Background: The importance of contraception use is immense for young girls of age 15–24 years. In literatures, there were significant attempts made to study factors associated with adolescent and young women contraception use in Africa. Despite the resulting interventions followed those studies, the contraception uses among youth population in Africa remained below average. Thus, this study is aimed to assess individual and community-level factors associated with contraceptive use in Ethiopian context to support further interventions. Methods: Our analysis was based on the secondary data from Ethiopia Demography and Health Survey (EDHS) 2016. Adolescent girls and young women (AGYW) aged 15–24 years were the target population. Means, standard deviations, and proportions were used to describe the study population. To control for the variations due to the differences between clusters, a series of multilevel logistic regression modeling steps were followed and determinants of contraceptive use were outplayed. All variables with bivariate p-value < 0.25 were included in the models and p-value < 0.05 was used to declare associations. Results: The prevalence of modern contraceptive use among AGYW in Ethiopia was 34.89% [95% CI, 0.32, 0.36]. Married adolescents were 2.01 times [AOR = 2.01, 95% CI = 1.39,3.16], having work was 1.36 times [AOR = 1.36, 95% CI = 1.06,1.71], living in urban areas was 1.61 times [AOR = 1.61, 95% CI = 1.16,2.45], being in middle wealth status was 1.9 times [AOR = 1.90, 95% CI = 1.32,2.65], being in rich wealth quintile was 1.99 time [AOR = 1.99, 95% CI = 1.35,2.68], and having TV exposure was 1.61 times [AOR = 1.6, 95% CI = 1.17,2.20] more likely associated with modern contraceptive uses. Conclusion: The use of modern contraception among AGYW in the country remained appealing and factors like region, residence, marital status, wealth index, religion, working status, parity, husband desire children, ever aborted AGYW, and the television exposures were attributed for the poor improvements. Therefore, the enhancements that consult those factors remained remarkable in improving contraception use, while further increasing in educational engagement, access to health services, and economic empowerment of the AGYW might be the good advantages for the improvements.
A cross-sectional survey data from EDHS 2016 were used for this study. EDHS 2016 is the fourth nationally representative survey conducted in Ethiopia. The EDHS 2016 data was collected using a two-level multistage stratified cluster sampling to pick eligible respondents from rural and urban areas of Ethiopia. Different questionnaires were employed to collect data from women, men, couples, and children. The survey was intended to collect and deliver data on several demographic indicators, including sexual and reproductive health data like marriage, pregnancy, fertility, family planning, sexual behavior, maternal health, STIs, and HIV/AIDS [9]. In the current analysis, we included only AGYW (15–24 years) who were sexually active and were not pregnant during the survey from the dataset. The EDHS data were collected from participant by direct face-to-face interviews. The dependent variable for the study is the current use of modern contraception. WHO defines ‘Adolescents’ as an individuals in the age group of 10–19 years and ‘Youth’ in the age group of 15–24-year [15]. We derived the dependent variable from the question that the women asked about the type of contraceptive methods she is using at the time of the survey. We then coded responses as “no method”, “folkloric method”, “traditional method” and “modern method”. Modern methods include male and female sterilization, injectables, intrauterine devices (IUDs), contraceptive pills, implants, female and male condoms, and emergency contraceptive methods. Periodic abstinence (rhythm, calendar method), withdrawal (coitus interruptus), lactational amenorrhea, and we labeled country-specific traditional methods. Locally and spiritually defined methods of unverified effective methods, such as herbs, amulets, and gris-gris methods were the folkloric methods. The existing EDHS data has already excluded women who were pregnant, and those who never had sex from the variables lists. For this study, we coded adolescents and young women using modern contraception methods as ‘1 = yes’ and recoded those not using any modern methods, those using traditional methods, and those using folkloric methods as ‘0 = no’. AGYW’s age at birth was obtained after subtracting the date of birth of AGYW in century month code (CMC) from date of birth of child in CMC. It was then grouped into 15–19 years, 20–24 years. AGYW’s educational status was categorized in to no education, primary, secondary, and technical/vocational or higher. Given the few respondents in vocational and higher categories, it was re-categorized in to: no education, primary, and secondary and above. Religion was categorized into the dominant religion groups as Protestant, Orthodox, Muslim