Factors associated with decision making power of women to use family planning among married women in sub-Saharan Africa: A multilevel analysis of demographic health surveys

Background: In sub-Saharan Africa there are several socio-economic and cultural factors which affect women’s ability to make decision regarding their own health including contraceptive usage. The main aim of this study was to determine factors associated with decision making power of women to use family planning in sub-Saharan Africa. Methods: The appended, most recent demographic and health survey datasets of 35 sub-Saharan countries were used. A total weighted sample of 83,882 women were included in the study. Both bivariable and multivariable multilevel logistic regression were done to determine the associated factors of decision making power of women to use family planning in SSA. The Odds Ratio (OR) with a 95% Condence Interval (CI) was calculated for those potential variables included in the nal model. Results: Those married women with primary education (AOR=1.24; CI:1.16,1.32),secondary education (AOR=1.31; CI:1.22,1.41), higher education (AOR=1.36; CI:1.20,1.53),media exposure(AOR=1.08; CI: 1.03, 1.13), currently working (AOR=1.27; CI: 1.20, 1.33), antenatal care visit of 1-3 (AOR=1.12; CI:1.05,1.20) and ≥ 4 ANC visit (AOR=1.14;CI:1.07,1.21), women who were informed about family planning (AOR=1.09; CI: 1.04, 1.15), women who had less than 3 children(AOR=1.12; CI: 1.02, 1.23) and 3-5 children (AOR=1.08; CI: 1.01, 1.16) had higher odds of decision making power to use family planning than their counter parts. Besides, mothers with age of 15-19 (AOR=0.61; CI: 0.52, 0.72), 20-24 (AOR= 0.69; CI: 0.60, 0.79), 25-29 (AOR=0.74; CI: 0.66, 0.84), determine associated factors of decision making power to use family planning among married women in sub-Saharan Africa. Accordingly this study; age, education level of women, women and their husbands’ occupation, wealth index, region, media exposure, ANC visit, fertility preference of the women, husband’s desire for the number of children and information about family planning were factors associated with decision making power of women to use family planning.


Background
Sub-Saharan Africa accounted 66% of the deaths globally and had the highest Maternal Mortality Ratio (MMR) at 546 maternal deaths per 100,000 live births(1). Unplanned pregnancy and short inter-pregnancy spacing are leading causes of maternal and child death in this region; use of modern family planning methods after delivery is considered an important part of interventional efforts (2,3). In developing countries, more than 222 million women's pregnancies are unplanned (4).
The 2030 Agenda for Sustainable Development Goal(SDGs) includes relevant targets for using contraceptive under the broader goals of health and well-being of the population and gender equality (5,6). Family planning service contributes not only to the reduction of morbidity and mortality of mothers and children, but also avoiding unintended pregnancy and its adverse consequence including preventing HIV/AIDS and high risk abortion, improving standard of living, and e cient use of natural resources (7).
Although 30% of maternal deaths and 10% of child death could be avoided by extending pregnancy (10), 41% of women in Sub-Saharan Africa who intended to use modern contraceptives were not using (11).
A woman's ability to choose the method of modern contraception she uses are likely to be affected by her self-image and sense of empowerment. A woman who feels that she is unable to control other aspects of her life may be less likely to feel that she can make decisions about fertility (12). Independent decision making or joint decision making with partners on family planning use has a substantial contribution to the improvement of maternal health (13). Although women's empowerment is a key to use contraceptives (12), most of husbands give inferior positions to women in all aspects of decision-making in developing countries including decision making power to use contraceptives (12,14).
Women often have less power due to their political, economic, and sociocultural status and may not be in a position to protect themselves from unwanted sexual intercourse and gender-based violence and resulting in sexually transmitted infections and other sexual and reproductive health (SRH) problems(15).
This indicates that gender inequalities worsen a difference in SRH well-being and ill health, and sometimes life and death (16).
One of the reasons women have reported for their nonuse of contraceptive is that they do not decide to use or not to use it (17). Evidences showed that women who have decision-making power are more likely to use contraceptive than those who had not (18,19). However, decisions for contraceptive use may be affected by unbalanced power relations between women and their partners, especially in more malecontrolled societies and cultural discrimination (20). Furthermore, previous studies showed that decisionmaking power of women to use family planning was associated with education(4, 21-23), age (4,(23)(24)(25)(26), better knowledge (24,27,28), working status of women (25,26,29), gender equitable attitude (27), number of living children (21,25,26), socio-economic status (22,23,(29)(30)(31), residence (25,26), husband desire number of children (28) and attitude towards family planning (24).
Despite decision making power of women to use family planning is a huge problem in SSA region, to the best of our knowledge, there is no study that investigates the factors associated with decision making power to use family planning among married women in the region. Hence, this study was conducted to ll the gap in literature in determining the associated factors of decision making power of women to use family planning in sub-Saharan Africa, which in turn highly relevant to recommend the design and implementation strategies of decision making power of women to use family planning in the region.

