Predictors of Modern Contraceptive Use Among Married Women in Angola: A Population Based Study

Background: Maternal mortality is unacceptability high in Sub-Saharan-African countries including Angola. Despite family planning is one of the vital intervention to reduce maternal death, the coverage of modern contraceptive use in Angola is extremely low and there is a paucity of evidence regarding current factors associated with contraceptive use in Angola. Therefore, this study aimed at investigating predictors of contraceptive use among married women in Angola using nationally representative data. Methods: The data were extracted from 2015/16 Angola Multiple Indicator and Health Survey for this study and approximately 8,033 married women aged 15 - 49 years were participated. Bivariate and multivariate logistic regression were performed using STATA version 14 software to identify the predictors, and p-value less than 0.05 was considered as statistically signicant. Results: The coverage of modern contraceptive use among married women was 8.9%. Maternal age, women’s educational level, maternal occupation, place of residence, media exposure, number of living children, desire for more children and subnational region were main predictors for modern contraceptive use. Conclusion: Modern contraceptive use among married women in Angola was very low. Married women who were educated, wealthiest, exposed for media, who had living children, interested to have child after 2 years and those who were not interested to have at all were more likely to use contraceptive. Therefore, the government of Angola and other concerned bodies need to emphasize in empowering women through education and economy as well as dissemination of contraceptive related information through media especially for women’s living in rural settings.

Worldwide, in 2017, 63% of married or partnered women of reproductive age were using some form of contraception, including any modern or traditional methods of contraception (12,13). On the other hand, contraceptive use was much lower in Africa (36%) compared to other regions in the world, where contraceptive use ranged from 58% in Oceania to around 75% in Northern America, Latin America and the Caribbean in 2017 (12,13).
The changes in Africa have been slow, where contraceptive prevalence increased from 8% in 1970 to 25% in 2000 and reached 36% in 2017; considerable gaps persist in the use of modern methods among couples who want to prevent pregnancy. Large gaps remain in the proportion of demand for family planning satis ed by use of modern methods in countries where overall contraceptive use is low or where many couples rely on traditional methods of contraception (12,13). In low-and middle-income countries (LMICs), contraceptive use differs broadly from country to country due to numerous socio-economic factors. For instance, modern contraceptive prevalence varied from below 10% in Chad, Guinea and South Sudan to 67% in Zimbabwe and 71% in the Democratic People's Republic of Korea in the year 2017 (12,13).
Prior studies in Angola showed maternal age, maternal educational status, prior information and knowledge about contraceptive use, discussion with partner, marital status, and accessibility to supply were the main factors for contraceptive use (14)(15)(16). According to 2015/2016 Angola Multiple Indicator and Health Survey, married women in urban areas are more likely to use modern methods of family planning (18%) than married women in rural areas (only 2%) (17). Provincially, Cuando Cubango has the lowest use of modern methods by married women (1%), and Luanda has the highest (30%). Use of modern methods increases dramatically with economic status (17). Only 1% of married women from the poorest households use modern methods, compared with 31% of married women from the wealthiest households (17).
Few studies have been conducted in Angola related to contraceptive use (14)(15)(16). However, these studies were focused on a single factor (15) and limited to some areas of the country (14,15). Moreover, the available studies might not indicate the current situations of modern contraceptive use. Hence, this study aimed at investigating predictors of modern contraceptive use among married women in Angola using the recent 2015/16 nationally representative data.

Data source
The data used for this study were from 2015/16 Multiple Indicator and Health Survey, which was conducted by Angola National Institute of Statistics (INE) in direct collaboration with Ministry of Health (MINSA) and Ministry of Planning and Territorial Development (MPDT) with technical assistance from United Nations Children's Fund (UNICEF) and ICF (17). It is a nationally representative survey designed to provide current information with regard to the demographic and health situation of women, men, and children including fertility levels, marriage, sexual activity, fertility preferences, family planning methods, childhood and maternal mortality, maternal and child health, breastfeeding practices, nutrition, malaria, HIV/AIDS, domestic violence, and child wellbeing (17). For sample selection, a two stage cluster sampling technique was employed involving the systematic selection of clusters at national level in the rst stage, and then nal selection of households from those clusters were conducted in the second stage for the survey (17). The 2015-16 IIMS survey included a sample of 14,379 women aged 15-49 years and 5,684 men aged 15-54 years. The sample design for the 2015-16 IIMS provides estimates at the national and provincial levels, and for urban and rural areas (17). We used the Individual Recode (IR) le and the analysis for this study restricted for married women.

