Inquest into the deaths of Ranjit Singh and Sandeep Thind

Wap

te

eet ey

A if

Beep erga ew Ws

CORONERS COURT NEW SOUTH WALES

Inquest: Inquests into the deaths of Ranjit Singh and Sandeep Thind Hearing dates: 12 May 2016 Date of findings: 10 June 2016 Place of findings: State Coroner’s Court, Glebe Findings of: Deputy State Coroner HCB Dillon Catchwords: CORONERS - Death in custody - homicide/ suicide -

remanded prisoner suicide following homicide of wife -- suicide risk assessment - placement in ‘one-out’ cell - CCTV equipment failure

File numbers: 2013/ 267697 & 2103/ 243783

Representation:

Findings:

Recommendations:

Set D Williamson (Advocate assisting coroner)

I find that Sandeep Kaur Thind died about 2:30pm on 10 August 2013 at McDonalds Family Restaurant, 1 Lasso Road, Gregory Hills, New South Wales as a result of blunt force trauma to the neck inflicted by her husband, Ranjit Singh.

I find that Ranjit Singh took his own life by way of hanging on Tuesday 03 September 2013 in cell 585 at the Metropolitan Remand and Reception Centre Silverwater, New South Wales.

I make the following recommendations to the Minister for Corrective Services and Minister for Health:

(i) Areview of the Mental Health Screening Unit be conducted to identify and, if reasonably practicable, remove hanging points and any other identifiable hazards to both staff and inmates.

(ii) An audit be carried out all CCTV equipment within the Mental Health Screening Unit area and a system of daily back-up of CCTV footage be installed.

IN THE CORONERS COURT

GLEBE

SECTION 81 CORONERS ACT 2009

REASONS FOR DECISION Introduction

These are joint inquests into the deaths of Sandeep Thind and Ranjit Singh, a married couple with a small child. Unfortunately, for reasons that are not clear, in the early hours of 10 August 2013, Mr Singh lost his temper and killed Ms Thind by strangling her.

About a month later, while remanded in custody awaiting trial on a charge of murdering Ms Thind, Mr Singh committed suicide.

Inquests are mandatory in homicide cases and also when a death occurs in custody: s 23 Coroners Act 2009. As the two deaths are closely related, the inquests are being dealt with together.

The coroner's functions and the nature of the inquest

An inquest is an independent judicial inquiry by coroner. When a person dies violently or unnaturally, the Coroners Act requires that it be reported to a coroner. If the death is a suspected homicide, and a person is not tried for the killing, an inquest will be held.

Both the bereaved family and the wider community have a profound interest in learning how the death came about and how it was investigated. This public inquest isa demonstration of the value placed on human lives - the lives of those who have died but also of those who mourn for the dead -- by our society and institutions.

Deaths in custody raise different considerations. When a person to whom the state owes a particular duty of care dies in the custody of the state, questions can and should be asked. Loss of liberty is the greatest punishment that our society can impose ona member of this community or visitors to it. Courts are only permitted to deprive a person of their liberty if they are proven to have committed serious criminal offences or ifthey are suspected of having committed serious offences and the safety and welfare of the community is reasonably considered to be in jeopardy if they remain free.

The corollary of this extraordinary state power to detain people in custody is the responsibility to care for and protect prisoners. If, for whatever reasons, the system intended to protect prisoners fails to do so, s 23 of the Coroners Act 2009 requires that an inquest be held.

At an inquest, a coroner is obliged to make findings, if possible, as to the identity of the person who has died, the date and place of death, the cause of death and the manner or circumstances of death. In this case, it is the manner and circumstances of these deaths that raise the difficult questions. lfit appears necessary or desirable to do so, a coroner may also make recommendations to relevant persons or organisations.

The death of Ms Thind

Sandeep Kaur Thind was a 30 year old woman. Ranjit Singh was a 38 year old man.

They were married in June or July 2007 at Patala Village, Punjab India, where Mr Singh and his family lived. Ms Thind had grown up in a nearby village. The marriage was arranged in traditional Indian fashion by Mr Singh’s uncle, the head of the family. By all all accounts both parties were happy with the marriage. On 6 December 2008, Ms Thind and Mr Singh arrived in Australia from India through Kingsford Smith Airport, Mascot on Student Visas. On 22 March 2011, Ms Thind gave birth in Australia to a child named Momi Sahib Singh, the first and only child of the couple.

On 29 June 2012, Ms Thind Jodged an application for Permanent Residency under the Regional Sponsored Migration Scheme with Mr Singh and Momi Sahib Singh listed as dependents on the application. The sponsor was Golden Scissors Hair Dressing Australia Pty Ltd. In October 2012, Ms Thind returned to India with Mori Sahib Singh. Due to work commitments of both Ms Thind and Mr Singh, Ms Thind left Momi in the care of Mr Singh’s family in India so that she could return to Australia and continue work without the need to arrange care for the child while she and Mr Singh worked.

