Some contraceptive methods, such as long-acting and permanent methods, are more effective than others in preventing conception and are key predictors of fertility in a community. This study aimed to determine which factors were linked to married women of childbearing age who no longer desired children using long-acting reversible contraceptives (LARCs) in Yemen. We used a population-based secondary dataset from Yemen’s National Health and Demographic Survey (YNHDS), conducted in 2013. The study analyzed a weighted sample of 5149 currently married women aged 15 to 49 years who had no plans to have children. Logistic regression analyses were used to investigate the parameters linked to the present use of LARCs. The final model’s specifications were evaluated using a goodness-of-fit test. An alpha threshold of 5% was used to determine statistical significance. Of the total sample, 45.3% (95% CI: 43.3-47.4) were using contraception. LARCs were used by 21.8% (95% CI: 19.6-24.1) of current contraceptive users, with the majority (63.8%) opting for short-acting reversible contraceptives (SARCs). In the adjusted analysis, maternal education, husbands’ fertility intention, place of residence, governorate, and wealth groups were all linked to the usage of LARCs. According to the findings, women whose spouses sought more children, for example, were more likely to use LARCs than those who shared their partners’ fertility intentions (AOR = 1.44; 95% CI: 1.07-1.94; P =.015). In this study, married women of reproductive age who had no intention of having children infrequently used contraception and long-acting methods. Improving women’s education and socioeconomic status could contribute to increasing their use of LARCs.
This study modified Andersen and Newman behavioral model of health service use, which has been widely utilized in research on the use of health services, including family planning.[20,21] The model illustrates how predisposing circumstances (also called psychosocial factors) influence one’s utilization of health care. In other words, four domains – attitudes, knowledge, social norms, and perceived control – influence the decision-making process related to an individual’s planned behavior. Enabling variables are those that make it easier to use the service, such as the availability of adequate individual and community-level resources. Overall, one’s access to and ability to pay for healthcare services may limit their utilization. The term “need” refers to how people describe their health and functional state, which can be negatively or positively influenced depending on how bad their health is. So, the researchers hypothesized that in the study environment, predisposing, enabling, and need factors influence the use of LARCs by married women of reproductive age.[22] This study had a cross-sectional design. The researchers used a secondary dataset from the 2013 YNHDS. This study relied on data from the 2013 YNHDS, which was implemented by the Ministry of Public Health and Population in collaboration with the Central Statistical Organization. The sample for the original survey was selected from 213 clusters in urban areas and 587 clusters in rural areas, giving a total of 800 clusters. The sampling frame used was taken from the 2004 General Population Housing and Establishment Census. Of the 19,517 households selected for inclusion, 18,027 were included in the study. The women’s file (dataset) was used in this study. The dataset contains information about women’s background characteristics such as age, education, type of place of residence, governorate, wealth quintile, and reproductive health data such as fertility and fertility preferences, as well as knowledge and use of FP methods. Information about how the 2013 survey was conducted, including the questionnaire that was used to collect data, is contained in the final report.[11] The original survey interviewed 25,434 women of reproductive age (15–49 year). There were 15,649 married people among those who participated in the study. The researchers eliminated 2166 pregnant women from the sample, leaving 13,483 non-pregnant women in the study. The current study focused on the use of LARCs by married women who no longer wanted to have children. Consequently, the research was limited to 6209 women who said they no longer wanted children to meet the study’s objectives. Records with missing data for any of the study’s explanatory factors (n = 1157) were also eliminated. The final sample of women in this study was made up of 5052 (weighted N = 5149) married women of childbearing age who were not pregnant. In this study, two outcome factors were investigated. The first step was to estimate the percentage of married reproductive women in the study population who were currently using any type of contraception. This was done to provide an estimate of contraceptive use among married women of reproductive age who no longer wanted children in the study setting. Current contraceptive use is a binary dummy variable, with “0” denoting non-users and “1” denoting current users. Participants’ self-reports of contraceptive use by themselves or their husbands provided this information. The current usage of LARCs, the second outcome variable,is also a binary variable coded as “0” for non-users and “1” for current users. Only participants who reported using any method of contraception at the time of data collection were classified in this manner; non-contraceptive users, in other words, were not included in this analysis. According to the information available in the dataset on the contraceptive methods used as reported by the participants, the contraceptive methods were classified as LARCs, which included intrauterine devices, implants, and norplant; SARCs, which included pills, injections, diaphragms, male and female condoms, lactational amenorrhea method, and other modern methods; PCMs, which included male and female sterilization; and lastly, traditional methods (TMs), which included periodic abstinence, withdrawal method, and other traditional methods. The researchers selected specific variables for inclusion in the study as potential determinants based on the current literature and variables accessible in the dataset. The predisposing factors in this study included maternal age, age at first marriage, maternal employment status, maternal educational level, maternal decision-making autonomy regarding health, and number of living children. The enabling factors were husband’s educational level, husband’s employment status, place of residence, governorate, wealth, media exposure to FP information (print media, audio, and audiovisual), and interaction with the health care system (whether the woman visited the health facility or was visited by a health worker in the last 12 mo). Need factors included the husband’s desire for children. Most variables were utilized exactly as they were in the demographic and health survey (DHS) dataset, including maternal age, place of residence, governorate, and wealth quintile (as a composite variable). Based on the existing DHS dataset, new variables were created, such as the number of living children and the employment of the woman and her husband/partner. STATA/IC 15.0 (StataCorp LLC, College Station, TX) was used to analyze the data. To account for the sampling design used by the DHS, weights were applied to the data to generate nationally representative statistics.[23] Descriptive statistics were used to report the distribution of the population analyzed by key characteristics, including sociodemographic and economic factors. A Pearson design-based Chi-square (χ2) test was used to assess differences between current LARCs users and non-users. Binary logistic regression was used to model the factors associated with the dichotomous dependent variable, current use of LARCs. All independent variables were forced into the model to assess their independent association with current use of LARCs. The final model for this study was built after controlling for the confounding factors. The specifications of the final model was evaluated using the “goodness-of-fit test” developed by Archer and Lemeshow for logistic regression models fitted with survey data.[24] There was no statistical evidence to ascribe a lack of fit to the final model, as evidenced by the probability value (P = .986). Statistical significance was set at a probability value (P value) of not more than .05. The original survey was approved by the Institutional Review Board of the Inner-City Fund International and ORC Macro. Before the interview, all respondents were provided information about the survey and agreed to participate by submitting written informed consent. The current study was a secondary analysis; therefore, approval by an institutional review board was not required. Permission to use the data for the current study was obtained from the DHS program.
N/A