Female adolescents’ reproductive health decision-making capacity and contraceptive use in sub-Saharan Africa: What does the future hold?

Introduction Given the social, economic, and health consequences of early parenthood, unintended pregnancy, and the risks of HIV infection and subsequent transmission, there is an urgent need to understand how adolescents make sexual and reproductive decisions regarding contraceptive use. This study sought to assess the association between female adolescents’ reproductive health decision-making capacity and their contraceptive usage. Materials and methods Data was obtained from pooled current Demographic and Health Surveys (DHS) conducted in 32 countries in sub-Saharan Africa (SSA). The unit of analysis for this study was adolescents in sexual unions [n = 15,858]. Bivariate and multivariable analyses were conducted using Pearson chi-square tests and binary logistic regression respectively. All analyses were performed using STATA version 14.2. Results were presented using Odds Ratios [OR] and adjusted Odds Ratios [AOR]. Statistical significance was set at p<0.05. Results The results showed that 68.66% of adolescents in SSA had the capacity to make reproductive health decisions. The overall prevalence of contraceptive use was 18.87%, ranging from 1.84% in Chad to 45.75% in Zimbabwe. Adolescents who had the capacity to take reproductive health decisions had higher odds of using contraceptives [AOR = 1.47; CI = 1.31-1.65, p < 0.001]. The odds of contraceptive use among female adolescents increased with age, with those aged 19 years having the highest likelihood of using contraceptives [AOR = 3.12; CI = 2.27-34.29, p < 0.001]. Further, the higher the level of education, the more likely female adolescents will use contraceptives, and this was more predominant among those with secondary/higher education [AOR = 2.50; CI = 2.11-2.96, p < 0.001]. Female adolescents who were cohabiting had higher odds of using contraceptives, compared to those who were married [AOR = 1.69; CI = 1.47-1.95, p < 0.001]. The odds of contraceptive use was highest among female adolescents from the richest wealth quintile, compared to those from the poorest wealth quintile [AOR = 1.65; CI = 1.35-2.01, p<0.001]. Conversely, female adolescents in rural areas were less likely to use contraceptives, compared to those in urban areas [AOR = 0.78; CI = 0.69-0.89, p < 0.001]. Conclusion The use of general and modern contraceptives among adolescents in SSA remains low. Therefore, there is a need to strengthen existing efforts on contraceptives usage among adolescents in SSA. This goal can be achieved by empowering these young females, particularly those in the rural areas where the level of literacy is very low to take positive reproductive health decisions to prevent unintended teenage pregnancy, HIV/AIDs and other sexually transmitted infections. This approach would help reduce maternal mortality and early childbirth in studied SSA countries. Copyright: The study used pooled data from current DHS conducted from January 1, 2010 and December 31, 2018 in 32 countries in SSA. Table 1 provides a full list of the countries included in the survey and their survey years. The 32 countries were used in the study because they had complete data on the variables of interest for this study. The study included the countries under the DHS program in order to provide a holistic and in-depth evidence of the relationship between reproductive health decision-making capacity and contraceptive use in SSA. DHS is a nationwide survey collected every five-year period across low- and middle-income countries. The survey is representative of each of these countries. The survey targets core maternal and child health indicators such as unintended pregnancy, contraceptive use, skilled birth attendance, immunisation among under-fives, and intimate partner violence. Stratified dual-stage sampling approach was employed. The same questions were posed to all women, making it feasible for a multi-country comparisons. Selection of clusters (i.e., enumeration areas [EAs]) was the first step in the sampling process, followed by systematic household sampling within the selected EAs. The population for the study was obtained from adolescents aged 15–19 from whom data were collected during the current individual surveys for each country. For the purpose of this study, only adolescents (15–19 years) in sexual unions (marriage and cohabitation) who had complete cases on all the variables of interest were used (N = 15,858). Detailed description of the study sample is shown in Table 1. We relied on the Strengthening the Reporting of Observational Studies in Epidemiology’ (STROBE) statement in conducting this study and writing the manuscript. Source: Authors’ computations The dependent variable in this study was “contraceptive use” which was derived from ‘current contraceptive method’. The responses were coded 0 = “No method”, 1 = “folkloric method”, 2 = “traditional method,” and 3 = “modern method”. The existing DHS variable excluded women who were pregnant and those who had never had sex. The modern methods included female sterilization, intrauterine contraceptive device (IUD), contraceptive injection, contraceptive implants (Norplant), contraceptive pill, condoms, emergency contraception, standard day method (SDM), vaginal methods (foam, jelly, suppository), lactational amenorrhea method (LAM), country-specific modern methods, and respondent-mentioned other modern contraceptive methods (e.g., cervical cap, contraceptive sponge). Periodic abstinence (rhythm, calendar method), withdrawal (coitus interruptus), and country-specific traditional methods of proven effectiveness were considered as traditional methods while locally described methods and spiritual methods (e.g., herbs, amulets, gris-gris) of unproven effectiveness were the