Council for the Advancement of Nursing Conference Intimate partner violence and condom use among South African adolescents . A.M. Teitelman1, J.B. Jemmott III2, L. Icard3, A. O'Leary4, G.A. Heeren2, Z. Ngwane5, S. J. Ratcliffe2, S. Bellamy2 1University of Pennsylvania, School of Nursing, Philadelphia, 2University of Pennsylvania, Perelman School of Medicine, Philadelphia, 3Temple University, Philadelphia, 4Centers for Disease Control & Prevention, Atlanta,, 5Haverford College, Haverford Council for the Advancement of Nursing Conference September 2012
Acknowledgements South African research team members: University of Fort Hare's Centre for Health Promotion Joanne Tyler, PhD Pretty Ndyebi, MSW Phelisa Mpulu, BHon Craig Carty, MSc Funding: NIMH 1 R01 MH065867 NIMH 1K01MH080649-01A1 The Penn Center for AIDS Research AI-045008 NIMH 1 R01 MH065867 and NIMH 1K01MH080649-01A1 and the Penn Center for AIDS Research AI-045008
Background Intimate partner violence (IPV) has been tied to HIV/STI sexual risks among adolescent and adult women globally In South Africa, this link is especially important because the prevalence of both HIV and intimate partner violence is high (Jewkes et al., 2002).
Background (cont’d) IPV includes physical abuse; sexual coercion, threats, and emotional abuse and each type of IPV may be related to sexual risk Little is known about gender comparisons of IPV and sexual risk among South African adolescents
Study Purpose Examine associations between IPV and condom use Multiple types of IPV were examined: physical sexual psychological threats combined measures of overall IPV Determine if associations differ by gender
Methods Sample: 702 sexually experienced adolescents in Eastern Cape, South Africa Urban township Rural settlement During the 54-month follow-up of an HIV risk-reduction intervention trial Paper and pencil self-report survey, collected by read aloud procedure
Methods- Measures Proportion of protected sex acts in prior 3 months Outcome Measure Proportion of protected sex acts in prior 3 months Background Data -Demographic and health information Age; lives with mother/father Alcohol/drug use Age of first sex, older sexual partner Transactional sex; history of forced sex
Methods- Measures (cont’d) Intimate Partner Violence Measure: Total IPV (as victim) (adapted from Wolfe, 2001) 22 item scale, alpha = .92 Comprised of 4 subscales Physical violence (4 items) Sexual violence (4 items) Threatening behavior (4 items) Psychological/Emotional Abuse (10 items)
Methods - Data Analysis: Descriptive statistics: Chi- Square tests were used to compare frequencies t-tests were used to compare means as appropriate Lowess plots were used to determine log- transformations needed to meet model assumptions Logistic regression: examined associations of condom use with IPV Total IPV 4 types of IPV and whether the associations differed by gender
Results - Descriptive Background data N= 702 (50% male; 50% female) Mean age = 16.9 years
Demographics and Partner Experiences by Gender
Results - Descriptive Intimate Partner Violence Girls more likely to experience Physical violence Psychological violence Threats Boys more likely to experience Sexual violence
Results- Physical Violence by Gender
Results- Sexual Violence by Gender
Results- Psychological Abuse by Gender
Results- Threats by Gender
Results- Descriptive (con’t) Proportion of protected sex acts in past 3 months | Percent No 38.99 Yes 61.01
Results – Multivariate analysis Threats, psychological and overall IPV were not associated with condom use For both males and females physical IPV was significantly associated with the proportion of protected sex acts (P= 0.002, P= 0.011 respectively) A gender x physical IPV interaction was significant (P = 0.001), such that females reporting more IPV were less likely to use condoms and males reporting more IPV were more likely to use condoms. Males reporting more sexual IPV were less likely to use condoms (P=0.009)
Multiple regression Odds of Having Protected Sex in Past 3 months Correlate Males Only Females Only Gender x Correlate Interaction Odds Ratio 95% CI p-value Physical abuse 1.99 1.08, 3.68 0.028 0.65 0.47, 0.91 0.011 0.002 Sexual abuse 0.72 0.56, 0.92 0.009 0.84 0.66, 1.07 0.156 0.355
Results Gender by Physical IPV Interaction For males: > physical abuse MORE condom use For females: physical abuse LESS condom use Highly significant (p<0.001) interaction of effect by gender.
Conclusion Sexual risk consequences of IPV may differ by gender among adolescents in this region of South Africa. For females, physical IPV may increase sexual risk through condom non-use For males, physical IPV may occasion greater condom use, but sexual IPV may increase sexual risk through condom non-use These findings highlight the importance of addressing multiple types of IPV in HIV prevention interventions for both males and females and tailoring by gender.