Theory-based determinants of modern contraceptive use in sub-Saharan Africa: an analysis of demographic and health surveys

Introduction Despite improved access to modern contraceptives in sub-Saharan Africa (SSA), the region has the highest fertility rate. Although modern contraceptive usage and its determinants in SSA have been assessed, most authors were not guided by behavioral change theories. This study sought to assess the modern contraceptive coverage in SSA and identify the theory-based determinants that need to be considered in demand creation interventions. Methods Data was obtained from the most recent demographic and health surveys conducted across 37 countries in SSA . Estimates of country-specific and pooled Regional modern contraceptive coverage were generated from 501,324 responses. Logistic regression was used to assess the relationship between modern contraceptive use and determinants selected based on the Health Belief and Social-Ecological behavior change models. Results Modern contraceptive coverage in SSA was 22.26% (95% CI: 17.91, 26.60) . The health belief model determinants of modern contraceptive use included last birth by caesarian section (AOR=1.44, 95% CI:1.31,1.59), hearing of family planning at the health facility (AOR=1.18, 95% CI:1.12,1.24), or from at least one media source, being able to negotiate condom use (AOR=1.65, 95% CI: 1.55,1.76), and having a previous terminated pregnancy (AOR=0.76, 95%CI: 0.71, 0.81). The social ecological model determinants of modern contraceptive use included being above 24 years, having at least primary education, non-urgent need for a last child, and being involved in decision-making concerning personal health (AOR=1.81,95% CI:1.71,1.92). Discussion Modern contraceptive coverage in SSA is low. Age, educational status, past obstetric history, exposure to family planning information, ability to negotiate condom use or make personal health-related decisions, and the need for a child were the determinants for modern contraceptive use in the region. Countries need to develop context-specific interventions considering these determinants to help improve coverage and reduce the poor maternal and child health outcomes and the developmental gaps resulting from unplanned pregnancies.


43
In recent times, the rapid population growth across the globe is leading to immoderate resource . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 22, 2022. ; https://doi.org/10.1101/2022.07.20.22277859 doi: medRxiv preprint residence was influential at the community level [22]. At the societal level, women's empowerment 81 and access to contraceptives were noticed by Paul to determine contraceptive use [23]. 82 Although a combination of the HBM and SEM constructs provides a comprehensive list of 83 determinants for contraceptive behavior, their use in identifying determinants of modern 84 contraceptive use in SSA has not been documented. Therefore, the authors sought to assess the 85 modern contraceptive coverage and identify the theory-based determinants of their use in SSA.  larger EA is more likely to be selected than a smaller one. A pre-determined number of households 99 are selected from each EA in the second stage. Eligible participants in those households (women 100 aged 15 to 49 years and men 15 to 59 years) are then interviewed on various topics using 101 comparable questionnaires across countries" [25]. For all 37 countries, data from Individual 102 Recode (IR) file was used in the analysis.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The independent variables for modern contraceptive use were selected based on four constructs 117 from the HBM and three from SEM. Table 1 lists the constructs and their respective variables.

118
Some of the variables were recoded to aid analysis. A new variable was created to determine the 119 number of media sources from which respondents receive family planning information. Row totals 120 of variables seeking to determine if respondents had heard of family planning on radio, TV, or in 121 a newspaper were obtained for each respondent and coded as found in table 1. The "very poor" 122 and "poor" categories in the wealth index variable, as well as the "rich" and "very rich" categories 123 of the variable, were recoded as "poor" and "rich," respectively. Regarding involvement in decision 124 making concerning contraceptive use and decision making concerning the respondent's health, the 125 responses were coded as "involved" if the respondent took those decisions alone or with their . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 22, 2022. coverage, standard errors, and sample population. Subgroup analysis was also conducted using the 140 sub-regional classification of the countries to obtain estimates for West, Central, Eastern, and 141 Southern Africa and was presented with a forest plot (Fig 1).

