Friday, 27 May 2016

Too many false positives

A study by Andy Bilson and Katie Martin at the University of Central Lancashire has found that more than 20% of children born in 2009-10 were referred to children’s services in England before their fifth birthday. Half of those referred were suspected of being abused or neglected. Child protection investigations were carried out in the cases of 5% of the children.


Bilson is quoted as saying:

“Children’s services are under considerable pressure to investigate more mainly because of government, media and public responses to child deaths and an Ofsted inspection regime that is covering its back…. Social workers are swamped by this growing tide of investigative work leaving little time to support victims and help families overcome the problems leading to referral.”

The authors conclude that the scale of statutory involvement and the growing focus on early investigative interventions results in “a considerable proportion” of families suffering “high levels of suspicion, fear and shame”. And that this is done “…without evidence that the individualised, investigative approach is effective in preventing further harm.”

These findings are deeply concerning. Absence of longitudinal data (e.g. in government statistics) does not allow year on year comparison, but there must be a strong suspicion that reaction to tragedies such as Baby Peter and Khyra Ishaq, and the relentless pressure on services not to make mistakes, has resulted in unwelcome net-widening.

One of the most important performance indicators of a child protection system seems to me to be how accurately it identifies children who need to be investigated and how well it excludes those that do not. Put another way, a test of the system is how well it minimises the number of false positives. Just as medics are concerned to spare people from unnecessary operations, procedures and invasive tests, so those of us in child protection should be constantly trying to avoid unnecessary investigations.


Of course, if we do not monitor how well we are doing in this regard it is no wonder that we don’t do very well. Bilson and Martin are to be congratulated for starkly laying out the facts on this issue. Hopefully those in authority will now put in hand collecting and publishing this kind of data on a routine basis.