U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Responding to Intimate Partner Violence and Sexual Violence Against Women: WHO Clinical and Policy Guidelines. Geneva: World Health Organization; 2013.

Cover of Responding to Intimate Partner Violence and Sexual Violence Against Women

Responding to Intimate Partner Violence and Sexual Violence Against Women: WHO Clinical and Policy Guidelines.

Show details

Evidence and recommendations

1. Women-centred care

Women who experience intimate partner violence or sexual violence can have very different needs, depending on their circumstances and the severity of the violence and its consequences. Furthermore, women in similar circumstances may need different types of support over time. There are, however, a minimum set of actions and principles that should guide the health-care response to women suffering from violence (physical, sexual or emotional), whether by an intimate partner, relative, acquaintance or stranger, regardless of the circumstances. This minimum supportive response is outlined in the recommendation below.

1.1. From evidence to recommendation

This recommendation is based on the experience of those working with survivors of intimate partner violence and sexual violence, and builds on the recommendations of the WHO publication, Psychological first aid (WHO, 2011), with the specific elements adapted by the GDG to deal with violence against women. Psychological first aid is aimed at individuals in community crisis situations; there is only indirect evidence for “psychological first aid” (see Psychological first aid. Geneva, World Health Organization, 2011, aimed at individuals in community crisis situations (whqlibdoc.who.int/publications/2011/9789241548205_eng.pdf).

1.2. Recommendation

  1. Women who disclose any form of violence by an intimate partner (or other family member) or sexual assault by any perpetrator should be offered immediate support. Health-care providers should, as a minimum, offer first-line support when women disclose violence. This includes:
    • ensuring consultation is conducted in private
    • ensuring confidentiality, while informing women of the limits of confidentiality (e.g. when there is mandatory reporting)
    • being non-judgmental and supportive and validating what the woman is saying
    • providing practical care and support that responds to her concerns, but does not intrude
    • asking about her history of violence, listening carefully, without pressuring her to talk (care should be taken during sensitive topics when interpreters are involved)
    • helping her access information about resources, including legal and other services that she might think helpful
    • assisting her to increase safety for herself and her children, where needed
    • providing or mobilizing social support.
    If health-care providers are unable to provide first-line support, they should ensure that someone else (within their health-care setting or another that is easily accessible) is immediately available to do so.

Quality of evidence: Indirect evidence was identified

Strength of recommendation: Strong

Remarks:

  1. Any intervention must be guided by the principal to “do no harm”, ensuring the balance between benefits and harms, and prioritizing the safety of women and their children as the uppermost concern.
  2. The privacy and confidentiality of the consultation, including discussing relevant documentation in the medical record and the limits of confidentiality with women, should be a priority. Therefore, good communication skills are essential.
  3. Health-care providers should discuss options and support women in their decision-making. The relationship should be supportive and collaborative, while respecting women's autonomy. Health-care providers should work with the women, presenting options and possibilities, as well as providing information, with the aim to develop an effective plan and set realistic goals, but the woman should always be the one to make the decisions.
  4. In some settings, such as emergency care departments, as much as possible should be done during first contact, in case the woman does not return. Follow-up support, care, and the negotiation of safe and accessible means for follow-up consultation should be offered.
  5. Health-care providers need to have an understanding of the gender-based nature of violence against women, and of the human rights dimension of the problem.
  6. Women who have physical or mental disabilities are at an increased risk of intimate partner and sexual violence. Health-care providers should pay particular attention to their multiple needs. Women who are pregnant may also have special requirements (see recommendation 8).

2. Identification and care for survivors of intimate partner violence

This section covers identification of survivors of intimate partner violence and the clinical interventions that address it.

2.1. Identification of intimate partner violence

There has been much debate about the safe and effective identification in health-care settings of women experiencing intimate partner violence. Some individuals, particularly in the United States of America (USA), advocate asking all women consulting health-care providers about partner violence (“universal screening” or “routine enquiry”), while others argue the case for a more selective approach on the basis of clinical and diagnostic considerations (“clinical enquiry” or “case-finding”) (see Glossary). In general, studies have shown that screening for intimate partner violence (i.e. systematically asking all women about violence) increases the identification of women with intimate partner violence, but, crucially, it does not reduce intimate partner violence, nor has it been shown to have any notable benefit for women's health.

The evidence was assessed to identify “the effects of interventions aimed at identifying women survivors of intimate partner violence delivered at the health system level”.

2.1.1. Evidence summary

An unpublished WHO review on screening and clinical interventions for intimate partner violence (available on request), an update of a previous systematic review (Feder et al., 2009), concluded there was insufficient evidence to support the idea that screening leads to a reduction in intimate partner violence or an improvement in quality of life or health outcomes, concluding that the connection between the two was complex. The current review therefore identified “screening” programmes that also offered post-screening action, in the hope that this would be more likely to go beyond increased detection rates and health-care provider acceptance, thus leading to improved outcomes for the women. In the screening programmes considered in this review, the most common type of “action” described was a prompt in the medical record of the screening test result provided to the health-care provider before the visit, or automatic referrals to social workers or professional advocates.

Of the four additional studies (Rhodes et al., 2006; Ahmad et al., 2009; MacMillan et al., 2009; Koziol-McLain et al., 2010) not reviewed in Feder et al. (2009), most were implemented in an emergency department setting (n = 4), two were conducted in family practice ambulatory settings (Ahmad et al., 2009; MacMillan et al., 2009), and two were conducted in obstetrics/gynaecology or antenatal clinic settings, or a family practice (MacMillan et al., 2009; Humphreys et al., 2011). In total, there were 1919 women in these studies; two studies were in Canada, one in New Zealand and one in the USA. MacMillan et al. (2009) did not report findings by study site.

No studies were found to have demonstrated an important or statistically significant reduction in recurrence of intimate partner violence. The MacMillan et al. (2009) and the Koziol-McLean et al. (2010) randomized-efficacy trials had similar findings of a small effect size (odds ratio of 0.82 and 0.86, respectively, not statistically significant) in family practice, obstetrics and gynaecology clinics and emergency departments in Ontario Canada, and the emergency department of an urban hospital in New Zealand respectively.

Only one study, a randomized controlled trial, was found to have assessed multiple health outcomes – quality of life, and symptoms of depression and post-traumatic stress disorder (PTSD) (MacMillan et al., 2009). At 18 months' follow-up, there were no important differences in these health outcomes. Hence, similar to the findings of earlier reviews, there remains insufficient evidence that screening plus action leads to a reduction in recurrence of intimate partner violence, or an improvement in quality of life or health outcomes.

A more recent randomized clinical trial of intimate partner violence screening in the USA (Klevens et al., 2012) that measured recurrence of intimate partner violence and improvements in quality of life and other health outcomes, strengthened the evidence for recommendation 2, which recommends against the screening of all women in health-care settings for intimate partner violence. Overall, the quality of the available evidence for screening studies was graded low to moderate, and showed that screening women for intimate partner violence in health-care settings did not fulfil the public health criteria for implementation of a screening programme. There are no studies measuring outcomes for women comparing identification through screening versus case-finding or clinical enquiry. While there is one trial (MacMillan et al., 2009) showing no harm caused by screening, another study conducted in antenatal settings detected potential harm (Bachus et al., 2010).

2.1.2. From evidence to recommendations

One of the public health criteria for screening is the availability of an effective response. The review of screening studies summarized above considered quality of life, recurrence of intimate partner violence, and referral rates as outcomes. In addition to the evidence, issues considered by the GDG included the opportunity cost and usefulness of screening in settings with very high prevalence and limited referral options and sustainability, and potential risks and concerns for women's safety. Based on the experience of some members of the GDG, these considerations included:

  • The high burden of universal screening where there is high prevalence, particularly in settings with limited referral options and overstretched resources/providers, which translates into opportunity costs of overstretched health-care providers and a limited capacity for responding to women who may be identified through screening. In these settings, focusing on selective enquiry based on clinical considerations is more likely to benefit women.
  • Women may find repeated enquiry difficult, particularly if no action is taken. This may potentially reduce their uptake of health services.
  • While screening increased detection, it also tends to increase resistance from clinicians, and rates of screening tend to decrease rapidly. It can easily become a tick-box exercise carried out without due consideration, or undertaken in an ineffectual way.
  • Training providers to ask all women about violence when there are limited options to offer them has an important opportunity cost. It is preferable to focus on enhancing providers' ability to respond adequately to those who do disclose violence, show signs and symptoms associated with violence, or are suffering from severe forms of abuse.

Minimum Requirements for asking about VAW

  • A protocol/standard operating procedure
  • Training on how to ask, minimum response or beyond
  • Private setting
  • Confidentiality ensured
  • System for referral in place

A minority of GDG members thought that the benefits of universal screening outweighed the disadvantages. Their reasons were:

  • it increases detection, without which intervention cannot take place (even though options for this are limited)
  • screening programmes do not seem to harm individuals, and most women do not object to being asked
  • health-care providers are not necessarily familiar with the signs and symptoms of intimate partner violence and may only ask women who they think may be at risk, increasing the potential for stereotyping
  • there is also a risk that if the enquiry is selective, it could mean that health-care providers will avoid asking if they feel uncomfortable (although this happens even when universal screening is the recommended approach).

2.1.3. Recommendations

2.

