‘Why the evidence does not support multi-agency child protection units’

By Andy Bilson 

Last month, Sir Alan Wood called for the establishment of specialist child protection teams to create “… a highly skilled high performing group of staff that can ‘smell’ the cases likely to lead to death and serious injury”.

Just days later, the Child Safeguarding Practice Review Panel’s report on the deaths of Arthur Labinjo-Hughes and Star Hobson similarly recommended that multi-agency child protection units (MACPU) be established in every local authority with the expertise of children’s social care concentrated in them.

The Department for Education (DfE) will decide whether to take up the idea later this year.

The panel’s proposal is that MACPUs would be responsible for convening and leading strategy discussions, carrying out section 47 child protection enquiries, chairing child protection conferences, overseeing, reviewing and supporting child protection plans, recommending court applications and advising other teams and agencies on child protection.

Major reorganisation required

This would require a major reorganisation of children’s social care as the new units would carry out a large proportion of its current workload.  However, it is questionable whether this proposed role would bring the MACPU staff within “sniffing” distance of a significant proportion of the children who are seriously harmed or who have died.

Whilst the panel reported that in 2020, 64.5% of children reported in serious incident notifications were known to children’s social care, most were at a lower level. Only 29% of the children subject of these notifications had been on a child protection plan or been looked after, said the panel’s annual report last year (p25). This raises significant questions about the development of MACPUs.

So, what is the evidence to support such a move and is it possible to have staff able to ‘smell’ out these high-risk cases?

There are no examples of MACPUs operating in England and thus no direct evidence is available. However, the panel suggests we can learn from the evidence base on multi-agency safeguarding hubs as MACPUs “share similarities with some MASH models”.

Evidence base for MASHs

However, it is not clear what or where this evidence base is for their effectiveness or otherwise. The panel cite a 2014 Home Office report on multi-agency working, but that report says it is not research but rather “an exchange of information, views and experiences” (p4).

Similarly, most research into MASHs is based on interviews with professionals and does not provide empirical measures of impact. Thus, a recent paper concluded that “whilst the theoretical benefits of implementing a MASH have been widely documented, the extent to which they transfer into everyday safeguarding practices has not”.

The same paper, a rare empirical study of a MASH, found that it did not prevent repeat victimisation, concluding that “practices and processes need to be reviewed if MASH is to proactively prevent repeat victimisation”. This is not a ringing research endorsement.

Looking at broader data, MASHs have been introduced in most local authorities since 2010, a period in which the number of child protection investigations has more than doubled, to a peak of almost 200,000 a year for the last three years. This is a higher rate of investigations than at any time since records were kept.

In 2010, half of these investigations were followed by a child being placed on a child protection plan. This rate fell to one in three by 2021.

‘Harmful’ enquiries increasingly badly targeted

The number of occasions families were put through a highly stigmatising, and harmful, investigation not leading to a child protection plan increased from 45,000 to 135,000 during this period, a threefold increase. Thus, the targeting of investigations has become considerably worse during the period in which MASHs have proliferated.

Does this increasingly high rate of investigations lead to better detection of children at risk of physical harm? Let’s look at the data.

Over the period of this increase, the number of investigations finding that a child has been physically or sexually abused have changed very little and child deaths and serious cases have not fallen.

I have compared rates of child protection enquiries and plans for Bradford and Solihull, the authorities responsible for Arthur and Star, with those for authorities with the most similar charcateristics to them, taken from the DfE’s children in need statistics.

Compared with their five closest “statistical neighbours”, Bradford and Solihull, both of which have MASHs, had the highest rates of section 47 investigations per 10,000 children, and the lowest proportions of these being followed by child protection plans, over the past three years.

Carrying out high rates of investigations does not mean we protect more children; it means instead that resources are spread thin, decisions are rushed and the likelihood of mistakes increases.”

Such an approach also has significantly negative real-world impacts for children and families. Thus, national statistics do not provide evidence that MASHs are able to ‘smell’ out cases and prevent physical or sexual abuse.

However, MASHs do vary in their operation and the national picture includes a few local authorities which do not have MASHs.

The panel cited Hampshire as its example of a successful authority that has “extended the MASH model or aspects of it to deliver a multi-agency response”, moving it closer to the panel’s proposed model. In Hampshire, strategy discussions were held in the MASH, leading to “comprehensive information-sharing”, found Ofsted’s 2019 inspection of the county.

I found that, in Hampshire, between 2013 and 2021, the number of investigations starting in the year rose more rapidly than in any of its five closest statistical neighbours. By 2020-21 it had a rate of 212 investigations per 10,000 children, while its statistical neighbours’ ranged from 70 to 189.

Less than one in four of Hampshire’s investigations in 2021 led to a child protection plan, a lower rate than any of its statistical neighbours or the national rate of one in three. Based on these figures, Hampshire does not demonstrate that an enhanced role for the MASH leads to well-targeted investigations.

No evidence, experience or plan

The panel’s report includes one paragraph on implementing MACPUs, in which it proposes that the government provides start-up funding and some areas test the model before wider implementation.

It is concerning that a proposal for a major reorganisation of children’s social care and partner agencies is being put forward with no evidence for it, no experience of it in operation, no plan to properly evaluate it before wide implementation, and a single paragraph to describe implementation.”

It is understandable that, following such tragic child deaths we should look to what went wrong and try to put it right. The mistakes that the panel’s report identifies in the way agencies worked together in these two cases chimes with previous findings of problems in inter-agency work.

Sadly, the suggested response looks like more of the approach which has caused the spiral of increasing investigations and that has left agencies drowning in responding to heightened concern about risk of harm. This is a pattern we have seen following inquiry after inquiry throughout this century.

Engaging – not alienating – families

To better protect children, we need to look for a different response that engages families and communities rather than alienating them.

We need to lift our eyes from constantly picking over failings that is the inevitable focus of inquiries into child deaths. Instead we should look to positive approaches where stronger inter-agency work is being developed.

For example, in its recent final report, the Independent Review of Children’s Social Care (pp 51-52) identified how Camden’s investment in help and support in partnership with families led to major improvements. There were falls of 47% in referrals and 29% in children in need per 10,000 children between 2014 and 2021. Children in care rates fell by 48% between 2012 and 2021.

Leeds, like Camden, promoted family group conferencing. It carried out a wide inter-agency programme (Family Valued) to change the culture of social work teams and local agency partners through promoting restorative practice. The 2017 evaluation of the programme showed statistically significant reductions in both the number and rate children looked after per 10,000 population; the number of child protection plans; and the number of children in need. In both these areas, the focus was on strengthening inter-agency work through promoting partnership with families.

If we want to improve children’s wellbeing, develop more effective inter-agency work, and engage communities so children are safer, we need a change from the culture of investigation and searching for blame and to develop expertise in working alongside families and communities.

Andy Bilson is emeritus professor of social work at the University of Central Lancashire and co-chair of the Parents, Families and Allies Network. He has been a senior manager in children’s social work and was director of the Council of Europe’s observatory on children’s rights.