|How far does screening women for domestic (partner) violence in different health-care settings meet criteria for a screening programme? Systematic reviews of nine UK National Screening Committee criteria
|Feder G, Ramsay J, Dunne D, Rose M, Arsene C, Norman R, Kuntze S, Spencer A, Bacchus L, Hague G, Warburton A, Taket A
This review assessed whether the current evidence fulfilled selected UK National Screening Committee criteria for the implementation of screening for partner violence against women in the UK. The authors found insufficient evidence to implement the screening programme in either health services generally or in specific clinical settings. These conclusions appear to be appropriate, given the limitations of the available data.
To assess the current evidence for clinical effectiveness and cost-effectiveness of selected UK National Screening Committee criteria for the implementation of screening for partner violence in the UK.
The electronic databases MEDLINE, EMBASE, CINAHL, DARE, NHS EED, NRR, HMIC, MIDIRS, British Nursing Index, the Cochrane Central Register of Controlled Trials (CENTRAL), Social Sciences Citation Index, IBSS, PsycINFO, and ASSIA were searched for studies and reviews from inception to December 2006. Studies reported in European languages were included and search terms were reported. Studies were also sought through contact with authors and consultation with partner-violence organisations and research networks in the UK, Western Europe, North America, and Australasia.
Seven questions were addressed:
1. What was the prevalence of partner violence against women and its health consequences?
2. Were screening tools valid and reliable?
3. Was screening for partner violence acceptable to women?
4. Were interventions effective once partner violence was disclosed in a health care setting?
5. Could morbidity and mortality be reduced following screening?
6. Was a partner violence screening programme acceptable to health professionals and the public?
7. Was screening for partner violence cost-effective?
This abstract focuses on the questions that address the effects of the treatment and screening interventions (questions four and five).
Randomised controlled trials (RCTs), non-randomised parallel-group studies, interrupted time-series studies, before-and-after studies, and case-control studies (and cohort studies for question four only) that were aimed at reducing abuse or improving the health of women aged 15 years or over (or aimed at evaluating the effectiveness of screening interventions for question four only) were eligible for inclusion. Studies conducted in health care settings, or outside health care systems to which referral could be made from a health care setting, were eligible for inclusion. Eligible outcomes were: the incidence of abuse, physical and psychosocial health, and proxy measures (and changes in identification, information giving and referrals from screening for question five only). Studies published in peer-reviewed journals and books by academic publishers and funder-published research reports were eligible.
The included studies that addressed question four were conducted in the USA, Canada, Spain, Mexico and Hong Kong, in various settings. Child-mother couples were recruited in several studies and the mean age of the mothers ranged from 27.97 to 47.33 years. The ethnic background of participants varied between studies. Interventions in the included studies were advocacy, support groups, psychological interventions (individual and group), and interventions with the children of the abused women. Outcome measures also varied.
The included studies that addressed question five were mainly based in the USA, in various settings. The mean ages of participants were not stated in the majority of studies and ethnicity varied. Follow-up ranged from six months to two years.
Studies were selected independently by two reviewers and if disagreements were not resolved by discussion a third reviewer adjudicated.
Assessment of study quality
Methodological quality was assessed by one reviewer using the published United States Preventative Services Task Force (USPSTF) quality appraisal framework, which assesses specific criteria to rate the quality of study methods as good, fair, or poor and the quality of study design as greatest, moderate, or least. The USPSTF framework also considers external validity. RCTs were also quality assessed using a modified version of the Jadad criteria to give a quality score out of four.
Data were extracted by one reviewer and checked by another; if disagreements were not resolved by discussion, a third reviewer adjudicated. Where necessary, the authors were contacted to assist in resolving a disagreement.
Methods of synthesis
Studies were pooled in a narrative synthesis. The level of evidence of effectiveness was assessed using the USPSTF criteria to class the evidence as strong, sufficient, or insufficient.
Results of the review
Question four: Were the interventions effective once partner violence was disclosed in a health care setting?
