Anda HIV Violence feb06

.: 

. . Connections between HIV and Intimate Partner Violence in Kenya Anda Samson February 23, 2006

Contents: 

Contents Background Methods Results Discussion Conclusion Recommendations

Background: 

Background Percentage of HIV positive women worldwide: 30% in 1990 40% in 1997 50% today In Sub-Saharan Africa, 60% Young women seem to be more affected Source: UNAIDS epidemic updates

Male-female HIV distribution S-SA: 

Male-female HIV distribution S-SA Source: UNAIDS Epidemic update 2005

Explanations for feminization: 

Explanations for feminization Women are physically more vulnerable to HIV Lack of empowerment Gender based violence Lack of economic power Lack of decision-making power Lack of political power

Gender-Based Violence and HIV: 

Gender-Based Violence and HIV Forced sex Limited capacity to negotiate preventive behaviours Risk behaviours linked to previous sexual abuse Violence as a result of HIV infection

Gender-Based Violence: 

Gender-Based Violence 15 – 71% of women in countries across the world experienced either sexual, physical or both types of violence by an intimate partner A Physical abuse accompanied by: Psychological abuse: almost always Sexual abuse: ⅓ – ½ of cases B,C A: WHO Multicountry Study on Women’s Health and Domestic Violence Against Women. WHO, Geneva, 2005 B: Ellsberg M. et al. Candies in hell: women’s experiences of violence in Nicaragua. So sci med 2000;51(11):1595 – 610 C: Coker AL et al. Frequency and correlates of intimate partner violence by type: physical, sexual and psychological battering. Am J Public Health 2000; 90 (4): 553 -9

Gender-Based Violence: 

Gender-Based Violence Source: Violence against women: the health sector responds. PAHO occ publication 12

Empowerment : 

Empowerment Education and employment do not equal empowerment Public vs. private empowerment Empowerment vs. gender defined duties http://ambassadors.net/archives/issue12/news.htm

Empowerment (II): 

Empowerment (II) 'People – both men and women – taking control over their lives (…) Women’s empowerment implies an expansion in their ability to make strategic life choices in a context where this was previously denied to them' 1 Being able to do something can be a sign of empowerment Choosing NOT to act can be a sign of empowerment too 1. KABEER N; Resources, Agency, Achievements: Reflections on the Measurement of Women’s Empowerment. Sida Studies no.3

Hypothesis: 

Hypothesis There is a relationship between HIV and Intimate Partner Violence. This relationship is strongly influenced by empowerment related factors.

Methods: 

Methods Kenya DHS 2000 Population based survey Final sample 1630 women

Methods (II): 

Methods (II) 4 categories of variables crossed with HIV: Intimate partner violence Risk behaviour Empowerment related factors Risk perception

Categories of variables explained: 

Categories of variables explained Emotional, less severe physical, severe physical and sexual violence Ever given or received money in exchange for sex, used condom at last sexual intercourse, non-primary partners during last 12 months Agree to wife beating: any of 5 reasons; Disagree with being able to refuse sex: any of 4 reasons; Who makes decisions in household, 5 questions; Wife justified in demanding condom when husband has STD; education; literacy. No /low /moderate – high perceived risk of contracting HIV

Methods: 

Methods 3 step analysis: (1) descriptive statistics, (2) Chi-square cross-tabulations, (3) bivariate logistic regression model Outcomes with pandlt;0.05 in Chi-square were used in logistic regression Final model: significance level of 0.05

Results: descriptive: 

Results: descriptive

Results: Chi-square/HIV: 

Results: Chi-square/HIV

Results: Chi-square/HIV: 

Results: Chi-square/HIV NB: overall HIV % = 8.7

HIV & IPV: similar patterns: 

HIV andamp; IPV: similar patterns

Results: chi-square: 

Results: chi-square Almost all groups of variables interrelated Within groups, most variables related Refusing sex not justified, any reason andgt; not significant with any other variable Only 2.9% reports transactional sex Condoms are rarely used

Results: Logistic regression model: 

Results: Logistic regression model Risk behaviours, decision about food to cook, and violence were tested and found non-significant.

Results : Errors: 

Results : Errors Sampling technique vs. analysis; design effect was not calculated. Average design effect over all variables in the KDHS was 1.78 HIV prevalence level in Kenya lower than that of surrounding countries: harder to establish relation Underreporting of violence and risk behaviour Violence measured as ever-experience of violence Only quantitative data

Discussion: 

Discussion Increased percentage of HIV in abused women, although not significant in regression model Many interactions between reported empowerment variables DHS gender based violence questionnaire significantly underestimates abuse Low report of risk behaviours: beliefs do not necessarily match behaviour

Discussion : 

Discussion Data cannot be extrapolated When repeating research, take into account other ways of measurement (WHO instrument, focus groups) There are many other possible factors influencing this relationship, which are not easy to measure

Conclusion: 

Conclusion There is interaction between HIV and IPV, but the relation is complex Many other factors play a role; to measure these adequatel,y a working definition of empowerment and related factors is needed Working with HIV services or IPV services only will not impact the epidemic enough to bring change

Recommendations: 

Recommendations Work on many fronts rather than focus on one Involve both men and women when addressing the issues of HIV and IPV Work on more standardized definitions for sexual and emotional violence to make comparison between countries easier Work as much as possible with messages tailored towards the beliefs and habits of the local community.

Thank you!: 

Thank you! PAHO - AD/GE PAHO - AIS LUMC andamp; Leiden University