A Review of the Incidence, Distribution and Characteristics of Homicides

A Review of the Incidence, Distribution and Characteristics of Homicides

in Avon and Wiltshire Mental Health Partnership NHS Trust:

Benchmarked against the National Confidential Inquiry into Suicide and Homicide by People with Mental Illness

April 2001 to September2007

Dr Bill Jerrom

Dr Tim Amos

January 2008

CONTENTS

AUTHORS DETAILS………………………………………………………………………………………………2

EXECUTIVE SUMMARY ……………………………………………………………………………………….…3

INTRODUCTION……………………………………………………………………………………………………4

RESULTS……………………………………………………………………………………………………………6

SUMMARY OF RESULTS ………………………………………………………………………………………11

DISCUSSION ……………………………………………………………………………………………………..13

REFERENCES …………………………………………………………………………………….…………..…14

APPENDIX ………………………………………………………………………………………………..………15

1 AUTHORS DETAILS

1.1 Dr Bill Jerrom is the Director of Psychology in AWPT. He is a Consultant Clinical Psychologist working in Adult mental Health with more than 30 years clinical experience across a range of services. He is Clinical Audit Lead for the Trust, and has led the Unexpected Death Audit process in the Trust since 1999, including undertaking 3 homicide investigations. He has received NPSA and Trust training in Root Cause Analysis. He has held a number of academic positions and has extensive research experience.

1.2 Dr Tim Amosis a Consultant Forensic Psychiatrist in West of England Forensic

Mental HealthService and aSenior Lecturer in Forensic Psychiatryat the University of Bristol. He has worked for a number of years on the National Confidential Inquiry into Suicide and Homicide by People with Mental Illness with Professor Lewis Appleby, now National Director for Mental Health. He has experience of most facets of inquiries, both as a clinician as well as sitting as a member of the inquiry panel.

We are grateful to Professor Glyn Lewis, Professor of Psychiatric Epidemiology,University of Bristol, for his advice on this Report.

2 EXECUTIVE SUMMARY

2.1Homicides committed by people with mental illness are a rare and shocking event. The occurrence of 4 homicides alleged to have been committed by service users in contact with AWPT Teams within a 6 month period in 2007 has caused great concern. This Review was requested by the SHA to benchmark the incidence of homicides in AWPT against the national rates, and to compare the characteristics of the cases with information from the National Confidential Inquiry into Suicide and Homicide by People with Mental Illness.

2.2The NCI judges that homicides committed by people with a mental illness are less preventable than suicides. In their most recent report, Avoidable Deaths (Appleby et al, 2006), they conclude that overall 13% of homicides committed by people in contact with mental health services were preventable.

2.3Very rare events, such as homicides committed by people with mental illness, are not distributed normally, either in timing or location. Examining the incidence of homicides committed, and alleged to have been committed, by AWPT service users since the Trust was established within current boundaries reveals that the AWPT rate is three quarters of the national average, even when the 4 cases occurring during 2007 are included. However the occurrence of 4 alleged homicides in a 6 month period is highly unusual, and is more than twice the average rate for the period. However this may not indicate any underlying change in incidence, and it should be noted that due to natural variation there will be annual variations above and below the range of the average rate.

2.4The 8 cases were distributed across PCTs and teams across the large demographically varied, geographical area covered by the Trust.Critical analysis of the events demonstrates no evidence of any clustering by geography or service.

2.5The characteristics of the 8 AWPT cases were broadly similar to the national sample reported inAvoidable Deaths. The Discussion reviews some conclusions to draw from the series of cases when compared to the national data and highlights 4 issues: use of the Enhanced ICPA in Schizophrenia, recent violence, Early Psychosis and risk, and dual diagnosis. The Trust has generated a combined Action Plan based on the Internal Investigations of the four 2007 cases, and it is recommended that these points are considered as part of the Action Plan.

3INTRODUCTION

3.1Homicides committed by people with a mental illness are a source of major public concern, and have shaped mental health policy in the UK (DoH, 1998; DoH 1999). Each homicide has traumaticeffects on the relatives of the victim and of the perpetrator, and also on the staff who have been involved in the care of the mentally ill person.

3.2International comparisons show that the rate of homicides committed by people with a mental illness is broadly similar across many countries, in marked contrast to the overall rates of homicide which vary widely. In countries with overall low rates of homicide, such as the UK, homicides by individuals with mental disorder stand out much more than in those countries with high rates of homicide such as the USA, where such homicides receive little attention (Coid, 1983).

3.3The National Confidential Inquiry into Suicide and Homicide by People with Mental Illness (NCI) has reviewed homicides in England and Wales since 1996. The Inquiry has issued a series of recommendations aimed at making Mental Health Services safer, and at the prevention of suicide and homicide committed by people with mental illness (Appleby et al, 1999; Appleby et al, 2001; Appleby et al, 2006).