Variable selection
The main outcome variable of this study was modern contraceptive methods. Married women who have used pills, injectables, implants, intra-uterine devices (IUDs), condoms, sterilization, diaphragms, spermicides, and sponges were categorized and coded as modern contraceptive method users (18), whereas non-users and a very few (less than 1%) traditional users were coded as non-users. Explanatory variables were categorized based on prior evidence (8,(19)(20)(21)(22)(23)(24)(25)(26). These include women's age, women's educational level, and women's occupation, place of residence, regions, religion, and husband educational level. Fertility and reproductive health speci c variables included number of living children, ideal number of children, desire for more children, and distance to health and exposure to media was assessed in terms of frequency.
Moreover, variables like wealth index, attitude towards wife beating and decision making (as women's empowerment) were included. The DHS wealth index is a composite measure of household's cumulative living standard Wealth index, which is derived from household assets and features was used to approximate economic status. Principal Component Analysis (PCA) is used to compute wealth index in DHS and classi es it as poorest, poorer, middle, richer, and richest (25). The DHS asks respondents about their decision-making power regarding several dimension: making decisions regarding own health care, large household purchases and visits to family or relatives. In this study, if the decision on all three dimensions was made by husband/partner alone or by other (i.e. mother in law), it was coded as no decision making; if respondent had decision-making power alone or with her husband/partner over one or two of the decision-making parameters, it was coded as moderate decision making; if the respondent had decision-making power either alone or with her husband on three of the decision-making parameters, it was coded as higher decision making. Women were asked if they agree that a husband is justi ed in beating his wife for the following reasons: (i) burning food, (ii) arguing with him, (iii) going out without telling him, (iv) neglecting the children and (v) refusing to have sexual intercourse with him. If the woman disagreed to all of these reasons, they were assumed to be empowered. Therefore, an overall binary variable was created to re ect attitude towards wife beating. Attitude towards wife beating was coded as 'no' if respondent did not agree with wife beating, and 'yes' if she accepted wife beating as normal/healthy behaviour.

Statistical analysis
Descriptive statistics and multivariate analysis were performed in this study. First, percentages (%) were used to show the distributional characteristics of all the variables. Second, we conducted bivariate logistic regression to select candidate variable for multivariate analysis using p-value less than 0.05 as cut of point. Then, we applied multivariate logistic regressions to examine the relationship between each explanatory variable and modern contraceptive utilization. Crude and adjusted Odd ratios with 95% con dence intervals (CI) were estimated. .
The complex sampling structure of the DHS requires speci c weighting to adjust for disproportionate sampling and nonresponse. Thus, the 'svyset' command was used to weight the data, accounting to the sample cluster and strata (27). All statistical analyses were performed In STATA 14 (Stata Corp, College Station, Texas, USA).

Ethical consideration
From publicly available DHS data set, data analysis were performed. Because the ethical clearance was approved by the institution that commissioned, funded and managed the overall DHS program, further ethical clearance was not required. Informed consent from the participants prior to survey was ensured. The ICF international and respective country's ethical review Board (IRB) also ensured that the protocols are in compliance with the U.S. Department of Health and Human Services regulations for the protection of human subjects.

Results
In this study a total of 8,033 married women were involved. Of them, about 686 (8.5%) and 1,650 (20.5%) of the participants were in the age group of 15-19 and 20-24 years, respectively. Nearly two sixth (33.2%) of the respondents had no formal education, while about 3,039 (37.8%) attended primary school, and more than one forth (26.2%) of the respondents attended secondary school. Similarly, one forth (25.7%) of their husband had no formal education. More than two fth (42.1%) of the participants were rural residents and about 2,753 (34.3%) of the respondents had no exposure for newspapers, radio or television for at least once a week. Nearly half (49.7%) of the participants ideally wanted to have six children and more than half of the participants (51%) reported distance to facility is a big problem. Regarding decision making, while 561 (7%) of the women had no decision making power on all of the decision making parameters, her own health, purchasing household expense and visiting family or relatives, about 2,370 (29.5%) of the women decided either alone or together with their husband on the two of the three above mentioned decision making parameters (Table 1).

