Family violence: Towards a holistic approach to screening and risk assessment in family support services

Family violence: Towards a holistic approach to screening and risk assessment in family support services

Elly Robinson, Lawrie Moloney

AFRC Briefing No. 17 – September 2010

Overview

Since the 1960s, violence between intimate partners, between family members and towards children1 has been increasingly recognised as a significant problem. Seminal work on male violence towards women within families was conducted in Britain (Pizzey, 1973), Australia (Scutt, 1983) and the United States (Walker, 1984). Prior to that, Kempe, Silverman, Steele, Droeghmeuller, and Silver (1962) found convincing (and at the time shocking) evidence of the extent to which children were being physically abused by parents and carers.

While knowledge about family violence and its effects has grown considerably since this time, services still grapple with the most effective ways of identifying family violence issues with which clients present and, just as importantly, of taking appropriate actions once family violence has been accurately identified. Research such as the evaluation of the 2006 family law reforms (Kaspiew, Gray, Weston, Moloney, Hand, Qu, and the family law evaluation team, 2009) suggested that family violence is not always recognised by practitioners working in this area and that even when it is recognised, appropriate actions aimed at creating or preserving safety are not always taken.

This paper reviews the current research and literature specific to family violence screening and risk assessment. It is hoped that the paper will assist service providers and practitioners to develop and evaluate tools for use within family support services.

1 In this paper, the generic term of family violence will be used except where studies are described which use different terms.

Read the publication

Background and definitions

Background and definitions

Questions of how family violence is defined, how commonly it occurs, and how gendered are its origins and its expression, go to the heart of our understanding of and our responses to this phenomenon. Though considerable progress has been made, none of these questions are settled. Therefore legislative definitions continue to vary, as do definitions employed by the social sciences and health and welfare service providers (Australian Law Reform Commission & NSW Law Reform Commission, 2010). Differing definitions also reflect differing assumptions and differing emphases regarding the broad nature of violence, particularly family violence. An Australian discussion paper published by the Domestic Violence Resource Centre (DVRC),2 made the following pertinent observation in this regard:

Usually researchers go into the field armed with a preferred definition of domestic violence, then ask research participants for their view on, or experience of, that form of violence. They do not generally seek from participants their own understanding of violence. (MacDonald, 1998, p. 7)

The multiple examples of "preferred positions" with respect to definitions, prevalence and reasons for family violence found in the research literature, has a tendency to "muddy the waters" whenever this issue is discussed.3 The confusion in the literature in part reflects the developmental nature of the field. Some definitions reinforce particular views of practitioners whose experience with largely clinical samples impress upon them the damage that family violence can do, especially to women and children. Other definitions are more likely to reflect the work of those who study violence across populations, for whom an emphasis on gender is not usually so prominent. MacDonald's (1998) observation is an important reminder not just to researchers but also to those who work as practitioners in this difficult field, that high quality practice begins with high quality assessment of exactly what has happened, how often, for what reasons, and the extent to which the behaviour is likely to continue.

Definitional, prevalence and causative debates are likely to continue for some time. We suggest nonetheless, that there are several core propositions, which are unequivocally supported in the literature. These are:

  • Family violence is a significant problem, which is associated with a broad range of poor outcomes for children and for other family members.
  • There is general consensus that useful definitions of family violence must encompass the range of ways in which violence is expressed and the range of ways in which one individual seeks to control the life of another. Clearly violence is not just physical and just as clearly, significant fear can be engendered by attitudes and behaviours that are not necessarily obvious to the naïve or untrained observer.
  • Whilst not all violence is gendered, for a variety of reasons, the role that gender plays in the institutionalisation and maintenance of violence is one that cannot be ignored.

We proceed on the basis of these propositions for the remainder of this paper.

2 Formerly the Domestic Violence and Incest Resource Centre (Victoria).

3 Note: Practitioners need to combine their clinical skills with an appreciation of the best research available. However, it is not the intention of this paper to propose a decisive definition of family violence; to be definitive about its prevalence; or to attempt to resolve the ongoing dispute in the literature with respect to differing types of family violence and the extent to which violence is primarily a gendered phenomenon. Readers who wish to access one summary of these complex issues in an Australian context are referred to Moloney et al. (2007), which addresses issues around definitions, critiques the “not all violence is the same” debate, and cites Australian Bureau of Statistics and other figures on prevalence in Australia. More broadly on these issues, the reader is referred to the excellent special edition of the Family Court Review (2008, 46(3)), which presented a series of papers arising out of the Wingspread Conference on Family Violence, including a consensus paper by Ver Steegh and Dalton (2008); and to Johnston, Roseby, and Kuehlne (2009), who propose a series of decision-making steps that should be taken when allegations of violence or child abuse are made and contested.

Family violence: Challenges for relationship services and decision makers

Family violence: Challenges for relationship services and decision makers

Data from parents reported in the Australian Institute of Family Studies evaluation of the 2006 family law reforms (Kaspiew et al., 2009, Table 4.15) suggested that a majority of individuals who seek assistance from most family relationship services in Australia are struggling with the impact of physical violence or emotional abuse on their relationships. This particularly applies to family members making use of post-separation services. For example, large majorities of both men and women clients of family dispute resolution (mediation) and children's contact centres reported the experience of some form of violence (See Table 4.5, Kaspiew et al., 2009). In line with these findings, several studies in Australia (e.g., Brown, Frederico, Hewitt, & Sheehan, 1998; Kaspiew, 2005; Moloney, Smyth, Weston, Richardson, Qu, & Gray, 2007) have found that amongst those separated couples who make applications to a family law court, allegations of violence are made in the majority of cases.4

On this basis, and with increasing knowledge regarding family violence and its risk factors, attention has turned to the best methods for screening and assessing clients. There are currently many more questions than answers regarding effective strategies to make a judgement about the existence of and level of risk that family violence poses to any individual client. In fact, a considerable lack of clarity and agreement exists regarding the use of relevant terminology alone, similar to the debate surrounding the definition of family violence itself. Questions exist regarding who should be conducting assessments, who is being assessed and for what, and how and why family violence should be assessed. The remainder of this paper explores these questions and the current status of research to address them.