and others. Marital status was defined as single, married, Widowed and divorced. Current working status (AGYW occupation status) was captured by AGYW who are currently have work or who have worked in the last 12 months and recorded as not working and other categories. Since there were other several working categories, it was re-categorized as not working and working. In EDHS, household wealth index was categorized in quintiles as: poorest, poor, average, rich and richest and for this category, principal component analysis was used. Then, we re-categorized the scale in to poor, middle, and rich for easy understanding. Age at first sexual intercourse was a continuous variable, but categorized into < 20 and ≥ 20. Parity is the number of viable children a woman might have. It was grouped in to no birth, one birth, two birth, and three and above births. Husbands’ desire for children was the plan of number of children by husbands. It was defined as husband want some, husband want more, husband want fewer, and don’t know. Abortion is any type of pregnancy ended before 28th weeks of gestation. It includes any spontaneous and non-spontaneous abortion performed for treatment or other purposes. Media exposure was described as hearing information from radio, TV, and newspaper. EDHS assessed exposure to media by asking “Do you listen to the radio or watch to television (TV) at least once a week, less than once a week or not at all?”. These variables first categorized into “yes”, “no”, and not at all Exposure to media variable was considered “yes” if the subject was exposed to one or two of the medias, and said “no” otherwise. The summary of definition of some of the variables were provided in Table 1. The summary of definition of some of the variables There are nine regions (Tigray, Afar, Amhara, Oromia, Somali, Benishangul, SNNPR (south nation nationalities people’s region), Gambela, and Harari) and two city administrations (Addis Ababa and Dire Dawa) in the country. AGYW were selected from any of the eleven areas. indicates whether an individual live in rural or urban or whether place is rural or urban. Descriptive statistics were applied to summarize the study variables as mean, standard deviation, percent or proportions. Before applying descriptive statistics weighting, technique was applied to account for disproportionate sampling and other segregations implemented during sampling. Bivariate analysis was conducted to identify variable that merit to be included in the model. Due to the sampling methods DHS apply, the dependence of responses from different levels of hierarchy was suspected. This implies that a single-level traditional statistical model might not be adequate to control for the clustering effect. Thus, since the data has individual and community levels, we applied multilevel binary logistic regression. The decision was made based on the Intra-Class Correlation (ICC), which showed high dependency due to the clustering of the data at the community level. Four consecutive models were built to identify predictors of modern contraceptives use. Model 1 is an empty (the intercept only model) employed before adding predictors [16]. model 2(fixed effect model) included all individual-level variables that were initially significant at p-value of < 0.25 to determine the level of variance explained by the model. Model 3 (random effect model) included cluster-level (community -level) variables and model 4(the mixed effect model) was the final model in which both the individual and community level variables introduced. All analyses were performed in STATA 14.2 and the output was presented using adjusted odds ratio (AOR) and 95% CI. To determine the community effect, Intra-community Correlation (ICC) was estimated by applying the community level and individual level variances. Likelihood Ratio (LR) test, Median Odds Ratio (MOR), and Proportional Change in Variance (PCV) were also examined to check the fitness of the model using the following statistical formula. ICC= σ2aσ2a+σ2b; where, σ2a is the community level variance and σ2b indicates individual level variance. The individual variance (σ2b) equal to π2/3 that is the fixed value. MOR= e 0.95* Va_1, where, Va_1 is the variance in the empty model. PVC = Va_1−Va_2Va_1, where, Va _ 1 is variance of the empty model and Va _ 2 is neighborhood variance in the subsequent model). Data for this study was accessed from the Demographic Health Survey (DHS) website (http://www.dhsprogram.com). The procedure was confidential and we avoided any ways exposing households or individuals. To collect the data, EDHS obtained permission from the Ethiopian Health Nutrition and Research Institute (EHNRI) Review Board and the National Research Ethics Review Committee (NRERC) from the Ministry of Science and Technology. During the data collection, verbal informed consents were collected from participants and data collectors explained the purpose of the study for participants as published in 2016 EDHS report.
N/A
DIMA AI Care
