Perceived Ideal Number of Children Among Adolescents in Sub-Saharan Africa: Does Exposure to Family Planning Messages Matter?


 BackgroundEven though evidence shows that fertility transition has begun almost everywhere in sub–Saharan Africa, the decline has been slower than in other parts of the world. Researchshows that there is a positive relationship between fertility levels and desired family size. Therefore, many countries in the region are implementing family planning education campaigns targeting at influencing reproductive behaviour of women. Thus,this studyaimed at examiningthe extent to which exposure to family planning communication is associated withdesired family size in adolescencein sub-Saharan Africa.MethodsThis study used data extracted fromthe most recent Demographic and Health Survey datasets for 28 countries in sub-Saharan Africa. Analyses were conducted on a pooled weighted sample of 87,950female adolescents’ aged 15-19 yearscaptured in respective country surveys. Multivariate logistic regression models were fitted in Stata version 16 software to examine the effects of exposure to family planning communication on adolescents desired family size.ResultsFindings show that regardless of the country,exposure to family planning communication among adolescentshad a significant effect on desire for fewer children. Average ideal family size among adolescents who had exposure to family planning communication was (3.8 children compared to 4.5 children;p<0.001) among those with no exposure.Marital status, wealth status, education level,contraceptive use and employment status were also significantly associated with adolescents’desiredfamily size.Country level differences were observed regarding desired family size among adolescents. Adolescents from West Africa Countries had higher desired family size.ConclusionExposure to family planning communication has shown the potential to influence adolescents’ fertility behavior in the region. Desire for large family size is predominantly high among adolescents from West and Central African countries. Suggesting that high fertility remains a social concern in these regions. Furthermore, the study found sub-regional variations in determinants ofdesired family size among adolescents. Calling for the need to scale-up family planning education programmes especially in West and Central Africa countries.

level,contraceptive use and employment status were also signi cantly associated with adolescents'desiredfamily size.Country level differences were observed regarding desired family size among adolescents. Adolescents from West Africa Countries had higher desired family size.

Conclusion
Exposure to family planning communication has shown the potential to in uence adolescents' fertility behavior in the region. Desire for large family size is predominantly high among adolescents from West and Central African countries. Suggesting that high fertility remains a social concern in these regions. Furthermore, the study found sub-regional variations in determinants ofdesired family size among adolescents. Calling for the need to scale-up family planning education programmes especially in West and Central Africa countries.
Background Population growth remains a major subject of concern among development practitioners, policy makers and demographers among other professionals the world over (1,2). It is forecasted that the world population will reach 9.73 billion around mid-century (2). This growth rate has negative consequences including food insecurity, environmental degradation, poverty, unemployment, low quality of life, uncontrolled urbanization, climate change and political turmoil (3,4). However, it is well-known that population growth and projections are not uniform across sub-regions in sub-Saharan Africa (5)(6)(7)(8). Sub-Saharan Africa (SSA) is one of the regions among the developing world where population size has been increasing faster than it can generate resources to support it (2,5). Recent statistics show that the average total fertility rate for sub-Saharan Africa is 4.8 children per woman compared to an average of 3.8 children per woman for all least developing countries worldwide (5). Evidence shows that even though fertility transition has begun almost everywhere, the decline has been slowerin SSAcompared to other parts of the world (8-12).
However, sub-Saharan African countries are not homogeneous as there are important regional variations in total fertility rate. As of 2015, the total fertility rate in many countries in Eastern and Southern Africa has been on the decline trajectory, whereas the total fertility rate in Western and Central Africa has remained stable at approximately 6 children per woman (1,13). Fertility has been predominantly higher in Western and Central Africa due to two major factors. In the rst place, there has been a low uptake of family planning, with little improvement over time (1,14,15). Another factor that has an adverse impact on family planning programming is fertility desire, which has been predominantly pro-natal.
There is consensus in literature that access to family planning education in uences reproductive outcomessuch as age at rst sex, contraception, age at rst birth, teenage pregnancy and fertility preferences among women of reproductive age (16)(17)(18)(19)(20). Given the observed high levels of fertility in SSA(6,8,21), family planning education campaigns have focused on disseminating messages on bene ts of smaller family sizes on both maternal and child health. Demographic evidence indicates that there is a positive relationship between fertility levels and desired family size(6,7). So that a decrease in desired family size is likely to result in a decline in total fertility rate.It is therefore prudent that such campaigns should target changing reproductive behaviors of people,especially adolescents. This is because adolescents' perceived ideal number of children is highly likely to impact on acountry's future fertility course (22)(23)(24)(25). Earlier studies conducted in Guinea,Ethiopia, Nigeria, Ghana and Zambia showed that family planning education in uenced women's reproductive behavior, including desire for smaller family size and increased contraceptive use (16,26-29).
However, there are several gaps inthe literature on fertility desire among female adolescents in sub-Sahara Africa. First, although family planning communication through media or health facility visit has the power to in uence fertility behavior especially among adolescents, there are fewer studies that have focused on the role FP communication plays in in uencing adolescents' ideal family size in SSA. Second, there is missing information on how socio-economic, cultural and demographic factors in uence adolescents'fertility desire at sub-regional levels in SSA. Consideringthat there is vast heterogeneity in socio-cultural norms and valuefor children acrosscountries in SSA (30), it is important to study determinants of fertility desire in adolescents to inform appropriate design of regional family planning strategies to reduce fertility in SSA. Furthermore, several studies on adolescents' fertility desire in SSA have focused on country level analysis. Even though studies on FP and fertility have shown a positive association in most SSA countries, it remains unclear how family planning communication in uence fertility desires among adolescents in SSA. A holistic understanding of howFP communication in uences adolescents' desired family size would produce information relevant to inform fertility policy and FP programming to effectively contribute to fertility decline in the region.
In this study, we used data from nationally representative cross-sectional surveys to have a comprehensive understanding of the extent of the association between exposure to family planning communication and cited ideal family size during adolescence in SSA. The ndings could inform strengthening of family planning policy suggestions tofurther reduce fertility desire among adolescents in SSA. Using multivariate logistic regression models, the study also sought to establish sub-regional and country level heterogeneity in the in uence of FP communication and desired family size among adolescents.