Data source
This study used the most recent appended demographic and health survey (DHS) datasets of 35 sub-Saharan countries which were conducted from 2009 to 2018. The DHS is a nationally representative survey, collected every ve years, to provide population and health indicators at the national and regional levels. A pretested standard Demographic and Health Survey questionnaires were used for data collection of the DHS surveys. The questionnaire was conceptualized to the different countries context and the data were collected by trained data collectors. The datasets of each sub Saharan country were obtained at https://dhsprogram.com/data/dataset_admin/index.cfm.Those countries with no data on decision making power of women to use family planning were excluded from the analysis. In this study, a total weighted sample of 83,882 women were included (Table 1). Data analysis procedure We used STATA 14 software to extract, recode and analyze the data. The data were weighted before doing any statistical analysis to restore the representativeness of the sample and get a reliable estimate and standard error. The whole procedure of weighting and its rationale is found on the guide of DHS statistics (32).
Due to the correlated nature of DHS data, measures of community variation/random-effects such as Median Odds Ratio (MOR), Interclass Correlation Coe cient (ICC), and Proportional Change in Variance (PCV) were calculated. Accordingly, the values of these measures were found to be signi cant, and hence the use of multilevel logistic regression model is more appropriate than using ordinary logistic regression.
To choose the best tted model, rst we developed four models and compared them with Deviance. These were: the null-model; a model with no independent variable, model I ; a model that have individuallevel factors only, model II; a model with community-level factors only and model III; a model that contain both community level and independent variables. Of the four models tted, model III was selected as best tted model (it had the lowest Deviance).
Then after, both bivariable and multivariable multilevel logistic regression was performed to determine the associated factors of decision making power of women to use FP in SSA. All variables which has a p value < 0.25 during bi-variable analysis were entered into the multivariable logistic regression model. In the nal model, we used p value ≤ 0.05 to declare statistically signi cant variables.

Sociodemographic characteristics of the respondents
The total weighted samples of 83,882 married women were included in this study. Of these, 22.9% of the respondents were in the age group of 25-29 years and while more than half (60 %) of the respondents were live in rural area. More than one-third of both the respondents (39.9%) and their husbands (34.4%) had primary education level. Majority of the respondents (73.3%) and their husbands (92.7%) were currently working /employed. Similarly majority of the study participants (67%) had media exposure ( Table 2).  Random effect analysis The random-effects model result showed that there is signi cant clustering of decision making power of women to use family planning across the communities (OR of community level variance = 0.07, 95% CI = 0.06-0.10). The value of ICC in the null model revealed that 2.16 % of the overall variation of decision making power of women to use family planning was attributed to cluster variability. The 1.23 MOR value of the null model also indicates the presence of variation in the decision making power of women to use family planning between clusters. It means if we randomly select households from different clusters, those households at the cluster with higher decision making power of women to use family planning had 1.23 times higher chance of decision making power to use family planning compared to their counterparts. As you can see in the Table 3 below, model III has the lowest Deviance value. Hence, it was selected as best tted model ( Table 4). Women who are currently working were 1.27 (AOR = 1.27; CI: 1.20, 1.33) times more likely to have decision making power to use contraceptive as compared to women who were not currently working. Women who had media exposure were 1.1 (AOR = 1.08; CI: 1.03, 1.13) times more likely to have decision making power on family planning use as compared to those women who did not had media exposure.
Similarly, the odds of decision making power on family planning among participants who had 1-3 and ≥ 4 ANC visit was increased by 12%(AOR = 1.12; CI:1.05,1.20) and 14% (AOR = 1.14;CI:1.07,1.21) than those who had not ANC visit, respectively. Besides, the odds of decision making power on family planning among participants who had informed about family planning was increased by 9% (AOR = 1.09; CI: 1.04, 1.15) than their counterparts. Women whose husband's desire want fewer children 14% (AOR = 0.86; CI: 0.79, 0.93) reduced chance of decision making power for family planning than their counterparts.
Women who had less than 3 and 3-5 children were 1.12 (AOR = 1.12; CI: 1.02, 1.23) and 1.08 (AOR = 1.08; CI: 1.01, 1.16) times higher odds of decision making power to use family planning than women who had > 5 children, respectively. Women who did not have children were 48% reduced odds of decision making power to use FP than women who want to have children (AOR = 0.52; CI: 0.47-0.58). Moreover, the odds of decision making power to use FP was increased by 1.12 (AOR = 1.10; CI: 1.04, 1.17) times among participants who do not want other children than those who wants to have other children (Table 5).