3.4The number of homicide convictions annually in England and Wales increased steadily from 1973 until 2003 (Appleby et al, 2006). The Avon and Somerset, Gloucestershire and Wiltshire police force areas have homicide rates below the national average.

3.5There are a number of ways to define homicides committed by people with mental illness, the NCI defines 3 main types:

Perpetrators with mental illness at the time of the offence
Perpetrators in contact with mental health services(within 12 months of the offence or lifelong)
Perpetrators defined as mentally ill by the criminal justice system, in terms of verdict or disposal

3.6Avoidable Deaths, the latest report from the NCI (Appleby et al, 2006), reports the number of homicides committed by people in contact with mental health services over the previous 12 months between 1997 and 2003 in England and Wales. The average annual rate over the period was 52 homicides, with a range between 67 and 42 deaths annually. There was no consistent trend of increase or decrease in the number of homicides committed by this group.

3.7The extent to which homicide committed by people with mental illness can be prevented is controversial (Munro and Rumgay, 2000). Avoidable Deaths reports that clinicians identified 21% of homicides committed by people in contact with services within the past 12 months as preventable. However reviewing the cases in detail the NCI team judged that 27% of homicides committed by inpatients, and 13% of homicides committed by community cases, were preventable (Appleby et al, 2006). Given the rarity of homicides among inpatients this means that overall the NCI judge 14%of homicides as preventable. This equates to 7 preventable homicides per year in England and Wales.

3.8The National Patient Safety Agency has recently issued updated guidance on the investigation of homicides committed by people with mental illness. The NPSA has selected the criteria of mental health service contact within the last 12 months as defining when an Independent External Inquiry should be conducted following a homicide.

3.9As homicides committed by people with mental illness are an extremely rare event there are wide variations in the local incidence. The occurrence of homicides within individual mental health Trusts needs to be benchmarked against the national rates over a period of years to provide a balanced view of the incidence rate.

3.10Avon and Wiltshire Mental Health NHS Partnership Trust was established in April 2001 following the merger of Avon and Western Wiltshire Mental Health Trust with services in Swindon and South Wiltshire. The Trust has a population of 1.6 million, and provides secondary and tertiary mental health services across the former Avon area and the whole of Wiltshire. The geography of the Trust varies from dense urban populations to rural areas.

3.11There have been 4 homicides leading to convictions in AWP Trust since 2001;they were all subject to Internal Investigations and one led to an External Independent Inquiry (Downham et al 2006). One case was followed by the suicide of the perpetrator, the remaining 3 all lead to convictions. During 2007 there have been 4 alleged homicides, none of which have yet come to court. Each of the 2007 cases has received an Internal Investigation reported to the SHA.

3.12The occurrence of 4 homicides committed by patients receiving care in one mental health Trust in one year, even in a large Trust, is an unusual event which has caused concern at all levels in the Trust, and at the Regional Health Authority. This report has been requested by the Trust Board and the SHA to review the homicides against national benchmarks drawn from the NCI. The review will consider:

the incidence rate of homicides in the Trust
the distribution of homicides both geographically and in the service
the characteristics of the perpetrators and offences

4 RESULTS

Incidence of homicides

4.1 The current incidence of homicide (Appleby et al, 2006) committed by people with a

mental illnessin England and Wales is on average 52.4 cases per year; the population of England and Walesis 52.04 million, and the population of AWPT Trust area is 1.6 million. This means that the projectedrate of homicides in the AWPT area is 1.6 annually. During the period covered by Avoidable Deaths (1997 to 2003), the national incidence rate varied by 41%. The national benchmark ratefor homicides in AWPT therefore ranges from 1.95 to 1.3 homicides annually.

4.2 The number of homicides occurring in AWP annually in the 6.5 years from April 2001 until September 2007 is shownin Table 1. The projected rate for the Trust during this period is 10.4 homicides; there were4homicides and 4 alleged homicides during the period. The overall rate during the period wastherefore 23% below the national benchmark rate.

4.3Analysing individual years in Table 1 shows that 5 years were below the benchmark rate

with 2 years, 2004/05 and 2005/06 when no homicides occurred. There were two periods which were above the national benchmark rate, 2001/02, and the period from April to the end of September2007. Three homicides occurring in a 6 month period was more than twice the maximum national benchmark projected rate for theTrust.