Predictors of modern contraceptive use
This study showed that about 91.1% of the married women were not used modern contraceptive methods (Fig. 1).
The results of the adjusted OR and 95% CI for the relationship between demographic, socioeconomic, woman-and husband/partner factors and modern contraceptive use are shown in Table 2.
We found women's age had signi cant association with utilization of modern contraceptive methods. For instance, the odd of married women within 45-49 years age groups were 92% (AOR = 0.08, 95% CI; 0.03-0.25) less likely to use modern contraceptive as compared to married women within 15-19 years age group.
Educational level of women were another predictor of modern contraceptive use. In fact, the present study showed the utilization of modern contraceptive among married women who attended secondary school were 2.5 times (AOR = 2.53, 95% CI; 1.28-5.01)) higher as compared to married women who had no formal education. Similarly, the utilization among married women who attended above secondary school were approximately 3 times (AOR = 2.99, 95% CI; 1.27-7.07) higher as compared to married women who had no formal education.
Maternal occupation had signi cant association with utilization of modern contraceptive. We found, the odd of modern contraceptive use among married women worked as agricultural self-employed were lower by 51% (AOR = 0.49, 95% CI; 0.30-0.79) as compared to married women who were not working. However, the odd of modern contraceptive use among married women whose occupation were in household and domestic are 3.4 times (AOR = 3.41, 95% CI; 1.04-11.16) higher as compared to married women who were not working.
Husband educational level was another predictor of modern contraceptive use among married women. For instance, the odd of modern contraceptive utilization among married women whose husband attended above secondary school level were 2.8 times (AOR = 2.83, 95% CI; 1.54-5.21) higher as compared to married women whose husband had no formal education.
Place of residence were also signi cantly associated with use of modern contraceptive. For instance, compared to married women living in urban setting, the odd of modern contraceptive utilization among married women living in rural setting were lower by 52% (AOR = 0.48, 95% CI; 0.32-0.73).
We found media exposure had signi cant association with utilization of modern contraceptive methods.
In fact, in our study the utilization of modern contraceptive methods among married women who had media exposure for at least less than a week, be it for newspapers, radio or television were 76% (AOR = 1.76, 95% CI; 1.01-3.08) higher as compared to married women who had no media exposure.
We observed more utilization of modern contraceptive methods among married women who had many living children. For example, in our study the odd of modern contraceptive utilization among married women who had 1-2 and 3-4 living children were 5.8 times (AOR = 5.85, 95% CI; 1.53-22.31) and 6 times (AOR = 6.04, 95% CI; 1.61-22.63) higher as compared to married women who had no child.
Similarly, the odd of contraceptive utilization among married women who had ve and above living children were approximately 8 times (AOR = 7.97, 95% CI; 1.97-32.19) higher as compared to married women who had no child.
Desire for more children was another predictor of modern contraceptive use. For instance, in our study, compared to married women who wanted to have children within two years, the odd of modern contraceptive utilization among married women wanted to have children after two years were 84% (AOR = 1.82, 95% CI; 1.24-2.68) higher.