“Universal screening” or “routine enquiry” (i.e. asking women in all health-care encounters) should not be implemented.

Quality of evidence: Low–moderate

Strength of recommendation: Conditional

Remarks:

  1. There is strong evidence of an association between intimate partner violence and mental health disorders among women. Women with mental health symptoms or disorders (depression, anxiety, PTSD, self-harm/suicide attempts) could be asked about intimate partner violence as part of good clinical practice, particularly as this may affect their treatment and care.
  2. Intimate partner violence may affect disclosure of HIV status or jeopardize the safety of women who disclose, as well as their ability to implement risk-reduction strategies. Asking women about intimate partner violence could therefore be considered in the context of HIV testing and counselling, although further research to evaluate this is needed.
  3. Antenatal care is an opportunity to enquire routinely about intimate partner violence, because of the dual vulnerability of pregnancy. There is some limited evidence from high-income settings to suggest that advocacy and empowerment interventions (e.g. multiple sessions of structured counselling) following identification through routine enquiry in antenatal care, may result in improved health outcomes for women, and there is also the possibility for follow-up during antenatal care. However, certain things need to be in place before this can be done (see Box 3, Minimum requirements).
Box Icon

Box 3

Minimum requirements for a health sector response to violence against women.

3.

Health-care providers should ask about exposure to intimate partner violence when assessing conditions that may be caused or complicated by intimate partner violence (see Box 1, Examples of clinical conditions associated with intimate partner violence), in order to improve diagnosis/identification and subsequent care (see recommendation 30).

Box Icon

Box 1

Examples of clinical conditions associated with intimate partner violence. Symptoms of depression, anxiety, PTSD, sleep disorders Suicidality or self-harm

Quality of evidence: Indirect evidence

Strength of recommendation: Strong

Remarks:

  1. A minimum condition for health-care providers to ask women about violence is that it is safe to do so (i.e. the partner is not present); they must be trained on the correct way to ask and on how to respond to women who disclose violence (see Box 3, Minimum requirements). This should at least include first-line support for intimate partner violence (see recommendation 1).
  2. Providers need to be aware and knowledgeable about resources available to refer women to when asking about intimate partner violence.
4.

Written information on intimate partner violence should be available in health-care settings in the form of posters, and pamphlets or leaflets made available in private areas such as women's washrooms (with appropriate warnings about taking them home if an abusive partner is there).

Quality of evidence: No relevant evidence was identified

Strength of recommendation: Conditional

2.2. Care for survivors of intimate partner violence

Evidence was reviewed for the question: “What effects do health-care provider-initiated interventions have for women survivors of intimate partner violence?”

Eight new studies published since the WHO systematic review were identified, as well as an additional publication (Kiely et al., 2010) of a new outcome from a previously reviewed study. The evidence for addressing the effectiveness of clinical interventions came from these new studies, as well as 14 prior studies that met the inclusion criteria from three previous systematic reviews (Sadowski and Casteel, 2010; Ramsay et al., 2006; and unpublished WHO 2009 update of Ramsay et al., 2006).

The evidence for the interventions in the following categories is summarized below:

  • Psychological/mental health interventions
  • advocacy/empowerment interventions (see Glossary for context-specific definitions)
  • mother–child interventions
  • other interventions (expressive writing and yogic breathing).

The 22 studies were all controlled trials; the majority were conducted in high-income countries: Australia (1), Hong Kong (3) and the USA (17), and one in a middle-income country: Peru. The settings varied and included community, health care, shelter/refuge or a hybrid of health-care and non-health-care settings. Challenges in interpreting studies of clinical interventions for intimate partner violence included lack of detail and the overlap between psychological and advocacy/empowerment interventions, in so far as the former often have components of non-psychological support and the latter may include psychological support such as counselling. In addition, in many studies, the interventions were not described in enough detail to distinguish between formal psychological interventions and psychological support.

The 20 studies of advocacy/empowerment and psychological interventions had both methodological strengths and limitations. The quality of evidence ranged from low (13) to moderate (7). All were randomized controlled trials using standardized assessment instruments and outcomes rated as “important” or “critically important”. The limitations included the lack of blinding of the randomization process or outcome assessment, high attrition rate of the sample, and, frequently, very small sample size, which rendered the study underpowered. The heterogeneity of the studies precluded data pooling of underpowered studies and meta-analysis.

Since the last systematic review, the field has been active and has expanded the body of evidence considerably. Also, the newer studies have been conducted in health-care settings, demonstrating the feasibility of studying intimate partner violence in this setting, presumably gaining increased support from the health sector in the process.

2.2.1. Psychological/mental health interventions

Evidence summary

There were five studies that evaluated this type of intervention (Kubany et al., 2004; Gilbert et al., 2006; Lieberman et al., 2006; Kiely et al., 2010; Zlotnick et al., 2011). In a few of these, the intervention was tailored for women who had experienced intimate partner violence, and contained elements of advocacy or empowerment. As it was not possible to tease the components apart, their findings contribute to the body of evidence in both the advocacy/empowerment and psychological intervention categories. These studies show that some form of individual cognitive behavioural therapy (CBT) interventions for women who have experienced intimate partner violence may reduce PTSD and depression (during pregnancy), compared with no intervention, with one study reporting better birth outcomes. However, there is no evidence to suggest these interventions have a beneficial effect on quality of life as measured in the studies. From a previous systematic review, nine controlled studies of group psychological interventions were identified, including four randomized controlled trials (Wingood, Gilbert, Laverde, Melendez), one case control (Arinero), and four parallel group studies (Cox, Limandri, Rinfert-Raynor, Kim). They included studies based in Colombia and Korea and evaluated highly heterogeneous, psychologically based interventions and outcomes. Although the majority reported some positive outcomes, the quality of the overall study design and conduct was poor.

From evidence to recommendations

For individual psychological interventions, CBT interventions are recommended for women who are no longer experiencing violence but are suffering from PTSD. The evidence for this specific population is low quality, but comes supported by a much larger body of evidence for CBT that is of moderate quality. There is insufficient evidence to recommend CBT for women who are still experiencing intimate partner violence.

There was insufficient evidence to recommend a group psychological intervention for women who have experienced intimate partner violence.

However, the GDG wished to remind health-care providers that women with diagnosed mental health disorders who suffered intimate partner violence should receive mental health care as advised in the 2009 WHO Mental Health Gap Action Programme (mhGAP) guidelines (WHO, 2010), provided by professionals with an understanding of the impact and management of violence against women (Howard et al., 2010).

There is a substantial existing body of literature on the treatment of a range of mental disorders, including depression and PTSD (Bisson et al., 2007). WHO already has guidelines on treatment of depression, psychosis and alcohol use disorders, among other conditions (WHO, 2010; Dua et al., 2011). WHO is currently working on mhGAP guidelines and clinical protocols for acute stress, bereavement and PTSD. The WHO mhGAP Guidelines Development Group has thus far approved the recommendations shown in Box 2 that are potentially relevant to victims/survivors of intimate partner violence.

Box Icon

Box 2

Abridged recommendations for depression (DEP 1–6) and other significant emotional or medically unexplained complaints (OTH 1–7).

Recommendations
5.

Women with a pre-existing diagnosed or partner violence-related mental disorder (such as depression, or alcohol use disorder) who are experiencing intimate partner violence should receive mental health care for the disorder in accordance with WHO mhGAP intervention guidelines (WHO, 2010), delivered by health-care professionals with a good understanding of violence against women.

Quality of evidence: Indirect evidence; variable (varies with intervention, see http://www.who.int/mental_health/mhgap/evidence/en)

Strength of recommendation: Strong

Remark:

  1. Use of psychotropic medications in women who are either pregnant or breastfeeding requires specialist knowledge and is best provided in consultation with a specialist where available. For details on management of mental health issues in these two groups please see the mhGAP guidelines (WHO, 2010).
6.

Cognitive behavioural therapy (CBT) or eye movement desensitization and reprocessing (EMDR) (see Glossary) interventions, delivered by health-care professionals with a good understanding of violence against women, are recommended for women who are no longer experiencing violence but are suffering from PTSD.

Quality of evidence: Low–moderate

Strength of recommendation: Strong

2.2.2. Advocacy/empowerment interventions

Evidence summary

Fifteen studies of interventions defined as advocacy or empowerment were identified (McFarlane et al., 2000, 2006; Sullivan et al., 2002; Constantino et al., 2005; Tiwari et al., 2005, 2010a, 2010b; Gillum et al., 2009; Cripe et al., 2010; Bair-Merrit et al., 2010; Kiely et al., 2010; Humphreys et al., 2011; Taft et al., 2011; Miller,). These interventions included multiple components such as linking women to services, empowering women, educating women on parenting, and safety behaviours. The implementation of the interventions was heterogeneous in terms of who delivered the services (lay to professional) and setting (i.e. home, community, telephone and health-care setting). There is evidence that intimate partner violence advocacy/empowerment interventions may reduce recurrence of intimate partner violence for some women, but there is insufficient evidence of an impact on quality of life or mental health outcomes. The strongest evidence came from three advocacy trials conducted in Hong Kong, which implemented similar empowerment-based interventions of brief duration to three (relatively small) samples of women – antenatal, community health centre based, and shelter based.1 Of these three studies, the two evaluating the intervention in health-care settings reported benefit in some health and abuse outcomes. There remains uncertainty about the intensity required for advocacy/empowerment to have an effect outside of antenatal settings.