TABLE 1: Homicides in AWPT: Annual rate and NCI Benchmark Rate

Number of Homicides /
Alleged Homicides* / Projected Rate / Difference
01/02 / 2 / 1.6 / +0.4
02/03 / 1 / 1.6 / -0.6
03/04 / 1 / 1.6 / -0.6
04/05 / 0 / 1.6 / -1.6
05/06 / 0 / 1.6 / -1.6
06/07 / 1* / 1.6 / -0.6
07 (6 months) / 3* / 0.8 / +2.2
Total / 8 / 10.4 / -2.4

4.4Trends in the occurrence of very rare events, such as homicides, are more accurately

analysed by reviewing incidence rates over a group of years. Table 2 presents 3 year moving average rates from April 2001 to September 2007. The projected rate is again drawn from the NCI national benchmark data. In 4 of the 5 time periods the homicide rate in the Trust was below the projected benchmark rate. In 2 periods, from April 2003 to March 2006, and April 2004 to March 2007, the Trust rate was substantially below the benchmark rate (1 vs 4.8 projected rate). During the most recent period, April 2005 to September 2007, the incidence equals the benchmark rate (adjusted to a 2.5 year period).

TABLE 2: Homicides in AWPT: 3-year rate and NCI benchmark rate

April 01 to
March 04 / April 02 to
March 05 / April 03 to March 06 / April 04 to
March 07 / April 05 to Sept 07
No. of Homicides / 4 / 2 / 1 / 1 / 4
Projected rate / 4.8 / 4.8 / 4.8 / 4.8 / 4
Difference / -0.8 / -2.8 / -3.8 / -3.8 / 0

Distribution of homicides

4.5Table 3 shows the distribution of homicides across the 6 PCTs in the AWPT area

(homicides committed in PCT’s which have subsequently merged are shown in the current PCT).Bath and North East Somerset was the only PCT which did not have a homicide occur during the period. Three PCT’s; Bristol, South Gloucestershire and Wiltshire, had 2 homicides occur during the period. Only 1 PCT, Wiltshire, had 2 homicides occur in one year.

TABLE 3: Homicides by PCT

April 01 -
Mar 02 / April 02 –
Mar 03 / April 03 – Mar 04 / April 04 – Mar 05 / Apr 05 – Mar 06 / April 06 – Mar 07 / Apr 07 – Sept 07 / Total by PCT
North Somerset / 1 / 1
Bristol / 1 / 1 / 2
South Glos / 1 / 1 / 2
BaNES / 0
Wilts / 2 / 2
Swindon / 1 / 1
Total by Year / 2 / 1 / 1 / 0 / 0 / 1 / 3 / 8

4.6The community teams working with the perpetrator at the time of the offence are shown

in Table 4 for each case. One team, the Inner City Mental Health Team in Bristol, had 2 homicides occurring during the period. This team covers the most socially deprived area of theTrust, which has the highest scores on indices of both general health needs and of mentalhealth needs.

TABLE 4: Community Teams working with Perpetrator and Specialism/SBU

Perpetrator / Team / Specialism/SBU
PMc / Inner City CMHT / Adult
GW / Marine Hill CMHT / Adult
RS / Kingswood CMHT / Adult
MN / Yate CMHT / Adult
1/07 / Swindon S and E CMHT / Adult
2/07 / Inner City CMHT / Adult
3/07 / Ridgeway Downs CMHT / Adult
4/07 / W Wilts SDAS Team / SDAS*

* SDAS = Specialist Drug and Alcohol Service

4.7 The 8 perpetrators were under the care of 8 different Consultant Psychiatrists in the Trust. Thedetails of the Consultants concerned are available in the Internal Investigation reports.

Characteristics of AWPT homicides

4.8 Avoidable Deaths describes a range of social, clinical and offencecharacteristics of

perpetratorswho had been in contact with Mental Health Services in the 12 months prior to the offence.These characteristics for the AWPT cases are shown in Table 5, with the comparison percentagefrom the national sample. Information on the AWPT cases was obtained from the Investigations and from post offence Forensic Reports. The results are shown by case in the Appendix. Initials of the perpetrators are used for the 4 cases which have had a legal outcome, the 4 cases which occurred in 2007, all of which are pending trial, are numbered. Given the very small sample ofcases great caution should be exercisedin drawing conclusions from this data.

