Article Text

other Versions

Randomised controlled trial
Primary care-based training and support interventions are successful in increasing identification and referral of women experiencing domestic violence
  1. Angela Taft
  1. Mother and Child Health Research, La Trobe University, Melbourne, Australia
  1. Correspondence to Angela Taft
    Mother and Child Health Research, La Trobe University, 215 Franklin Street, Melbourne, VIC 3000, Australia; a.taft{at}

Statistics from

Commentary on: OpenUrlCrossRefPubMedWeb of Science

Implications for practice and research

  • A comprehensive systems approach in primary care (using supportive interdisciplinary collaboration linked to domestic violence (DV) services) could increase identification and referral of women for DV support.

  • DV training for primary care nurses should include building clinician/advocate partnerships.

  • Nursing intervention research on DV for primary care should include examination of outcomes for women referred to DV services.


Between 15–71% of women are thought to experience lifetime DV with highest rates among women living in disadvantaged communities.1 Chronic depression, anxiety and suicide, sexually transmitted infection and adverse pregnancy outcomes are among harms associated with DV. Most primary care providers are inadequately trained and supported to identify and refer women to appropriate DV help. Screening all women, or routinely asking only symptomatic women (case-finding), continue as competing policy responses with little evidence supporting either.2 3 The Identification and Referral to Improve Safety (IRIS) study evaluates a comprehensive systems approach to routine inquiry of symptomatic women located in UK general practices.


IRIS is a cluster randomised controlled trial conducted in primary care practices in Bristol and London, UK. A total of 51 general practices were computer randomised, stratified by proportion of female doctors; postgraduate training; low-income patients and numbers of registered patients. In 24 intervention practices, Feder et al provided an intensive whole practice programme: clinician training by clinician/advocate pairs; computer prompts for inquiry about abuse among symptomatic female patients; a clinical pathway to specialist advocacy referrals; and a practice champion – for 12 months. Advocate/educators also attended clinical practice meetings, offered secondary consultation and provided referral feedback.

Identification and referral outcomes were measured by data extraction from practice records (and referral agency data) after 12 months and compared with those in 24 control practices. Poisson regression analysis adjusted for cluster and stratification variables.


After 12 months, 641 disclosures of DV were documented in intervention practice records compared with 236 in control practices (Adj rate ratio of 3.1, 95% CI 2.2 to 4.3). Only 12 referrals were identified in control practice records in contrast to 223 for intervention practices (Adj rate ratio 22.1, 95% CI 11.5 to 42.4). Specialist agencies recorded 40 referrals from control practices and 238 from intervention practices (Adj rate ratio 6.4, 95% CI 4.2 to 10.0).


The threefold increase in recorded identifications and sixfold increase in validated referrals represent substantial clinician behaviour change within intervention practices. However, the question remains whether these results are clinically significant given the absence of measured benefit for women and their families.4 Furthermore, are the results achievable in other settings within and outside the UK? Feder et al argue that the intervention was at a minimum revenue neutral, therefore observed improvement in rates of identification and referral were worthwhile in the context of universally low baselines in UK primary practice settings. Findings from earlier systematic reviews2 5 suggest that advocacy interventions are beneficial for women's health and well-being, however future research is needed to measure their impact on health outcomes.

Process evaluation of IRIS should shed light on which features of this complex intervention were critical to its success.6 This question is central for scholars in middle or low-income countries wishing to evaluate advocate/clinician collaboration yet, are unable to implement the computer elements of the IRIS programme.7 With limited evidence of effective interventions, nurse scholars should continue to evaluate systems' approaches to identify DV among patients in primary care. The question of clinical benefit to women should be addressed through explicit measurement of health outcomes. IRIS offers an additional piece in the complex puzzle of effective health system responses to DV, but gaps remain.

View Abstract


  • Competing interests None.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.