In December 2012, Ms Thind made arrangements to be sponsored and work for Breeze Hair and Beauty Salon, Ingleburn. Temporary Sponsored Work Visas were applied for by Ms Thind and Mr Singh and were granted on 07 January 2013. The visas were valid through to 07 January 2017.

On 16 January 2013 the Department of Immigration received notification via Migration Agent, Mr Nishant Malik, that the application for permanent residency had been withdrawn. On 11 February 2013 Mr Singh was offered employment as a subcontractor, cleaning McDonald’s family restaurant at Gregory Hills. The manager of SBS Maintenance Service, Wisam Sadik, was aware that Ms Thind would help Mr Singh to clean the store. In March 2013, Ms Thind and Mr Singh moved into a unit at 18/4 Dotterel Place, Ingleburn. The unit was shared accommodation with Ms Thind and Mr Singh occupying one room; Matreet Kaur, Parmjeet Mr Singh and their child Harsiman Mr Singh occupying a second room and Chamkaur Mr Singh, the nephew of Ms Thind and Mr Singh, who occupied the final room.

On 20 March 2013, Mr Nishant Malik advised the Department of Immigration that he had ceased representing Ms Thind in the application for permanent residency. During April 2013, Wisam Sadik spoke with Ms Thind about cleaning the McDonalds Restaurant at Narellan. Ms Thind agreed to clean the Narellan store with Mr Singh. The normal routine involved Ms Thind and Mr Singh cleaning the Narellan store after closing at 11:00pm on week days and 12:00am on weekends. They would then attend the Gregory Hills store between 2:00am and 2:30am to clean.

On 12 April 2013, Golden Scissors Hair Dressing advised that they had withdrawn sponsorship of Ms Thind. On 4 June 2013, the Department of Immigration received information that Ms Thind had paid a salon owner $25,000 for sponsorship. It was unable to substantiate or verify which sponsor was alleged to have been paid. The Department was also told that Ms Thind was undertaking other work on a casual basis.

14,

On the morning of Friday 9 August Ms Thind attended her workplace at Breeze Hair and Beauty. Whilst at work Ms Thind asked her employer if she could obtain a price for flights to and from India. She explained that Mr Singh’s mother was ill and that he was planning on returning to India to visit his sick mother. Ms Thind was upset by this as she did not want to be left alone and be apart from Mr Singh. She did not want to travel with Mr Singh as she was saving money so Momi could return to Australia to live with them.

Ms Thind also spoke with Wisam Sadik who approved leave for Mr Singh. Throughout the remainder of the day, Ms Thind appeared in good spirits.

At about 6:00pm on Friday 9 August 2013, Ms Thind finished work at Breeze Hair and Beauty and was picked up by Mr Singh. They returned to their residence and later that evening went to bed. Mr Singh was unable to sleep due to working long hours and feeling confused. He woke Ms Thind and told her he wanted to go to work early and start cleaning. This made Ms Thind angry as she wanted to remain in bed and go in to work later that morning. Nevertheless, she accompanied Mr Singh to McDonalds Gregory Hills.

They arrived at the restaurant about 12:21am on Saturday 10 August 2013. At the time there were three employees working: the Manager Gabriella Zavaglia as well as Sean Byers and Kelsea Heinjus.

At 1:08am on Saturday 10 August 2013 Mr Singh was captured on CCTV footage entering the rear female staff toilet cubicle with Ms Thind following close behind. Once in the cubicle, Mr Singh locked the toilet indicator bolt door lock. He immediately grabbed Ms Thind around the neck with his right hand, squeezing her throat and neck.

Ms Thind said "Why are you doing this? What are you doing this for? Let me go, let me go”.

Mr Singh did not reply and continued to squeeze Ms Thind’s neck. He held Ms Thind’s throat for several minutes until he felt her stop breathing. He then placed her on the ground face up. Ms Thind started to shake her head and convulse. He placed his right foot across her neck and throat using his body weight to push down. He held his foot on her throat for about two or three minutes until she stopped moving. Then he removed his jacket and tied it around his own neck, attempting to strangle himself.

At about 1:55am Ms Zavaglia asked Mr Byers to look for the cleaners. He heard noises behind the locked toilet door and alerted Ms Zavaglia. She manipulated the lock to open the door. She found Ms Thind tying on her left side. Mr Singh was in a seated position with his legs over her body. He had a maroon jacket tied around his neck which he was twisting in a further attempt to strangle himself. Ms Zavaglia screamed for help. A number of customers assisted with CPR and contacted “000”.