Discussion
The main aim of this study was to determine associated factors of decision making power to use family planning among married women in sub-Saharan Africa. Accordingly this study; age, education level of women, women and their husbands' occupation, wealth index, region, media exposure, ANC visit, fertility preference of the women, husband's desire for the number of children and information about family planning were factors associated with decision making power of women to use family planning.
As this study showed older women were more likely have decision making power to use FP. This nding is similar to a study conducted in Ethiopia(26), Mozambique (33) and Bangladesh (34). This might be due to if a woman gets more aged, she may feel more self-con dent to decide on family planning use individually and by discussing with her husband (35). In most African countries context, young women might not be expected to argue with their older husbands and are required to respect their opinions at all time which might be responsible for low decision making power of women to use FP.
The present study revealed that education level of women was associated with decision making power of women to use FP. Consistently, other studies also showed that women who were more educated were higher odds of decision making power to use family planning (4,25,35,36). Education improves women's control over their reproductive choices by increasing women's position within the family and educated women are more likely to desire smaller families than others and hence have a stronger motivation to practice contraceptives (37).
This study showed that those women and their husbands who were currently working contribute to decision making power of women to use FP. This nding is similar with other studies in Malawi (38), Ethiopia (4), Nigeria (39) and South Asia (40). Women who had occupation might have power and resources and consequently independent in making decisions since they may not depend on their spouses for every resource compared to those who had no occupation. Besides, women whose husbands had occupation may improve the family life generally and this may contribute in women decision making power to use FP indirectly.
Similarly wealth index was positively associated with decision making power of women. Those women from a household with richest wealth index had higher chance of decision making power to use FP than the poorest one. This nding is in line with other previous studies which explain that women's economic status impacts their health and decision-making power on contraceptive usage (41)(42)(43). Women who had more income might have access and exposure to mass media about contraceptive and hence increase the likelihood of women decision making power to use contraceptive. Furthermore, in this study, media exposure was also associated with women decision making power to use FP which is in line with other studies (26,39). Mass media helps to increase the decision making power of women to use contraceptive (27).
In the present study we observed that women who had more children were less likely to have decision making power on contraceptive use as compared to those who had fewer children. This nding seems odd and contrast with other studies (21,25). This might be due to related to some religions which teach their followers no to use any modern family planning methods. On the other hand, in this study we also observed that women whose husbands' had high desire number of children had poor decision making power to use FP. This nding was similar to a study conducted in Hondurans (44) and Ethiopia (28). This could be related to husbands might in uence women not to use FP especially in developing countries.
Hence, awareness on the importance of FP should be given not only for women but also for their husbands.
In this study women who were informed about FP at health facility had more decision making power to use FP compared to their counterparts. This nding is consistent with other studies (45,46). The implication of this nding is those women who have information and knowledge about family planning could help them to discuss and in uence their husbands. Similarly this study showed that those women who had ANC visit had more likely to have decision making power to use family planning for next. This nding was also consistent with other studies (22,47). This might be due to as women go to health facilities for ANC service, they could be getting health information including family planning.
One of the strength of this study is using a very huge and representative datasets of 35 sub-Saharan countries which makes the nding of the study to be generalizable for the region. The other strength of this study is the use of multilevel modeling, a model that accounts the nested/hierarchical nature of the demographic and health survey to get reliable estimates. However, the study has also limitation. Because of the secondary nature of the study, we are unable to assess self-stem and others similar psychological variables which could have strong association with the decision making power of the women to use family planning.

Conclusion
Age, women level of education, women and their husband's occupation, wealth index, media exposure, ANC visit, fertility preference, husband's desire for the number of children, region and information about family planning were factors associated with decision making power to use family planning among married women. Behavior change interventions including health education and promotion in this region should target on young married women, women who are not educated, women who are not currently working and whose husband's desire more number of children to improve the decision making power of women to use family planning. Since we have used a secondary analysis of DHS data, obtaining ethical approval was not needed. However, we have received a permission letter to download and used the data les from DHS Program. Anyone who wants to use the DHS data can obtain from the DHS Program website (https://dhsprogram.com/data/dataset_admin/index.cfm) after reasonable request.

Consent for publication
Not applicable Availability of data and materials All data related to the study were included in the manuscript.

Competing interests
All authors declare that they have no competing interests Funding There was no funding for this study.
Authors' contributions GDD and YY designed the study, analyzed the data and drafted the manuscript. YA, WA and AAG were involved in the analysis of the data and critically reviewed the article. All authors read and approved the nal manuscript