4 In the Moloney et al. study (2007), the severity of the allegations was also assessed. Most were found to be of a serious nature. The study contains a summary description of each set of allegations (pp. 148–171).

Terminology

Terminology

A significant difficulty in considering the literature on screening and assessment for family violence is the common conflation of the terms, and a consequent lack of distinction between the two different types of tools. For the purposes of this paper, we will adopt the view that screening and assessment for family violence are different processes, and use definitions adapted from Braaf and Sneddon (2007):

  • Screening is a process by which the identification of victims of family violence occurs, in order to take further action or intervention. Routine screening implies that all clients attending a service should be asked questions related to the existence of family violence.
  • Risk assessment refers to ongoing efforts to assess the degree of harm or injury likely to occur as a result of past, present or future family violence.

Where possible, a distinction has been made between screening and assessment, when discussing the existing literature and research in this paper.

Who should screen for/assess family violence?

Who should screen for/assess family violence?

There is little in the literature that specifically discusses the issues inherent in whether the best approach to screening and assessment is to separate out the tasks and have them undertaken by two separate people, or whether one person should do both. The Framework for Screening, Assessment and Referrals in Family Relationship Centres and the Family Relationships Advice Line (FRC & FRAL Framework) (Winkworth & McArthur, 2008) provides guidance around establishing a "first point of contact" worker, who identifies (screens) whether or not the client requires a more in-depth assessment, which would normally be undertaken by another practitioner, or whether the client's needs can be met through information provision and/or early referral elsewhere.

In some cases, assigning the screening and assessment processes to two separate workers may generate efficiencies; in other cases it may produce fragmentation of effort. It cannot be assumed that a second worker will be able to simply build on where the other left off. Revelations, even at the screening phase, are made by a client within the context of some level of trust in the competence and integrity of the individual conducting the screening. This dynamic will not always repeat itself with another individual. Difficulty in re-establishing trust and rapport is also likely to be greater if there is a delay between the screening and the assessment phase and/or if there is no active handover or facilitated referral between workers. It can indeed be argued that the very fact of beginning a screening process brings clear ethical and professional obligations on the part of the practitioner and the organisation to ensure that risks that are thought to be there even at this phase are acted upon and not "left" to the assessment phase.

This leads in turn to questions about the level of professional skill and knowledge needed to undertake these processes effectively. According to the FRC & FRAL Framework, significant skills are required by the professional undertaking screening to respond sensitively and respectfully to clients or callers, while supporting or "holding" clients and simultaneously not engaging in a deeper conversation about their concerns. Spangaro, Zwi, and Poulos (2009) also suggested that screening itself can be a therapeutic experience, but insensitively or incompetently handled, it can be traumatising and increase the risk that was present at the outset. So, should screening be regarded as a less specialised task than the task of assessment? Can it be effectively accomplished mainly by the use of validated self-reporting instruments administered independently? If used, should a validated self-reporting instrument be administered only in the presence of an empathically engaged worker, or is it acceptable to invite an individual to complete the task in isolation from such support? And whether or not an instrument is used, should other problems that frequently co-exist with family violence - especially substance abuse and certain forms of mental illness - also be screened in or out from the outset?

We suggest that there are considerable dangers associated with the use of a screening instrument in isolation from empathic engagement with a worker. Such a procedure would not, for example, pass an ethics application for a research project. This is because once begun, even for the purposes of research, there can be no guarantee that screening can be neatly concluded. The problem of decoupling early screening interventions from knowledge of and accessibility to "what comes next" has been addressed in the context of family law by Jaffe, Crooks, and Bala (2006). According to Jaffe and colleagues, safety concerns need to be given higher prominence at the early stage of intervention, the very stage at which "adequate information to evaluate the safety of children and adults" is more likely to be lacking (p. 47).

There are no easy solutions to these service delivery issues, which are essentially problems of triage. It can be argued that the person at the beginning of the triage process bears the greatest responsibility because a failure to detect violence or associated issues at this stage can reverberate throughout the service delivery system. At the same time, such an individual usually cannot take on the full burden of the case. Does this person need to have the assessment skills to be employed when a client passes a screening threshold? If not, should there always be a formally facilitated referral to an individual more formally trained in assessment? If so, when and how should these referrals be made? Should that person, who may be inside or outside the organisation, also have skills in the next phase of the intervention, whether it is safety planning, education, counselling, or mediation? Answers to these questions cannot be universal. They will, for example, depend on local conditions and availability of staff. But the governing principle, we suggest, should be high levels of skill at all stages, combined with the minimum number of new interventions possible.

Returning to the question of formal assessment, there is a clear emphasis in the literature on the need for the exercise of professional judgement to be part of the process (Kropp, 2008). Kropp calls for the following skills:

  • expertise and experience in interviewing and assessing offenders and victims;
  • considerable knowledge of the dynamics of domestic violence; and
  • completion of assessments with the assistance of risk assessment guidelines or tools that have some acceptance in the scientific and professional communities.

