Minister admits domestic abuse screening tool doesnt work

The main screening tool used to decide which domestic abuse victims get urgent support "doesnt work", Safeguarding Minister Jess Phillips has admitted to the BBC.

Police, social services and healthcare workers across the UK have largely relied on the Dash questionnaire to assess risk since 2009.

But there are now mounting concerns from academics and those working in the sector that the checklist does not correctly identify victims at the highest risk of further harm.

Phillips told the BBCs File on 4 that she was reviewing the entire system supporting victims but said it wouldnt "change overnight".

The Dash (Domestic, Abuse, Stalking, Harassment and Honour-Based Violence) assessment is a list of 27 questions put to victims, to 24 of which they answer yes, no or dont know.

The questions include things like: "Has the current incident resulted in injury?" and "is the abuse getting worse?"

The resulting score typically determines what happens next. If a victim is classified as "high" risk, they are referred on for specialist, intensive support.

There were 108 domestic homicides in England and Wales in the year to March 2024, according to the Office for National Statistics.

But one analysis of 135 recent domestic homicide reviews, shared with the BBC, found just 20% of the Dash forms completed for the victims identified them as high risk.

Now, families of women who were murdered after not being graded as high risk are exploring legal action against the institutions they believe failed their loved ones.

"These arent just statistics, theyre preventable tragedies and it leaves families devastated when risks are inadequately assessed and victims are left exposed," lawyer Matthew Jury, whose firm has been approached by relatives, told the BBC.

Jess Phillips, whose remit also covers violence against women and girls, admitted there were "obvious problems" with the Dash questionnaire but said: "Until I can replace it with something that does [work] we have to make the very best of the system that we have."

"Any risk assessment tool is only as good as the person who is using it," she said, adding that practitioners needed to be trained to understand that risk was dynamic.

People were killed even when deemed to be at high risk, she added. "The grading system wont immediately protect you… It is the systems that flow from those risk assessments that matter much, much, much more than the score."

The Home Office is now reviewing how all agencies handle domestic abuse cases - including risk assessments - as part of its wider strategy to tackle violence against women and girls, which is expected to be published in the autumn.

Multiple academic studies, some dating back almost a decade, have looked into how Dash is performing and have found that it is unable to accurately identify high-risk victims.

A study from the London School of Economics published in 2020, analysing Greater Manchester Police data, found that in almost nine out of 10 repeat cases of violence, victims had previously been classed as standard or medium risk by officers using Dash.

In 2022, academics from Manchester and Seville Universities analysed anonymous police force data and found 96% of victims who were retrospectively judged as "high risk" had previously been classed as "standard" or "medium" risk by Dash.

"When it comes to the question of the reliability of Dash as a predictive tool, there is a growing consensus that Dash does not do that job at all well", Dr Heather Strang, director of the Jerry Lee Centre for Experimental Criminology at the University of Cambridge told the BBC.

"It wasnt designed to do that, it has become the primary tool over the past several years used by police to risk assess and predict future behaviour", she said.

"It is possible to do much, much better," said Tori Olphin, a former police officer and data scientist who worked with Thames Valley Police to analyse their Dash performance and who came up with an alternative tool, using a much wider collection of police data.

"This is a group of people who we are making decisions about who are incredibly vulnerable, who may go on to suffer life-changing, life altering or life-ending outcomes from their domestic abuse."

Details of organisations offering support for victims of domestic abuse are available at BBC Action Line.

The National Police Chiefs Council (NPCC) says that about 20% of all police recorded crime relates to domestic abuse, accounting for over 400,000 recorded incidents each year.

It decided in 2022 that forces should use a new questionnaire, Dara, which was developed by the College of Policing following its concerns that Dash wasnt producing consistent information on risk when used by officers.

But 20 out of 39 police forces across the UK which responded to freedom of information requests sent by the BBC, said they were still currently using Dash.

The NPCC told the BBC that risk assessment tools were not designed to be used in isolation.

"We also expect and encourage officers to use their own professional curiosity, training and experience to bring offenders to justice and ensure the effective safeguarding of victims," it said in a statement.

Charities who operate in the sector have also raised concerns about Dash, which many are required to use as part of their contracts to provide domestic abuse services on behalf of public bodies, the BBC found from freedom of information requests.

"We have always believed that it falls short of being competent, specifically with the demographic that we represent," said Djanomi Robinson, operations manager at Sistah Space, a charity working with black African and Caribbean heritage women affected by domestic and sexual abuse.

"Theres a lot of nuances that are missed, cultural specifications that are missed.

"Weve had many instances where service users of ours fall short, they dont rank very high in their risk assessment, but we as specialists can tell actually their situation is quite serious", she said.

Ellen Miller, chief executive of SafeLives, the charity that initially developed Dash and promotes its use, told the BBC that the "beauty" of the questionnaire was "its simplicity, its universality and its accessibility".

She said the apparent failings of Dash to correctly identify high-risk cases were down to "a combination of the tool not being updated to reflect everything we know now, and it is also an issue about how it is being used."

The government has commissioned SafeLives to run an initial review into the full domestic abuse risk process across agencies.