CAFCASS: Learning from Serious Case Reviews



Every so often yours truly suffers from an attack of optimism. Do not be alarmed. Thus far it has never lasted long. Whilst we continue to flounder in the usual ocean of falsity, some odd things have been happening recently. My last post advertised the Ministry of Justice’s ‘fessing up to the greater likelihood of men being imprisoned for the same category of offence as women. And I note that in the 2017 Parliamentary debate for International Men’s Day, Philip Davies was not the only MP who noted that men are both more likely to be imprisoned than women, and for longer sentences. Eh? Eh? But that’s the sort of thing that only hateful, misogynistic, conspiracy nutjobs think, surely?

And now this report from Cafcass: Learning from Cafcass submissions to Serious Case Reviews. There is some even more unexpected ‘fessing up being done. I’ll get to that.

First a brief reprise.


Firstly there was the 2016 report from Women’s Aid: Nineteen Child Homicides. This report highlighted “the tragic stories of 19 children and 2 women in 12 families that were killed by perpetrators of domestic abuse in circumstances related to unsafe child contact within a ten year period.”  This report used a few specially selected Serious Case Reviews to form the basis of a campaign culminating in a ‘debate’ in Parliament on 15th September 2016. It was not a debate, as reading the Hansard record testifies. A debate has two sides. This was a monoculture of opinion. Every MP reiterated the same perspective – the perspective put in their mouths by Women’s Aid’s carefully crafted campaign.

The purpose of Women’s Aid’s report is exposed by this claim,

“It’s possible that these deaths could have been prevented if the domestic abuse had been considered as an ongoing risk factor.”

And hence,

Women’s Aid believes that to prevent further avoidable child deaths, lessons must be learned from the deaths of these 19 children.  This report highlights key failings that need to be addressed in child contact cases involving domestic abuse.”

The report sought to cement a link between allegations of domestic abuse and a father being a danger to his child. The purpose was to further strengthen the strategy of alleging domestic abuse as a means of frustrating fathers’ contact with their children after separation. Women’s Aid called for,

Independent, national oversight into the implementation of “Practice Direction 12J – Child Arrangement and Contact Orders: Domestic Violence and Harm”

A review of Practice Direction 12J was duly carried out, its wording tightened in many places, and Women’s Aid have claimed a campaign win.

The efficacy of the domestic abuse strategy for frustrating fathers’ contact is based on,

  • The fact that roughly half* of child contact cases in the family courts involve allegations of domestic abuse, and this has become normalised and accepted despite its intrinsic improbability; and,
  • The linking of domestic abuse allegations with a father’s danger to his children – the claim which the Child First: 19 Child Homicides campaign was designed to consolidate.

*Aside: The proportion of child contact cases in private family law which involve allegations of domestic violence is estimated to lie in the range ~40% to 62% (see here and here). Even allowing for an increase in the incidence of DA during separation, such rates of DA are far larger than would be expected based on general surveys, e.g., the CSEW. As argued here, this suggests that more than half of these allegations are false.  

The 19 Child Homicides report raised in my mind an obvious question: if one reviews all the Serious Case Reviews, instead of ‘cherry picking’ those involving male killers, what do they show? I did not know what I would find, though I had a suspicion in view of earlier studies, especially those in the USA. Accordingly 330 Child Homicides was published on 2nd October 2016 (updated to 332 Child Homicides in February 2017). These reports were based on reviewing all the SCRs in the 7 year period 2009 to 2015.

The claims of the Women’s Aid report were fatally undermined by the following conclusions,

  • One or both parents are culpable in the killing of at least 65% of children – and probably a much greater percentage if culpability could be assigned in every case.
  • Mothers are more likely to be responsible for a child death than fathers and male partners combined.
  • Single mothers are the demographic most likely to be responsible for the deaths of children.
  • Six cases of mothers executing ‘spite killings’ or ‘family annihilations’ have been identified (involving 10 children’s deaths) in the seven year period 2009 – 2015, of which two occurred whilst the mother was “on contact”.
  • Nine cases, involving 15 children’s deaths, have been identified as either spite killings by mothers or murder-suicides by mothers, in the period 2009 – 2015.
  • Four cases have been identified in this period of mothers killing children whilst the father was subject to a restraining order or had been arrested for assault of the mother.

Why did 19 Child Homicides not pick up these facts, despite its claim to be gender neutral in its selection of cases? Answer: because its selection methodology required there to be an allegation of domestic abuse against the ultimate perpetrator of the killings. They sought to link allegations of domestic abuse with child killings. But, in fact, they did the opposite. By filtering on the basis of allegations of domestic abuse, they filtered out the bulk of the child killings.