This links in turn to the need for ongoing training related to family violence and the use of screening and assessment tools - another issue highlighted in the literature. In the NSW Health pilot project (Irwin & Waugh, 2001), all staff members were expected to undertake training in the use of the screening tool. In Victoria, an extensive state-wide cross-sectoral training program in the use of the Family Violence Common Risk Assessment Framework (CRAF) has been undertaken. An evaluation, including impact on practice, has been completed, although not publicly available at the time of writing. An emphasis on skills and knowledge in family violence is also evident in accreditation standards for family dispute resolution practitioners in Australia.5 Family dispute resolution practitioners who were included in the Family Dispute Resolution Register6 before 1 July 2009 must demonstrate competency in three specific units of study, one of which is "Responding to Family and Domestic Violence in Family Work".

Ongoing assessment

There are strengths and weaknesses in the institutionalisation of screening and assessment procedures. On the positive side, these procedures formalise the process of determining the nature of the issues being presented by clients. They increase the chances that the service offered would be the most appropriate one. The CRAF, however, states that "assessment of risk based on a single tool alone will not deliver the desired outcome or guarantee victim safety. In fact, such an approach may endanger a victim because no tool currently available is 100% accurate".(p. 30). This raises the question of the possible link between assessment tools and the generation of too many false positive findings (that violence occurred when in fact it did not) or false negative findings (that violence did not occur when in fact it did). There is a risk of becoming complacent about false negatives when too much focus is placed on formal procedures.

As Gould (1981) has demonstrated with respect to the measurement of "intelligence", the results of formalised and institutionalised assessment procedures can too easily be seen as truths or "entities" that become fixed in time. An assessment of risk may be seen to have been "completed", at which point a practitioner or a service might remain less alert to the possibility that from a statistical point of view, past or present violence must always be regarded as a possibility. As such, the assessment process by practitioners must be ongoing. Although the principle of continuous assessment is also well established in the literature, putting it into clinical practice once a formal screening and assessment phase has "ended" can feel counter-intuitive, especially to the novice practitioner.

But like a pilot who seeks take-off clearance once all systems have been checked, a practitioner cannot simply assume that there are no dangers ahead simply because screening and risk assessment have been attended to. Like the pilot who continues to check systems in flight, the practitioner is required to focus on the task now at hand (for example addressing relationship tensions, attending to parenting issues, assisting in the resolution of a dispute) whilst simultaneously having an "ear open" to the possibility that these matters may be secondary to questions of more critical importance that have not been revealed (see Box 1: Case Study).

Box 1 - Case study - Attuning to violence in a Family Dispute Resolution (FDR) session

Towards the end of three sessions of FDR (mediation), a woman wanted to make adjustments to a draft agreement. Her ex-partner was not willing to make the change and a stalemate ensued. When the woman held her ground, her former partner quietly "reminded" her that this might not be in her interests. The FDR practitioner felt uncomfortable about the atmosphere that had developed in the room and called a private session. At that session the woman revealed an incident that had occurred shortly after their marriage. She recalled that during a dinner party, her then husband asked to see her in the kitchen. He had been angered by a remark she had made about him in front of the dinner guests. He put his fist through a wall and told his wife that she should be very careful in future. The woman had lived in fear of a recurrence of this incident and in fear of her own safety for many years.

This revelation significantly changes the nature of the work of a practitioner. Suddenly, the FDR practitioner is faced with many challenging decisions, and must slow the process down to allow himself/herself as well as the clients to adjust to this new situation. The revelation is likely to have placed the woman in a more vulnerable position, at least in the short term. Some questions that need to be addressed include:

  • How is her immediate safety to be ensured?
  • Can the changed circumstances, with respect to if and how FDR might continue, be communicated to the former husband in a way that ensures the safety of the wife?
  • What further input is needed from the wife and what further assistance may she need before the husband is engaged further?
  • How will the FDR practitioner deal with the possibility (perhaps likelihood) of denial or minimisation?
  • Can parts of the draft agreement about future parenting arrangements be salvaged?
  • In the event that the husband acknowledges the impact of his behaviour, how can the children's need for healing be incorporated into the parenting plan?
  • What are the wife's views on continuing with a parenting plan and who else should she speak with about this?

Another set of questions arises when other professionals and their possible roles are considered. For example:

  • Is there somebody who can advocate clearly for the wife in a way that is not punitive towards the father?
  • Who can advocate for the children in a way that prioritises their safety without necessarily "removing" them from the care their father might provide?
  • What services may be of assistance to the husband at this time?
  • What other community-based referrals should be considered?
  • Who should facilitate engagement with services such as those above, in what sequence and how?
  • Do the circumstances warrant calling the police?
  • What role can or should the legal representatives now play?
  • If there are no legal representatives for one or both the parties, should they be advised to seek such representation?
  • How active should the FDR practitioner be in promoting legal representation if it does not exist?
  • If the woman has a legal representative, should the FDR practitioner consider seeking the client's permission to call her?
  • Should the FDR practitioner ask the client if the legal representative has been told of this or similar incidents? If not, what might this say of the current relationship between the client and her lawyer. If so, how can this be reconciled with the fact that the legal representative appears to be supportive of FDR?
  • Should the FDR practitioner seek permission to also speak to the former husband's legal representative, or if he does not have a legal representative, advise him to seek one?

The above list is not exhaustive. These and other possible actions require careful contextual analysis. They call for consideration and "judgement calls" regarding not only what to do and not do, but as noted, the sequencing of actions. Decisions at such a critical moment should not be rushed, and whenever possible should be made in consultation with a colleague or supervisor. On the other hand, action may need to be taken to ensure immediate safety. The former wife for example may need to be escorted when she leaves. A safety plan may also need to be devised for both the mother and her children. The former husband may need to be linked in quickly with a men's relationship or similar service.

5 Family dispute resolution practitioners need to have met accreditation standards based on new competency-based qualifications developed specifically for the family relationships sector (Vocational Graduate Diploma of Family Dispute Resolution).