Data source
The study used data extracted from the mostrecent Demographic and Health Survey (DHS) datasets for 28 countries in sub-Saharan Africa conducted between 2009 and 2018 ( Table 1). The DHS programme draws national representative samples of households which are usually selected via two-stage strati ed cluster sampling technique (31). Women aged 15-49 and 15-59 are selected for interviews. The interviews are conducted using three main questionnaires namely; household questionnaire, woman questionnaire and men questionnaire. Participants in the DHS survey were interviewed by eld workers who were wellversed in a wide range of sexuality and family planning and reproductive health topics. DHS data are typically weighted to account for the complexity of survey design and response bias, with the goal of ensuring that the sample is truly representative of the general population (32).

Study sample
The analysis samples for this study comprised female adolescents' aged 15-19 years extracted from each country's recent DHS. The data came from the women individual recode les (IR dataset) for each country.The samples included all adolescent who were not declared infecund or sterile. This resulted in aweighted pooled sample of 87, 884 adolescents included in the analysis.The samples ranged from 1,505 adolescents' in South Africa to 8,423 in Nigeria. Adolescents who reported non numeric ideal number of children were excluded from the analysis.

Outcome variable
The outcome variable of interest in this study is ideal number of children. The DHS program usually collects information onideal number of children from all interviewed women aged 15-49 years. For our analysis, the outcome variable was classi ed in two levels; Inthe rst level, the outcome variable was classi ed as a discrete distribution to facilitate computation of overall average number of ideal children in SSA and across countries included in the study. In second rst level, we classi ed the outcome as binary, such that threshold of 3children or less was classi ed as "0" representing preference fora small family and adolescents who desired more than 3 childrenwere classi ed as "1" representing desire for a large family size. This choice for the cut-off was informed by existing literature on determination of low or high fertility (33-35).

Independent variables
Based on literature review, we identi ed individual and household level predictors that could potentially be associated with fertility desire of adolescents in SSA. These variables are classi ed as socio-economic and demographic factors. DHS reference materials and data collection forms were used to identify the independent variables of interest presented here. The main predictor variable for this study was exposure to family planning information. This is a composite variable, constructed by merging 3 related variables (that is, exposure to mass-media FP messages, exposure to FP messages at health facility and exposure to FP messages via visit a community health work). Other control variables included in the study were; age of adolescent categorised as (15, 16, 17, 18 and 19); current marital status (categorised as never married, currently married/living with partner and formally married; residence (urban; rural); education (no education, primary, secondary, tertiary); Household wealth index (categorised as poor, middle, rich); religion (catholic, protestant, Muslim, other); employment status (categorised as employed, unemployed) and contraceptive use (not using a method, using a method) andvisited health facility in the last 12 months (yes, no),

Statistical analysis
Statistical software Stata SE version 16.0 was used to perform complex survey analysis by taking into account sample weight. Descriptive analysis was performed to summarize study samples for each countryincluded in the study. Categorical variables were presented using frequencies and percentages while means were computed for continuous data. Cross tabulations were conducted to explore bivariate association between exposure to family planning messages and ideal number of children for each country. Furthermore, analysis was conducted to statistically assess mean differences in perceived ideal number of children between adolescents who were exposed to family planning messages and those who were not for each country and for SSA in general. We also conducted multivariate binary logistic regression to examine the determinants of perceived ideal number of children among adolescents in SSA. The choice of this model was informed by the dichotomous distribution of the dependent variable. The multivariate models were tted in two steps. In the rst model (model I) we only included our main explanatory variable (exposure to FP messages). This was followed by model II where allcontrol variables were entered into the model. On the basis of both models, the odds ratio (OR) were calculated and presented along with their respective 95 percent con dence intervals (95 percent CI).