Discussion
Contraception is one of the key components for family planning and vital means to reduce mortality and morbidity of both child and mother (1-3, 8, 28). In this study, using the recent 2015/2016 Angola Multiple Indicator and Health Survey, several personal and community level predictors for utilization of modern contraceptive methods among married women were analyzed. We found that women's age, women's and husband educational level, place of residence, women's occupation, economic status, media exposure, number of alive children, desire to have more children, and subnational region were the main determinants for utilization of modern contraceptive methods.
In line with previous literature (19,21,24,25,28), our ndings also revealed that women were less likely to use modern contraceptives with increasing age when controlling for other factors. Modern contraceptive use among women 45-49 years of age was lower than adolescent women (women within 15-19 years of age). Higher uptake of contraceptive use among young women may be due to the in uence of communication among young women and their husbands/partners about family planning, as seen in a 2016 study in Angola (29).
We found women's educational level in uenced uptake of modern contraceptive use as supported by previous studies (30)(31)(32). It highlights, especial attention should be given for married women with less educational level or not have formal education at service provision area such as during counseling and need to taking account during awareness creation such as using audio-visual aid instead of lea ets distribution.
Modern contraceptive also varied based on their occupation types. Compared to married women who were not working, women with agriculturally self-employed were less likely to use. However, those worked in household and domestic were more utilize modern contraceptive methods.
In our study, we found that utilization of modern contraceptives varied based on husband educational level. In fact, married women whose husband attended primary school and higher level were more likely to use contraception as compared to women whose husband did not attend formal education. Similar ndings in previous studies (26,33), suggest husband's educational level increased the uptake of contraceptive use for family planning because of increased communication between the couple, that lead to increased support of contraceptive use by the husband, which also in uenced his actions through is adoption of condom use and/or sterilization (34,35).
Having exposure for media signi cantly increased the uptake of modern contraceptive methods. In our study, married women who either read the newspaper, listened to radio or watched television for at least once a week were more likely to utilize modern contraceptive methods than women with no media exposure. Previous studies in Senegal and West Bengal (36,37), showed similar ndings. Mass media is a key strategy for increasing demand for use in health services (38). It is a process that helps communities to identify their own needs and to respond to and address these needs (38). Gaining the participation of community members can help providers raise awareness both of health issues at the community level and of social and cultural issues that may promote or inhibit use of information and services, as well as improve clients' understanding of the methods or services being offered (39). Speci c barriers to service access and use can be addressed and service utilization increased (39). Community mobilization encourages concern of the needs of speci c populations and localities. In particular, underserved populations, such as youth and men, can be reached more effectively through community mobilization (39).
Consistent with previous studies (18,40,41), the number of living children had signi cant association with the use of modern contraceptive methods. Compared to married women with no child, the likelihood of utilization was higher among women with children. A possible explanation might be related to certain myths that prohibit women from seeking family planning, such as the belief that contraceptive use is associated with women who are "promiscuous" (42). Cultural and traditional beliefs hold strong in many rural areas in Africa, holding on to beliefs that in uence lack of change (i.e. use of modern contraception) (43). Another justi cation for less use of modern contraceptive among women with no child especially if it is due to death of child, it could be due to higher desire to have more children as means of lling the gap brought on by the loss of that particular child (44).
The other main nding of the study was, the in uence of women's desire to have more children on the utilization of contraceptive methods. Women wanted to have children after two years were more utilized than women who desired to have children in a shorter period. Additionally, those women who undecided when to have children and those who had no more desire or sterilized (themselves/their husband) better utilized the modern contraceptive methods as compared to women who wanted to have more children in less than two years. Our nding is in line with previous study (45). Evidence suggested that desire for more children might be related to religion (43). A study with similar results found that in Kenya and Tanzania, Muslim women desire more children than Christians (43,46). Additionally, It might be related to educational level of the women and overall perception towards family size (43). A study in Ethiopia showed, women with an educational attainment less than grade one were found to have a larger ideal number of children, compared to illiterate women as well as those who have attained grade one or higher (43). Another justi cation related to perception is, women who attach a "high-bene t" economic value to children have a larger reported ideal number of children (i.e., family size) than women who perceive children to be a high cost. Couples view their children as an investment and that these children are due for pay-up when their parents reach old-age (43). In African traditions, it is almost always predictable that once the child grows older, he/she will be indebted to nancially pay back (i.e., support) the parent(s) for his/her (47)a high value to having children will aim for larger families who will later t the role of being a provider and caregiver to the parents in old age (43).
Women living in urban setting were more likely to utilize modern contraception than their counterparts in rural regions as documented in previous studies (47,48). This could be related to knowledge about contraceptive methods; which is highly related with its uptake (27,49,50) and rural residents had relatively lower level of knowledge than urban residents (47). Furthermore, uptake of contraception is strongly related with education (34), and formal education remains low among women and men in rural areas (47).
We found difference in utilization of modern contraceptive across regions as suggested in previous studies (25,33,51). The disparities might be due to existed social norm and culture (45,52) and difference in accessibility of health facility and quality of the services given in the facilities across regions (33).
This study has its own strength and limitations. Using the recent and nationally representative data can help to understand the current nationwide predictors of modern contraceptive use, believed to be the main strength of the study. However, the study has the following limitations. Since the data is cross-sectional it does not measure causality between explanatory factors and the outcome variable-modern contraceptive use. The authors are only able to use the variables available in the secondary dataset for analysis and other variables such as cultural and perception factors as well as quality of care were not included.

Conclusions
Numerous personal and community level predictors such as women's educational level, women's occupation, place of residence, media exposure, economic status, number of living children, desire for more children and region were identi ed in this study.
Strategies to increase family planning should consider changing societal myths and community cultural perceptions while also engaging key community leaders. It is also important to reach the women living in deprived regions (i.e. rural) where access to health facilities and resources are low. Mass media would be an approach to upsurge awareness, education and community mobilization regarding utilization of modern contraceptive methods. Empowering women through education and economy could be the possible policy implication.