Police arrived at 2:09am and continued CPR. Mr Singh was found to be breathing and was placed in the recovery position. An ambulance arrived shortly afterwards. Mr Singh was escorted from the restaurant and was placed under arrest. He was seen to by paramedics before being conveyed to Liverpool Hospital for treatment where he remained under police guard.

At 2:25am Ms Thind’s bady was removed from the toilet cubicle and conveyed to Campbelltown Hospital where she was pronounced life extinct at 3:31am.

21,

24,

Asa result of the investigation, Ranjit Singh was charged with the murder of Sandeep Kaur Thind. He was transferred to Silverwater Correctional Centre on and housed in Cell 585 within the Mental Health Screening Unit (MHSU) under the care of a psychiatrist.

Up to this time, it seems that to all outward appearances that Ms Thind and Mr Singh had had a loving relationship. There is no known history of violence by him towards her.

Nor is there any evidence of bad blood between Mr Singh and Ms Thind. Why he acted in such an extreme manner is therefore a mystery. There is nothing known to the investigators that might have predicted this terrible event. Clearly a person like Mr Singh does not act like that without a great deal of underlying pressure having built up.

The only thing known to have caused tension in the relationship was having to leave Momi Sahib Singh in India. It is reasonable to assume that the impermanency of their immigration status was another stressor on both Mr Singh and Ms Thind and their relationship. The long hours of work, the constant night shifts, the crowded accommodation and the sheer drudgery of cleaning work also probably added to whatever psychological load Mr Singh felt he was bearing. And there had been an argument earlier during the night.

For all of that, there is no clear explanation for Mr Singh’s behaviour that morning.

There is no evidence of pre-meditation. It appears that, for whatever reasons, Mr Singh simply “snapped”, completely losing his temper with Ms Thind.

The death of Mr Singh

At about 09:45pm on Tuesday 3 September 2013, Correctional Officers were conducting head checks in Pod 20 at the MHSU. Whilst attempting to communicate with Mr Singh, Correctional Officers looked through the small window on the cell door and saw Mr Singh’s feet in the doorway of the bathroom. He did not respond to any verbal direction so Correctional Officers opened the cell door. They found Mr Singh hanging by his neck from a handrail in the adjacent bathroom. A piece of material, later identified as part of a pillow slip, had been secured to the top disability hand railing in the bathroom and the other end around his neck. Mr Singh was in a seated position on the floor below the hanging point.

The material was cut by Correctional Officers before placing Mr Singh in the recovery position. His face was pale, his lips were blue and he was cold to touch. Justice Health Staff attended a short time later. No signs of life were found. NSW Ambulance and Police were notified and a crime scene was established. There was no indication that the death was suspicious in nature.

Issues The main question concerning the death of Ms Thind is why Mr Singh killed her.

The main issues surrounding the death of Ranjit Singh are to do with the risk

assessment, his classification for a one-out cell placement and why he was placed ina disabled cell.

Risk assessment

Directly after the murder of his wife, Mr Singh was found trying to harm himself with the intent of taking his own life.

In custody, however, during subsequent placement interviews, he denied suicidal ideation and thoughts of self-harm. He guaranteed his own safety and denied thoughts of harming others. It appears that his classification was changed to one-out cell placement on the basis that his assurances seemed plausible and truthful.

It is possible that Mr Singh was giving a truthful account at the time he was asked whether he was thinking of suicide and self-harm. He may, however, have been concealing his real thinking and intentions. It is also possible that he made the decision and formed the plan to take his own life relatively spontaneously despite having previously decided against such action and guaranteed his own safety.

Suicide risk assessment is an inherently very difficult task. In a report presented to the Coroners Court concerning suicide risk in psychiatric units, an independent psychiatric consultant, Dr Christopher Ryan described the problems of suicide risk assessment:

The first and most important thing to understand is that it is not possible to usefully categorise patients, who are admitted to a psychiatric unit, into those at relatively higher and those at relatively lower likelihood of future suicide.’ In reality all psychiatric inpatients are at a very greatly elevated likelihood of dying hy suicide. Although it is possible to use some features of such patients’ presentations ~ especially a past history of suicide attempt and a diagnosis of major depression — to categorise patients into those at a statistically higher risk of suicide than those who do not exhibit such features, the degree to which these features increase the suicide rate of this “high-risk” group is so small as to be of no utility in guiding management decisions.

The idea that it is not possible to usefully identify psychiatric inpatients at a particularly high risk of suicide strikes most people as counter-intuitive, at least at first blush. However, it is possible to confidently make this statement because it is based on a number of large studies that have examined possible risk factors in large numbers of people who have been psychiatric inpatients, some of whom have gone on to suicide?