Ethical approval
The datasets used in the analysis for this study are publicly available on DHS program website (https://dhsprogram.com/). Permission to use the dataset was obtained through registration of the study at DHS program. No ethical approval was required since the study used secondary dataset which do not contain any personal identifying information.

Results
Twenty-eight DHS datasets were included in the study. They outline description of the sample information for the study in presented in Table 1 and Figure 1. Findings show that the average desired family size among adolescents in SSA was 4.6 (95% CI: 4.5 -4.7). The average perceived ideal number of children among adolescents ranged from a low of 2.1 children in both Lesotho and South Africa (95% CI: 2.0 -2.2) to highs of 9.5 children (95% CI: 8.7-10.4) in Mali and 8.1 (95 CI: 7.9-8.3) in Niger. Furthermore, our study found that, overall, six in every ten adolescents in SSA preferred a large size. Niger and Chad had the highest proportion of adolescents who desired large family sizes (96.0% and 95.6%) respectively, while Lesotho and South Africa had the lowest percentage of adolescents preferring large family sizes 7.5% and 11.1% respectively.
Generally, average ideal number of children among adolescents in SSA is highest in countries from Western Africa. Ghana had the lowest proportion of adolescents who had desire for large family size in the subregion (60%) while Niger had the highest at 96%. Countries from Southern Africa recorded the least average ideal number of children. Lesotho had the lowest percentage of adolescents who had desire for a large family size (7.5%) and Zambia had the highest at 60%.  Furthermore, use of contraception and visiting the health facility in the last 12 months were positively associated with reduced odds of desire for large family size as twenty-nine of adolescents who were using contraception and 14% who visited a health facility were less likely to desire a large family size.

