A meta-analysis of the association between intimate partner violence and age disparity in sub-Saharan Africa

Date
2017-12
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT : Intimate partner violence (IPV) is the most common form of violence against women and a worldwide human rights and public health problem. IPV against women can be emotional, physical, and sexual in nature. The current body of research has identified multiple risk factors for IPV including age disparity between women and their partners. Studies that have looked at age disparity as a risk factor of IPV show conflicting results. We conducted a meta-analysis to examine the association between IPV against women experienced within 12 months before the survey and age disparity using survey data from Demographic and Health Surveys (DHS) collected in 21 sub-Saharan African countries. Two-stage sample weights were proportionally used to represent the different countries. The age disparity was divided into two groups: relationships where the age difference (age of the male partner minus age of the woman) was less than five years and those where the age difference was five or more years. Three generalized linear models were used to estimate relative risk (RR) of the association between IPV and age disparity. These models were: modified Poisson regression, logistic regression model where RR was estimated from odds ratio, and log-binomial model. Modified Poisson regression proved to be a better model after comparing the three models using DHS data. The RR from modified Poisson regression model were pooled in a meta-analysis. A random-effects model was used in the meta-analyses. We found that weighted relative risk averages (WRRAs) from the meta-analysis were less than 1, and significant at alpha = 0.05 level for less severe and severe physical violence, which suggest that having an older partner has a protective effect against physical IPV. This study also indicates high country heterogeneity. For instance, results from Burkina Faso, and Sierra Leone indicated that having an older partner is protective against emotional and less severe IPV, while the results show that having an older partner puts a woman at a higher risk of experiencing emotional and less severe IPV in Kenya. We also calculated the prevalence of IPV against women and found that there is high prevalence of IPV against women in sub-Saharan Africa. In conclusion, this study shows an association between IPV and age disparity but the association is country dependent. We speculate that heterogeneities in underlying socio-cultural and economic histories and current realities explain dependence. Therefore, IPV prevention programmes should be country dependent when considering age disparity as a risk factor for IPV.
AFRIKAANSE OPSOMMING : Intieme lewensmaat geweld (IPV in Engels) is een van die mees algemene vorme van geweld teenoor vroue en is wêreldwyd ‘n menseregte- en openbare gesondheidsprobleem. IPV teenoor vroue kan emosioneel, fisies en seksueel van aard wees. Die huidige navorsingsliggaam het risikofaktore vir IPV geïdentifiseer, insluitende ouderdomsongelykheid tussen paartjies. Studies wat ouderdomsongelykheid as risikofaktor ondersoek het, toon teenstrydige resultate. In hierdie studie is ‘n meta-analise gedoen om die verband tussen IPV teenoor vroue in die 12 maande voor die opname en die ouderdomsverskille tussen paartjies te bepaal, met behulp van die Demografiese en Gesondheidsopname (DHS in Engels) in 21 sub-Sahara lande. Twee-stadium steekproefgewigte was proporsioneel gebruik om die verskillende lande te verteenwoordig. Ouderdomsongelykheid is in twee groepe verdeel: verhoudings met ouderdomverskille (ouderdom van manlike metgesel minus ouderdom van die vrou) minder as vyf jaar en ouderdomsverskille van vyf en meer jare. Drie veralgemeende lineêre modelle is gebruik om relatiewe risiko (RR) van die verband tussen IPV en ouderdomsverskil te beraam. Hierdie modelle was: gewysigde Poisson regressie, logistieke regressie model waar RR benader is deur die odds verhouding, en log-binomiaal model. Gewysigde Poisson-regressie was ‘n beter model nadat die drie modelle met behulp van DHS data vergelyk is. Die RR van gewysigde Poisson regressiemodel is in ‘n meta-analise saamgevoeg. ‘n Ewekansigeeffekmodel is in die meta-analises gebruik. Ons het gevind dat die geweegde relatiewe risiko gemiddeldes (WRRAs in Engels) van die meta-analise minder as 1 was, en betekenisvol by ‘n alfa = 0,05 vlak vir minder ernstige en ernstige fisiese geweld, wat aandui dat ouer metgeselle ‘n beskermende effek teen fisiese IPV het. Hierdie studie dui ook op hoë heterogeniteit tussen verskillende lande. Byvoorbeeld, resultate van Burkina Faso en Sierra Leone het aangedui dat ‘n ouer lewensmaat beskermend is teen emosionele en minder ernstige IPV, terwyl die resultate toon dat ‘n ouer lewensmaat dui op ‘n hoër risiko vir emosionele en minder ernstige IPV in Kenia. Ons het ook die voorkoms van IPV teen vroue beraam en het gevind dat daar ‘n hoë voorkoms van IPV teen vroue in sub-Sahara Afrika is. Ten slotte toon hierdie studie ‘n verband tussen IPV en ouderdomsverskil, maar die sterkte van die verband is afhanklik van die land. Daar word dus gespekuleer dat die heterogeniteit onderliggend aan die sosio-kulturele en ekonomiese geskiedenisse, asook huidige realiteite, hierdie afhanklikheid verduidelik. Gevolglik word voorgestel dat IVP voorkomings programme landsgebonde moet wees wanneer ouderdomsongelykhede binne verhoudings as risikofaktor vir IPV beskou word.
Description
Thesis (MSc)--Stellenbosch University, 2017
Keywords
Intimate partner violence -- Mathematical models, Intimate partner violence -- Sub-Saharan Africa, Age disparity in sexual relationships -- Sub-Saharan Africa, Mathematical analysis, Generalized linear model, UCTD, Poisson regression
Citation