2017 Cafcass Review of SCRs

The report Learning from Cafcass submissions to Serious Case Reviews by Richard Green and Emily Halliday was issued internally to Cafcass in June 2017, but the linked version is that made available publicly in November 2017. The report derives lessons to be learnt from the 97 Serious Case Reviews (SCRs) to which Cafcass has contributed from 2009 to 2016. These SCRs involve known or suspected abuse or neglect of a child where the child has died or the child has been seriously harmed.

The period covered by the Cafcass review is similar to that covered by my 332 Child Homicides (2009 to 2015). However, Cafcass review only those SCRs in which they themselves were involved, and include non-fatal cases as well as fatal cases. Of the 97 cases Cafcass review, about 60% involved the death of a child and 40% serious harm. My review, in contrast, was a complete trawl of all SCRs over the 7 year period, but confined to fatal cases. Ten of Cafcass’s 97 cases involved suicide, whereas I excluded suicide from my review.

In my review I warned that “these terrible cases have little or nothing to tell us about the bulk of society. In these SCRs we are almost always dealing with the sump of society: addicts, alcoholics, the seriously mentally ill, parents with severe learning difficulties, and people damaged by their own background”. Cafcass also warn against extrapolating from these severe, and atypical, cases to the wider public, thus,

It is important to bear in mind that SCRs represent a tiny fraction of the workload of Cafcass and other agencies. Their tragic outcomes commonly entail a degree of chance in line with the inherent unpredictability of much human behaviour, rather than a predictable or preventable external factor……Tragic outcomes are extremely rare.”

Already, and before examining the findings, there is excellent reason to reject the Women’s Aid position that Family Court policy should be based upon these exceptional cases.

Cafcass have summarised the findings of their review on-line: Learning from 97 Serious Case Reviews. Headline findings are as follows,

  • The vast majority of known, or suspected, perpetrators were family members;
  • Men and women were suspected perpetrators of a similar number of incidents of homicide;
  • Almost all allegations of domestic abuse were against men, usually fathers.
  • Allegations of domestic abuse had been made in 71% of cases, but in almost half of these cases the person thought to have killed or harmed the child was not the alleged domestic abuse perpetrator.
  • SCR submissions showed that in some cases the risk posed by the alleged violent adult may have masked the risk posed by the other parent.

The first two conclusions, above, are consistent with the conclusions of 332 Child Homicides.

It is worth pausing to digest the implications of the last two bullets. Despite the fact that the Cafcass report continues to refer to allegations of domestic abuse as a ‘risk factor’ – as if by habit – their own conclusion implies that allegations of domestic abuse are not a risk factor. Indeed, their own conclusion implies that allegations of domestic abuse are just as likely to detract attention from the true risk. This is exactly in line with the overall message of 332 Child Homicides.

Contrary to the claims of Women’s Aid’s Child First campaign, the laudable objective to “stop avoidable child deaths and make sure children are put first in the family courts” will not be met by further restrictions on child contact by those accused of domestic abuse (usually fathers), because unbiased review of the empirical evidence does not support such allegations being a risk factor. Indeed, in as far as this false perspective detracts attention from at least half the perpetratiors, the Child First campaign’s recommendations are seen to be the opposite of putting the safety of children first.

It is worth noting that not only this Cafcass report, but also the DoE’s Triennial Review (next section), “found that equal numbers of fathers and mothers are responsible for deliberate homicides”.

The Cafcass review notes that, “in some cases where index incidents were perpetrated by the mother, SCRs found that the mother’s history had not been sufficiently analysed, concerns about her being overshadowed by concerns about the father or other male. It is interesting to note that such SCRs do not show a simple relationship between male domestic abuse and the fatal/serious maltreatment of children.” The Cafcass review cites examples of cases where the overt risk was derived from one adult but the fatal or serious harm was perpetrated by somebody else.

