Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Psychiatry
http://ccp.sagepub.com/
Published by:
http://www.sagepublications.com
Additional services and information for Clinical Child Psychology and Psychiatry can be found at:
Subscriptions: http://ccp.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://ccp.sagepub.com/content/6/1/9.refs.html
What is This?
A B S T R AC T
A group of children (N = 95) seen by our team at least one year previously for
assessment after one parent had killed the other, was followed up by a postal ques-
tionnaire to the original referrer. Through this we examined a number of outcome
variables including placement effects, the frequency of their contact with the
surviving parent, the referrer’s view of the difference our intervention had made
and their view on the child’s adjustment over time. We analysed the data to deter-
mine any associations between these factors to help us understand the difficulties
these children face and to aid clinical decisions.
K E Y WO R D S
bereavement, children, homicide, outcome, trauma
Introduction
THERE CAN BE few more appalling events for a child than the violent death of one
parent where the assailant is the other parent. The orphaned child has to cope with trau-
matic bereavement, numerous other losses and changes of carer and home, as well as
with the shame and ignominy of being the child of a killer. Since declaring an interest in
the predicament of these children (Black & Kaplan, 1988), we have, as an NHS clinic,
formerly based at the Royal Free Hospital and now at the Traumatic Stress Clinic,
London, seen more than 370 children from all parts of the UK, where one parent has
killed the other. Our book When Father Kills Mother: Guiding Children Through Trauma
and Grief (Harris Hendriks, Black, & Kaplan, 1993) describes this work and summarizes
research findings on 95 children from 45 families who were referred consecutively to our
clinic for assessment.
In that study, 40% of the children were under 5 years old at the time of their parent’s
killing, 41% were between 6 and 11 years and only 18% were already teenagers. Ninety
per cent of cases involved the killing of the mother by the father; fathers were the victims
in only 10% of cases. Jealousy was the commonest reason cited for the homicide; other
factors acting in conjunction with previous domestic violence, which occurred in many
of the cases, were alcohol abuse and the spouse’s threats to leave. About one-half of the
children had witnessed previous domestic violence; one-third witnessed the actual killing
and a further 10% heard the killing or saw the dead body immediately afterwards. There
was a strong association between witnessing the homicide and the development of post-
traumatic stress disorder (PTSD). Moderate to severe PTSD affected 25% of the chil-
dren; a further 25% had mild or isolated symptoms of post-traumatic stress. At the time
of assessment, almost 60% of the children had behavioural problems in the moderate to
severe range and 40% had symptoms of emotional disorders.
The group of children we saw was heterogeneous, referred for different reasons and
at different times after the homicide, ranging from days to years. It might be said that all
that they shared in common was an event, the tragic loss of one parent at the hands of
the other. Otherwise the children were more different than they were alike. However,
there were some significant findings, in particular that this was a highly symptomatic
group of children and that these children, who required security and stability more than
most, seemed to have many changes of placement. We wished to examine further these
two observations by asking the original referrers by means of a postal questionnaire,
what had happened to the children following our assessment and intervention.
Most of our assessments had included a brief crisis intervention, particularly if the chil-
dren were referred soon after the homicide, and in some cases we engaged the child in
a more comprehensive therapeutic programme. We wanted a rough estimate, from the
referrer’s perspective, of the effects of the intervention or therapy on the child and caring
system. We also wished to know whether the children we saw soon after bereavement,
when we were more able to influence placement decisions, had had fewer placements
AC K N OW L E D G E M E N T S :
We are grateful to Lisa Strickland-Clark and Tracey Thorns who
performed the data entry and helped with the statistics. Our thanks also to Jean Harris-
Hendriks for her comments and suggestions. This article is adapted from Chapter 12 of the
revised second edition of the book When Father Kills Mother: Guiding Children Through
Trauma and Grief authored by J. Harris Hendriks, D. Black, and T. Kaplan (2000), published
by Routledge.
D O R A B L AC K , M B, F R C P S Y C H , F R C P C H , D P M ,
is Honorary Consultant Child and
Adolescent Psychiatrist, lately Director, at the Traumatic Stress Clinic, Camden and Isling-
ton Community NHS Trust, London. She is also Honorary Senior Lecturer, University of
London at Royal Free Hospital School of Medicine and UCL, and Honorary Consultant, Gt
Ormond Street Hospital, Royal Free Hospital and the Tavistock Clinic.