There are two important caveats, particularly in the context of developing an evidence base for international guidelines for health-care services. First, the strongest evidence for individual advocacy or support comes from trials of women in shelters, refuges, or safe houses with no direct connection to health-care settings, although there is more recent evidence from small-sample studies in antenatal care settings. Secondly, the evidence is based on studies in high-income countries, so other considerations need to be taken on board before making an extrapolation to the majority of the world's population.

Two studies investigated harm or distress caused by the topics of discussion, or breaches in confidentiality (McFarlane et al., 2006; Tiwari et al., 2005), and found none.

From evidence to recommendations

After reviewing the evidence, the GDG thought there was some uncertainty about (i) the effectiveness of advocacy for quality of life and mental health outcomes, and (ii) extrapolation of benefit to women not residing in shelters, or women who are not pregnant. A vote was taken on recommendation 8 and the caveats were represented in the remarks.

Recommendations
7.

Women who have spent at least one night in a shelter, refuge, or safe house should be offered a structured programme of advocacy, support, and/or empowerment (see Glossary).

Quality of evidence: Low

Strength of recommendation: Conditional

Remarks

  1. The extent to which this may apply to women leaving the household in situations where shelters do not exist is not clear.
  2. This may be considered for women disclosing intimate partner violence to health-care providers, although the extent to which this may apply in circumstances outside of shelters is not clear and should be researched further.
  3. In populations where the prevalence of intimate partner violence is high, priority should be given to women experiencing the most severe abuse. (The GDG did not agree whether this should extend to severe psychological abuse.)
  4. Interventions should be delivered by trained health-care or social care providers or trained lay mentors, tailored to the woman's personal circumstances and designed to combine emotional support and empowerment with access to community resources.
8.

Pregnant women who disclose intimate partner violence should be offered brief to medium-duration empowerment counselling (up to 12 sessions) and advocacy/support, including a safety component, offered by trained service providers where health systems can support this. The extent to which this may apply to settings outside of antenatal care, or its feasibility in low- or middle-income countries is uncertain.

Quality of evidence: Low

Strength of recommendation: Conditional

Remarks

  1. Information about exposure to violence should be recorded unless the woman declines, and this should always be conducted in a discreet manner (i.e. not with labels or noticeable markings that can be stigmatizing for women, especially when health-care professionals label them as “battered”). Women may not wish to have information recorded in their clinical history files, in the fear that their partner may find out. Women's preferences need to be balanced against the need to ensure adequate forensic evidence in circumstances where women decide to pursue a legal case.
  2. A woman should be helped to develop a plan to improve her safety and that of her children, where relevant.
  3. Attention should be paid to self-care for providers, including the potential for vicarious trauma (see Glossary).

2.2.3. Mother–child interventions Evidence summary

Four studies were identified that evaluated mother–child interventions (Jouriles et al., 2001; Sullivan, 2002; Lieberman et al., 2005, 2006). Three randomized controlled trials (one involving follow-up of an earlier trial) of intensive interventions (at least 20 sessions) focusing on the mother–child dyad, found improvements in either the children's behaviour problems (Jouriles et al., 2001; Lieberman et al., 2005, 2006), their sense of competence and self-worth (Sullivan, 2002) and/or traumatic stress symptoms in children (Lieberman et al., 2005, 2006). One intervention showed reduction in some, but not other maternal post-traumatic stress symptoms (Lieberman et al., 2005, 2006). Of the two studies that considered maternal distress related to general psychiatric symptoms, one showed benefits (Sullivan et al., 2002) while the other showed no effect (Jouriles et al., 2009). A randomized controlled trial of community-provided trauma-focused CBT (TF-CBT; Cohen, 2011), with sessions provided to children and parents that focused exclusively on child outcomes, showed improvements in children's intimate partner violence-related PTSD and anxiety. This lends further support to the evidence for effectiveness of psychotherapeutic mother–child interventions, but specifically within the context of high-income countries.

From evidence to recommendations

The GDG judged the evidence for specific intensive mother–child dyad interventions was sufficiently strong to recommend this intervention, although the applicability of this to low-income settings is uncertain.

Recommendation
9.

Where children are exposed to intimate partner violence, a psychotherapeutic intervention, including sessions where they are with, and sessions where they are without their mother, should be offered, although the extent to which this would apply in low- and middle-income settings is unclear.

Quality of evidence: Moderate

Strength of recommendation: Conditional

Remark

  1. The cost of intensive interventions focusing on the mother–child dyad makes it challenging to implement them in resource-poor settings.
  2. The lack of providers trained to provide this type of interventions also poses challenges in resource-poor settings.

2.2.4. Other interventions

Studies evaluating expressive writing (Koopman et al., 2005) and yogic breathing (Franzblau et al., 2008) were reviewed. Both were community based without linkage to health-care services and were poor quality studies. The GDG did not regard the evidence as strong enough to make any recommendations.

Figure 1 summarizes the care pathway for intimate partner violence and should help guide providers in their response to women survivors of intimate partner violence

Figure 1. Care pathway for intimate partner violence (IPV = intimate partner violence).

Figure 1

Care pathway for intimate partner violence (IPV = intimate partner violence).

3. Clinical care for survivors of sexual assault

Sexual assault is a potentially traumatic experience that may have a variety of negative consequences on women's mental, physical, sexual and reproductive health, meaning they may require acute and, at times, long-term care, particularly mental health care. In certain situations such as where there is a breakdown of law and order, armed conflict and post-conflict, or displacement, sexual violence may be exacerbated. In prisons, mental health facilities and other settings where people are institutionalized, sexual violence also appears to be more prevalent.

The gathering of forensic information is not covered by these guidelines, but is a critical element of post-rape care for those women who may want to pursue legal action. Please refer to the WHO Guidelines for medico-legal care for victims of sexual violence (2003) and the WHO/UNHCR Guidance on clinical management of rape guidelines (2004) and e-learning programme (2009) for further information on this.

3.1. Interventions during the first 5 days after the assault

3.1.1. First-line support

Recommendations
10.

Offer first-line support to women survivors of sexual assault by any perpetrator (see also recommendation 1), which includes:

  • providing practical care and support, which responds to her concerns, but does not intrude on her autonomy
  • listening without pressuring her to respond or disclose information
  • offering comfort and help to alleviate or reduce her anxiety
  • offering information and helping her to connect to services and social supports.

Quality of evidence: Indirect evidence was identified2

Strength of recommendation: Strong

11.

Take a complete history, recording events to determine what interventions are appropriate, and conduct a complete physical examination (head-to-toe including genitalia).3 The history should include:

  • the time since assault and type of assault
  • risk of pregnancy
  • risk of HIV and other STIs
  • mental health status.

Quality of evidence: Indirect evidence was identified (WHO, 2002; WHO/UNHCR/UNFPA, 2009)

Strength of recommendation: Strong

3.1.2. Emergency contraception

Sexual assault may place women of reproductive age at risk of unwanted pregnancy. Although little research exists documenting the likelihood of pregnancy as a result of sexual assault, research conducted in the USA (namely, the National Women's Study) estimated a rape-induced pregnancy rate of 5% per rape of women of reproductive age (Holmes, 1996). Rape-induced pregnancy may be even more common among women who are sexually assaulted by intimate partners, with one small study finding that 20% out of 100 women sexually assaulted by intimate partners reported becoming pregnant as a result of this violence (McFarlane et al., 2005). Analysis of data from the WHO multi-country study on women's health and domestic violence against women shows that intimate partner violence is significantly associated with unwanted pregnancy and abortions (Pallitto et al., 2013).

Evidence summary

A search of the scientific literature did not identify any research studies that focused on the effects of emergency contraception used by survivors of sexual assault.

Given the absence of evidence for this particular PICOT question, and since there is no reason to believe that the effects of emergency contraception would differ in women who have been sexually assaulted compared to non-assaulted populations, four sets of evidence-based guidelines concerning emergency contraception for general populations of women were reviewed to help inform the recommendations, including those of:

Comparison of these four guidelines, in terms of recommended drug regimens, shows that all groups except for Cheng and colleagues (2008) suggest that progestogen-only emergency contraceptive pills are the first choice of a drug regimen, with combined oestrogen–progestogen pills the secondary choice.By contrast, Cheng and colleagues (2008) recommend mifepristone as the first choice, followed by progestogen-only emergency contraceptive pills, and then a combined course of oestrogen–progestogen pills. It should, however, be noted that mifepristone in the dosage needed for emergency contraception is available in only four countries and is not recommended by WHO. All four groups agree that copper-bearing intrauterine devices (IUDs) may also be used as emergency contraception if not contraindicated. All four groups also agree that, if used, emergency contraception pills should be initiated as soon as possible after the unprotected sexual intercourse (or rape), to maximize effectiveness, with Cheng and colleagues (2008) specifying that this should occur within the first 24 hours. All groups also agree that these drugs may be initiated up to 5 days after the unprotected sexual intercourse (or rape), even though their effectiveness decreases with time. Two groups (the American College of Obstetricians and Gynecologists and the FIGO Working Group on Sexual Violence/HIV) recommend that anti-emetics be used to prevent nausea when using a combined oestrogen–progestogen regimen as emergency contraception. By contrast, WHO (2004) recommends that anti-emetics should not be routinely used; instead, they recommend that this decision be based on clinical judgment and availability. Cheng and colleagues (2008) did not make a specific recommendation on anti-emetics, although they did find that the combined oestrogen–progestogen regimen was commonly associated with the side-effects of nausea and vomiting.

