A database of 20,000 to 30,000 vulnerable children selected by a "predictive risk model", proposed in the Government's long-awaited white paper on child abuse, is raising alarm in surprising quarters. Social issues reporter Simon Collins investigates

Dr Patrick Kelly is puzzled by the key elements of the White Paper on Vulnerable Children.

This is a surprise. As head of the country's main child abuse unit at Auckland's Starship Children's Hospital, he was part of an eight-person "expert forum" in 2009 that laid out the first outlines of a plan to tackle New Zealand's high rate of child abuse.

But when the white paper finally emerged last week, with a proposed shared database of 20,000 to 30,000 "vulnerable" children to be selected partly by a new "predictive risk model" untried anywhere in the world, he felt "kind of blindsided".


"I'm all for better collaboration, that's what I've been trying to work for for 20 years, but I guess I thought that could be achieved by evolution rather than revolution," he says.

"The danger of a revolution is that if you do a revolution and it's not successful, you can undo an awful lot of stuff in the process."

It's for the best of intentions. Social Development Minister Paula Bennett says the shared database will help doctors, teachers, social workers and the justice system "build up a clear picture of what is happening with a child".

"This is what has been missing for years - it is how children have fallen through the many gaps in multiple systems," she said.

But experts are deeply divided over whether, by "stigmatising" some families as potentially abusive and undermining the parents' authority, the database could actually do more harm than good.

Overseas experience is not encouraging. Britain created a national database - of all children rather than just the most "vulnerable" - in 2004, and disbanded it in 2010 because of concerns about inaccurate data and security.

In New Zealand, this week's revelation that a blogger could walk into two public Work and Income kiosks and access thousands of personal welfare files shows the difficulty of keeping any database confidential.

And, if Kelly is right, a new database may not even be necessary. His 2009 forum called for more data sharing, but added: "The forum did not interpret data sharing to require any expenditure on new IT. It is possible to share data simply by allowing individuals in different agencies to talk to each other."


The problem

A 2003 Unicef league table rated New Zealand's death rate from maltreatment of children under 15 as third-highest in the developed world after Mexico and the United States.

A 2007 update, including deaths from accidents as well as injuries in children and young people under 19, rated us worst in the rich world.

Ministry of Social Development data suggests the trend may be improving slightly. The homicide rate dropped from 1.07 for every 100,000 children under 15 in the 1990s to 0.85 in the five years up to 2006, but this was based on small numbers that can fluctuate wildly - averaging nine children killed each year in the 1990s and 7.6 a year in the first years of this century.

Bennett's response

The 2009 expert forum identified lack of information sharing as a big part of the problem. It recommended giving the health and education sectors a legal responsibility to protect children and requiring Child, Youth and Family Services (CYFS) to work with the other two sectors.

Last year Bennett issued a "green paper", written by consultant Dr Jo Cribb who has just been appointed head of the Ministry of Women's Affairs, suggesting teachers, social workers, doctors, police and other professionals should be able "to freely share information about children they work with".

Both the 2009 forum and Cribb's green paper estimated that 15 per cent of children, about 163,000, could be considered "vulnerable" at any one time.

In contrast, the white paper, written by a team in the Ministry of Social Development apparently without consulting experts like Kelly, is much more prescriptive but for a much smaller group. It proposes:

- A national "Vulnerable Kids Information System", cutely shortened to "Viki", that will be "a mechanism for extracting and combining information on children (and their caregivers) from existing databases" once a child reaches a "threshold of concern". Professionals across the sectors "will be able to both view information about these children and enter information about them" by late 2014.

- Two paths into the system: one through reports of suspected abuse or neglect to a "Child Protect" call centre; and one through a "predictive risk model" that predicts the children most likely to be maltreated, based on their families' past histories. Together, the two paths add to "around 20,000- 30,000 children and families who will need to be worked with intensively each year".

- All agencies working with children, including schools, health services, housing providers and police as well as CYFS, will have to have policies on recognising and reporting child abuse and neglect by the end of next year.

- Urgent cases where children are at immediate risk will still be referred to CYFS.

- Other families, where children are not at immediate risk but need help from more than one agency, will be referred to multi-agency "children's teams", organised by regional children's directors possibly based in district health boards.

- The children's teams will ensure a plan is developed for each child in which all parties sign up to agreed responsibilities. Parents will be told that they may be referred to CYFS if they refuse to participate.

The critics

Social Workers Association chief executive Lucy Sandford-Reed notes a major change of tone from the green paper to the white paper.

"The green paper certainly seemed to be signalling a much welcome look across the whole policy system to ensure that kids have the things they need to get a good start in life," she says.

"The white paper really has come from a broad definition of vulnerability - a lack of access to good health, education, housing and so on - down to focusing on a really quite small group of the hard-end families where abuse is highly likely. I find that quite surprising and disappointing."

The best clue to what has driven officials towards monitoring a small group of families through a national database lies in the Government's drive to get better value for money out of all public spending.

Its 10 outcome targets across the public service include getting 98 per cent of children into preschools before they start school by 2016, and immunising 95 per cent of infants and cutting physical child abuse cases by 5 per cent by 2017.

It has already used welfare reform to help with the first two targets by making benefits conditional on parents enrolling their children in preschool and having their Well Child checks on time from next July.

A database of children at risk, backed by a predictive model and a threat to report families to CYFS if they don't comply with plans, is one way to achieve the third target of cutting physical child abuse.

Sandford-Reed says the resulting policy is "harsh" and "rigid", driving social workers to trust a computer model more than their own on-the-ground judgment.