Thirty-three studies were included, but only those published since a previous review (see Other Publications of Related Interest) were analysed. The authors reported that 14 studies were analysed, with at least 1,483 participants. The previously reviewed studies were described separately.
The level of evidence for the effectiveness of individual psychological interventions was sufficient (five studies) and the level of evidence for advocacy interventions was on the border between insufficient and sufficient (four studies). The effect sizes for post-traumatic stress disorder ranged from 0.10 to 1.23 (both individual psychological interventions); depression ranged from 0.16 to 1.77 (both individual psychological interventions); self-esteem ranged from 0.10 to 2.55 (both individual psychological interventions); and physical abuse ranged from 0.02 to 0.48 (both advocacy). The strength of evidence for the effectiveness of interventions with children of abused women (three studies) and for the effectiveness of support groups (one study) was insufficient.
Question five: Could morbidity and mortality be reduced following screening?
Eight studies were included (n=at least 16,272 participants); seven were before-and-after studies (n unclear) and one was a RCT (n=3,795). The quality of execution was rated poor for the before-and-after studies and fair for the RCT. Design quality was rated moderate for the before-and-after studies and greatest for the RCT. The RCT had a Jadad score of three.
No trials of screening programmes measuring morbidity and mortality were found. The proxy outcome measure of referral rates ranged from a difference of 4 to 67% between control and intervention sites. The proxy outcome measure of identification showed little change, ranging from 3 to 25% between intervention and control sites.
Sensitivity analyses were also reported.
No cost-effectiveness studies of partner-violence screening interventions were found. A Markov model of a pilot intervention suggested that it was potentially cost-effective.
There was insufficient evidence to implement a screening programme for partner violence against women in either health services generally or in specific clinical settings.
The research question was supported by inclusion criteria for participants, intervention, outcomes and study design. Language restrictions were applied, which may have increased the risk of language bias, but the authors' justification was that they wanted to gather information from countries culturally similar to the UK. Publication bias could not be ruled out as unpublished studies were not sought. Study selection and data extraction were performed by two reviewers, reducing the risk of error and bias, however validity assessment was performed by only one reviewer. Study quality was assessed using appropriate tools and it was taken into consideration in the analysis. A narrative synthesis appeared to be appropriate, given the heterogeneity of the interventions and the poor quality of many of the included studies. It was unclear why the studies found in a previous review were presented separately to those published since the previous review, for question four.
The authors' conclusions appear to be appropriate given the limitations of the available data.
Implications of the review for practice and research
Practice: The authors did not state any implications for practice.
Research: The authors stated that the following trials were needed: system-level interventions to improve the response of health services to survivors of partner violence; psychological and advocacy interventions, after disclosure, in health care settings; and explicit investigations of what works, for whom, and in what context. Qualitative studies exploring what women want from interventions after disclosure of partner violence; cohort studies measuring risk factors, resilience factors and the trajectory of partner violence throughout life; and longitudinal studies of the long-term prognosis of survivors of partner violence after identification in health care settings were also required.
NIHR Health Technology Assessment programme.
Feder G, Ramsay J, Dunne D, Rose M, Arsene C, Norman R, Kuntze S, Spencer A, Bacchus L, Hague G, Warburton A, Taket A. How far does screening women for domestic (partner) violence in different health-care settings meet criteria for a screening programme? Systematic reviews of nine UK National Screening Committee criteria. Health Technology Assessment 2009; 13(16): 1-113
Other publications of related interest
Ramsay J, Rivas C, Feder G. Interventions to reduce violence and promote the physical and psychological well-being of women who experience partner abuse: a systematic review. London: Department of Health; 2005.
Subject indexing assigned by NLM
Attitude of Health Personnel; Female; Great Britain; Health Services; Humans; Mass Screening /methods /standards; Patient Acceptance of Health Care; Spouse Abuse /diagnosis /prevention & control
Date bibliographic record published
Date abstract record published
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.