In addition, information from recent randomized controlled trials of uliprisal acetate was reviewed (including studies by Creinin et al., 2006; Glasier, 2010). These studies suggest that ulipristal acetate is as effective as (and possibly more effective than) levonorgestrel in pregnancy prevention when taken close to ovulation, with somewhat similar side-effects.

From evidence to recommendations

The GDG accepted that guidance on emergency contraception for the general population would apply to women who had been sexually assaulted, and made recommendations accordingly, based on the review of the guidelines presented above. The FIGO and the American College of Obstetricians and Gynecologists classified the strength of their evidence and recommendations, but used different systems to do so.4 The Cochrane review (Cheng et al, 2008) did not classify the strength of evidence. The strength of evidence in the following recommendations is based on the best assessment of the evidence provided in the guidelines reviewed.

Recommendations
12.

Offer emergency contraception to survivors of sexual assault presenting within 5 days of sexual assault, ideally as soon as possible after the exposure, to maximize effectiveness.

Quality of evidence: Moderate

Strength of recommendation: Strong

Remarks

  1. If used, emergency contraception should be initiated as soon as possible after the rape, as it is more effective if given within 3 days, although it can be given up to 5 days (120 hours).
13.

Health-care providers should offer levonorgestrel, if available. A single dose of 1.5 mg is recommended, since it is as effective as two doses of 0.75 mg given 12–24 hours apart.

  • If levonorgestrel is NOT available, the combined oestrogen–progestogen regimen may be offered, along with anti-emetics if available.
  • If oral emergency contraception is not available and it is feasible, copper-bearing intrauterine devices (IUDs) may be offered to women seeking on-going pregnancy prevention. Taking into account the risk of STIs, the IUD may be inserted up to 5 days after sexual assault for those who are medically eligible (see WHO medical eligibility criteria, 2010).

Quality of evidence: Moderate

Strength of recommendation: Strong

Remarks

  1. The GDG discussed some of the contraindications and side-effects of the drugs. Emergency contraceptive pills on the market are extremely safe and well tolerated and meet the criteria for over-the-counter provision.
  2. Ulipristal acetate is a relatively new drug that appears to be as effective as, or more effective than, levonorgestrol. While the side-effect profile seems similar to that of levonorgestrol, it is not yet included in the WHO essential medicines list (WHO, 2011), although further evidence may change this. Levonorgestrel remains cheaper and is relatively widely available.
  3. The higher risk of STIs following rape should be considered if using a copper-bearing IUD. IUDs are an effective method of emergency contraception and should be made available to women seeking emergency contraception.
  4. A pregnancy test is not required, but if one was done and the result was positive, emergency contraception would not be necessary or effective.
14.

If a woman presents after the time required for emergency contraception (5 days), emergency contraception fails, or the woman is pregnant as a result of rape, she should be offered safe abortion, in accordance with national law.

Quality of evidence: No relevant evidence was identified

Strength of recommendation: Strong

Remarks

  1. Where abortion is not permitted, other options such as adoption should be explored with the survivor.

3.1.3. HIV post-exposure prophylaxis: treatment and adherence

Sexual assault may be associated with the transmission of HIV. While the rate of sexual transmission of HIV is low (Boily et al., 2009), it is difficult to establish risk and there are several characteristics of sexual assaults (potential for tears, multiple perpetrators) that can affect this risk. Therefore, particularly in high-prevalence settings, there are strong ethical arguments to support the provision of post-exposure prophylaxis (PEP) for HIV infection.

Evidence summary: treatment

A search of the scientific literature did not identify any studies that examined the effects of HIV PEP for survivors of sexual assault that met all of the criteria specified in the PICOT question; however, four studies (Wiebe et al, 2000; Drezett, 2002; Garcia et al., 2005; Roland et al., 2012) were identified that focused on the effects of HIV PEP among survivors of sexual assault, even though they did not meet all the PICOT criteria. These studies were reviewed to shed some light on this important topic.

Two of the four studies used prospective double cohort follow-up study designs to compare HIV seroconversion among survivors of sexual assault who were prescribed HIV PEP, with survivors of sexual assault who were not prescribed HIV PEP (Drezett, 2002; Garcia et al., 2005). The other two studies (Wiebe et al., 2000; Roland et al., 2012) used prospective cohort follow-up study designs to examine seroconversion among survivors of sexual assault who all were prescribed HIV PEP (these studies did not include a comparison group).

Each of these studies had important methodological limitations, such as lack of a comparison group, small sample sizes, and low follow-up rates. In addition, three of the studies (Wiebe et al., 2000; Garcia et al., 2005; Roland et al., 2012) included men, with no subgroup analysis that reported exclusively on the findings from women; an analysis of interest for this WHO review focused on female survivors of sexual assault. In addition, the research was conducted in only three countries: Brazil, Canada and South Africa, which limits the generalizability of the findings.

Results showed that only one of the two prospective double cohort follow-up studies found that, when compared to no HIV PEP (Drezett, 2002), HIV PEP reduced the probability of HIV seroconversion. In the two prospective cohort follow-up studies, seroconversion rates ranged from 0% to 3.7% (Wiebe et al., 2000; Roland et al., 2012).

From evidence to recommendations: treatment

Since research on HIV PEP for survivors of sexual assault is limited, and as it is extremely unlikely that clinical trial studies will be undertaken on this issue (because of ethical and logistical reasons), recommendations need to be established by extrapolating the findings from other research, including animal studies and research on people other than survivors of sexual assault. For example, there is evidence from a case-control study that a short course of antiretroviral therapy effectively reduces HIV transmission following needle-stick exposure (Cardo et al, 1997). Thus, the recommendations developed for these guidelines considered other relevant guidelines concerning this topic (CDC 2010; Jina R et al., 2010; WHO 2008).

The GDG discussed the generalizability of the evidence in all situations. Many women do not complete the 28 days of HIV PEP treatment required for it to be effective. This may be due to the side-effects of taking some of the drugs, and also related to the emotional consequences of the sexual assault. In addition, there are resource and logistic implications of providing HIV PEP. Taking this into consideration, the GDG questioned whether HIV PEP should be used routinely in locations where the prevalence of HIV is predicted to be low. There was a suggestion that health systems may wish to set a prevalence cut-off point, below which HIV PEP is not routinely offered. Additionally, in some circumstances, the risk of the individual perpetrator infecting the woman is low. It was therefore agreed that, particularly in low-prevalence settings, the risk should be considered in consultation with the woman before HIV PEP is offered.

Evidence summary: adherence

A search of the scientific literature did not find any studies that examined the effectiveness of interventions aimed at enhancing sexual assault survivors' adherence to HIV PEP, and that met all of the criteria specified in the PICOT question. However, one study was identified (Abrahams et al., 2010) that did examine this topic, even though it did not meet all the PICOT criteria (i.e. the study population included female children who were sexually assaulted, as well as adult women who were sexually assaulted, with no subgroup analyses conducted exclusively on the adult survivors). This study was reviewed to shed some light on this important topic.

Abrahams and colleagues (2010) examined whether psychosocial support via telephone would enhance adherence to HIV PEP. The study population included female child and adult survivors of sexual assault who were HIV negative when they presented at four sexual assault services in an urban and a rural site in South Africa. Participants were randomly allocated to receive a leaflet including an adherence diary, or the aforementioned plus psychosocial support by telephone. Adherence to HIV PEP was assessed during an interview that occurred 1–5 days after the 28-day period (the time during which the patients were supposed to take the HIV PEP).

Although this study used a strong research design to address the question, it also had several methodological limitations. One important limitation is that for nearly one third of participants the assessment of the primary outcome of interest (adherence to HIV PEP) was based on patients' reports of the amount of medication they did and did not take, a potentially unreliable measure, despite all study participants being provided with a diary to record when they took their medication. In addition, there were no subgroup analyses that reported exclusively on the findings from the adult participants, an analysis of interest for this review. Results showed that the intervention was not found to be effective. There were similar levels of extremely poor adherence to HIV PEP in both the intervention and comparison groups.

From evidence to recommendations: adherence

Given that only one study was identified on this topic, and showed a negative outcome in that the intervention did not enhance adherence to HIV PEP, there is a lack of good research evidence on this topic on which to make recommendations.

The view of the GDG was that, while adherence is an important issue to address in relation to HIV PEP, the current evidence did not show an effective approach to enhancing adherence.

Recommendations
15.

Consider offering HIV post-exposure prophylaxis (PEP) for women presenting within 72 hours of a sexual assault. Use shared decision-making (see Glossary) with the survivor to determine whether HIV PEP is appropriate (WHO, 2007).

Quality of evidence: Very low, based on indirect evidence (see WHO/ILO [International Labour Organization], 2008)

Strength of recommendation: Strong

Remarks

  1. PEP should be initiated as soon as possible after the assault, ideally within a few hours and no later than 72 hours after the exposure.
  2. In low-prevalence settings, policies on offering routine HIV PEP will need to consider the local context, resources and opportunity and other costs of offering it.
16.

