonstrates the ways unequal levels of power between men and women in intimate relationships prevent women from making decisions regarding their sexual and reproductive health [
14]. Frequently, unequal control over and access to economic resources, unequal relationship power, and limited ability to make sexual decisions (including whether, when, how often, and with whom to have sex; and negotiating condom use, contraceptive or other protective practices) make women vulnerable to SRH risks [
15].
Empowerment is one such characteristic that may influence a woman’s experience of pregnancy, delivery, and postnatal care. Women’s empowerment leads to significant positive changes in many domains. Studies have found an association between increased empowerment and reduced mortality and morbidity [
16,
17]. In terms of reproductive health, empowerment has been associated with reduced rates of unintended pregnancies [
18] and sexually transmitted diseases, such as chlamydia and gonorrhea in high-risk populations [
19]. Other reports show the benefits of empowerment for health-related behaviours such as obtaining nutritional supplements and participating in health education sessions [
20]. The benefits of empowerment are not necessarily limited to women themselves but have the potential to extend to those around her, including but not limited most prominently to her own children.
Transformative strategies between gender helps to address gender inequalities while promoting health. These techniques support awareness of gender roles and intervene in the distribution of resources and allocation of responsibilities between male and female, handle power relationships and promote the position of women [
17]. Women’s empowerment should be regarded as a crucial element and an indicator of maternal health. In this study, we investigated whether empowerment could help women utilize contraceptive services.
Results
Sample characteristics
We analyzed information on 474,622 respondents (Level 1) nested within 16,748 neighbourhoods (Level 2) from 32 countries (Level 3) in Sub-Saharan Africa (Table
1). Table
1 shows the countries, year of data collection, and the surveys characteristics. The median number of neighbourhoods sampled was 455, ranging from 214 in Senegal to 1593 in Kenya. The median number of respondents was 10,990 (range: 5329 to 39,948). Seven countries could be classified as post-conflict countries. As shown in Fig.
1, there was a wide variation in the percentage of women who reported ever used contraceptive in the 32 countries studies. It ranged from as low as 6.7% in Chad to as much as 72% in Namibia. The characteristics of the pooled sample are shown in Table
2. Almost 40% of the respondents were aged between 15 to 24 years. More than one-third of the respondents had no formal education and more than half were in labor force. Contraceptive use was significantly more common among respondents from the richest households (28.5% versus 18.9%). In addition, women with partners with secondary or higher education reported more contraceptive use (48.6% versus 27.1%).
Table 2
Summary of pooled sample characteristics of the Demographic and Health Surveys data in sub-Saharan Africa
| 441,098 | 272,705 | 167,547 | |
Age (%) | | | | < 0.001 |
15–24 | 174,203 (39.6) | 127,692 (46.8) | 46,510 (27.8) | |
25–34 | 140,294 (31.9) | 73,701 (27.0) | 66,592 (39.7) | |
