Study area and data sources
Subjects were enrolled in the ZDHS via a two-stage sampling procedure to select households. A total of 400 ZDHS sample locations were selected. The study population was limited to 5291 women aged 15 to 49 years who were ever in a marital union, administered the violence module and also tested for HIV. A trained interviewer conducted a woman’s Questionnaire to participants in either English, or the two major languages that are spoken in Zimbabwe; Shona and Ndebele. Anonymous HIV testing was performed with the informed consent of all sampled individuals. HIV serostatus was determined by testing with the enzyme-linked immunosorbent assay (ELISA) Vironostika Uniform 2 Ag/AB. All those individuals who tested positive were retested with a follow-up ELISA, the Enzygnost® HIV Integral II assay (Siemens). The samples that tested positive to the two tests were classified as HIV positive. When the first and second tests are discordant, a confirmatory test, the HIV 2.2 western blot (DiaSorin), was then used as a tie-breaker.
Gender based violence variables
The GBV variables used in this manuscript have been described elsewhere [1]. Briefly, the 2015 ZDHS measured violence committed by spouses and by other household members . Accordingly, information was obtained from ever married women on violence by spouses and others, and from never married women on violence by anyone, including boyfriends. International research on violence shows that intimate partner violence is one of the most common forms of violence against women. Spousal or partner violence was measured in more detail than violence by other perpetrators through the use of a shortened, modified Conflict Tactics Scale (CTS). Specifically, spousal violence by the husband or partner for currently married women, and the most recent husband or partner for formerly married women was measured by asking all ever-married women the following set of questions:
Did your (last) husband/partner ever:
a) Say or do something to humiliate you in front of others?
b) Threaten to hurt or harm you or someone you care about?
c) Insult you or make you feel bad about yourself?
Did your (last) husband/partner ever do any of the following things to you:
d) Push you, shake you, or throw something at you?
e) Slap you?
f) Twist your arm or pull your hair?
g) Punch you with his fist or with something that could hurt you?
h) Kick you, drag you, or beat you up?
i) Try to choke you or burn you on purpose?
j) Threaten or attack you with a knife, gun, or any other weapon?
k) Physically force you to have sexual intercourse with him even when you did not want to?
l) Physically force you to perform any other sexual acts you did not want to?
m) Force you with threats or in any other way to perform any sexual acts you did not want to?
Domestic Violence
When the answer to any of these questions was “yes,” women were asked about the frequency of the act in the 12 months preceding the survey. A “yes” answer to one or more of items (a) to (c) above constitutes evidence of emotional violence, a “yes” answer to one or more of items (d) to (j) constitutes evidence of physical violence, and a “yes” answer to items (k) to (m) constitutes evidence of sexual violence. This approach of asking about specific acts to measure different forms of violence has the advantage of not
being affected by different understandings of what constitutes a summary term like violence. By including a wide range of acts, the approach also has the advantage of giving the respondent multiple opportunities to disclose any experience of violence. In addition to these questions asked only of ever-married women, all women were asked about physical violence perpetrated by others with the question: From the time you were 15 years old, has anyone [other than your current (last) husband/partner] hit, slapped, kicked, or done anything else to hurt you physically? Respondents who answered this question in the affirmative were asked who had done this to them. A similar question asked women who had ever been pregnant about violence during pregnancy. Women were also asked about sexual violence by anyone other than the current husband/partner with the following question: At any time in your life, as a child or as an adult, has anyone ever forced you in any way to have sexual intercourse or perform any other sexual acts? Although this approach to questioning is generally considered optimal, the possibility of underreporting of violence exists in any survey.
Statistical analysis
STATA Version 15.1, Texas USA, was used to conduct statistical analysis for this cross-sectional study. The analysis began by using simple proportions to describe the characteristics of the women included in the analysis. Associations between GBV variables and HIV positivity were explored using the simple chi-square test for categorical variables and t-test for continuous variables. Odds ratios and their 95% confidence intervals were used to calculate the risk estimate for HIV positivity among women experiencing various forms of violence compared to women not abused. Statistical significance cut-off for purposes of describing the significant GBV factors associated with HIV positivity was set at p<0.05.
To describe the geographical structure of the female HIV prevalence and the prevalence of each GBV variable, and to identify potential areas where these variables had similar spatial distribution, maps of the HIV prevalence in females and each of the GBV variables reported using the geographic coordinates for each primary sample unit (PSU) where the survey was conducted were generated. First, the prevalence of HIV in females was estimated at each PSU. HIV prevalence (HIVp) at PSU location i was defined to be HIVpi = ni / Ni, where ni denotes the number of HIV-positive females and Ni denotes the total number of females at location i. The prevalence of each GBV at each PSU was calculated in similar fashion as the HIV prevalence, in which the number of female participants with “yes” responses was divided by the total number or responders at each PSU. A kernel smoothing method to generate a continuous kernel density surface to illustrate the local spatial variations of female HIV and GBV prevalence was employed [16].