Discuss HIV risk to determine use of PEP with the survivor, including:

  • HIV prevalence in the geographic area
  • limitations of PEP5
  • the HIV status and characteristics of the perpetrator if known
  • assault characteristics, including the number of perpetrators
  • side-effects of the antiretroviral drugs used in the PEP regimen
  • the likelihood of HIV transmission.

Quality of evidence: Indirect evidence was identified (WHO, 2008)

Strength of recommendation: Strong

17.

If HIV PEP is used:

  • start the regimen as soon as possible and before 72 hours
  • provide HIV testing and counselling at the initial consultation
  • ensure patient follow-up at regular intervals
  • two-drug regimens (using a fixed-dose combination) are generally preferred over three-drug regimens, prioritizing drugs with fewer side effects
  • the choice of drug and regimens for HIV PEP should follow national guidance.

Quality of evidence: Indirect evidence was identified (WHO, 2008)

Strength of recommendation: Strong

Remark

  1. The choice of PEP drugs should be based on the country's first-line antiretroviral regimen for HIV.
18.

Adherence counselling should be an important element in PEP provision.

Quality of evidence: Very low, based on indirect evidence

Strength of recommendation: Strong

Remark

  1. Many female survivors of sexual assault provided with HIV PEP do not successfully complete the preventive regimen because HIV PEP results in physical side-effects such as nausea and vomiting, may trigger painful thoughts of the rape, and may be overtaken by other issues in the lives of survivors. Health-care providers should be aware that adherence is very difficult to attain and efforts should be made to ensure that it is maintained. As yet, no effective intervention to promote adherence has been identified.
General remarks
  1. It is important to determine the circumstances of the rape and whether HIV PEP is appropriate. The Joint WHO/ILO guidelines on post-exposure prophylaxis (PEP) to prevent HIV infection (WHO, 2007, p.52) recommend the following eligibility criteria for HIV PEP post-sexual assault:
    • rape (penetration) took place less than 72 hours ago
    • HIV status of perpetrator positive or unknown
    • exposed individual not known to be HIV infected (need to offer HIV testing at time of consultation)
    • defined risk of exposure, such as:
      • receptive vaginal or anal intercourse without a condom or with a condom that broke or slipped; or
      • contact between the perpetrator's blood or ejaculation and mucous membrane or non-intact skin during the assault ; or
      • recipient of oral sex with ejaculation; or
      • the person who was sexually assaulted was drugged or otherwise unconscious at a time of the alleged assault and is uncertain about the nature of the potential exposure; or
      • the person was gang-raped.
  2. HIV testing is recommended prior to giving PEP but should not preclude PEP being offered. However, people with HIV infection, should not be given PEP and should be linked to care and provided with antiretroviral therapy.
  3. Health policy-makers should consider whether to routinely offer HIV PEP for post-rape care, based on local prevalence, ethical and resource considerations.

3.1.4. Post-exposure prophylaxis for sexually transmitted infections

Evidence summary

A search of the scientific literature did not identify any studies that examined the effects of STI PEP provided by health-care providers to women survivors of sexual assault. In light of the absence of evidence on this particular PICOT question, and since there is no reason to think that the effects of PEP for STIs would work differently in women who have been sexually assaulted compared to non-assaulted populations, health organizations/groups, including the Centers for Disease Control and Prevention (CDC) (2010) and the FIGO Working Group on Sexual Violence and HIV (Jina et al., 2010), have based their recommendations concerning this topic on research evidence from other populations, expert opinion, or reports of expert committees. Therefore, these two recent sets of evidence-based guidelines concerning PEP for STIs were reviewed to help inform the recommendations.

Comparison of the CDC and FIGO guidelines on this topic found them to be in general agreement. Both recommend that survivors of sexual assault be provided with prophylaxis/treatment for Chlamydia, gonorrhoea and Trichomonas; however, the FIGO guidelines also recommend prophylaxis/treatment for syphilis. Both the CDC and FIGO guidelines recommend that survivors of sexual assault receive vaccination for hepatitis B; however, the CDC guidelines also specify that this should be hepatitis B vaccination without hepatitis B immune globulin.

From evidence to recommendations

Similar to the earlier section on emergency contraception, evidence was extrapolated from studies gathered from the general population, on the grounds that the effectiveness of STI PEP was unlikely to be different for survivors of sexual assault. The view of the GDG was that to test first and then treat for a positive result would necessitate a time lag, and risk the woman not returning for the result or treatment. Therefore, the GDG recommended presumptive treatment for STIs without prior testing.

As the recommendations below were based on the CDC guidelines (which did not include strength of evidence or strength of recommendation), and the FIGO guidelines (which did include this information), the strength of evidence in the recommendations below is based on the best assessment of the evidence provided in these guidelines.

Recommendations
19.

Women survivors of sexual assault should be offered prophylaxis/presumptive treatment for:

  • chlamydia
  • gonorrhoea
  • trichomonas
  • syphilis, depending on the prevalence in the geographic area.

The choice of drug and regimens should follow national guidance.

Quality of evidence: Indirect evidence; low–very low

Strength of recommendation: Strong

20.

Hepatitis B vaccination without hepatitis B immune globulin should be offered as per national guidance.

  • Take blood for hepatitis B status prior to administering the first vaccine dose.
  • If immune, no further course of vaccination is required.

Quality of evidence: Indirect evidence; Very low

Strength of recommendation: Strong

Remark

  1. Presumptive treatment is preferable to testing for STIs, in order to avoid unnecessary delays. Therefore, the GDG does not recommend testing prior to treatment.

3.2. Psychological/mental health interventions

3.2.1. Interventions during the first days after the assault

Evidence summary

A search of the scientific literature identified nine studies examining the effects of mental health interventions provided by health-care providers for women survivors of sexual assault, with one of these studies meeting all of the PICOT question criteria (Echeburua et al., 1996) and eight of these studies (Rothbaum, 1997; Resick et al., 1988, 2002; Resick and Schnicke, 1992; Foa et al., 1991; Rothbaum et al., 2005; Galovski et al., 2009; Anderson et al., 2010) meeting most of the PICOT question criteria. Each of these studies was reviewed.

The nine studies evaluated 10 types of mental health therapies (assertion training, clinician assisted emotional disclosure, cognitive processing therapy, cognitive restructuring and coping skills, eye movement desensitization and reprocessing (EMDR) (see Glossary), prolonged exposure, progressive muscular relaxation, stress inoculation therapy, supportive counselling, and supportive psychotherapy and information). The first four and prolonged exposure and stress inoculation therapy represent different forms of cognitive behavioural therapy (CBT) (see Glossary). Seven studies focused on individually delivered therapies, while two studies focused on group therapies. The therapeutic interventions were delivered in sessions provided over a relatively short time (from 10 days to 12 weeks), with the total treatment time ranging from 5 to 18 hours.

The studies had both methodological strengths and limitations. In terms of strengths, six of the nine studies used a randomized controlled trial study design, one was a secondary analysis of data from a randomized controlled trial, and two were non-randomized controlled trials. Standardized assessment instruments were used in all studies, and multiple outcomes were assessed in most studies. The limitations included frequently not using a blinded assessment, having high loss to follow-up, no intent-to-treat analysis, and no control for potentially confounding variables. In addition, the studies often had multiple exclusion criteria, including psychological/psychiatric co-morbidity, substance abuse/dependence, and/or having experienced, or currently experiencing various forms of intimate partner violence and/or incest. Since many survivors of sexual assault have these types of problems, the generalizability of the findings of these studies to the greater population of survivors of sexual assault may be open to question. Moreover, most studies focused on survivors of sexual assault whose most recent assault had occurred at least 3 months before the study, with many of the study participants having been assaulted several years prior to the study. Although this inclusion criterion may be justified in that the investigators were trying to include only survivors of sexual assault whose levels of rape-related symptoms had not “naturally” decreased with time (or fulfilled the Diagnostic and statistical manual of mental disorders [DSM-IV; American Psychiatric Association, 1994] diagnostic criteria for PTSD, which include for the exposure to have occurred at least 1–3 months previously), this inclusion criterion may limit the generalizability of the findings, especially for women who seek care soon after the assault. In addition, most of the studies had extremely small sample sizes. Finally, it should be noted that eight of the nine studies were conducted in the USA, and eight of the nine studies focused on clinic samples.

Taken together, the study findings appear to suggest that relatively brief mental health interventions, in particular several forms of CBT as well as EMDR, may improve the psychological health of many adult female survivors of sexual assault. Moreover, although these mental health interventions appear to be more helpful than receiving no treatment, the research does not unequivocally demonstrate that one particular type of therapy is clearly superior to all others.

Finally, given the few studies on this specific population and the variety of mental health interventions assessed (with some interventions being assessed in only one study), as well as the aforementioned methodological concerns, caution is urged in making recommendations on these limited study findings. Recommendations therefore also considered the wider body of research evidence concerning mental health interventions for all trauma victims, not just victims of sexual assault (Bisson et al., 2007).

From evidence to recommendations

The strength of recommendations for the effects of mental health interventions for the care of survivors of sexual assault is limited by the relatively few studies on this topic and the methodological limitations of this research. However, this research does provide a modicum of evidence that offering survivors of sexual assault particular types of mental health interventions, such as CBT (in particular, cognitive processing therapy, prolonged exposure, and stress inoculation therapy), as well as EMDR leads to improved psychological health, including improvement of PTSD. This is supported by more general evidence on the effectiveness of these approaches for trauma survivors.