35–49 | 125,759 (28.6) | 71,312 (26.1) | 54,445 (32.5) | |
No of children (mean (sd)) | 2.87 (2.75) | 2.65 (2.89) | 3.23 (2.46) | < 0.001 |
Wealth (%) | | | | < 0.001 |
Poorest | 87,727 (19.9) | 63,640 (23.3) | 23,908 (14.3) | |
Poorer | 82,941 (18.8) | 54,940 (20.1) | 27,839 (16.6) | |
Middle | 83,487 (18.9) | 52,480 (19.2) | 30,861 (18.4) | |
Richer | 87,309 (19.8) | 50,005 (18.3) | 37,112 (22.2) | |
Richest | 99,634 (22.6) | 51,640 (18.9) | 47,827 (28.5) | |
Partner’s education (%) | | | | < 0.001 |
No education | 107,685 (36.6) | 85,067 (49.9) | 22,419 (18.2) | |
Primary | 80,264 (27.3) | 39,234 (23.0) | 40,815 (33.2) | |
Secondary+ | 106,313 (36.1) | 46,298 (27.1) | 59,774 (48.6) | |
Labour force participation (yes, %) | 243,911 (55.3) | 140,671 (51.6) | 102,918 (61.4) | < 0.001 |
Acceptance of wife beating (%) | | | | < 0.001 |
Low | 249,706 (56.6) | 147,288 (54.0) | 101,795 (60.8) | |
Medium | 68,756 (15.6) | 45,831 (16.8) | 22,924 (13.7) | |
High | 122,636 (27.8) | 79,586 (29.2) | 42,828 (25.6) | |
Women’s knowledge level (%) | | | | < 0.001 |
Low | 197,536 (44.8) | 135,993 (49.9) | 61,043 (36.4) | |
Medium | 131,254 (29.8) | 76,920 (28.2) | 53,989 (32.2) | |
High | 112,308 (25.5) | 59,792 (21.9) | 52,515 (31.3) | |
Decision making power | | | | < 0.001 |
Low | 47,485 (37.3) | 27,943 (40.8) | 19,485 (33.3) | |
Medium | 56,469 (44.4) | 26,870 (39.3) | 29,515 (50.4) | |
High | 23,210 (18.3) | 13,642 (19.9) | 9568 (16.3) | |
Neighbourhood SES (%) | | | | < 0.001 |
Tertile 1 (least disadvantaged) | 148,852 (33.7) | 78,012 (28.6) | 70,635 (42.2) | |
Tertile 2 | 145,590 (33.0) | 88,626 (32.5) | 56,680 (33.8) | |
Tertile 3 (most disadvantaged) | 146,656 (33.2) | 106,067 (38.9) | 40,232 (24.0) | |
Human Development Index (%) | | | | < 0.001 |
Low HDI | 159,306 (36.1) | 119,123 (43.7) | 40,183 (24.0) | |
Moderate HDI | 162,673 (36.9) | 94,496 (34.7) | 68,173 (40.7) | |
High HDI | 119,119 (27.0) | 59,086 (21.7) | 59,191 (35.3) | |
Measures of associations (fixed effects)
The results of the different models are shown in Table
3. In the fully adjusted model controlling for the effects of individual, neighbourhood and country level factors, all the factors remained significantly associated with odds of contraceptive use.
Table 3
Individual compositional and contextual factors associated with contraceptive use in sub-Saharan Africa identified by multivariable multilevel logistic regression models, Demographic and Health Surveys data
Fixed-effect | | | | | |
Control variable | | | | | |
Survey years | | | | | |
2008 | | 1 (reference) | 1 (reference) | 1 (reference) | 1 (reference) |
2010 | | 0.07 (0.04 to 0.12) | 0.43 (0.20 to 0.76) | 0.86 (0.44 to 1.91) | 0.22 (0.07 to 0.62) |
2011 | | 0.37 (0.12 to 0.67) | 1.00 (0.76 to 1.25) | 0.52 (0.37 to 0.65) | 0.28 (0.16 to 0.39) |
2012 | | 0.23 (0.16 to 0.33) | 0.85 (0.58 to 1.19) | 0.36 (0.29 to 0.43) | 0.31 (0.14 to 0.52) |
2013 | | 0.20 (0.11 to 0.39) | 0.73 (0.51 to 0.94) | 0.62 (0.53 to 0.76) | 0.18 (0.11 to 0.31) |
2014 | | 0.48 (0.31 to 0.74) | 1.86 (1.55 to 2.16) | 0.83 (0.69 to 1.00) | 1.14 (0.68 to 1.85) |
2015 | | 0.79 (0.11 to 2.04) | 0.74 (0.53 to 1.10) | 0.77 (0.52 to 1.05) | 0.20 (0.07 to 0.37) |
2016 | | 0.66 (0.17 to 1.01) | 1.38 (0.92 to 1.95) | 0.34 (0.19 to 0.53) | 1.68 (0.36 to 4.21) |
Individual-level factors | | | | | |
Age (%) | | | | | |