6 Established by the Attorney-General’s Department, see: Family Dispute Resolution Register <http://fdrregister.familyrelationships.gov.au/Search.aspx>.

What is used to screen for and assess family violence?

What is used to screen for and assess family violence?

Screening tools

At this stage, there is no easy answer to the most effective tool to screen for family violence, or the most appropriate length of time to devote to screening, particularly in family services. Many studies consider the use of tools in broader health care settings, such as hospital emergency departments. A systematic review of 21 intimate partner violence screening tools used in a health care setting found that even the most common screening tools, such as HITS (Hurt, Insult, Threaten, Scream) and the Woman Abuse Screening Tool, have only been evaluated in a small number of studies. Further validity and reliability testing is needed (Rabin et al., 2009). The average number of screening items in the study tools was 4.2, with four tools using only one question. The single question tools performed inconsistently in identifying intimate partner violence (IPV) victims.

The FRC & FRAL Framework promotes the use of three broad screening questions, synthesised from a range of international instruments, with a fourth question if the contact is related to setting up a joint session:

  • Do you have any reason to be concerned about your own safety or the safety of your children?
  • Do you have any other concerns about your children's wellbeing at the moment?
  • Do you have any reason to be concerned about the safety of anyone else?
  • How do you think your partner/ex-partner would answer these questions? (optional)

While such questions represent an understandable compromise in attempting to establish the presence or absence of safety across a range of domains (violence, child abuse, self-harm), what they gain in covering this broad ground, they lose in directness and are open to interpretation. On the other hand, further questions undermine the attractiveness of a brief and focused screening protocol. The FRC & FRAL Framework does suggest that the three screening questions provide entry points into deeper questioning on relevant topics, using a range of suggested tools. Rabin et al. (2009) suggested that providers should consider the balance between brevity and comprehensiveness. Similarly, Gawande (2010) discussed an inherent tension between brevity and effectiveness when developing checklists. The balance itself will be partially informed by careful consideration of what it is that the practitioner believes he or she needs to know in order to proceed.

On the question of specificity, in one of the most significant and long-running routine screening programs for domestic violence, NSW Health Area Health Services7 use direct questions related to domestic violence:

  • Within the last year have you been hit, slapped or hurt in other ways by your partner or ex-partner?
  • Are you frightened of your partner or ex-partner?

If the woman answers yes to either or both questions:

  • Are you safe to go home when you leave here?
  • Would you like some assistance with this?

Between 62-75% of eligible women were screened in the years 2003-2006, with 6-7% of these women identified as experiencing family violence. The main reason given by professionals for not undertaking screening was the presence of a partner or others (NSW Department of Health, 2007). This may in turn suggest the need for a consideration of more routine one-to-one intake procedures as a prelude to work with couples or families. Services such as family dispute resolution, for example, routinely conduct separate assessments of each individual involved in the dispute. Relationship counselling services are also increasingly adopting such practices. Even with such a protocol however, consideration needs to be given to the subtle influence of circumstances such as a family member, partner or ex-partner being in the building, or an expectation of joint sessions at a later time.

While universal screening has its benefits in terms of identifying clients who are experiencing family violence, problems with implementation include increased workloads (Irwin & Waugh, 2001), and a common lack of protocols and training regarding use of the tool (Braaf & Sneddon, 2007). Additionally, the application of such a tool does not automatically increase protection for the victim; nor is there convincing evidence about long-term improvements for clients as a result of using a screening tool (Braaf & Sneddon, 2007).

Risk assessment

An important element of family violence risk assessment highlighted in the literature is the degree to which professional judgement and/or more formal tools are used. Three approaches have been considered in the literature - unstructured (clinical) decision-making, actuarial decision-making and structured professional judgement. These approaches are compared in Table 1.

Table 1. Comparison of methods of family violence screening and assessment
  Description Pros Cons
Unstructured (clinical) decision-making
  • Assessment is based on professional discretion, intuition or “gut” feelings, justified via qualifications and/or experience.
  • No constraints or guidelines for the practitioner.
  • Focused on individual, allowing for person- and context-specific tailoring of interventions.
  • Criticised due to lack of validity, reliability and accountability - decisions regarding responses may be based more on training, preferences, specialisations and biases of the evaluator.
  • Because it relies on discretion, may miss important factors requiring intervention.
Actuarial decision-making
  • Integration of statistical evidence into assessment, using tools based on evidence-based risk factors associated with the outcomes of interest. Risk is determined via scales or matrices.
  • Aids the worker in assessing if violence has occurred and the risk of further harm.
  • Violence is predicted relative to a norm-based reference group.
  • Provides a precise and probabilistic estimate of further violence.
  • Has been shown to correlate with various measures of violent behaviours.
  • Increasing accuracy due to empirical approaches finding better predictors and reliability.
  • Practitioners may resist their use, due to their lack of utility.
  • Less emphasis on the unique, unusual or context-specific factors.
  • Historically focused on identifying immediate and visible harm, but may be less successful in identifying concerns associated with neglect or emotional harm, or supporting vulnerable families.
  • Existing instruments are not precise enough to discriminate types of risk (e.g., who will commit homicide vs less serious violence).
Structured professional judgement
  • Attempts to bridge the gap between clinical and actuarial decision-making, with a primary goal of preventing violence. Guidelines are used to conduct the assessment, but assessment also includes information gathering, communicating opinions, and implementing violence prevention strategies. The flexibility is in the final step of combining risk factors and tailoring management strategies.
  • More prescribed than clinical decision-making but more flexible than actuarial decision-making.
  • Does not impose restrictions on the inclusion, weighting or combining of risk factors.
  • Allows for a logical, visible and systematic link between risk factors and responses, as well as the ability to identify those who are at higher or lower risk for violence.
  • Some evidence of reliability and validity (i.e., the Spousal Assault Risk Assessment Guide).
  • Still allows considerable professional discretion, which can open up the approach to criticism.
  • Was initially met with optimism as a way of meeting halfway with clinical and actuarial judgement, but inter-rater reliability has been reported as poor compared to actuarial methods.