The GDG discussed the availability of resources. Complex therapies, delivered by specialists, are likely to be unavailable or would involve a long waiting time in many countries, although the costs of such services must be considered against the social and human costs of sexual violence. It is noted that there is a move to simplify and test these therapies in general health settings (Rahman et al.'s 2008 study on CBT by “lady health workers” for maternal depression). This approach needs to be tried and evaluated beyond research settings, which tend to involve higher levels of supervision and oversight to prove efficacy.

In addition to the evidence summarized above, the GDG reviewed the WHO (2011) publication Psychological first aid, which provides guidance for crisis situations, the WHO (2010) mhGAP intervention guide for treating mental, neurological and substance use disorders for non-specialist health settings, and the WHO (2009) guidelines for pharmacological treatment of mental disorder in primary care, which aim to enhance the provision of mental health services at primary care level. Psychological first aid is a very basic form of psychological support (see recommendations 1 and 10), suitable for primary care level, with limited or no referral possibilities.

Recommendations
21.

Continue to offer support and care described in recommendation 10.

Quality of evidence: Indirect evidence was identified (WHO, 2011, Psychological first aid)

Strength of recommendation: Strong

22.

Provide written information on coping strategies for dealing with severe stress (with appropriate warnings about taking printed material home if an abusive partner is there).

Quality of evidence: No relevant evidence was identified

Strength of recommendation: Strong

23.

Psychological debriefing should not be used.

Quality of evidence: Very low–low (WHO, 2011, Psychological first aid)

Strength of recommendation: Strong

3.2.2. Interventions up to 3 months post-trauma

24.

Continue to offer support and care described in recommendation 10.

Quality of evidence: Indirect evidence was identified (WHO, 2011, Psychological first aid)

Strength of recommendation: Strong

25.

Unless the person is depressed, has alcohol or drug use problems, psychotic symptoms, is suicidal or self-harming or has difficulties functioning in day-to-day tasks, apply “watchful waiting” for 1–3 months after the event. Watchful waiting involves explaining to the woman that she is likely to improve over time and offering the option to come back for further support by making regular follow-up appointments.

Quality of evidence: Very low–low (WHO, 2010)

Strength of recommendation: Strong

26.

If the person is incapacitated by the post-rape symptoms (i.e. she cannot function on a day-to-day basis), arrange for cognitive behaviour therapy (CBT)or eye movement and desensitization and reprocessing (EMDR), by a health-care provider with a good understanding of sexual violence.

Quality of evidence: Low–moderate

Strength of recommendation: Strong

27.

If the person has any other mental health problems (symptoms of depression, alcohol or drug use problems, suicide or self-harm) provide care in accordance with the WHO mhGAP intervention guide (WHO, 2010).

Quality of evidence: Indirect evidence, variable (varies with the intervention, see http://www.who.int/mental_health/mhgap/evidence/en/)

Strength of recommendation: Strong

3.2.3. Interventions from 3 months post-trauma

28.

Assess for mental health problems (symptoms of acute stress/PTSD, depression, alcohol and drug use problems, suicidality or self-harm) and treat depression, alcohol use disorder and other mental health disorders using the mhGAP intervention guide (WHO, 2010), which covers WHO evidence-based clinical protocols for mental health problems.

Quality of evidence: Indirect evidence; variable (varies with intervention, see http://www.who.int/mental_health/mhgap/evidence/en)

Strength of recommendation: Strong

29.

If the person has been assessed as experiencing post-traumatic stress disorder (PTSD), arrange for PTSD treatment with cognitive behaviour therapy or eye movement and desensitization reprocessing.

Quality of evidence: Low–moderate

Strength of recommendation: Strong

3.2.4. General remarks

  1. Consider the potential harms of psychotherapy (including CBT) when not administered properly to potentially vulnerable survivors. Informed consent and attention to safety is essential. A trained health-care provider with a good understanding of sexual violence should implement therapy.
  2. Pre-existing mental health conditions should be considered when making an assessment and planning care and, where necessary, treatment or referral provided as per the WHO mhGAP intervention guide (WHO, 2010). Women with mental health and substance abuse problems may be at greater risk of rape than other women, so there is likely to be a disproportionate burden of pre-existing mental health and substance abuse problems among rape survivors. Similarly, pre-existing traumatic events (e.g. sexual abuse in childhood, intimate partner violence, war-related trauma, etc.) should be considered.
  3. It is important to recognize that sexual assault is sometimes perpetrated by a person the woman lives with. This can include not just a partner but other family members, such as a stepfather, in-law, friend of the family, or other.
  4. Most women should have access to group or individual lay support, ideally based on the principles of Psychological first aid (WHO, 2011).

Figure 2 shows the care pathway for a woman presenting for sexual assault and should help guide the providers' response to a survivor of sexual assault.

Figure 2. Woman presents following sexual assault.

Figure 2

Woman presents following sexual assault.

4. Training of health-care providers on intimate partner violence and sexual violence

4.1. Evidence summary

Evidence was searched for the

  • effects of training interventions for health-care providers on intimate partner violence and sexual violence that improves: (i) providers' skills/ability and/or (ii) outcomes for women
  • elements of training courses that improve the skills and ability of providers to respond appropriately to women exposed to violence and/or improve outcomes for women
  • effectiveness of training for health-care providers on intimate partner and sexual violence at pre-qualification level.

4.1.1. Training interventions for intimate partner violence

The review of the evidence for the effects of training health-care providers in intimate partner violence found that most studies showed some improvement in knowledge of providers following a training intervention. However, there is little support for interventions where health-care providers are only trained in the identification of intimate partner violence, without adequate training in care and referral (Coonrod et al., 2000). Yet many interventions currently focus only on training health-care providers in identification. Interventions involving training in multicomponent aspects of intimate partner violence (identification, clinical skills, documentation and provision of referral) that used interactive techniques, appear to improve identification rates and changes in attitude and behaviour of health-care providers. Very few studies evaluated the impact of training on outcomes for women survivors of intimate partner violence (Campbell et al., 2001; Dubowitz et al., 2011; Feder et al., 2011).

Most of the evidence comes from high-income countries, with fewer, lower quality studies of the effectiveness of training interventions in low- and middle-income countries (PRIME, 2002; Grisurapong, 2004; Bott et al., 2005). In the studies carried out in high-income countries, there was some support in favour of brief multicomponent (20-minute to 1.5-day training) interactive, multimedia-based interventions that involved discussions, simulations and role-plays to train health-care providers in all aspect of intimate partner violence, including identifying, managing and providing links to agencies in the community. However, there is no conclusive evidence on what impact these interventions have on attitudes and beliefs regarding intimate partner violence, referral to services for intimate partner violence, or patient outcomes. The latter are rarely measured in evaluations of training.

There is some evidence that training in intimate partner violence, alongside other changes in systems of care and referral pathways, may be more beneficial in improving identification, and possibly even outcomes, for women survivors of intimate partner violence than training on its own (Lo Fo Wong et al., 2006; Garg et al., 2007).

Most studies were of low to very low quality and did not report on lasting effects of the interventions.

4.1.2. Training interventions for sexual assault

Only four studies were identified (Parekh et al., 2005; McLaughlin et al., 2007; Donohoe, 2010; Milone et al., 2010) that focused on examining the effects of health-care provider training in sexual violence against women. Each study focused on outcomes of the health-care providers who were trained, but none examined whether this training translated into improved outcomes for the survivors of sexual assault. Moreover, each study had one or more important methodological limitation, such as the lack of a comparison group, no assessment made either before or after the training, lack of psychometrically sound assessment tools, and extremely small sample sizes. In addition, even though the research was done in three countries (Australia, UK and USA), no research studies on training in low-income countries were identified.

Given this limited evidence base, no firm conclusions can be drawn concerning the effects of training health-care providers in responding to sexual violence against women. However, the results from these studies do indeed provide some evidence that training health-care providers in sexual assault against women may have some positive impacts. In particular, these studies suggest that such training may lead to positive changes in health-care providers' knowledge about sexual assault, and better equip them to care for survivors. This includes their attitudes towards survivors of sexual assault, and beliefs that particular groups of patients should be asked about sexual violence, as well as behaviours (in particular, clinical practices with sexual assault patients, including improved care for survivors of sexual assault, improved evidence collection, and improved writing of emergency department notes).

4.2. From evidence to recommendations

Although evidence was not found on the impact of training from low- and middle-income countries, consistent impact on knowledge, and, to some extent, on behaviours of health-care providers, was seen in studies from high-resource settings. The GDG agreed that training of health-care providers in intimate partner violence and sexual assault needs to be added to the curriculum of basic professional education and, at a minimum, provided in the form of continuing education of those providers most likely to encounter women. Although most of the interventions assessed through well-designed studies used resources such as computers, access to video players, etc., there was no evidence to indicate whether the success of the training in high-resource settings depended on these elements or not, which is something that might be difficult to replicate in a low-resource setting. The GDG agreed that the training should be tailored to requirements and provided on-site. The minimum training for staff should involve learning how to provide first-line support in response to women exposed to intimate partner violence and/or sexual violence and when to suspect and identify situations of violence, in order to provide appropriate clinical care and diagnosis.

4.3. Recommendations

30.

