Evaluation of the Safer Pathway Program and the DVSAT screening tool

Release date: Thursday 3 May 2018

The Safer Pathway Program

Full report: Assessing the impact of NSW’s Safer Pathway Program on recorded crime outcomes – an aggregate-level analysis (pdf, 903Kb)

The Safer Pathway program has only had a limited effect on the incidence of domestic violence (DV) in NSW, according to a report released by the NSW Bureau of Crime Statistics and Research today.

Under the Safer Pathway program, all women reporting domestic violence to police are assessed to determine whether they are at serious risk of repeat victimisation using a questionnaire known as the Domestic Violence Safety Assessment Tool (DVSAT). The cases of those found to be at serious risk of repeat victimisation are referred to a Safety Action Meeting (SAM), where a team of government and non-government officials develop a 'safety action plan' for the victim to reduce the risk of further victimisation.

The program was rolled out in two stages. The first stage involved the Local Area Commands (LACs) of Waverley (Eastern Suburbs, Eastern Beaches, Rose Bay and Botany Bay) and Orange (Canobolas LAC). The second stage involved the LACs of Bankstown, Broken Hill, Parramatta and Tweed Heads/Byron.

To evaluate the program, BOCSAR matched each of the treatment LACs just referred to with a similar LAC where the SAMs were not operating. It then conducted two sets of analyses on these matched pairs. The first compared the treatment and comparison groups as a whole within each stage. The second examined differences in outcomes between the matched pairs. 

Seven different measures of domestic violence were examined; including the number of reported domestic violence (DV) related assaults, the number of people (POIs) arrested for a DV-related incident and the number of police call-outs to DV related incidents.

No significant improvements were observed for any of the Stage 1 sites. In the Stage 2 sites, the number of DV incidents, the number of POIs proceeded against for DV and the number of DV victims all fell by 0.8% per month.

The matched pair analysis revealed few significant improvements in any of the LACs except Canobolas and Rose Bay. In Canobolas LAC, three of the seven DV indicators (those measuring total DV incidents, POIs proceeded against for DV incidents and DV victims) exhibited a faster drop than Griffith LAC (the matched control LAC), by 2.2%, 1.6% and 2.2% per month respectively.

The number of POIs proceeded against for DV offences in Rose Bay LAC declined by 2.9% per month faster than in Ku-ring-gai LAC (the matched control). However, significant upward differences in trend changes were found for four different outcomes in Botany Bay LAC over Albury (the matched control).

The DVSAT screening tool

Full report: The Domestic Violence Safety Assessment Tool (DVSAT) and intimate partner repeat victimisation (pdf, 854Kb)

In a separate but related investigation, BOCSAR evaluated the DVSAT screening tool (which is a key component of the Safer Pathway program), to see how reliable it is in identifying victims of domestic violence at high risk of repeat victimisation.

The DVSAT consists of 30 questions and is made up of two parts: Part A and Part B. Part A involves 25 questions dealing with the background/current environment of the offender, the threat of violence, the dynamics of the specific relationship, the presence of children and any experience of sexual behaviours/assault. This part is only used for victims in intimate partner relationships.

Part B is completed for victims in both intimate and non-intimate relationships. The investigating officer provides responses to five questions relating to the level of fear felt by the victim, the reasons for those fears, whether there are children at risk of harm, and whether there are any additional factors that cause the officer to believe there is a threat or serious threat to the safety of the victim and/or children. Victims are assessed as being 'at serious threat' if:

  1. They give 12 or more 'yes' responses to the 25 questions in Part A or
  2. A police officer concludes on the basis of responses to Part B that the victim is at serious threat.

To evaluate the DVSAT BOCSAR tracked 24,462 victims of intimate partner violence who were administered the DVSAT between 1 January 2016 and 30 June 2016 to see which of them (a) experienced any repeat incident of domestic violence and/or (b) experienced a physical incident of domestic violence (including homicide, assault, sexual assault or robbery).

The DVSAT turned out to be a very poor instrument for measuring the risk of repeat domestic violence victimisation, often performing little better than chance.  For example, while the rate of repeat victimisation overall in females who responded 'yes' to 12 or more items was higher than the rate in those who responded 'yes' to fewer than 12 items (45.0% vs. 34.2%), over one-third of those who responded 'yes' to fewer than 12 items actually went on to experience repeat victimisation.

Again, of the females who experienced repeat victimisation involving any type of intimate partner violence, only 10.0 per cent had responded 'yes' to 12 or more items. Put another way, 90.0 per cent of those who experienced repeat victimisation had responded 'yes' to fewer than 12 items in the DVSAT.

Commenting on the results of the research, the executive director of BOCSAR, Dr Don Weatherburn, said that the unreliability of the DVSAT may be one reason the Safer Pathway program has had only limited success in reducing domestic violence. "Many victims of domestic violence who are at serious risk of repeat victimisation are not getting the benefit of a safety action plan. BOCSAR is working on developing an improved screening tool to overcome this problem", he said.

Further Enquiries: Dr Don Weatherburn 02 8346 1100

Copies of the report: www.bocsar.nsw.gov.au