Source: Department for Victorian Communities (2007); Hilton, Harris, & Rice (2006); Kropp (2008); Winkworth & McArthur (2009).

There are many assessment tools in circulation, including those independently developed by services. As previously stated, however, there is a considerable lack of attention to evaluation of these tools. Kropp (2008) has suggested that four risk assessment tools currently hold the most promise, although they still only yield moderate associations with recidivism. These tools are outlined in Table 2.

Table 2: Family violence risk assessment tools and their existing strengths and limitations
Tool Description Comment
Danger Assessment (DA)
  • Designed to specifically address the likelihood of lethality or near lethality occurring in the context of intimate partner violence (Campbell et al., 2009)
  • Weighted scoring system, but developer does not recommend cut-off scores for decision-making.
  • One of the oldest measures still commonly used (Hanson et al., 2007).
  • One of the better-tested tools, with acceptable internal consistency and good test-retest reliability.
  • Further independent testing is needed, including its applicability to different cultural groups (Campbell et al., 2009).
Domestic Violence Screening Inventory (DVSI & DVSI-R)
  • Designed to be a brief risk assessment tool, to be completed alongside a criminal history review.
  • Produces an overall score indicating likelihood of imminent risk of violence.
  • Uses 12 social and behavioural factors found to be statistically related to domestic violence recidivism.
  • No independent validity studies to date. Authors have reported statistically significant predictive validity of DVSI, and concurrent and predictive validity of the DVSI-R.
Ontario Domestic Assault Risk Assessment (ODARA)
  • 13-item tool derived empirically from a list of potential risk factors gleaned from over 500 police files.
  • Uses actuarial scores of risk of repeated domestic violence. Male offenders are placed in one of seven categories of risk (as such it is a tool used with perpetrators rather than victims).
  • Correlated with Danger Assessment and Spousal Assault Risk Assessment.
  • Significantly discriminates between recidivists and non-recidivists of wife assault.
  • Authors caution against use for predicting lethal domestic violence.
Spousal Assault Risk Assessment (SARA)
  • Widely used structured judgement tool, used to guide professional judgement rather than a test in itself.
  • Twenty items were developed as identified by a review of the empirical literature on wife assault and clinicians evaluations of male wife abusers. Authors evaluated reliability and validity, and it is one of the few tools for which validity is supported by independent studies.
  • Uses an inclusive definition of spousal assault. It is not limited to acts that involve injury or death, nor particular legal status of relationship or gender of victim or perpetrator.
  • There is also a brief version - Brief Spousal Assault Form for the Evaluation of Risk (B-SAFER), which contains 10 items.

Frameworks

Specific tools for the screening and assessment of family violence are often discussed within the context of a broader framework. Most current frameworks for screening and assessment of family violence promote a structured judgement approach. The FRC & FRAL Framework (Winkworth & McArthur, 2008) considers both standardised tools and instruments, but also the role of evidence-based practice, practitioner and client knowledge and collaborative practice with other service providers (Winkworth & McArthur, 2009). A number of tools are suggested as a basis for developing screening and assessment tools and processes, rather than a recommendation of a particular tool. The structured professional judgement approach is favoured in the FRC & FRAL Framework, due to the limitations and criticisms of the other two approaches. However, it is also recognised that practitioners need a high level of skills and knowledge regarding family violence to make use of the structured professional judgement approach.

The Victorian Family Violence CRAF (Department for Victorian Communities, 2007) draws on three elements to determine risk: victim's own assessment (see below), evidence-based risk indicators and practitioners' professional judgement. Similarly, in South Australia, experienced workers who utilise the Family Safety Framework (Office for Women, SA, undated) are encouraged to act on professional judgement, even though the use of a common actuarial risk assessment tool is proposed. The importance of listening to the woman's experience of violence and taking into account her assessment of risk and safety is emphasised.

How good are victims of family violence at predicting future risk?

There is wide acceptance in the literature that risk assessments should be informed by victims' perceptions of their own risk, and that this is a reasonably accurate predictor of re-assault (Campbell et al., 2009; Heckert & Gondolf, 2004). This is both in isolation and also by improving the predictive value of other risk assessment instruments (Heckert & Gondolf, 2004). Hanson, Helmus, and Bourgon (2007) found in their meta-analysis of risk assessment that victim assessment has similar predictive value to the other, more formal approaches.

Caution in the use of victim's risk assessment, however, is suggested in the literature. Women who are at greatest risk may be the ones who communicate a feeling of safety, possibly because they have some uncertainty or uneasiness but not enough to take adequate precautions (Heckert & Gondolf, 2004). Kropp (2008) also suggested a consideration of the risk of danger minimisation by victims, due to issues such as denial or fear. He also pointed out that domestic violence is unique compared to many other forms of violence in that potential victims are known.

7 NSW Health has introduced routine screening for domestic violence in all Area Health Services, following a successful pilot program in 2001 (Irwin & Waugh, 2001). Women who attend antenatal, early childhood, mental health and alcohol/drug services are screened as part of a routine assessment.

What is being measured?

What is being measured?

A further consideration is the purpose or goal of assessment - what is, in fact, being measured? Tools may aim to elicit risk without clearly identifying:

  • the nature of the risk, such as re-assault or lethality;
  • the type of violence, such as physical or sexual violence; and/or
  • the severity, frequency or imminence of violence.