Training at pre-qualification level in first-line support for women who have experienced intimate partner violence and sexual assault (see recommendation 1) should be provided to health-care providers (in particular doctors, nurses and midwives).

Quality of evidence: Very low

Strength of recommendation: Strong

Remark

  1. The health-care provider may have experience of gender-based violence, as either a victim or a perpetrator. This needs to be addressed in their training. Attention should be paid to self-care for providers, including the potential for vicarious trauma.
31.

Health-care providers offering care to women should receive in-service training on violence against women, ensuring it:

  • enables them to provide first-line support (see recommendations 1 and 10)
  • teaches them appropriate skills, including:
  • addresses:
    • basic knowledge about violence, including laws that are relevant to victims of intimate partner violence and sexual violence
    • knowledge of existing services that might offer support to survivors of intimate partner violence and sexual violence (this could be in the form of a directory of community services)
    • inappropriate attitudes among health-care providers (e.g., blaming women for the violence, expecting them to leave immediately, etc.), as well as their own experiences of partner and sexual violence.

Quality of evidence: Low–moderate

Strength of recommendation: Strong

Remark

  1. Training should be intensive and content-appropriate to the context and setting.
32.

Training for health-care providers on intimate partner violence and sexual assault should include different aspects of the response to intimate partner violence and sexual assault (e.g. identification, safety assessment and planning, communication and clinical skills, documentation and provision of referral pathways).

Quality of evidence: Low

Strength of recommendation: Strong

Remarks

  1. Intensive multidisciplinary training (e.g. involving different kinds of health-care providers and/or police and advocates) delivered by domestic violence advocates or support workers should be offered to health-care professionals where referrals to specialist domestic violence services are possible.
  2. Using interactive techniques may be helpful.
  3. Training should go beyond the providers and include system-level strategies (e.g. patient flows, reception area, incentives and support mechanisms) to enhance the quality of care and sustainability.
33.

Training for both intimate partner violence and sexual assault should be integrated in the same programme, given the overlap between the two issues and the limited resources available for training health-care providers on these issues.

Quality of evidence: No relevant evidence was identified

Strength of recommendation: Strong

4.4. General remarks

  1. Priority for training should be given to those most likely to come into contact with women survivors of intimate partner violence and/or sexual assault, for example health-care providers in antenatal care, family planning or gynaecologic services, and post-abortion care, mental health and HIV, as well as primary care providers and those in emergency services.
  2. Training should include clinical examination and care for intimate partner violence and sexual assault, as well as attention to cultural competency, gender equality and human rights.
  3. Training should take place within the health-care setting, to promote attendance.
  4. There should be reinforcement of initial training and the provision of continual support. Regular follow up and quality supervision are extremely important.
  5. A clear care pathway of management and referral, a designated and accessible (domestic) violence against women worker, and regular reminders (e.g. computer prompts) were shown in one study to be helpful in sustaining the benefit of training.

5. Health-care policy and provision

Evidence was searched for the following questions:

  • “What are the effects of health system-level programmes/services for women survivors of partner violence?”
  • “What are the effects of the components/features of health system-level interventions/programmes for women survivors of partner violence?”
  • “What are the effects of integrating a sexual assault nurse examiner (SANE) programme, or another type of sexual assault programme, into a health-care setting, on the care of sexual assault survivors?”

5.1. Evidence summary

Health system-level programmes or interventions that could deliver care for intimate partner violence and/or sexual assault survivors, effectively and efficiently, were identified, particularly keeping in mind the needs in resource-poor settings. As most of the published evidence on health-system response to intimate partner violence and sexual assault came from high-income countries, a separate search of the grey literature was carried out, to identify practices and programmes that had not been published in peer-reviewed journals.

The evidence reviewed here overlapped substantially with that for training interventions, particularly as no study identified which components of multi-intervention studies (which included training) were found to be effective.

In high-income countries, there was some evidence indicating that a system-level intervention involving staff training does increase referral to other services. Five out of 10 studies (Coyer et al., 2006; Fanslow et al., 1998, 1999; Harwell et al., 1998; Spinola et al., 1998; McCaw et al., 2001; Muñoz et al., 2001; Ramsden and Bonner, 2002; Feder et al., 2011) reported an increase in referral rates. A recent cluster randomized controlled trial in general practices in the UK presented the strongest evidence of increased referrals to, and appointments with, specialist intimate partner violence agencies, as well as increased rates of disclosure (Feder et al., 2011).

In studies with positive findings, it was difficult to identify the actual factor or component responsible for the positive outcome, since interventions usually comprised various small components that were not evaluated independently. Studies also differed in the way interventions were constituted and offered, making comparisons across interventions difficult. There is a need to understand the reasons why projects showed significant improvements or otherwise.

Many of the studies were of low quality, owing to study design (often observational), small sample sizes, or a lack of sample sizes being presented, high loss to follow-up of participants or short follow-up period, inadequate presentation of data (e.g. in graphs without the presentation of actual percentages), among other factors. Overall, there is a shortage of robust evaluations of the effectiveness of health-system interventions for intimate partner violence and sexual assualt. Despite the work that has been done, there remains insufficient evidence that specific policies, protocols or models of care are more effective than others in the delivery of care to women exposed to intimate partner violence and sexual violence.

In resource-poor countries, various models of care were described. The models for service delivery often depended on the availability of financial and human resources, and varied across settings. It was apparent that “one-stop crisis centres” were a popular approach, albeit not well evaluated, although these centres were constituted very differently in different countries. Almost all had a permanent nurse, who sometimes had other service obligations, but the involvement of other staff differed and included having a doctor, counsellor, advocate, psychiatrist or psychologist on-call or on-site. The role of NGOs also varied, from initiating the service to delivering it. Linkages with other government agencies were in some instances very formal and saw them playing a central role, but more often this occurred through the development or improvement of referral systems. The involvement and linkages with the police, social services and legal services also varied in “one-stop crisis centre” models.In many instances, the greatest challenge reported was in obtaining management and administrative support, mostly linked to financial support, especially for the long-term sustainability of the centres. There was not always a smooth replication of “good” models of service delivery, even within the same country. In addition, low- and middle-income countries face the challenge of not having sufficient skilled personnel, especially for counselling, mental health and advocacy/support services. Where staff, such as counsellors, social workers, psychologists and psychiatrists, are in short supply, there is a bigger dependence on NGOs. Table 1 summarizes the advantages and disadvantages of different sites for delivering care to survivors of intimate partner violence.

Table 1. A comparison of different models of delivering care for survivors of violence against women.

Table 1

A comparison of different models of delivering care for survivors of violence against women.

With regard to integrating a SANE or other type of sexual assault care programme into health-care settings, only four studies (Derhammer et al., 2000; Crandall et al., 2003; Kim et al., 2009; Sampsel et al., 2009) were identified that focused on this. Although each study had its strengths, they all had important methodological limitations (such as small sample sizes, low response rates, and much missing data on important variables). Moreover, three were conducted in North America (one in Canada and two in the USA), with only one from a middle-income country (South Africa). They all used a historical comparison group (comparing health-care services for survivors of sexual assault, before and after implementation of a new programme). The results from these studies were generally consistent and positive. They suggest that integrating sexual assault care programmes into health-care settings leads to greater percentages of sexual assault patients receiving potentially vital health-care services, including emergency contraception, STI prophylaxis, HIV counselling and PEP, and post-care referrals. Integration of SANEs into health-care settings does not always result in nurses becoming more involved in the sexual assault examination. Where they are involved, however, integration of SANEs into health-care settings appears to enhance the collection of forensic evidence often required for successful prosecution of perpetrators of sexual assault.

5.2. From evidence to recommendations

The GDG discussed the evidence and reached the conclusion that there is no evidence to accept or refute any one model of intimate partner violence or sexual assault service provision. There is also evidence that a model that appears to be working effectively in one setting may be rapidly adopted in another site and not necessarily be effective. This has been true of “one- stop crisis centres”. Therefore, the advantages and disadvantages of the various models have been summarized to help policymakers with decisions, taking into account local human and financial resources (see Table 1). In addition, an attempt was made to answer the following questions to help guide policy-makers when making decisions:

  1. “What are the strengths and weaknesses of different models (one-stop centres, integrated services and well-articulated referral networks) for delivering services to women who have experienced violence by an intimate partner?”

These are presented in Table 1. One should consider these while also looking at the local infrastructure, resources, capacity and financial situation. The expected case-load should also be considered when setting up the service.

2.

“Which sectors of the health-care delivery system (e.g. emergency department, primary care, antenatal care or other sexual and reproductive health services, HIV counselling and testing) are better sites for interventions for women suffering intimate partner violence or sexual violence? Do they require different approaches?”

Every site has some advantages and disadvantages. Some tend to be better equipped to deal with women's health problems as compared to others. Ideally, women experiencing partner violence should be identified at the point of contact with health services, although these settings are not always conducive to providing such services. When setting up the service, one should consider the strengths and weaknesses of each site within the facility, looking at the infrastructure, resources, capacity and financial situation. No matter what site within the facility is selected, the minimal requirements (see Box 3, p. 39) need to be in place, including training and support to staff to provide the service appropriately.

3.

“Which patient groups (e.g. pregnant women, women admitted to the emergency department, women with mental health problems, women attending prevention of mother-to-child transmission [PMTCT] or HIV testing and counselling services) may benefit most from health-sector interventions?”