142
To identify the determinants of modern contraceptive use, the HBM and SEM variables, as between the determinants was assessed using pair-wise correlation, and those with absolute 147 correlation coefficients above 0.7 were excluded from the models. During the multivariable . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 22, 2022. ; https://doi.org/10.1101/2022.07.20.22277859 doi: medRxiv preprint logistic regression, variables with p≤ 0.05 were declared as significant determinants of modern 149 contraceptive use in SSA.

150
Three models were fitted. Model 1 was a multivariable model adjustment for the HBM variables, 151 and Model 2 adjusted for the SEM variables. In Model 3, all the statistically significant variables 152 from Models 1 and 2 were fitted with the outcome variable to obtain the final determinants.    . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 22, 2022. within the West African sub-region (14.42%, 95% CI:11.42-17.41), (Fig 1).  The study sought to find the pooled prevalence of modern contraceptive coverage and its 228 determinants among women in fertility age in SSA. The pooled prevalence was 22.26%. With 229 respect to the determinants, women in fertility age (WIFA) who are more than 24 years and have 230 a minimum of primary education were more likely to use modern contraceptives. Last birth by 231 caesarian section, being exposed to family planning information on at least one media source or at 232 the health facility, and the ability to negotiate condom use were also found to increase modern 233 contraceptive use. WIFA who had no urgent need for a last child were more likely to use modern 234 contraceptives than those who needed one urgently. Finally, those involved in decision-making 235 concerning their health were more likely to use modern contraceptives than those not involved.

236
Being the second most populous and the region with the fastest-growing population, the low 237 coverage of modern contraceptives is of significant concern because only Namibia nearly met the 238 global SDG 3.7.1 thresholds of at least 50% modern contraceptive coverage among WIFA [26].

239
For the region to achieve its growth potential and harness all its resources for development, Aliyu 240 calls for modern contraceptives to be vigorously promoted, not only for their demographic 241 dividends but also for their socio-economic and health-related benefits that will accelerate the 242 attainment of the sustainable development goals [27]. Evidence, however, needs to be used to guide 243 the contraceptive promotion process, and the theory-based determinants identified in the study can 244 serve as a good starting point.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 22, 2022. ; https://doi.org/10.1101/2022.07.20.22277859 doi: medRxiv preprint The finding that exposure to family planning information in newspapers, on TV, or on radio . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 22, 2022. ; https://doi.org/10.1101/2022.07.20.22277859 doi: medRxiv preprint While attempts are made to empower the WIFA in SSA, there is the need to prioritize those 267 between 15 to 24 years who were found in this study to be less likely to use a modern contraceptive

276
The increased odds of modern contraceptive use among women whose last delivery was by 277 caesarian in this study was observed by Bryant to be due to the education provided to them to delay . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 22, 2022. Countries in the region need to invest in developing context-specific interventions considering 295 these determinants to help improve their contraceptive coverage and reduce the poor maternal and 296 child health outcomes that result from unplanned pregnancies. Interventional studies need to be 297 conducted in SSA to assess the effectiveness and cost-benefit ratio of interventions developed from 298 these theory-based interventions in improving modern contraceptive use in the region.

299
Strengths and limitations 300 The study's strengths lie in using nationally representative data pooled together to obtain a large

308
Modern contraceptive coverage in SSA is low. Age, educational status, past obstetric history, 309 exposure to family planning information, ability to negotiate condom use or make personal 310 health-related decisions, and the need for a child were the determinants for MC use in the region.

311
To improve its coverage, interventions should be focused on empowering women through . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 22, 2022. ; https://doi.org/10.1101/2022.07.20.22277859 doi: medRxiv preprint increased enrolment of girls in school, providing sexual and reproductive rights education, and 313 increasing access to family planning information.
Planning. InTech; 2018. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 22, 2022. ; https://doi.org/10.1101/2022.07.20.22277859 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.