C O N TA C T : Dora Black, Traumatic Stress Clinic, 73 Charlotte Street, London W1P 1LB, UK.
P H I L L I PA H Y M A N , B A ( H O N S ) is a Research Psychologist.
10
than the children referred late. We know of no other specific study and certainly no
outcome studies on a cohort of children who have lost one parent at the hands of the
other.
Method
Every professional referrer (in most cases, the allocated social worker) of the 45 families
(95 children) in the original cohort was sent a 12-item questionnaire for each child in the
family we had assessed originally. A reminder letter was sent out after three months, in
some cases followed by a telephone call. Data were analysed using the statistical package
of the SPSS with Pearson’s Chi-squared analysis of association (verified by Fisher’s exact
two-tailed test because of the small sample size).
All children had had their initial assessment with our service more than 18 months
before follow-up (the median was 3.5 years, with a range of 19 months to 15 years). Many
of the children had not yet reached adolescence.
Results
Response rate
Approximately three-quarters (74%) of the referrers returned our questionnaires. Ques-
tionnaires were, in some cases, returned incompletely filled in. The information we
requested was not always available to the respondents. We obtained useful but incom-
plete data on 61 children from 33 families whom we had assessed, and in some case
treated. (Similarly, data from our original study were not uniformly complete.) Thus
‘missing data’ varies across different tables.
Characteristics of non-responders
The children lost to follow-up (n = 34) were not significantly different from those we
were able to follow-up with regard to age or whether they had witnessed the killing. Most
successfully followed up were the children who were, at the time of the assessment, living
in children’s homes or other residential institutions – questionnaires were returned on
15 children of a total of 18. This compares with 8/13 who had been staying with paternal
(in all cases, perpetrator’s) relatives at the time of our original assessment, 10/24 who
had been with maternal (in all cases, victims) relatives and 16/34 who had been with
foster parents. It was not possible, from an examination of the case notes to determine
where six children in the follow-up sample were living at the time of the assessment.
Placements
At the time of assessment there were more children living with mother’s (i.e. the victims)
relatives (27% of the original sample) than with father’s (15%), 38% were in foster care
and 20% were in children’s residential establishments. At the time of follow-up approxi-
mately one-third of children (21) were living with their mother’s relatives, 16% were with
their father or his relatives (six and four children, respectively), and almost one-half were
not living with relatives. Most of these were in foster care.
The three children whose mother killed their father were all living with non-related
carers. Therefore, all children living with maternal relatives had lost their mother and so
had all children living with father or his relatives.
Only approximately half of the children (30) on whom follow-up data was available,
were in the same placement as they had been in at the time of our original assessment.
Fourteen children (approximately one-quarter) had had more than one change of
11
Table 1. Children with three or more placements since the death of a parent
Number of placements Number of children Number of families
3 15 8
4 6 4
5 6 3
6 2 2
Total 29 17
placement following the assessment. Twenty-nine children (that is nearly one-half) had
had three or more placements after the homicide of their parent (Table 1).
Of the first placements after the homicides which became permanent, 75% were with
the mother’s kin. In all of these cases, the mother was the victim (Table 2).
Children who were in foster care or in adoptive placements at follow-up had had more
placements than the other children, although the only significant difference within this
domain was between the number of placements experienced by those in foster care or
adopted, and those in the mother’s family. Our early involvement did not reduce the
number of placements.
12
boys and one girl) were now living with their father. In one other case, a daughter (living
independently at follow-up) had lived with her father for a while but eventually had to
be moved because of his violence. Of the other 13 sibling groups whose fathers had been
released, only two had regular contact with him, and both these sibling groups were living
with long-term foster parents. The remaining 11 sibling groups had infrequent or no
contact. Of these, seven were living with maternal kin.
13
Placement effects
We examined aspects of outcome for the children according to where they were living
at the time of our initial assessment and, separately, at follow-up. In all three cases in
which the mother had killed the father, the children were placed with non-related carers.
At follow-up we found that children who had, at assessment, been living with the
perpetrator’s (in all these cases, father’s) family were doing relatively worse than chil-
dren in any other placement with regard to behavioural and emotional problems.
