Psychosocial interventions for intimate partner violence in low and middle income countries: A meta-analysis of randomised controlled trials

摘要

BackgroundIntimate partner violence (IPV) is prevalent worldwide and presents pernicious consequences for women in developing countries or humanitarian settings. We examined the efficacy of psychosocial interventions for IPV among women in low- and middle-income countries (LMICs).MethodsSeven databases were systematically searched for randomised controlled trials (RCTs) examining psychosocial interventions for IPV in LMICs. Thirteen RCTs were included in random-effects meta-analyses. Risk ratios (RR) and risk difference were calculated as pooled effect sizes. Risk of bias was assessed using an adapted version of the Cochrane tool accounting for cluster RCTs. Sensitivity analyses were conducted for risk of bias and design characteristics. Publication bias and heterogeneity were assessed.ResultsPsychosocial interventions reduced any form of IPV by 27% at shortest (relative risk (RR)?=?0.73) and 25% at longest (RR?=?0.75) follow up. Physical IPV was reduced by 22% at shortest (RR?=?0.78) and 27% at longest (RR?=?0.73) follow up. Sexual IPV was reduced by 23% at longest follow up (RR?=?0.77) but showed no significant effect at shortest follow-up. Sensitivity analyses for risk of bias led to an increase in magnitude of the effect for any form of IPV and physical IPV. The effect on sexual IPV was no longer significant. Heterogeneity was moderate to high in the majority of comparisons.ConclusionsPsychosocial interventions may reduce the impact of IPV in humanitarian or low and middle income settings. We acknowledge heterogeneity and limited availability of RCTs demonstrating minimal risk of bias as limitations.Estimates suggest that 30% of women worldwide have been subject to violence perpetrated by a partner [1]. Intimate partner violence (IPV) – including physical, sexual and emotional abuse- is a pertinent problem in low and middle income countries (LMICs) [2]. There has been increased focus on efforts to reduce IPV [3]. Sexual IPV may exacerbate the spread of HIV/AIDS and may adversely impact women by severely limiting empowerment and quality of life [4]. There are also well-established links between IPV victimisation and mental and physical health outcomes including posttraumatic stress disorder (PTSD), depression, substance abuse, chronic diseases, chronic pain and gastro-intestinal or gynaecological complications [5-8], Violence reduction efforts have also focused on post-conflict and disaster settings since humanitarian crises may exacerbate vulnerability among women in affected communities [9].The past two decades have witnessed the development of a number of interventions attempting to reduce intimate partner violence in LMICs using psychological and/or social approaches. Primary outcome-focused examples include the targeted adaptation of established approaches such as counselling [10] or educative interventions [9]. IPV has also been increasingly included as a secondary outcome measure in RCTs targeting other outcomes related to women’s health in LMICs [11] and has been included as an outcome in RCTs implementing psychological therapy in LMICs [12]. IPV has further been a target of wider community-based interventions providing a variety of resources, often in combination with HIV prevention [13]. These psychosocial interventions share common characteristics in attempting to reduce IPV or related outcomes by psychological, social or educative methods.A number of systematic reviews have investigated IPV interventions although none provide meta-analytic evidence regarding the reduction of IPV in LMICs [14-19]. Existing reviews often include quasi-experimental research and trials from high income countries. One non-meta-analytic systematic review provided preliminary findings regarding the efficacy of psychosocial intervention for IPV and primarily highlighted the necessity for better evidence on the topic [14]. We identified systematic reviews of IPV interventions in the specific populations of pregnant women and adolescents which did not focus on LMICs [15,16]. Another review focused on summarising evidence for the prevention of IPV in LMICs although included non-randomised designs [19]. A large systematic review of reviews on all forms of violence against women and girls applied the AMSTAR criteria to assess review quality although 77% of the included research- both experimental and quasi-experimental- was conducted in North America resulting in limited focus on LMICs [17]. A further publication focusing specifically on the LMIC evidence from this review concluded that there was promising evidence for interventions including community mobilisation, group training and livelihood strategies [18]. However, the inclusion in each review of non-experimental research limits the validity of findings. Since the methodology the aforementioned reviews did not allow for pooling of effect sizes, a statistical estimation of the overall impact of IPV interventions is lacking. In recent years, the availability of RCTs