Table 5: Characteristics of AWPT and NCI Perpetratorsand Offences

Characteristics / AWP No / AWP % / NCI %
Demographic features
Age of Perpetrator: (median, range) / 29 (21 to 46) / 31 (15 To 80)
Male Gender / 8 / 100 / 87
Ethnicity: proportion ethnic minority / 0 / 0 / 19
Not currently married / 7 / 87.5 / 71
Unemployed / 7 / 87.5 / 62
Long term sick / 6 / 75 / 12
Living alone / 3 / 37.5 / 37
Homeless / 0 / 0 / 6
Priority Groups and Care
Under enhanced CPA / 1 / 14 / 27
Missed last contact / 2 / 25 / 39
Non-compliance in the last month / 3 / 37.5 / 25
Discharged from hospital in the last 3 months / 0 / 0 / 18
Previous care elsewhere / 1 / 12.5 / 23
Receiving pharmacotherapy / 8 / 100 / 66
Previous admission to hospital / 4 / 50 / 68
Previous admission under MHA / 2 / 25 / 31
Clinical features
Primary diagnosis (lifetime):
Schizophrenia & other delusional disorders / 4 / 50 / 30
Affective disorder (bipolar disorder & depression) / 1 / 12.5 / 15
Alcohol dependence / 0 / 0 / 14
Drug dependence / 3 / 37.5 / 11
Personality disorder / 3 / 37.5 / 17
Other mental health diagnosis / 2 / 25 / -
Alcohol or drug misuse / 5 / 62.5 / 85
Onset more than 5 years earlier / 3 / 37.5 / 52
Offence variables
Age of victim: (median, range) / 60.5 (21 to 79) / 38 (0 to 86)
Male victim / 5 / 62.5 / 60
Female victim / 3 / 37.5 / 40
Victim was a stranger / 2 / 25 / 11
Victim was a family member / 2 / 25 / 17
Victim was a current or former spouse/partner / 0 / 0 / 17
Victim was an acquaintance / 4 / 50 / 28
Blunt instrument was used / 2 / 25 / 10
Sharp instrument was used / 5 / 62.5 / 47
Previous violence / AWP No / AWP % / NCI %
Previous recorded violence in case file / 5 / 62.5 / 46
Physical assault committed in past year / 2 / 25 / 20
Threats of homicide or serious violence in past year / 4 / 50 / 20

Demographic Features

4.9 The median age of the perpetrators was similar to the national sample. None of the AWPT perpetrators was from an ethnic minority group, in comparison to 19% nationally.

4.10 The AWPT cases were similar to the national sample on all other demographic features

except for the proportion registered as long term sick, which was 75% in AWPT compared to only 12% of the national sample. This finding will be explored further, as it may relate to differences in the definitions used.

Priority Groups and Care

4.11 One AWPT patient was subject to the Enhanced ICPA, a lower proportion than the national rate(27%).Four AWPT cases were on Standard ICPA; this is not reported for the national sample. A further 2 were not registered on the ICPA (PMc and MN), and 1 case was under the care of Specialist Drug Services alone,and therefore was not subject to ICPA.

4.12 All 8 AWPT cases were receiving some form of pharmacotherapy, in comparison to 66% in the national sample. Three AWPT cases were non-compliant with medication in the month prior to the offence, similar to the national rate for non-compliance.

4.13 Two AWPT cases (25%) missed their last appointment with services prior to the homicide, a similar proportion to the national rate (39%). The breakdown of last contact by AWPT cases, and by the national sample, is shown in Table 6. All 8 AWPT cases had a service contact in less than 13 weeks of the homicide occurring, in comparison to the national rate of 77%. Four of the last AWPT contacts were urgent and 4 routine, a similar proportion to the NCI sample.

Table 6: Timing of last contact with mental health services.

AWPT % / NCI %
Within 7 days / 3 37.5 / 29
1 – 4 weeks / 3 37.5 / 29
5 – 13 weeks / 2 25 / 19
14 weeks – 6 months / - / 9
7 months – 12 months / - / 14

4.14 Four (50%) AWPT cases had never been admitted to hospital, compared to 32% of the national sample. 2 AWPT cases (25%) had previous admissions under the Mental Health Act, a similar proportion to the NCIsample (32%).

4.15 None of the AWPT perpetrators were inpatients at the time of the homicide, compared to 6% nationally. None of the AWPT homicides occurred within 3 months of hospital discharge, in comparison to the national rate of 18%.

Clinical Features

4.16 Four (50%) AWPT cases had a diagnosis of Schizophrenia or Delusional Disorder, a slightly higher proportion to the NCI sample. Three AWPT cases had a formal diagnosis of Drug Dependence (including 2 with a Schizophrenia diagnosis), a higher proportion than in the national sample. However none of the AWPT cases had diagnoses ofAlcohol Dependence, andthe combined Alcohol/Drug Dependence rate was similar to the NCI level.

4.17 In 7 of the AWPT cases the responsible service was the general adult mental health service, rather than a specialist service, this is similar to the national picture. One AWPT case wasunder the care of Specialist Drug and Alcohol Services. In 3 of the AWPT sample (37.5%) theonset of mental illness had been more than 5 years earlier, compared to 52% of the NCI sample.

4.18 Five(62.5%) of the AWPT cases were known to be misusing alcohol and/or drugs; a slightly lower proportion to the NCI (85%). Three AWPT cases had a formal dual diagnosis (severe mental illness andalcohol/drug misuse or dependence), a similar proportion to the NCI (29%).