Other key features or findings of the Cafcass review were,

  • 69 out of the 97 SCRs (71%) involved allegations of domestic abuse: 46 against men, 9 involving reciprocal abuse, and just one female lone perpetrator (13 unknown or unspecified).
  • 11 of the 97 SCRs involved the child’s suicide or overdose or the child was otherwise the perpetrator;
  • Of the remaining 86 SCRs, 49 involved death of the child. Of these, the father was responsible for the death in 13 cases, the mother in 14 cases, both parents in 3 cases, the mother’s partner in 4 cases and the mother and her partner in 2 cases, and one or both parents and another in 4 cases. 9 cases involved none of the above.
  • Hence, where death occurred, the father was not involved in the death in at least 59% of cases (excluding suicides) or 67% of cases (including suicides).
  • In nearly half of cases featuring domestic abuse, the person thought to have killed or harmed the child was not the alleged domestic abuse perpetrator.
  • As was found in 332 Child Homicides, neglect or co-sleeping were frequent causes of child death, and these were more commonly associated with the mother.

Despite these clear results, the Cafcass report authors cannot help but regress to the usual chivalric narrative of women incapable of sin and men made to bear the burden of culpability. For example, quoting the DoE’s Triennial Review, and agreeing with it, they write,

fathers and mothers may have different motives; the former being driven by a loss of control and a need to exact revenge, the latter by ‘altruism’, such as a belief that the child’s suffering must end.”

And again,

in the few cases of homicides, the report notes how fragile the mental health of some of the female perpetrators seems to have been and how the male perpetrators had histories of domestic abuse and control.”

It is impossible to interpret this other than as a desperate attempt to cling to the old gender-stereotypical presumptions, in the teeth of their own data. In those particularly distressing cases of ‘family annihilations’, there is every reason to believe the motivations of men and women to be much the same. And in other cases of culpable child killing, it is not the word ‘altruism’ which comes to mind, for either sex of perpetrator, as the cases recorded here testify.

DoE Triennal Review of SCRs 2016

In May 2016 the Department of Education (DoE) published their triennial review of SCRs, “Pathways to harm, pathways to protection: a triennial analysis of serious case reviews 2011 to 2014”. Although funded by, and published under the aegis of, the DoE, the nine authors (8 women, one man) were academics from the universities of Warwick and East Anglia. The study considers a total of 293 SCRs relating to incidents which occurred in the period 1 April 2011- 31 March 2014. Of these, 197 related to child deaths, and 96 to serious harm.

Salient findings were,

  • 120 of the 293 children (41%) were aged under one year at the time of their death, or incident of serious harm; and nearly half of these babies were under 3 months old.
  • Maternal ambivalence towards her child (both during and after pregnancy) was highlighted in many SCRs as a potential indicator of a child’s vulnerability.
  • A key and recurring theme is the extent and significance of neglect in the affected children’s lives (55%).
  • For children on a protection plan, the proportion for whom the protection plan arose from the risk of physical abuse has fallen in 2011-14 to 17%. Neglect accounts for 63%; emotional abuse for 18%; and sexual abuse 14%.
  • Final reports were obtained for a subset of 175 SCRs from the 2011-14 period, including both fatal and non-fatal abuse. From these the findings were,
  • The perpetrator was: the mother (23%), the father (17%), both parents (12%), mother’s partner (10%), the child himself (15%), others or unknown (23%).
  • Parental alcohol or drug misuse were recorded as present in 47% of cases. In 27% of cases both factors were present.
  • Parental mental health problems were found in 53% of cases.
  • Domestic abuse was reported in 54% of cases.

The report discusses the role of domestic abuse in these incidents. It rather followed the predictable lines and one wonders why the authors were not more concerned about the conflict between the conventional male-power-and-control perspective of DA and the SCR data itself. However, they did at least include this observation,

In accepting the reality of coercive control as the dominant issue in domestic abuse, we have to recognise that it is usually women in a relationship who are the victims of men’s controlling behaviour. However, coercive behaviour may also be exhibited by women on their partners, or, to different extents by both partners in a relationship. An uncritical attribution of control to a male partner may lead to gender bias and a failure to appreciate the full complexity of a relationship. In the following example, a psychiatrist recognised the coercive behaviour of the mother in the context of significant ongoing mental health needs: ‘It is significant that this psychiatrist who knew mother better than most identified her as being the aggressor as opposed to her partner.’”

Horses and Water

Perhaps there are signs that factual data are beginning to penetrate previously impermeable crania. Unfortunately, whilst the horse may admit to seeing the water, he may still refuse to drink.

Take education, for example. It is universally accepted that boys and young men are falling by the wayside, and have been for decades. But there is no sign that anything is being done about it. The mantra remains “girls in STEM”. And the recent signs that some people are beginning to accept that men are indeed treated more harshly in the criminal justice system implies no sense of obligation to do anything about that either.

