Background: The prevalence of domestic violence remains unacceptably high with numerous consequences ranging from psychological to maternal and neonatal mortality and morbidity outcomes in pregnant women. The aim of this study was to identify factors that increased the likelihood of an event of domestic violence as reported by ever married Ghanaian women. Methods: Data from the 2008 Ghana Demographic and Health Survey (GDHS) was analysed using a multivariate logistic model and risk factors were obtained using the forward selection procedure. Results: Of the 1524 ever married women in this study, 33.6 % had ever experienced domestic violence. The risk of ever experiencing domestic violence was 35 % for women who reside in urban areas. Risk of domestic violence was 41 % higher for women whose husbands ever experienced their father beating their mother. Women whose mother ever beat their father were three times more likely to experience domestic violence as compared to women whose mother did not beat their father. The risk of ever experiencing domestic violence was 48 % less likely for women whose husbands had higher than secondary education as compared to women whose husbands never had any formal education. Women whose husbands drink alcohol were 2.5 times more likely to experience domestic violence as compared to women whose husbands do not drink alcohol. Conclusion: Place of residence, alcohol use by husband and family history of violence do increase a woman’s risk of ever experiencing domestic violence. Higher than secondary education acted as a protective buffer against domestic violence. Domestic violence against women is still persistent and greater efforts should be channelled into curtailing it by using a multi-stakeholder approach and enforcing stricter punishments to perpetrators.
This is a secondary data analysis from the household questionnaire of the 2008 GDHS. A detailed description of the GDHS study design and methods is available elsewhere [7]. Notably, this study was a nationally representative cross-sectional survey that sampled about 12,000 households using a weighted approach. Half of these households were selected for individual interviews and the domestic violence module was administered to women in two-thirds of households selected for the individual interview. Subsequently, only one person was administered the domestic violence module in each selected household. Informed consent was obtained at the beginning of the individual interview and at the beginning of the domestic violence module and additional information was given for domestic violence. Access to demographic and health survey data is managed and provided by MEASURE DHS following an online registration (http://www.dhsprogram.com). Of the households selected for individual interview, 2,563 women were eligible for the domestic violence module, 17 women were excluded because of lack of privacy, 23 women refused to be interviewed with the domestic violence module and 81 women were not interviewed for other reasons. A total of 2442 (unweighted) women agreed to be interviewed. We excluded never married women as well as participants with missing data (n = 765) on covariates included in the multivariable model such as partner’s education level, respondent’s alcohol use, husband’s alcohol use, history of mother beating father and vice versa. This resulted in a sample size of 1524 women for analysis of risk factors for intimate partner violence against ever married women after sampling weight was applied. The outcome variable, domestic violence, as defined for this study included violence perpetrated by intimate partners against women and manifested through acts of physical, sexual, and emotional violence. The following seven (7) questions were used to create the variable for physical violence: (Did) your (last) husband/partner ever i. Slapped you? ii. Twisted your arm or pulled your hair? iii. Pushed you, shook you, or threw something at you? iv. Punched you with his fist or with something that could hurt you? v. Kicked you, dragged you or did beat you up? vi. Tried to choke you or burned you on purpose? vii. Threatened or attacked you with a knife, gun, or any other weapon [7]. A “yes = 1” to any of these questions constituted physical violence. If a woman scores from 1 to 7 then physical violence was coded as “1” to represent an event of “physical violence” and if a woman scores “0” then physical violence was coded as “0” to represent an event of “no physical violence”. Furthermore, sexual violence was measured using the following set of questions for women: (Did) your (last) husband/partner ever i. physically forced you to have sexual intercourse with him even when you did not want to? ii. Forced you to perform any sexual acts you did not want to? [7]. A “yes = 1” to either questions constituted sexual violence; as such if a woman gets a score of “1” or “2”, then a code of “1” was assigned to represent an event of “sexual violence”. If a woman scores “0”, then a code of “0” was assigned to represent the event of “no sexual violence”. Subsequently, spousal violence was created as per its definition in the GDHS report by combining physical and sexual violence [7]. Emotional violence was measured in a similar way, using the following set of questions: (Did) your (last) husband ever: i. Said or did something to humiliate you in front of others? ii. Threatened to hurt or harm you or someone close to you? iii. Insulted you or made you feel bad about yourself? [7]. A “yes = 1” to any of these questions constituted emotional violence. Scoring from 1 to 3 was coded as “1” to represent the event “emotional violence”. Otherwise, a code of “0” was assigned to represent the event of “no emotional violence”. The outcome variable, domestic violence was then created as per the definition of domestic violence for this study by combining spousal violence and emotional violence. The event of “no domestic violence” was coded as “0” for participants who did not experience either spousal or emotional violence. For those who experienced only emotional violence, only spousal violence and both spousal and emotional violence, a code of “1” was assigned to represent the event of “ever experienced domestic violence”. Covariates considered as risk factors were selected on the basis of causal assumption derived from subject matter knowledge. These included age of respondent, place of residence, educational level of respondent and partner, religion, wealth index, marital status, employment status of both responded and partner and alcohol use by both respondent and partner [1, 4, 14, 17]. Distribution of categorical variables were reported as frequency counts whilst associations were tested using chi-square or fisher’s exact test. Univariate logistic regression analysis was initially performed to evaluate the ability of each covariate to predict the event “ever experienced domestic violence”. Predictors with some degree of association from the univariate analyses (p < 0.25) were entered into a preliminary multivariate logistic model [18] either as continuous variables or categorized as quartiles and those that showed some degree of association (p < 0.25) were added one by one until no remaining variable produces a significant F statistic (forward selection). The forward selection model was chosen over simultaneous model as this study was designed to select from a group of independent variables, the one variable at each stage which makes the largest contribution to R2. To ensure that the predictor variables included in the model were independent of each other, variance inflation factor was used as a measure colinearity and none of the predictor variables in final model was highly associated with each other. Data were analysed using SAS version 9.2 (SAS Institute) and all statistical tests were two tailed and a p < 0.05 was considered statistically significant.
N/A
DIMA AI Care