Women in any patient group can be exposed to violence. Women with unexplained injuries, signs or symptoms associated with depression, PTSD or other anxiety disorders (see Box 1 and recommendation 3, page 19), or those more likely to experience abuse, such as women with mental health disorders or other disabilities, may benefit from being asked about violence and receiving attention and care for sexual violence (whether by a partner or other perpetrator), or other forms of intimate partner violence. However, some health-care delivery sites may more easily lend themselves to integrating issues of violence into their routine provision of care.

4.

“What kinds of surveillance, monitoring and quality control systems are required?”

It is important to keep accurate records, since data that are properly collected, managed and analysed can both improve the services provided to women and help raise awareness about the issues. For example, it would be useful to have improved data collection on the nature of injuries and the perpetrator–victim relationship, although it must be recognized that intimate partner violence is not a disease that is easy to identify and record. For the effective inclusion of intimate partner violence indicators into the health information system, all health-care providers need training and sensitization to be able to document such cases, while ensuring this is done in a confidential way that does not put women at risk. In high-income countries with well-functioning electronic health information systems, this is easier to implement, compared to the paper-based systems in most low- and middle-income countries.

In many programmes, the major challenges faced are in monitoring referrals across sectors and maintaining the accuracy of data. Having standardized protocols/standard operating procedures/guidelines, including regular case-reviews and, if possible, monitoring of the clients' experience, can help to improve the quality of care provided.

5.3. Recommendations

34.

Care for women experiencing intimate partner violence and sexual assault should, as much as possible, be integrated into existing health services rather than as a stand-alone service (see Box 3).

Quality of evidence: Very low

Strength of recommendation: Strong

Remark

  1. A multicomponent programme including training of health-care providers to make them aware of factors that would raise clinical suspicion and of how to provide first-line support is preferable. A clear referral pathway may also increase effectiveness. This training needs to be repeated regularly, in order to sustain the benefit (see section 2, Identification and care for survivors of intimate partner violence).
  2. Offering vertical stand-alone services may be difficult to sustain and have potential harmful effects. For instance, there might be a risk that a currently under-staffed mental health service would be further weakened if it had to provide services specifically for victims of violence, rather than ensuring that all clients (including survivors of violence) get the best possible care.
  3. Providing support to the carers and the possibilities of debriefing should also be part of the health-systems response, although this requires additional human resources.It is also important for the health services to meet regularly with other agencies such as police or social workers, to ensure that there is coordination and coherence across services and that referrals are working effectively.
35.

A country needs multiple models of care for survivors of intimate partner violence and sexual assault, for different levels of the health system (see Table 1, p. 37). However, priority should be given to providing training and service delivery at the primary level of care.

Quality of evidence: Very low

Strength of recommendation: Strong

36.

A health-care provider (nurse, doctor or equivalent) who is trained in gender-sensitive sexual assault care and examination should be available at all times of the day or night (on location or on-call) at a district/area level.

Quality of evidence: Very low

Strength of recommendation: Strong

5.4. General remarks

  1. Until there is further evidence, countries need to have multiple models to provide care, but evaluation should be promoted to identify what works best and is most cost effective in different settings.
  2. One-stop centres, where appropriate, are best located within health services, where the priority for provision of services is women's health rather than being based on legal outcomes. They appear to be best suited for areas with high population density, whereas integrated services within or across health facilities may be more cost effective in rural areas.
  3. Whatever model is used, it should aim to reduce the number of services and providers that a woman has to contact (and tell her story to), and facilitate access to services she may need, in a manner that respects her dignity and confidentiality and prioritizes her safety.
  4. Violence against women is also a violation of a woman's human rights. Policies and laws need to be revised to ensure they do not discriminate against women and that they adequately penalize acts of violence, including those that take place within the home.

6. Mandatory reporting of intimate partner violence

Evidence was searched for the question: “What are the effects on women and their children of mandatory reporting of intimate partner violence to the police?”

6.1. Evidence summary

In total, 23 studies were reviewed, although only two studies (Sachs et al., 1998; Glass et al., 2001) attempted to measure the impact of mandatory reporting quantitatively. Five studies (Tilden et al., 1994; Rodriguez et al., 1998; Gerbert et al., 1999; Feldhaus et al., 2003; Smith et al., 2008) aimed to ascertain views of health-care providers, while 16 attempted to obtain the perspective of women.

Two studies quantitatively assessed the impact of introducing laws on mandatory reporting for intimate partner violence. Of these, a study (Sachs et al., 1998) that assessed the impact of mandatory reporting by health-care professionals on police dispatches to medical facilities in response to intimate partner violence, reported no significant effect of the mandatory reporting on the number of police dispatches to the facilities. In another study in the USA, the medical records of 36 acutely abused patients were reviewed. Notification to the police was the most consistently documented intervention for intimate partner violence in the medical records. However, only one in four cases was referred to domestic violence community resources such as shelters and hotlines.

The remaining studies attempted to obtain the views of health-care providers and women on the impact of mandatory reporting laws, as well as on barriers and facilitators to mandatory reporting. From the perspective of health-care providers, the advantages of mandatory reporting include improved collection of statistics, prosecution of the perpetrator and improved physician responsiveness.

Concerns shared by health-care providers included the time and resource requirements, the possibility that women may be discouraged from disclosing information, confidentiality and autonomy being compromised, the risk of retaliation, and the consequences of unsuccessful prosecutions.

From the perspective of the women, the advantages include: enabling them to get help while taking away the responsibility to report it themselves, making them feel less alone and less to blame, teaching partners the seriousness of abuse, and a potentially positive interaction with the police, with the incident being on record if needed in the future.

Women's concerns included the risk of retaliation, the fear their children would be taken away, anxiety about interacting with a social worker or other people in authority, being victimized by the health system, and being left with bills to pay as a result of the intimate partner violence report, as well as worries over autonomy and confidentiality.

While a number of women supported mandatory reporting, there appears to be an equally large number who do not. In particular, abused women appear to be against mandatory reporting, especially if it involves the police. Women in these studies suggested that the decision about reporting should be up to the woman; and that the safety of the woman and her children should be the first priority. Furthermore, recovery should focus on healing for the victims, including through counselling. If a restraining/protection order is in place, and the partner presents at the health visit, the relevant authorities should be called.

6.2. From evidence to recommendations

The evidence does not support mandatory reporting of intimate partner violence to police because it can impinge on women's autonomy and decision-making.6 While some women recognize there may be some benefits to legal action being taken on their behalf, it does not appear to be the preference for abused women. It is important to note that there may be differences between the reporting mandated by law and professional obligations/codes of conduct for health-care providers that mandate confidentiality and “do no harm”. Health-care providers need to understand their legal obligations (if any), as well as their professional codes of practice, to ensure that women are informed fully about their choices and limitations of confidentiality where this is the case.

6.3. Recommendations

37.

Mandatory reporting of intimate partner violence to the police by the health-care provider is not recommended. However, health-care providers should offer to report the incident to the appropriate authorities (including the police) if the woman wants this, and is aware of her rights.

Quality of evidence: Very low

Strength of recommendation: Strong

38.

Child maltreatment and life-threatening incidents must be reported to the relevant authorities by the health-care provider, where there is a legal requirement to do so.

  1. It is noted, however, that there is growing consensus that countries with mandatory child reporting laws should allow children and families greater access to confidential services where they can receive support on a voluntary basis.
  2. Furthermore, the usefulness of mandatory reporting is particularly questionable in situations where there is no functioning legal or child protection system to act on a report.7

Quality of evidence: Very low

Strength of recommendation: Strong

6.4. General remark

  1. The issue of mandatory reporting is intertwined with that of child protection (which was outside of the scope of these guidelines).

Footnotes

1

The strength of the evidence is labeled as “indirect evidence” when no direct evidence was identified for this population and the recommendation was therefore based on evidence extrapolated from another appropriate population.

2
3

See WHO, 2003; WHO/UNHCR, 2004 and WHO/UNHCR/UNFPA, 2009.

4

The FIGO guidelines used the Canadian Task Force on Preventive Health Care approach and classified this evidence and recommendation as I-A, denoting evidence from at least one properly controlled randomized controlled trial, so they felt there is good evidence to recommend clinical preventive action. The American College of Obstetricians and Gynecologists used the US Preventive Services Task Force approach and classified this as Level B, a recommendation based on limited or inconsistent scientific evidence.

5

In two cohort studies of HIV PEP, seroconversion rates ranged from 0% to 3.7%.

6

This is different from reporting the potential exposure of children to abuse in the home to child welfare authorities.

7

Butchart A, Harvey A, Mian M, Furniss T. Preventing child maltreatment: a guide to taking action and generating evidence. Geneva: World Health Organization; 2006. It's on the web at: http://www​.who.int/violence​_injury_prevention​/publications/violence​/child_maltreatment/en/index.html.

Copyright © World Health Organization 2013.

All rights reserved. Publications of the World Health Organization are available on the WHO web site (www.who.int) or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: tni.ohw@sredrokoob). Requests for permission to reproduce or translate WHO publications –whether for sale or for non-commercial distribution – should be addressed to WHO Press through the WHO web site (www.who.int/about/licensing/copyright_form/en/index.html).

Bookshelf ID: NBK174251
Contents

Views

In this Page

Other titles in this collection

Recent Activity

ClearTurn OffTurn On

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...