15–24 | | 1 (reference) | | | 1 (reference) |
25–34 | | 1.21 (1.16 to 1.27) | | | 1.19 (1.12 to 1.24) |
35–49 | | 0.78 (0.74 to 0.82) | | | 0.75 (0.71 to 0.79) |
No of children | | 1.16 (1.14 to 1.17) | | | |
Wealth (%) | | | | | |
Poorest | | 1(reference) | | | 1 (reference) |
Poorer | | 1.37 (1.30 to 1.45) | | | 1.24 (1.17 to 1.31) |
Middle | | 1.62 (1.53 to 1.72) | | | 1.33 (1.26 to 1.41) |
Richer | | 2.17 (2.04 to 2.32) | | | 1.60 (1.50 to 1.71) |
Richest | | 2.93 (2.73 to 3.15) | | | 1.98 (1.82 to 2.13) |
Partner’s education (%) | | | | | |
No education | | 1 (reference) | | | 1 (reference) |
Primary | | 1.69 (1.61 to 1.77) | | | 1.62 (1.55 to 1.69) |
Secondary+ | | 2.02 (1.93 to 2.12) | | | 1.91 (1.83 to 2.00) |
Labour force participation (yes, %) | | 1.14 (1.07 to 1.22) | | | 1.14 (1.06 to 1.21) |
Acceptance of wife beating (%) | | | | | |
Low | | 1 (reference) | | | 1 (reference) |
Medium | | 1.13 (1.08 to 1.19) | | | 1.13 (1.08 to 1.18) |
High | | 0.98 (0.96 to 1.04) | | | 1.01 (0.97 to 1.04) |
Women’s knowledge level (%) | | | | | |
Low | | 1 (reference) | | | 1 (reference) |
Medium | | 1.51 (1.45 to 1.57) | | | 1.46 (1.40 to 1.52) |
High | | 2.04 (1.94 to 2.14) | | | 1.96 (1.87 to 2.05) |
Decision making power | | | | | |
Low | | 1 (reference) | | | 1 (reference) |
Medium | | 1.23 (1.19 to 1.27) | | | 1.21 (1.16 to 1.26) |
High | | 1.25 (1.20 to 1.31) | | | 1.23 (1.18 to 1.29) |
Neighbourhood factor | | | | | |
Neighbourhood SES (%) | | | | | |
Tertile 1 (least disadvantaged) | | | 1 (reference) | | 1 (reference) |
Tertile 2 | | | 0.65 (0.63 to 0.68) | | 0.76 (0.71 to 0.81) |
Tertile 3 (most disadvantaged) | | | 0.31 (0.29 to 0.32) | | 0.43 (0.40 to 0.46) |
Country-level factor | | | | | |
Conflict (yes vs no) | | | | 0.82 (0.66 to1.05) | 1.20 (0.80 to 1.91) |
Human Development Index (%) | | | | | |
Low HDI | | | | 1 (reference) | 1 (reference) |
Moderate HDI | | | | 2.83 (2.59 to 3.19) | 1.56 (1.22 to 2.47) |
High HDI | | | | 3.75 (3.21 to 4.32) | 1.75 (0.95 to 2.68) |
Random-effect | | | | | |
Country-level | | | | | |
Variance (95% CrI) | 1.12 (0.67 to 1.87) | 1.85 (1.06 to 3.17) | 1.00 (0.60 to 1.69) | 0.74 (0.44 to 1.23) | 1.90 (0.99 to 3.47) |
VPC (%, 95 CrI) | 20.9 (13.7 to 30.4) | 30.5 (20.3 to 42.6) | 19.9 (13.1 to 29.5) | 31.2 (19.2 to 45.0) | 31.2 (19.2 to 45.0) |
MOR (95% CrI) | 2.75 (2.18 to 3.68) | 3.66 (2.67 to 5.46) | 2.59 (2.09 to 3.45) | 3.73 (2.58 to 5.91) | 3.73 (2.58 to 5.91) |
Neighbourhood-level | | | | | |
Variance (95% CrI) | 0.95 (0.92 to 0.98) | 0.93 (0.87 to 0.98) | 0.72 (0.70 to 0.74) | 0.95 (0.92 to 0.98) | 0.90 (0.85 to 0.94) |
VPC (%, 95 CrI) | 38.6 (32.6 to 46.4) | 45.8 (37.0 to 55.8) | 34.3 (28.2 to 42.5) | 33.9 (29.2 to 40.1) | 46.0 (35.8 to 57.3) |
MOR (95% CrI) | 2.53 (2.50 to 2.57) | 2.51 (2.43 to 2.57) | 2.25 (2.22 to 2.28) | 2.53 (2.50 to 2.57) | 2.47 (2.41 to 2.53) |
Model fit statistics | | | | | |
DIC | 474,622 | 122,624 | 473,685 | 474,632 | 122,165 |
Sample size | | | | | |
Country-level | 32 | 32 | 32 | 32 | 32 |
Neighbourhood-level | 16,748 | 15,344 | 16,748 | 16,748 | 15,344 |
Individual-level | 440,052 | 123,258 | 440,052 | 440,052 | 123,258 |