"The danger is it's an easy solution: this person is not on the database, therefore they are not at risk," she says. "These tools can come up with lots of false negatives and false positives, putting people through enormous amounts of stress when they don't need to. That in itself, generating stress, is going to generate some risk for children."

Kelly says the predictive model came "completely out of left field" and made him "profoundly sceptical".

Using 132 risk factors from a family's history, the model predicted that 37 per cent of the most at-risk fifth of children from welfare families would be maltreated in the next five years.

"So two-thirds even in that high-risk group are not going to abuse their children," Kelly says. "So the question is what intervention do you have that is going to reduce that risk without doing more harm than good?"

The Auckland University economists who developed the model assumed high-risk families might be offered a home visiting programme such as the US Nurse Family Partnership, which roughly halved the rate of child maltreatment from 54 per cent to 29 per cent.

But replicating that success in New Zealand is not so simple. A similar programme in Christchurch, Early Start, has had good success rates, but attempts to roll that out nationally through Family Start have had mixed results because the programme was contracted out to numerous local providers with inadequate training and support.

Kelly says those programmes already select families to participate from a short list of known risk factors.

"Why not just implement that?" he asks.

"I just don't understand why there is a need to re-create this."

"Essentially what this is is an experiment. They have analysed a set of data, they have come up with a set of risks, and they are now proposing to implement an intervention. They have no idea whether that intervention will work or not."

Professor Eileen Munro, a British academic who led a review of England's child protection system for the British Government last year, told the Weekend Herald by email that any predictive model built on child welfare data would capture a higher share of children at risk in certain groups than in others.

"If you have a problem in New Zealand of over-representation of Maori families in the child protection system, then this is likely to be worsened," she says.

The white paper says "close to half" of the 20,000-30,000 children targeted each year are expected to be Maori.

Munro warned the British Government back in 2004 against creating a national children's database. She said both children and parents would be less likely to seek help because they could not trust any professional to keep their information private.

She warned of security risks inherent in giving access to multiple agencies.

"One possible abuse is that it would help paedophiles identify vulnerable children," she said. "Another is that violent fathers, whose families are hiding from them, might be able to find their children's new school and so track them down."

The NZ white paper itself says the British database was shut down in 2010 because of "concerns about the security of the database and questions about the reliability of the data entered".

An AUT senior lecturer in computing, Dr Barry Blundell, warns that no database can be 100 per cent secure, and that the British one actually took professionals away from families.

"People were spending ever greater amounts of time sitting in front of a computer screen rather than actually out there and solving the problems," he says.

Munro's review found that "some social workers were spending up to 80 per cent of their time on paperwork".

The defence

Despite all these issues, many experts see value in the new system. Dr Emma Davies, whose research led to the multi-agency child abuse centre where Kelly is now clinical director, cites the way CYFS used a predictive model a few years ago to generate automatic reminders to social workers to check up on young people at risk of suicide.

"That got a reduction in youth suicide," she says. "Technology can help us do our jobs better."

Law Society family law chairman Garry Collin says that as a lawyer for children he often has to ring around multiple agencies to get information, so a shared database is "a great idea".

"You go to a family group conference and you have six or eight professionals all sitting around sharing information and a lot of it you have never heard before," he says.

"It seems to me that should never happen. That information should be readily available."

Ian Hyslop, a former CYFS practice manager in Waitakere now lecturing at Unitec, says CYFS already uses computerised tools to calculate risk factors, and sharing data with other agencies would be useful.

"Yes, we need some shared information system. It's about making that as useful as possible and being very careful that it doesn't become an impediment to professional decision-making and quality practice," he says.

Dr Irene de Haan, an Auckland University social work lecturer who worked with economists on the new predictive model, says it aims to help at-risk families before any abuse occurs.

"If there is a limited amount of money, which of course there is, it is better generally for that money to be channelled towards the preventive end," she says.

"Then the number of children needing to be attended to by CYFS would go down."

Labour Party welfare spokeswoman Jacinda Ardern and Greens co-leader Metiria Turei support building a shared database, but argue that it should be for all children, not just the most at-risk.

Former Children's Commissioner Dr Cindy Kiro also advocates a universal database, with full checks of a child's "wellbeing" at key transition points such as entering primary school and high school.

"I think this moves in the right direction," she says. "But it misses a huge group. There is an opportunity to do a lot better."

A wider independent inquiry into child abuse and domestic violence, funded by millionaire Owen Glenn and directed by former bureaucrat Ruth Herbert, may keep the issue alive over the next year.

In a written response to a request for an interview, Bennett told the Weekend Herald that targeting did not mean other children would not receive services.

"They will - but it does mean we can concentrate critical interventions on the children most at risk," she said.

The shared database would keep track of mobile families, allow teachers and doctors to report concerns about children, and let call centre staff go straight to the relevant lead professional when a new call came in.

"Security of the information will be paramount. Controlled and appropriate access will be used to ensure that professionals are only able to access the appropriate level of information required by them." She said the system would not repeat Britain's mistakes. It would cover only the most at-risk children, would not restrict professional judgments, and would "build on New Zealand's practice where the whole family is the 'case' rather than the UK model where the child was separated from the family".

"In the UK, young people expressed concern about data being held without their knowledge or consent. We are looking at safe ways in which young people would be able to have control and access to some of their information," Bennett said.

"The UK studies identified a problem of ambiguous or differential language between agencies and authorities, leading to misunderstanding. The proposed information system would overcome this issue by establishing a common lexicon and 'translation' between contexts."

On the web: Visit www.childrensactionplan.govt.nz for more information.