Thanks Mithran – the comment wasn’t about your coverage of the report though, it was about the WWC report (and the associated ‘Implications for policy and practice’ report they have also published).

My concerns are also reflected in the partial reporting of the study itself. My best guess is that this partiality reflects a current will to roll out FGCs in policy circles given the desire to grasp any policy lever that looks like it could reduce care numbers and also, probably even more crucially, to proclaim the success of a Randomised Control Trial (RCT) in children’s social care so the case can be made for more of these. While the aim of safely reducing care numbers is sound, we must also be wary of promoting the illusion that there are ‘intervention magic bullets’, rather doing the hard work of telling policy makers that without reducing poverty and inequality, and without substantially increasing support for families, any attempts to reduce the numbers in care will be like pushing a large boulder up a very steep hill.

By partial reporting of the study I mean that the study was unable to evidence any difference between families having FGCs and those who did not at the longest (18 month) data point. That finding is being given very little coverage in favour of the 12 month data which showed some favourable results for the impact of FGCs, and on which basis the claimed cost savings from FGCs are being made by the What Works. That the 12 month data show favourable impact can be welcomed. However, for those who demand clear evidence of efficacy, the 18 month data is unlikely to be the incidental footnote current reporting is treating it as. The lack of evidenced impact at 18 months may be more a reflection of the study execution (the actual sample size of children was around 1/6th of the planned sample size making smaller effects harder to pick up) than the impact of FGCs themselves. Nonetheless, we simply do not know. That is a significant gap in the study on its own terms given one of its principal aims was to provide clear evidence on the impact of FGCs on children’s entry into care.

The longest data points in a study (here the 18 month point) would usually be reported as the final, and key, outcome findings in an intervention study. The What Works Centre (now Foundation) previously took final outcome data from other FGC studies to argue that the evidence base for FGCs was inadequate and that this RCT was needed to address that gap. Judged by the same criteria the What Works used to assess the prior FGC evidence, the evidence base for FGCs is almost certainly little clearer following this costly study. That is before we even start to get on to the ethics of denying some families the choice of an FGC during the evaluation period on the grounds that that was the only way generating ‘causal evidence’ of FGCs’ efficacy.

I strongly welcome the proposal to roll FGCs out across England. But I am clear I welcome it as support for families’ rights of participation when children in their network are at high risk of being placed in the care of the state. Arguing for FGCs on instrumental rather than rights-based grounds, and also on questionable interpretations of evidence, is only likely to mean that their use is re-questioned further down the line.

Thanks for that, Gavin.
I think they covered this in the full report, which I’ve hopefully reflected in this section:

The two groups were similar aside from the fact that mothers in the control group were significantly more likely to have had a child taken into care previously.

Among those in the intervention group, 34.7% had gone into care within 12 months of the FGC, compared to 46.6% children in the control group. After researchers adjusted for the different care histories in the two groups, they calculated that 36.6% of those in the intervention group were in care after 12 months, compared with 44.8% of the control group.

Apologies if that doesn’t cover it.
Thanks,
Mithran

But unless I’ve misunderstand it, there were differences in maternal ‘care history’ between the two groups in the trial (those who got FGCs tended not to have had previous children removed, and those that didn’t get FGCs were more likely to have had previous children removed).

Given that we know having previous children removed is itself a ‘predictor’ of having more children removed (see Karen Broadhurst’s research on repeat removals), this seems like it could be important. Because families with previous children removed were less likely to get an FGC and more likely to have children removed, whereas families without previous children removed were more likely to get an FGC (or be offered one – less than half of families in the ‘intervention’ group actually got one) and less likely to have children removed.

If so, that seems a potentially important caveat to the headlines being reported here – and probably indicates that SW is more complicated than saying ‘hold this particular meeting at a particular point in time and 2,000 fewer children per year will end up in care’.

Maybe this doesn’t mean anything and / or I’ve misread the findings – but I do find it a bit worrying that there is NO mention of this at all in the “Implications for policy and practice”.

Oh and the report refers to ‘section 20 orders’.

Ps. I am pro-FGC, whether or not the child ends up in care or at home or with kinship carers etc., but I am anti-any suggestion that there can be simplistic solutions for families and SWs, however well-intended.