Without information regarding the validity and reliability of the tools, it may also be difficult to confirm whether any of these things are, in fact, being measured, and to what extent (Hays & Emelianchik, 2009). Hanson et al. (2007) questioned whether tools that specifically measure family violence are in fact needed, as risk factors such as substance abuse and unemployment are similar to those for general criminal recidivism.

A consideration of clients' cultural background is also a significant issue that is yet to be addressed in any comprehensive way in the literature. Hays and Emelianchik (2009) pointed to four indications of cultural bias in a tool - the extent to which:

  • a definition of violence is relevant for a culturally-diverse individual;
  • the severity and frequency do not match a particular norm;
  • they provide availability of culturally-appropriate resources; and
  • a particular item is clear and concrete for a particular culture.

The NSW Health pilot project (Irwin & Waugh, 2001) highlighted the importance of using interpreters for routine screening needed for those whose first language is not English, and the need for written material to be translated.

Creation of a specific organisational tool

Family support services within Australia have shown great initiative to create a tool or tools that fits the needs of the service, often mixing items from different tools and frameworks. This approach is supported by the Standards New Zealand Committee (2006), although it is advised that underlying principles regarding the use of tools remain consistent across tools. This flexible approach gives credence to the different circumstances, including cultural factors, which will occur for different groups accessing services in multiple areas.

The FRC & FRAL Framework offers a range of different items and tools from which service providers can and have developed their own approach to screening and/or assessment. Both the Victorian CRAF and the WA Department of Health Guidelines for Responding to Family and Domestic Violence (Women and Newborn Health Service, 2007) also offer a number of suggested questions, although the origins of the items and their psychometric properties are not specified. The CRAF offers three practice guides based on differing practitioner roles: mainstream professionals who may encounter victims of violence (an identification/screening tool); professionals who work with victims of violence but whose core work is not violence-related (risk assessment tool); and specialist family violence professionals (risk assessment tool). In this framework, there is a clear delineation between the tasks of screening and assessment, with some flexibility in terms of the questions used in practice.

Evaluation

Evaluation

There are a number of key dilemmas in evaluating the use of screening and assessment tools, some of which are reflected in the discussion so far. These include:

  • How do we know that the screening and/or assessment tool is measuring what it is supposed to measure?
  • How do we ensure there is consistent use of an organisation's chosen tools between practitioners, now and over time?
  • How do issues such as rates of refusal to answer questions, or inability to answer questions for reasons such as being accompanied by a partner, impact on evaluation?
  • How do we measure whether screening and/or assessment has made a difference? Is it enough to increase the number of cases identified, or is there an onus to see what the outcomes of referral or service provision were?

The last question requires follow-up beyond the screening/assessment period. The difficulty with this is: what is used to measure a difference? For example, Heckert and Gondolf (2004) used re-assault rates in their study, which examined the accuracy of women's perceptions of risk. Re-assault rates, however, were complicated by the fact that participants were chosen using a database of batterers from a multi-site evaluation of batterer intervention systems. In other words, the men were attending an intervention program, designed specifically to have an impact on their use of future violence. As such, it is difficult to know if the re-assault rates would have been higher if the men had not attended an intervention.

Spangaro (2007) suggested that further evidence is also needed to answer the following pertinent questions regarding screening and its routine use to detect family violence:

  • Does screening reduce abuse?
  • Does screening cause harm?
  • What outcomes can be expected?

Ethical issues also exist regarding asking women about violence when it is not the reason for their presentation to the service. Spangaro (2007) highlighted strategies to offset some of these dilemmas, such as a preamble warning women about the limits of confidentiality, the offer of an information card, and a policy of screening women alone. Interestingly, in the NSW Health pilot study (Irwin & Waugh, 2001) the overwhelming majority of women who completed a post-screening questionnaire were okay or relieved about being asked specific questions about domestic violence.

In a case such as that described in Box 1, a practitioner may also need to undertake a reflective process to answer questions such as:

  • What questions were asked and what formal assessment tools were employed at the intake stage?
  • Do procedures need to be tightened?
  • Is this a case of a person feeling sufficiently supported only now and therefore "ready" and willing to reveal her fears?

These questions suggest a need to visit an organisation's screening and risk assessment process on an ongoing basis.

Is it the tool that is important?

As Kropp (2008) has pointed out "any agency considering the implementation of risk assessment protocols (for family violence) must ... recognize that this is an imperfect enterprise" (p. 215). It seems that a consensus regarding the best tools for the job may be ambitious. There may be merit in a corresponding macro-level analysis of the issues involved in identifying and responding to family violence, and consideration of whether the search should focus on tools. For example, concentrating on the minimum, necessary steps to ensure client safety, and communicating between professionals to ensure that none of these steps are missed, may be a better way to manage the complexity associated with family violence. A service-specific checklist may be suitable for this purpose, similar to those used routinely by pilots and surgeons (Gawande, 2010).

If the focus remains specifically on tools, the current imperfections can be reduced in a number of ways, two of which deserve special mention. One is the provision of quality ongoing staff training and supervision, for both professionals and quasi-professionals, who may be involved in screening potential clients. The other is the sharing of information regarding the effectiveness and limitations of existing evaluation frameworks, and how to address associated issues of validity and reliability.

Don't forget the kids...

Although outside of the scope of this paper, it is critically important to acknowledge the effects of family violence on children (see for example McIntosh, 2003). This is not only in terms of violence directed at children, but also the effects of inter-parental violence and abuse on a range of physical and psychological factors that impact on parental capacity to remain attuned to the needs of their children. One of the key ways in which practitioners may be able to make a difference in this respect is by clearly communicating information to parents that summarises the extensive knowledge gained from social science research into the impact of family violence on children - impacts that are likely to exist and continue even if there are no signs of physical harm. A range of child-focused resources exist to support practitioners in this task.8

The mechanisms whereby family violence leads to a range of poor outcomes for children are also clearly and engagingly described by Johnston and colleagues, especially in Chapter 2 of In the Name of the Child (Johnston, Roseby, & Kuehnle, 2009), which explores "the prism and the prison" of the child caught up in these circumstances.