There is growing evidence – supplied from unexpected quarters – that mothers pose a similar or greater risk to children than fathers – and that allegations of domestic abuse are no guide to risk. But can the agencies that count be made to inwardly digest the evidence? That’s a whole different battle.

9 thoughts on “CAFCASS: Learning from Serious Case Reviews

  1. AB

    “it is not the word ‘altruism’ which comes to mind, for either sex of perpetrator, as the cases recorded here testify.”
    Link doesn’t work in case you wish to post the file somewhere else.

  2. Nick Langford

    Women’s Aid doesn’t want a presumption of no contact only in cases where there are allegations but in ALL cases unless the father can prove that contact will be supportive of the mother and of demonstrable benefit to the child. Any application by a father for contact should be treated with extreme scepticism and regarded as a violent intrusion into the mother’s autonomy. They have been quite stealthy in working towards this, but it has long been an aim of feminists such as Audrey Mullinder, former principal of Ruskin College, who advocated a ‘rebuttable presumption of no unsupervised contact unless evidence can be produced that it will be safe’, or the authors Jalna Hanmer and Catherine Itzin who argued for the same in their 2001 book Home Truths about Domestic Violence.

    In 2002 Women’s Aid was taken to task in the report Making Contact Work for misrepresenting the consultation which led up to the report by seeking to represent domestic violence as a factor in ALL cases; Making Contact Work was concerned only with those cases in which DV was NOT a factor and DV had been considered in the earlier report by Sturge and Glaser.

    Ultimately, an application by a father for contact will be seen immediately as putting the child at risk of experiencing parental conflict and will automatically be dismissed.

    And if they achieve that, WA certainly won’t stop there.

  3. Groan

    Two observations. The first that the MoJ is trying to deal with the sharp upswing in Domestic Abuse allegations in Family Court proceedings following the ending of legal aid (mainly given to mothers/women) for all cases except where there is an allegation of DA. The MoJ is not Altruistic in this the upswing effectively wiped out the expected “savings” on legal aid that would have been made had DA accusations remained at their pre change levels.
    The second: Please please please any reader do not presume officials read their own Gov. funded research or data. After over 30 years of swimming in social and healthcare I cannot emphasise too much that these are full of people who are completely ignorant of the information other than Guardian articles. It is tedious I know but do take any and every opportunity to push official data/reports at officials. Ten to one it will be the first time they’ve seen it.
    Finally thank you Mr. Collins for your work in bringing this information together.

  4. AJ

    Meant to say at the end of my previous comment.

    The quality of all your blog posts and research are fantastic and brings strong evidence to bear on subject areac which is almost entirely debated in ignorance and prejudice.

    1. William Collins Post author

      Thanks for your kind words.

      In respect of your comment, recall that DA is alleged in about half of ALL cases of child arrangement disputes in the family courts – of which these nasty cases resulting in SCRs are an extremely tiny percentage. The fact that a DA allegation ‘predicts’ the perpetrator in about half of SCRs is therefore what one would expect if DA were completely uncorrelated with risk to the child.

      1. AJ

        If DA is alleged in half of all cases and 71% of serious case reviews then a DA allegation is more likely in cases the results in SCRs and less likely in those that do not so there is some correlation although not very much. If there was no correlation then the probability of DA in SCRs would be the same as in the general population. The practical use of this is nil because allegations of DA are so common and it has no use at all in predicting who is the threat to the child. It suggests that the accuser in cases of DA should be subject to the same level of scrutiny as the accused from a child protection point of view.

  5. AJ

    You say that the evidence from the CAFCASS review is that allegations of domestic abuse is not a risk factor because in roughly half the cases in which domestic violence was alleged the perpetrator of child abuse was not the person accussed of domestic violence.

    I think this is incorrect for two reasons:
    It can be a risk factor for a child while not being a good indicator of the source of the risk. The fact that an allegation was made in 71% of cases suggest this although this depends on the distribution of allegations generally.
    It is a risk factor if the proprotion of children subject to a serioss case review for which there is a related domestic abuse accusation is higher than the numbe rof children not subject to a serious case review with an associated accusation. This can’t be assessed by just looking at one of the two groups. This depends very much on the populations being compared which starts to get quite difficult. The appropriate population is not all children but children in which CAFCASS have involvement.

    The key point is that an allegation of domestic abuse is not useful in determining how has custody or access from a child protection point of view and that child abuse is often perpetrated by the accuser not the accused.


Leave a Reply to Nick Langford Cancel reply

Your email address will not be published. Required fields are marked *