Screening women for intimate partner violence in healthcare settings

Abstract

Intimate partner violence (IPV) damages individuals, their children, communities, and the wider economic and social fabric of society. Some governments and professional organisations recommend screening all women for intimate partner violence rather than asking only women with symptoms (case-finding); however, what is the evidence that screening interventions will increase identification, and referral to support agencies, or improve women's subsequent wellbeing and not cause harm? To assess the effectiveness of screening for intimate partner violence conducted within healthcare settings for identification, referral to support agencies and health outcomes for women. We searched the following databases in July 2012: CENTRAL (2012, Issue 6), MEDLINE (1948 to September Week June Week 3 2012), EMBASE (1980 to Week 28 2012), MEDLINE In-Process (3 July 2012), DARE (2012, Issue 2), CINAHL (1937 to current), PsycINFO (1806 to June Week 4 2012), Sociological Abstracts (1952 to current) and ASSIA (1987 to October 2010). In addition we searched the following trials registers: metaRegister of Controlled Trials (mRCT) (to July 2012), and International Clinical Trials Registry Platform (ICTRP), ClinicalTrials.gov, Australian New Zealand Clinical Trials Registry and the International Standard Randomised Controlled Trial Number Register to August 2010. We also searched the reference lists of articles and websites of relevant organisations. Randomised or quasi-randomised trials assessing the effectiveness of IPV screening where healthcare professionals screened women face-to-face or were informed of results of screening questionnaires, compared with usual care ( which included screening for other purposes). Two review authors independently assessed the risk of bias in the trials and undertook data extraction. For binary outcomes, we calculated a standardised estimation of the risk ratio (RR) and for continuous data, either a mean difference (MD) or standardised mean difference (SMD). All are presented with a 95% confidence interval (CI). We included 11 trials that recruited 13,027 women overall. Six of 10 studies were assessed as being at high risk of bias.When data from six comparable studies were combined (n = 3564), screening increased identification of victims/survivors (RR 2.33; 95% CI 1.40 to 3.89), particularly in antenatal settings (RR 4.26; 95% CI 1.76 to 10.31).Only three studies measured referrals to support agencies (n = 1400). There is no evidence that screening increases such referrals, as although referral numbers increased in the screened group, actual numbers were very small and crossed the line of no effect (RR 2.67; 95% CI 0.99 to 7.20).Only two studies measured women's experience of violence after screening (one at three months, the other at six, 12 and 18 months after screening) and found no significant reduction of abuse.Only one study measured adverse effects and data from this study suggested that screening may not cause harm. This same study showed a trend towards mental health benefit, but the results did not reach statistical significance.There was insufficient evidence on which to judge whether screening increases take up of specialist services, and no studies included economic evaluation. Screening is likely to increase identification rates but rates of referral to support agencies are low and as yet we know little about the proportions of false measurement (negatives or positives). Screening does not appear to cause harm, but only one study examined this outcome. As there is an absence of evidence of long-term benefit for women, there is insufficient evidence to justify universal screening in healthcare settings. Studies comparing screening versus case finding (with or without advocacy or therapeutic interventions) for women's long-term wellbeing would better inform future policies in healthcare settings.