Menu
Menu
Menu
Menu
Background: Female genital mutilation/cutting (FGM/C) has been recognized as a gross violation of human rights of girls and women. This is well established in numerous international legal instruments. It forms part of the initiation ceremony that confers womanhood in Sierra Leone. Girls and women who are subjected to this practice are considered to be ready for marriage by their parents and communities and are rewarded with celebrations, gifts, and public recognition. Following this, we examined the relationship between education and women’s FGM/C intention for their daughters in Sierra Leone. Methods: We used cross-sectional data from the women’s file of the 2013 Sierra Leone Demographic and Health Survey (SLDHS) to explore the influence of education on FGM/C intention among women in the reproductive age (15-49). A sample of 6543 women were included in the study. Our analysis involved descriptive computation of education and FGM/C intention. This was followed by a two-level multilevel analysis. Fixed effect results were reported as Odds Ratios and Adjusted Odds Ratios with their respective credible intervals (CrIs) whilst results of the random effects were presented as variance partition coefficients and median odds ratios. Results: Our findings showed that women who had no formal education were more likely to intend to circumcise their daughters [aOR = 4.3, CrI = 2.4-8.0]. Among the covariates, women aged 20-24 [aOR = 2.3, CrI = 1.5-3.4] were more likely to intend to circumcise their daughters compared to women between 45 and 49 years old. Poorest women were more likely to report intention of circumcising their daughters in the future compared with the richest [aOR = 2.1, CrI = 1.3-3.2]. We noted that, 63.3% of FGM/C intention in Sierra Leone is attributable to contextual factors. Conclusion: FGM/C intention is more common among women with no education, younger women as well as women in the lowest wealth category. We recommend segmented female-child educational and pro-poor policies that target uneducated women in Sierra Leone. The study further suggests that interventions to end FGM/C need to focus on broader contextual and social norms in Sierra Leone.
Data for this study was obtained from the 2013 Sierra Leone Demographic and Health Survey (SLDHS) [38]. The 2013 SLDHS happens to be the second version of the Demographic and Health Survey following the first one in 2008. The survey intended to gather data for monitoring the population and health circumstances and to serve as a follow-up to the first survey. It captured information on female genital mutilation, fertility preferences, maternal and child health as well as sexual activity. It was implemented by the Statistics Sierra Leone (SSL) under the auspices of the Ministry of Health and Sanitation. The National Technical Committee and National Steering Committee provided technical and policy guidance for the conduct of the survey. Inner City Fund (ICF) International also offered technical assistance via the MEASURE DHS Program [38]. Sample for the survey was designed to generate dependable estimates for essential variables for the entire country in both rural and urban locations as well as across the four regions and the fourteen districts [38]. To ensure representation for all the aforementioned demarcations, the sample was stratified and selected in two stages. Primary Sampling Units (PSUs), also known as clusters were selected based on the list of enumeration areas (EAs) created in the preceding Population and Housing Census in the first stage [38]. The EAs offered the principal frame for selecting 435 clusters (158 from urban and 277 from rural) through probability proportional to their sizes. Excluded from the sampling frame were persons residing in collective housing units including hospitals, hotels and boarding schools [38]. Thirty (30) households were systematically selected from each of the clusters in the second stage. All eligible persons in the subsample of households participated in the study. This comprised women aged 15–49 and men aged 15–59. Although, 17, 1323 eligible women (15–49 years) were identified in all, complete interviews were conducted with 16,658 and this culminated in a 97% response rate. Nearly half of these women had at least one daughter (46.7%). The three questionnaires used (household questionnaire, woman’s questionnaire, and man’s questionnaire) emerged from the models developed by the MEASURE DHS Program. However, the questionnaires were modified to fit the context of Sierra Leone [38]. Intention to circumcise daughter in the future was the dependent variable for the study. As part of the survey, the women were asked “Do you intend to have daughter(s) circumcised in future?” The corresponding responses were ‘No’ coded as 0, ‘Yes’ coded as 1, and ‘Don’t Know’, coded as 8. To arrive at findings that reflect precision in thought and offer meaningful recommendations, women who responded ‘Don’t Know’ were excluded from the study. The principal independent variable was formal education. In the DHS survey, formal education is categorised into ‘No Education’, ‘Primary Education’, ‘Secondary Education’ and ‘Higher Education’. Some socio-demographic characteristics of these women were included in the study to access their association with intention of the women to circumcise their daughters in the future. These are age, wealth, religion, region, residence, healthcare decision making, number of daughters already circumcised, frequency of reading newspaper, frequency of listening to radio and frequency of watching television (TV). Following the hierarchical structure of the datasets, all the variables were categorised into individual (education, age, wealth, religion, healthcare decision making, number of daughters, number of daughters circumcised, frequency of reading newspaper, frequency of listening to radio, frequency of watching television) and contextual (family head, region and residence) level factors. To ensure clarity, two of the variables were recoded. These are religion-recoded as ‘Christian = 0’, ‘Islam = 1’, ‘Other = 2’ and number of daughters circumcised recoded as ‘No = 0’, ‘1 = 1’, ‘2 = 2’, ‘3 or more=3.’ We carried out descriptive analysis as well as two-level multilevel analysis. At the descriptive level, we calculated women’s education and the proportion of women who intended to circumcise their daughters in the future. Chi-square test was conducted in order to determine the socio-demographic variables that significantly relate with intention to circumcise daughter in the future. Due to the clustering and hierarchical nature of the datasets, these analyses were followed by the multilevel analysis conducted with the MLwinN command version 3.05 [39]. Prior to modelling, the dataset was ordered to account for the clustering nature of the survey. A total of four models were constructed with the first one being an empty model in order to ascertain the variance in FGM/C intention at the contextual level (model 1). The second model comprised fixed effects at the individual level (model 2) whereas model three accounted for fixed effects at the contextual level (model 3). In the final model (model 4), fixed effects at the individual and contextual level were fitted. Results of the fixed effects were presented as odds ratios (ORs) and adjusted odds ratios (aOR) together with their corresponding credible intervals (95% Crl). Credible interval was based on the Bayesian statistical approach whereby probability allocation for association measurements are derived whilst 95% credible intervals (95% Crls) are used for summarisation. The 95% CrI implies the possibility of the parameter assuming a value in the specified range [40]. Variance partition coefficient (VPC), also referred to as intraclass correlation coefficient (ICC), and median odds ratio (MOR) were reported for the random effects [41, 42]. The VPC assesses the magnitude of variance in the likelihood of FGM/C intention that is explained by or attributable to contextual factors. The MOR, on the order hand, accounts for the contextual variance in terms of odds ratio and calculates the propensity of FGM/C intention that is explained by contextual factors. Multicollinearity was assessed with the variance inflation factor before the models were developed [43]. Sample weight was applied and the entire analysis was executed with Stata version 13.0. Ethical approval was granted by the Ministry of Health and Sanitation of Sierra Leone as well as the ethics committee of the DHS Program. During the survey, informed consent was sought from all participants. We applied for and were granted access to the dataset by the Measure DHS Program. The dataset is freely available through https://dhsprogram.com/data/available-datasets.cfm.
N/A