Women aged 25 to 34 years old were more likely to have used contraceptive compared to those aged 15 to 24 years old (OR = 1.19, 95% CrI 1.12 to 1.24). Women from the richest households were as twice as likely to have used contraceptive than those from poorest households (OR = 1.98, 95% CrI 1.82 to 2.13). Women whose partner had secondary or higher were also almost as twice as likely to have used contraceptive than those whose partner had no education (OR = 1.81, 95% CrI 1.83 to 2.00). Women in labour were 14% more likely to have used contraceptive than those not working (OR = 1.14, 95% CrI 1.06 to 1.21). Women with medium acceptance of wife beating were 13% more likely to have used contraceptive than those with low acceptance of wife beating (OR = 1.13, 95% CrI 1.08 to 1.18). Women with high knowledge level were as twice as likely to have used contraceptive than those with low knowledge level (OR = 1.96, 95% CrI 1.87 to 2.05). Women with the high decision-making power were 23% more likely to have used contraceptive than those with the low decision-making power (OR = 1.23, 95% CrI 1.18 to 1.29).
Women living in the most SEP disadvantaged neighbourhood were 57% less likely to have used contraceptive than those in the least SEP disadvantaged neighbourhood (OR = 0.43, 95% CrI 0.40 to 0.46). Women from countries with moderate human development index were 1.56 times more likely to have used contraceptive than those from countries with low human development index (OR = 1.56, 95% CrI 1.22 to 2.47).
Measures of variations (random effects)
As shown in Table
3, in Model 1 (unconditional model), there were substantial variations in the odds of contraceptive use across the 32 countries (
σ2= 1.12, 95% CrI 0.67 to 1.87) and across the neighbourhoods (
σ2= 0.95, 95% CrI 0.92 to 0.98). Having 20.9 and 38.6% VPC estimates for the model implied that the variance in odds of contraceptive use could be attributed to country and neighbourhood level factors, respectively. Results from the median odds ratio (MOR) also confirmed evidence of neighbourhood and societal contextual phenomena shaping contraceptive use. From the full model (Model 5), it was estimated that if a women moved to another country or neighbourhood with a higher probability of contraceptive use, the median increase in their odds of contraceptive would be 3.73 (95% CrI 2.58 to 5.91) and 2.47-fold (95% CrI 2.41 to 2.53) respectively.
Discussion
The results of this study showed large disparities in contraceptive use across Sub-Saharan Africa countries. The findings also revealed low contraceptive use from several countries including Chad, Gambia, Guinea, Angola, Mali, Burundi, Comoros, Burkina-Faso and Nigeria where less than one-quarter of the women had ever used a contraceptive method. Similarly, in recent studies, there were reports of low contraceptive use in developing countries [
18,
36‐
38]. Here, our study also investigated the association between contraceptive use and various components of women’s empowerment. The significant association between women’s empowerment and contraceptive use is consistent with the results from previous studies where women’s empowerment was positively associated with the use of health care services in 67 developing countries [
38]. While several studies from individual countries in Africa have consistently shown the demographic and socio-economic factors associated with contraceptive use, this study employed a multi-country approach in Sub-Saharan Africa region to establish the association between women’s empowerment and contraceptive use.