8 See for example publications from the National Child Protection Clearinghouse <www.aifs.gov.au/nch>.

Conclusion

Conclusion

Both clinical practice and research in this area are currently hindered by a lack of quality evaluations of the psychometric properties of existing tools (Rabin et al., 2009). More particularly:

  • Screening and assessment tools are not always identified or considered separately in research.
  • Though a growing number of tools have been tested for validity and reliability, few have been tested independently (Hanson et al., 2007).
  • No gold standard currently exists by which to test sensitivity, specificity and overall effectiveness of such tools (Rabin et al., 2009). At the same time, detection of and responses to family violence cannot wait until the gold standard research has been completed.
  • Little research has been undertaken on risk assessment for family violence relative to violent and general criminal recidivism.

Anecdotal evidence suggests that family relationship service providers are drawing on an enormous wealth of experience and judgement to create a range of tools that meet the varied needs of their clients. A comprehensive summary of existing screening and assessment tools in the family relationship services field, and their evaluation, was outside the scope of this paper. The authors, however, are aware of several projects being undertaken throughout the family support and allied sectors that are addressing the use and evaluation of these tools (See Box 2 for examples). The next step will be to widely disseminate information on these projects and the findings, in an effort to inform the field in an area in which high quality empirical research is currently limited and in which high quality interventions and clinical practice will continue to pose a range of challenges.

Box 2: Current initiatives relevant to family violence screening and assessment tools

Mediator's Assessment of Safety Issues and Concerns (MASIC)

Dr Jennifer McIntosh (La Trobe Universty) and colleagues Amy Holtzworth Munroe (Indiana University) and Connie Beck (University of Arizona) are currently piloting a large screening tool for use in Family Relationship Centres and other family dispute resolution contexts. The Mediator's Assessment of Safety Issues and Concerns (MASIC), currently in a piloting phase, is specifically designed for screening ex-couples who present with family law disputes about their children.

Family Relationship Centres - Townsville, Rockhampton and Mackay

Dr Sue Rice, Manager, Family Relationship Services and Research Projects

Anecdotal evidence suggests that a significant number of cases are presenting to the FRCs with issues of family and domestic violence. This information is being captured by practitioners and recorded on the individual clients' intake and assessment form at the initial screening interview (start of the FDR process). To date, there has been little analysis of this information beyond the respective FRCs using it to determine suitability of individual cases for FDR. The focus of this research is on identifying the prevalence, characteristics, and impacts of the reported violence on clients attending the FRC for family dispute resolution. By collating and analysing the data obtained from the intake and assessment forms it is anticipated that a picture or "map" can be created that realistically shows the nature of the violence presenting to the FRCs. The study is a collaborative venture conducted across three regional FRCs in Queensland and utilising the data analysis capabilities of the School of Psychology at Central Queensland University.

The research design uses a mixed methodology approach to collate and measure clients' quantitative responses and thematic coding for the qualitative data. Some case studies will also be drawn from the data to allow for a deeper understanding of the issues presented. A focus of the data analysis will be twofold. Firstly, to identify clients' self-identified fears and concerns regarding the violence and their associated safety needs. Secondly, to examine the impact of the violence on clients' intentions, capacities and preferences for participation in joint family dispute resolution sessions. Findings from the research will potentially assist service provision and planning in several ways, including: gaining a better understanding of the nature of domestic violence cases that FRC practitioners encounter; informing the construction and evaluation of safety planning for clients reporting violence; assessing the effectiveness and limitations of the intake and assessment forms; targeting training requirements for practitioners; and possible areas for further research. The research is due for completion the second half of 2010 and will be disseminated via practice seminars, conferences and publication.