We identified children with pronounced bullying or victim behaviour. On this dimen-
sion, the children who were in foster families at the time of assessment were doing much
worse than the other groups at follow-up. Children living with their mother’s family at
assessment had less problematic peer relationships than children placed with their
14
father’s family. The children who were in children’s homes when we first saw them,
tended to come from large sibling groups who could not be placed together elsewhere.
It may be that being kept together helped to protect these children from extremes of
social maladjustment as they tended to be better adjusted on the bullying/victim dimen-
sion than the children in foster care.
These trends appeared to be persistent. Thus, children living with the victim’s family
(in all cases, the mother’s family) at follow-up were doing better with regard to their peer
relationships than children in foster care, and both groups were doing slightly better than
children in the perpetrator’s (father’s) family. Children in adopted placements were
doing better than children still in foster care. If this finding is confirmed by further studies
on our larger number of cases, then the results are important for long-term planning for
these children.
Furthermore, children living with their father’s family at the time of follow-up were
less likely to be showing improvements in their behavioural and emotional adjustment
following our intervention than those children living with their mother’s family (Table
5) or in foster or adoptive families (Table 6). The numbers are too small and the level of
significance is not high enough to justify far-reaching conclusions. However, if these
tentative findings are correct, they would call into question – at least in these unusual
circumstances – the received wisdom embodied in the Children Act, 1989, which assumes
that, when children cannot live with their parents, it is better for them to be placed with
any family members than placed outside the family.
15
16
category, 4–6 on the scale, was excluded from the calculation. According to the refer-
rer’s view, children in either of the placements with relatives did not seem to form
meaningful relationships as easily as children in foster placements or adoptive place-
ments, and there seemed to be no particular difference, at least at a statistical level,
between children living with mother’s family or with father’s family (Table 8).
17
Discussion
The original research was carried out post hoc on a clinical sample. The children were
interviewed within a clinical framework, not a research one. We felt that it would be
unethical to approach the children and their carers directly for follow-up information
without careful preparation. This would, in all probability, have also introduced a selec-
tion bias, as the children and carers who would have given their permission for a follow-
up research interview, would have been more likely to be those who felt that the children
had benefited from our initial intervention. Although there are clear limitations to the
methodology we chose, i.e. a questionnaire to the referrers, we thought that this would
give us a wider and more balanced sample. With hindsight, we have some reservations
about the form of some of the questions we posed to referrers and recognize that we did
not test the questions for reliability or indeed for validity. This of course applies in
particular to the psychological outcome data. We did not anonymize the responses, and
even though the correspondence was carried out through a research assistant and our
covering letter encouraged frankness, some referrers may still have erred on the side of
politeness when asked to evaluate the effect of our intervention. It was noticeable that
respondents answered with least frequency the questions that required the most in-depth
knowledge about the children, that is, regarding their openness about their mother’s
death with their carers and the extent to which they were able to form meaningful
relationships. It seems likely that they would have had less contact with those children
who did not seem problematic, and this may have limited their responses (recorded as
‘don’t know’) on some children whose adjustment was satisfactory.
However, the information about the nature of the children’s placements and their
contact with their surviving parent could be regarded as relatively hard data. We were
able to confirm that in the long-term, these children did indeed have a greater number
of changes of placement than was desirable or expected. For some of the analyses, cells
were so small that results should be regarded, at the most, as suggestive. None of our
data should be regarded as sufficiently robust to dictate important social decision-
making for such children. That being said, the referring social workers’ responses suggest
that placement within the victim’s family appeared to have advantages over placement
with that of the perpetrator’s, and was the most stable (i.e. fewest moves) placement
type. Surprisingly, since in the original sample the children in the children’s homes were
the most globally disturbed, at follow-up they appeared to be doing reasonably well in
terms of their social adjustment, at least as well as, or better than children in foster
homes. One of course needs to question whether there might be a greater degree of toler-
ance and optimism by their carers towards these particularly disadvantaged children
growing up in a children’s home.