The pooled multicountry analyses accounted for the heterogeneity across different levels of factors included in the model. We extensively examined women’s empowerment as a strong factor in contraceptive utilization, considering individual-level, neighbourhood-level, and country-level evidence. Women with higher knowledge level, those who participated in the labour force, with higher wealth status, and having more decision-making power were found to have increased in the odds of contraceptive use. These findings are consistent to previous studies which identified maternal socioeconomic status such as education and wealth index as significantly associated with service utilization [
36,
38,
39]. Partner’s education was also associated with improved contraceptive use. In terms of improved decision-making autonomy, access to economic resources, enhanced knowledge, low disadvantaged neighbourhood and countries with high human development index, such women would have greater chances to effectively cope with issues presented by socio-cultural, religious and health systems factors. The could be the basis for underscoring the findings of this study. High rates of maternal and child mortalities can be prevented through lower fertility rates and prolonged birth spacing, which involves contraceptive use through enhanced participation of women in labour force, protection against abuse and violence, higher knowledge, improved decision making power and socioeconomic status.
This study builds on the vast literature based on DHS data to describe components of women’s empowerment by considering standard indicators of women’s empowerment across Sub-Saharan Africa countries [
40,
41]. The validated measure becomes a crucial instrument for global health researchers in determining the influence of women’s empowerment on contraceptive use and other health outcomes. In the same vein, the measure allows for comparisons across various countries in Sub-Saharan Africa while accounting for unique features of each context. Therefore, concerted efforts are paramount to improve women’s economic empowerment and promote business enterprises. Particularly, local savings groups, community and microfinance banks will aid African women who may not have surety for loans from financial institutions [
42]. Also, the scaling up of lucrative medium and small-scale businesses should be supported by favourable trade policies and financial institutions. Promoting women’s economic empowerment through local support groups, charitable organizations or well-off individuals, trade ventures and effective policies would result in increased empowerment [
43,
44].
Strengths and limitations
The major strength of this study is the use of current nationally representative datasets from 32 different Sub-Saharan Africa countries, which makes the findings of the study generalizable to women of reproductive age in Sub-Saharan Africa countries. However, this analysis has some drawbacks. Prominently, the analyses utilized cross-sectional data, hence, only associations and no causal relationships are established. More so, our inability to measure sources of demand-side unobserved heterogeneity across the secondary data might have biased our estimates of correlates of contraceptive use. The unavailability of relevant variables was a major limitation in the DHS data. Hence, we considered supply-side limitations of data to basically those issues related to health care delivery. DHS did not report availability, accessibility, and frequency of utilization of contraceptive methods. Furthermore, recall bias could have occurred in this study that deals with life time contraceptive use.
Conclusion
Based on the findings of this study, it is imperative to note that enhancing contraceptive use among women could help to ameliorate their health care services. This study showed that contraceptive use is associated with women’s empowerment and other proximate determinants; including partner’s education, maternal age, wealth or socioeconomic status, individual-level, neighbourhood-level and country-level factors. Improving women’s participation in labour force, which implies creating employment opportunities, reduction in gender-based violence, enhancing the decision making the power of women and increasing their knowledge level can improve contraceptive use and therefore achieve better maternal health in Sub-Saharan Africa countries. Furthermore, the development of community-based women’s empowerment programmes, such as women’s access to media and health information could be useful interventions to empower women. The subject of women’s empowerment on maternal health care could consider various pathways from empowerment to action such as maternal autonomy to sexual and reproductive health care including family planning services to expedite achievement of the fifth sustainable development goal. However, further policy research aimed at evolving plans and strategies to promote widespread contraceptive utilization as an integral part of an overall programme to improve efficiency need to focus on two key areas; availability and accessibility issues. In addition, future research should explore reasons for low prevalence in contraceptive use using qualitative approach or should consider prospective studies using panel data, asking respondent and partners about their contraceptive use, or repeating questions in different ways to check the consistency of the answers to solve the problem of recall bias that is inherent in this study.