Discussion
The study focused on examining the in uence of exposure to family planning messages on desired family size among adolescents in SSA. Study also examined sub-regional variations in factors in uencing desire for family size among adolescents. Our review of the literature reveals that there has been no known comprehensive study of this nature conducted before in sub-Saharan Africaand thus bolstering the importance of our ndings. We focused on understanding adolescents'ideal number of children because future fertility of a country is highly likely to be in uenced by present adolescents' reproductive behavior. Sincefamily planning services are mostly offered during antenatal and under ve clinics, adolescentsmay prefer getting family planning information via mass-media channels (36).Our study revealed that exposure to family planning messages wassigni cantly associated with adolescents' desire for family size in sub-Saharan Africa. Other variables found to be important included age, marital status, household wealth status, education level, contraceptive use and visiting health facility in the last 12 months prior to data collection.
Study ndings revealed that adolescentswho had exposure to family planning messages were 17% less likely to desire a large family size comparable to those who had no exposure. This is can be attributed to the appreciation of bene ts of family planning education gainedby adolescentsthrough via mass-media, visit to health facility and community visits by community health workers. Family planning messages are usually targeted at in uencing individuals' reproductive behavior towards contraception, limiting of births, spacing of children and choice of small family. This ndingimplied a signi cant contribution of family planning education interventions to social and reproductive behavior change among adolescents in SSA. Similar results were reported in a study conducted in Rwanda in 2016 (37) where reduction in fertility desire among women of reproductive age was attributed to massive family planning education through mass-media and community-led sensitization programmes.
The study found that adolescents who were married or living with a partner were twice as likely to prefer a large family size compared to the never married. In most African culture, women may be expected to start having children soon after marrying, causing partnered women to stop using contraception(38). Such practices could explain why teenage pregnancy and early motherhood are high in sub-Saharan Africa.
There may also be sub-regional disparities in the norms surround timing of having a child after marriage. This result implies the urgent need for community led family planning education programmes aimed at in uencing reproductive behavior change of married adolescents, especially those coming from rural settings.
Earlier studies have shown that education and wealth status are strongly associated with fertility desire(9,39,40) such that individuals with higher level of education and those from higher wealth groups have a tendency to desire low family size. This is because this demographic has adequate information about bene ts that accrue with smaller family sizes. Our study con rms the ndings presented by earlier studies. We established that adolescents with secondary and higher-level education were 70% and 7% less likely to desire a large family size, respectively. These ndings are consistent with similar studies conducted in Zambia, Rwanda, Burkina Faso, Niger, Mali and Egypt that reported education and wealth status as signi cant predictors of ideal number of children (23,27,39,41). This implies that education is an important component for reducing fertility in SSA. Therefore, education policies should propagate the implementation of strategies that improve education access to girls and young women, especially those in marginalized communities.
Another notable nding from this study was the effect of contraceptive use on desired family size.
Contraception in adolescence was associated with a low chance of desiring large family sizes. Twentynine percent of adolescents who used any form of contraception preferred a small family size. This nding signi es the importance of family planning programmes in sub-Saharan Africa's agenda to reduce population growth rate through fertility reduction. Many studies conducted in SSA (17,39,40,42) have found that contraception enables women in the reproductive age to make an informed decision about planning their births and decide their desirable number of children. Such decisions have far-important health bene ts for couples and their households at large.
Access to family planning communication is another issue that needs to be addressed in most countries in SSA. Places such as schools, community youth friendly corners, private pharmacies/drug stores and traditional ceremonies can also serve as distribution points for family planning information to adolescents. The merits of disseminating family planning information in schools through introduction of comprehensive sexually education need to be explored further. Improving the demand for family planning information among adolescents should be stressed in the country's population policies as key priority strategy to reduce fertility further. Increasing access to family planning information is essential and has been shown to have a signi cant impact on decision making to use contraception, postponing of marriage and limiting the number of children, thus reducing fertility(43,44). Health education on limiting family size through family planning programming will surely assist in changing reproductive behavior of adolescents, but it will only be effective if adolescents will embrace the advantages of having smaller families.
Although the study has provided useful ndings to inform strengthening of family planning education programmes targeting at changing adolescents reproductive behaviour. There are a few limitations that could make the conclusions from thestudy to be interpreted with caution. First, because our data is crosssectional, we cannot conduct causality analyses, which limits our ability to understand the complexities of adolescents' experiences regarding their desire for children through their life cycle. As a result, our ndings highlight the need for additional research, particularly qualitative and longitudinal research, to further our understanding of the complex interplay between the various individual and community factors that shape adolescents' reproductive behavior and desire for children, as well as how these factors change over the course of their lives. Finally, because of the lack of other country level data on the DHS program website, the study did not contain data for all countries in SSA. As a result, the conclusions should not be extrapolated beyond the sub-sample of nations included in our analysis.

Conclusion
This study has shown that family exposure to family planning messages has the potential to in uence adolescents' future reproductive behavior in SSA. Although desire for large family size in adolescence is high SSA, Western and Central African countries had the highest percentages of adolescents desiring large family size. This suggests that high fertility remains a social concern in the region. Furthermore, the major factors associated with preferred ideal number of children among adolescent in SSA include; marital status, wealth status, education level, contraceptive use, visit to health facility in last 12 months and exposure to family planning messages. There is a need for governments and stakeholders especially in Central and West African countries to prioritise strengthening of family planning communication programmes that provides education on bene ts of smaller family sizes,targeting mostly in and out-ofschool adolescents in order to reduce fertility further in the regions. It is important to incorporate sexual reproductive health education into early primary and secondary level curriculum to maximize bene ts of family planning programmes. Further research is needed to examine how exposure to family planning messages operates through community level factors to in uence desiredfamily size among adolescents across different countries in SSA. The study utilized secondary data extracted from recent country DHS datasets for 28 countries in sub-Saharan Africa. The DHS program allowed for permission to use survey datasets. The datasets have no personal identities for research participants. All DHS studies were approved by the respective country ethical review boards and the Centers for Disease Control and Prevention (CDC) Atlanta. The surveys data collection procedures required consent from participants aged 18 and older. The survey protocol also required authorization from parents/guardians for all participants aged 15-17 years before seeking accent from teenagers.

Consent for publication
Not applicable Availability of data and materials Data used in our study is publicly available upon request from DHS program website. (https://dhsprogram.com/).

Competing Interests
Authors declare no competing interest.

Funding
No funding was received.
Author contributions MP developed the concept for this study, performed data analysis and wrote data interpretation text for the analysis and wrote discussion section. MLprepared the methodology. MS and SS prepared the background and conclusion section.MP and SS performed overall review and editing of the manuscript for intellectual content. All authors have read and approved the nal version of this manuscript.
42. Ariho P, Kabagenyi A. Age at rst marriage, age at rst sex, family size preferences, contraception and change in fertility among women in Uganda: analysis of the  Figure 1 Description of Adolescents by exposure to FP messages and ideal family size