References

References

  • Australian Law Reform Commission & NSW Law Reform Commission. (2010, April). Family violence: Improving legal frameworks (PDF 8.0 MB) (Consultation Paper, ALRC CP1, NSWLRC CP9). Sydney: ALRC & NSWLRC. <https://www.alrc.gov.au/publication/family-violence-improving-legal-frameworks-cp-1/>
  • Bala, N., Jaffe, P., & Crooks, C. (2007, January). Spousal violence and child related cases: Challenging cases requiring differentiated responses. Paper prepared for presentation at the Ontario Court of Justice, Judicial Development Institute, Toronto, Canada.
  • Braaf, R., & Sneddon, C. (2007). Family Law Act reform: The potential for screening and risk assessment for family violence (Issues Paper No. 12). Sydney: Australian Domestic and Family Violence Clearinghouse.
  • Brown, T., Frederico, M., Hewitt, L., & Sheehan, R. (1998). Violence in families: Report number one. The management of child abuse allegations in custody and access disputes before the Family Court of Australia. Melbourne: Family Violence and Family Court Research Program, Monash University and Australian Catholic University.
  • Campbell, J., Webster, D., & Glass, N. (2009). The Danger Assessment: Validation of a lethality risk assessment instrument for intimate partner femicide. Journal of Interpersonal Violence, 24(4), 653-674.
  • Department for Victorian Communities. (2007). Family violence risk assessment and risk management: Supporting an integrated family violence service system. Melbourne: Family Violence Coordination Unit, Department for Victorian Communities.
  • Gawande, A. (2010). The checklist manifesto. London: Profile Books.
  • Gould, S. (1997). The mismeasure of man. London: Penguin.
  • Hanson, R., Helmus, L., & Bourgon, G. (2007). The validity of risk assessments for intimate partner violence: A meta-analysis. Ottawa: Public Safety Canada. <https://www.publicsafety.gc.ca/cnt/rsrcs/pblctns/ntmt-prtnr-vlnce/index-en.aspx>
  • Hays, D., & Emelianchik, K. (2009). A content analysis of intimate partner violence assessments. Measurement and Evaluation in Counselling and Development, 42, 139-153.
  • Heckert, A., & Gondolf, E. (2004). Battered women's perceptions of risk versus risk factors and instruments in predicting repeat reassault. Journal of Interpersonal Violence, 19(7), 778-800.
  • Hilton, N., Harris, G., & Rice, M. (2006). Sixty-six years of research on the clinical versus actuarial prediction of violence. The Counseling Psychologist, 34(3), 400-409.
  • Irwin, J., & Waugh, F. (2001). Unless they're asked: Routine screening for domestic violence in NSW Health. Sydney: NSW Health Department.
  • Jaffe, P. G., Crooks, C. V., & Bala, N. (2006). Making appropriate parenting arrangements in family violence cases: Applying the literature to identifying promising practices (PDF 1.3 MB). Ottawa: Family, Children and Youth Section Department of Justice Canada. <https://www.justice.gc.ca/eng/rp-pr/fl-lf/parent/2005_3/2005_3.pdf>
  • Johnston, J., Roseby, V., & Kuehnle, K. (2009). In the name of the child: A developmental approach to understanding and helping children of conflicted and violent divorce. New York: Springer Publishing Company.
  • Kaspiew, R., Gray, M., Weston, R., Moloney, L., Hand, K., Qu, L., & the Family Law Evaluation Team. (2009). Evaluation of the 2006 family law reforms. Melbourne: Australian Institute of Family Studies.
  • Kempe, C., Silverman, F., Steele, B., Droegemueller, W., & Silver, H. (1962). The battered-child syndrome. Journal of American Medical Association, 181, 17-24.
  • Keys Young. (1996). Research/evaluation of family mediation practice and the issue of violence: Final report. Canberra: Attorney-General's Department.
  • Kropp, R. (2008). Intimate partner violence risk assessment and management. Violence and Victims, 23(2), 202-220.
  • MacDonald, H. (1998). What's in a name? Definitions and domestic violence. Melbourne: Domestic Violence and Incest Resource Centre.
  • McIntosh, J. (2003). Children living with domestic violence: Research foundations for early intervention. Journal of Family Studies, 9(2), 187-199
  • Moloney, L., Smyth, B., Weston, R., Richardson, N., Qu, L., & Gray, M. (2007). Allegations of family violence and child abuse in family law children's proceedings. Melbourne: Australian Institute of Family Studies.
  • New South Wales Department of Health. (2007). Routine screening for Domestic Violence Program: Snapshot 3 November 2005. Snapshot 4 November 2006 (PDF 661 KB). North Sydney: NSW Department of Health. <https://www.health.nsw.gov.au/parvan/DV/Pages/dvrs.aspx>
  • Office for Women, Government of South Australia. (n. d.). Family Safety Framework practice manual (PDF 1.2 MB). Adelaide: Office for Women.
  • Pizzey, E. (1973). Scream quietly or the neighbours will hear. London: Enslow.
  • Rabin, R., Jennings, J., Campbell, J., & Bair-Merritt, M. (2009). Intimate partner violence screening tools: A systematic review. American Journal of Preventive Medicine, 36(5), 439-445.
  • Scutt, J. (1983). Even in the best of homes: Violence in the family. Melbourne: Penguin.
  • Spangaro, J. (2007). The NSW Health routine screening for domestic violence program. NSW Public Health Bulletin, 18(5-6), 86-89.
  • Spangaro, J., Zwi, A., & Poulos, R. (2009). The elusive search for definitive evidence on routine screening for intimate partner violence. Trauma, Violence and Abuse, 10(1), 55-68.
  • Standards New Zealand Committee. (2006). Screening, risk assessment and intervention for family violence including child abuse and neglect. Wellington: New Zealand Police.
  • Ver Steegh, N., & Dalton C. (2008). Report from the Wingspread Conference on domestic violence and family courts. Family Courts Review, 46(3), 454-475.
  • Walker, L. (1984). The battered woman syndrome. New York: Springer.
  • Winkworth, G., & McArthur, M. (2008). Framework for screening, assessment and referrals in Family Relationship Centres and the Family Relationships Advice Line. Canberra: Attorney-General's Department.
  • Winkworth, G., & McArthur, M. (2009). A practice framework to guide screening and assessment in the Australian Family Relationship Centres and Advice Line. Child and Family Social Work, 14, 410-419.
  • Women and Newborn Health Service. (2007). Guidelines for responding to family and domestic violence (PDF 484 KB). Perth: Women's Health Policy and Projects Unit, Department of Health, Western Australia. <www.health.wa.gov.au/publications/documents/Gudielines_for_responding_to_Family_and_Domestic_Violence.pdf>

Authors and acknowledgements

Elly Robinson is Manager of the Australian Family Relationships Clearinghouse and Research Fellow at the Australian Institute of Family Studies.

Lawrie Moloney has had a long association with the Australian Institute of Family Studies and currently has a part-time appointment as a Senior Research Fellow. He is an Adjunct Professor at La Trobe University, where he was formerly Director of the Department of Counselling and Psychological Health in the Faculty of Health Sciences.

The authors would like to thank Margaret Anderson, Malia Dewse and Tibor Mokany for their comments and feedback.

Publication details

AFRC Briefing
No. 17
Published by the Australian Institute of Family Studies, September 2010
18 pp.
ISBN:
978-1-921414-49-7
ISSN:
1834-2434

Download Publication

Publication meta