folkloric methods. To obtain a binary outcome, all respondents who said they used ‘no method’ were put in one category and were given the code “0 = No” whereas those who were using either folkloric, traditional, or modern method were also put into one category and given the code “1 = Yes.” The main explanatory variable was reproductive health decision-making capacity. Two derivative variables that focused on decision-making on sexual intercourse and condom use were used to generate this variable. For decision-making on sexual intercourse, female adolescents were asked whether they can refuse their partners sex while for condom use, female adolescents were asked whether they can ask their partners to use condoms during sexual intercourse. Reproductive health decision-making capacity was then generated from the combination of the decision-making on sexual intercourse and condom use variables. Following previous studies [18,19] on reproductive health decision-making capacity, the original response category of these variables (1 = Yes, 2 = No and 3 = Don’t know/Not sure) were re-categorised, whereby “No and Don’t know/Not sure” were recoded as “No” and recoded as 0, with “Yes” recoded as 1. Responses coded as “0” represented female adolescents’ inability to make reproductive health decision whilst those coded as “1" were labelled as those capable of making reproductive health decisions. Apart from the main explanatory variable, eight other variables were considered in the study as covariates. These variables were age, place of residence, wealth quintile, employment status, educational level, marital status, age at first sex, and survey country. Again, apart from survey country, these variables were not determined a priori; instead, they were selected based on their significant associations with the outcome variable—contraceptives use as reported in previous studies [27,28]. Four of the covariates were recoded to make them meaningful for analysis and interpretation. Marital status was recoded into “cohabiting (1)” and “married (2).” Employment status was captured as “not working (0)” and “working (1).” Educational level was recoded as “no education (0),” ‘primary (1),” and “secondary/higher (2).” Age at first sex was coded as “less than 16 years (1)” and “16–19 years (2).” The analysis begun with a computation of contraceptive prevalence and proportion of female adolescents who could refuse a partner sex, ask a partner to use condom during sex, and those who had the capacity to make reproductive health decisions in the 32 SSA countries. The syntax “metaprop” in STATA version 14.0 (StataCorp, College Station, TX, USA) was used to generate forest plots for each of these indicators. Each forest plot showed the prevalence of an indicator in individual countries and its corresponding weight, as well as the pooled prevalence in all the countries and its associated 95% confidence intervals (CI). A test of heterogeneity of the DHS data obtained for the different countries showed a high level of inconsistency (I2 > 50%), thereby warranting the use of a random effect model in all the meta-analysis (see Figs ​Figs1,1, ​,2,2, ​,33 and ​and4).4). Secondly, the datasets were appended and a total sample of 15,858 was generated. After appending, contraceptive prevalence across the socio-demographic characteristics with their significance levels and chi square values [χ2] were calculated. Multicollinearity test was also performed and with a mean VIF of 1.21, there was no evidence of multicollinearity between the variables. Using the explanatory variables which were significantly associated with contraceptive use (p<0.05) among female adolescents from the chi-square test, a binary logistic regression analysis in a hierarchical order was performed. Model I looked at a bivariate analysis of the main independent variable (reproductive health decision-making capacity) and the outcome variable (contraceptive use). Model II was a complete model comprising all the explanatory variables and the outcome variable (see Table 3). In line with research evidence that modern contraceptive methods are the most effective [29,30,31], a further analysis was done to examine the association between reproductive health decision-making capacity and modern contraceptive use (see Table 4). All frequency distributions were weighted using v005/1000000 while the survey [svy] command in STATA version 14.2 was used to adjust for the complex sampling structure of the data in the regression analyses. Missing values were treated by using complete cases for our analysis. Results for the regression analysis have been presented as Crude Odds Ratios (COR) and Adjusted Odds Ratios (AOR), with their corresponding 95% confidence intervals (CI) that signify precision and significance of the reported OR values. Statistical significance was set at p<0.05. Source: Authors’ computations. Source: Authors’ computations. Source: Authors’ computations. Source: Authors’ computations. * p<0.05 ** p<0.01 *** p<0.001; Ref = Reference, CI = Confidence Intervals, COR = Crude Odds Ratio, AOR = Adjusted Odds Ratio Source: Authors’ computations * p<0.05 ** p<0.01 *** p<0.001; Ref = Reference, CI = Confidence Intervals, COR = Crude Odds Ratio, AOR = Adjusted Odds Ratio Source: Authors’ computations The DHS surveys obtain ethical clearance from the Ethics Committee of ORC Macro Inc. as well as Ethics Boards of partner organisations of the various countries such the Ministries of Health. During each of the surveys, either written or verbal consent was provided by the women. Since the data was not collected by the authors of this manuscript, official permission was sought from MEASURE DHS website and access to the data was provided upon the request that was assessed and approved on 3rd April, 2019. Data is available on https://dhsprogram.com/data/available-datasets.cfm.