In the majority of cases, referrers felt that our intervention had done some good (56%)
or at least no harm (28%). This was important information for us as our brief inter-
vention sometimes produced distress at the time. In the majority of cases the assessments
were requested by a civil court to help with decisions about where the children should
live and with whom they should have contact and the frequency of contact. In many of
the cases referred the children travelled great distances; this made it difficult for us to
offer treatment ourselves. We often questioned whether a brief ‘preventive’ intervention
without the opportunity to review and extend treatment, might do more harm than good.
This did not appear to be the case. However, there was more observable improvement
in children who had a longer period of therapeutic work. It is possible that the children
who persisted with treatment came from more supportive families and that may be an
influential factor in their better outcome.
18
We would tentatively conclude from our findings that, especially for children who
witnessed the killing who have a high risk of developing PTSR, a brief crisis intervention
is not sufficient, and the children should be assessed for their treatment needs following
any brief intervention. The children who did best in our study were those who received
longer therapeutic input, even though one might have predicted that these children, who,
on the whole, were witnesses to the killing, would have done worse. We treat these results
with caution, as internalizing symptoms are notoriously unreliably rated by untrained
observers (Barrett et al., 1991; Klein, 1991). The children’s post-traumatic symptoma-
tology may well have continued in some cases without this having come to the attention
of the carer or referrer. It would be necessary for a clinician to examine children in order
to elicit symptoms and signs of post-traumatic stress disorder as children rarely volun-
teer them, and indeed may actively conceal them from their carers in order to protect
them emotionally. It was not possible using our method of data collection to ascertain
whether PTSD persisted. Our clinical experience and that of others in this field, is that
PTSD is a chronic, remitting and relapsing psychiatric disorder which is difficult to treat
once it is established. There is controversy about whether early intervention following a
disastrous event can prevent it (Raphael, Meldrum, & McFarlane, 1995).
Approximately 10% of our larger clinical group is of mothers killing fathers (41 chil-
dren). We are currently evaluating this group of children separately. Most of them have
not yet reached adolescence but our clinical impression to date is that these children do
well during their pre-adolescent years, becoming very difficult and defiant when peri-
pubertal.
The children placed with relatives of the perpetrator are more likely to return to their
surviving parent’s care when he is released from prison. Even if the relatives are horri-
fied at the actions of the father, and form a firm intention to rear the child to adulthood,
they find it very difficult to resist the pressure from father for the child’s return, and the
child is more likely to return to his care. Our clinical experience is that the children who
do return to their surviving parent’s care on the whole do badly. This has led us to be
very wary of recommending placement with paternal relatives or the perpetrator parent.
When that option seems the ‘least detrimental’ (Goldstein, Freud, & Solnit, 1980) then
it may help to recommend some statutory oversight either by wardship or a care order.
Our finding from this study, that there was an persistence of an apparently avoidant
attachment style, especially in young children living with relatives, is likely to have
significance for later mental health. Insecure attachment patterns have been described
as ways for a child to deal with particular abnormal parenting situations and although
they may represent the only adaptation possible, they are maladaptive in that they do
not prepare the child for successful adaptation in the world outside their families and
they are risk factors which increase the likelihood of later psychiatric disorders (Zeanah
& Emde, 1994).
We had an hypothesis that traumatically bereaved children, placed either with rela-
tives who are grieving (victim’s family) or who avoid making close attachments to the
child, because they anticipate returning the child to the (perpetrator) father on his
release from custody, would be less able to talk freely about the dead parent, would be
less likely receive the therapy they needed, would fail to make secure attachments
following the homicide and would be more prone to develop disorders of attachment.
Although our numbers are too small and the methodology insufficient to amount to a
proof, our findings are in keeping with our hypothesis. Furthermore these attachment-
related problems are compounded by the high number of moves of home and carer that
many of the children have.
The findings from this limited study do not enable us to make any definitive
19
Summary
A postal questionnaire was sent to the referrers of 95 children whose one parent had
killed the other, on whom data had been collected at assessment at least one year before.
The response rate was 74% with data on 61 children. The main findings were as follows:
1. This young group of children experienced many changes of home and carer. Half
the children were living at time of follow-up with unrelated carers. Of those living
with relatives, the fewest changes of carer occurred when they were initially placed
with relatives of the victim, rather than the perpetrator.
2. Most children had little contact with the surviving parent. They were more likely to
have maintained regular contact if they were in care.
3. Children who had treatment sessions as well as assessment interviews, appeared to
have fewer problems on follow-up than those whose contact was briefer. This
included the children who had witnessed the killing. About half the children who
had had a brief intervention were rated as being able to talk more openly to carers
about the death of their parent, especially if the carers were unrelated to them. The
small number of children who had been adopted seemed (to their referring social
workers) most able to talk freely about the dead parent. (The small numbers
involved make this finding suggestive only.)
4. Few children received further therapy in their locality, even though it had been
recommended in most cases.
5. At least one-third of the children had continuing overt emotional and/or behavioural
problems. Children who witnessed the killing were more likely to develop post-trau-
matic stress disorder, and were more likely to receive treatment, but were, it seems
20
from the referrer’s feedback, also more likely to have persistent emotional and
behavioural problems at follow-up.
6. Although numbers are too small for results to be regarded as definitive, children
placed with the perpetrator’s (father’s) family did worse on ratings of overt
emotional, behavioural and social identity problems than did those placed with the
victim’s (mother’s) family, who, in turn, did worse than children living in non-related
households, except with regard to social adjustment.
7. Even though most children had witnessed domestic violence before the killing, the
majority were not having problems at follow-up, in the referrer’s opinion.
8. Children originally assessed clinically by our team as having attachment-related
problems with their new carers, were more likely to be having similar problems at
follow-up.
9. Children living with relatives at the time of follow-up may have been having more
difficulty in forming meaningful relationships than children living with non-related
carers. (However, this question about the child produced the fewest responses from
the referrers and this limits the conclusions that can be drawn. Because the impli-
cations of this finding are likely to be of such importance, this bears further research
with larger numbers.)
10. Children living with relatives of the perpetrator are more likely to return to their
surviving parent’s care. The clinical evidence is that they behave well as children but
may show disturbance in adolescence. The numbers of those who returned to their
parents’ care in our sample were too small to either verify or disprove this.
We conclude that, while good practice requires that each child and family are assessed
on their merits, placement of these children with the victim’s family appears to lead to
fewer changes of placement than placing with the perpetrator’s family, and that the treat-
ment needs of these children who carry an enhanced risk of developing emotional and
behavioural problems, attachment disorders and problems of social identity, need to be
given more consideration.
Note
A version of this article appears in J. Harris-Henricks, D. Black, & T. Kaplan (2000). When
Father Kills Mother; Guiding Children Through Trauma and Grief (2nd ed.). London: Rout-
ledge.
References
Ainsworth, M.D.S., Blehar, M.C., Waters, E., & Wall, S. (1978). Patterns of attachment.
Hillsdale, NJ: Erlbaum.
Barrett, M.L., Berney, T.P., Bhate, S., Famuyiwa, O., Fundudis, T., Kolvin, I., & Tyrer, S.
(1991). Diagnosing childhood depression: Who should be interviewed – parent or child?
The Newcastle Child Depression Project. British Journal of Psychiatry, 159 (Suppl. 11),
22–27.
Black, D., & Kaplan, T. (1988). Father kills mother: Issues and problems encountered by a
child psychiatric team. British Journal of Psychiatry, 153, 624–630.
Goldstein, J., Freud, A., & Solnit, A. (1980). Beyond the best interests of the child. London:
Burnett Books.
Harris-Hendriks, J., Black, D., & Kaplan, T. (1993) (rev. 2nd ed., 2000). When father kills
mother: Guiding children through trauma and grief. London: Routledge.
21
Klein, R.G. (1991). Parent–child agreement in clinical assessment of anxiety and other
psychopathology: A review. Journal of Anxiety Disorders, 5, 187–198.
Pynoos, R., & Eth, S. (1986). Witness to violence; the child interview. Journal of the
American Academy of Child and Adolescent Psychiatry, 25, 306–319.
Raphael, B., Meldrum, L., & McFarlane, A.C. (1995). Does debriefing after psychological
trauma work? British Medical Journal, 310, 1479–1480.
Zeanah, C.H., & Emde, R.N. (1994). Attachment disorders in infancy and childhood. In M.
Rutter, E. Taylor, & L. Hersov (Eds.), Child and adolescent psychiatry: Modern
approaches (pp. 490–504). Oxford, UK: Blackwell.
22