Based on the evidence of these large studies my colleagues and I have spent a number of years campaigning against the sort of check-box risk assessment tools that Ms GC has found so confounding when reviewing her husband's medical records. We have argued simply that, since it is literally impossible to usefully

• Matthew M. Large and Christopher J. Ryan, ‘Suicide Risk Categorisation of Psychiatric Inpatients: What It Might Mean and Why It Is of No Use' (2014) 22 Australasian Psychiatry 390; Christopher J. Ryan and Matthew M. Large, ‘Suicide Risk Assessment: Where Are We Now?" (2013) 198 Medical Journal of Austratia

• Matthew Large, et al., ‘The Validity and Utility of Risk Assessment for Inpatient Suicide’ (2011) 19 Australasian Psychiatry 507.

assign psychiatric inpatients into groups who are at meaningfully (as opposed to statistically) high or low risk, it makes no sense whatever to compel staff te make check-box assertions about a patient's “suicide risk”.

All that can be meaningfully said about the likelihood of future suicide regarding a person who has been admitted to a psychiatric unit is that he or she is at a very high relative risk of future of suicide during that admission compared to the population risk of suicide, but, notwithstanding this, he or she is at very low absolute risk of suicide, since only around one in 700 admissions end in suicide.

Compounding this problem ts the fact that the forms that were devised by New South Wales Health for the mandatory? documentation ofa “risk assessment” are poorly designed. As a result of the poor design it is not at all clear what staff members are affirming when they check the various boxes associated with the various statements on the forms. For example ... tt is not at all clear what a “significant past history of risk’ actually means - arguably the phrase is without meaning, certainly it is ambiguous. Patients do not have a past history of “risk” per se; at best they may have a past history of one or more events that might impact on their future risk, though as [ have noted above, this is only true in a strictly statistical sense, and then only with very a limited number of events. (Incidentally the offending phrase on the form aiso provides no indication of what is meant by “significant” in this context).

31. Similar observations apply in respect of persons in the custody of Corrective Services.

Nevertheless, Mr Singh was known to have attempted to take his own life immediately have killing Ms Thind, That was an act that went beyond mere suicidal ideation. In the immediate context, it was a very significant matter. He almost certainly acted out of remorse and shame when he made the attempt on his own life. This was probably an additional risk factor te be taken into account because he had not been able, at the time he kifled Ms Thind, to extinguish his sense of shame and remorse by extinguishing his own life.

32. At my request, Justice Health provided a research paper on the question of whether persons who have committed domestic violence homicides are at greater risk of suicide than others in custody. (The full paper is annexed to this decision as Appendix 1.) A very helpful literature review was conducted by Associate Professor Kimberlie Dean, a consultant forensic psychiatrist and chair of Forensic Mental Health in the School of Psychiatry at the University of New South Wales and consultant with justice Health.

Professor Dean’s review reveals that “these events are rare and there has been limited research devoted to understanding their occurrence.”

3g. Walsh, et al., ‘Meta-Analysis of Suicide Rates among Psychiatric in-Patients' (2015) 131 Acta Psychiatrica Scandinavica 174

“NSW Department of Health, "Mental Health Clinical Documentation” 2010 available at htto://www0.health.nsw.gov.au/policies/pd/2010/pat/PD2010 018.pdf {accessed: 19 December 2015).

• NSW Department of Health, “Redesigned Mental Heaith Clinical Documentation: Notification of Availability” 2008 available at http://www0.health.nsw.gov.au/policies/ib/2008/pdf/182008 047.pdf (accessed: 19 December 2015).

35,

The phenomenon of homicide-suicide is, fortunately, rare but it is common enough to warrant close study. One of the obvious practical problems, however, in conducting such research is that the principal witnesses are no longer alive to be reviewed and assessed by relevant health professionals. Studies and research must, therefore, be conducted indirectly and, to some extent, speculatively.

The studies show, among other things, that not only is the killing of a domestic partner the most common form of homicide, itis also, unsurprisingly, the most prevalent form of homicide-suicide. Men are the most usual perpetrators of homicides followed by their own suicides. In contrast with this case, some studies show that previous physical abuse tends to be prevalent in homicide-suicide cases.6

Anumber of theories have been developed to explain homicide-suicide. Social stress and strain that results from a person being unable to achieve desired positive goals combined with occurrence of negative stimuli is thought by some to be a plausible explanation.’ This seems a common sense analysis of some cases and is apposite here.

Other theorists are more interested in the direction of the violence than its origins. One researcher surmised that a homicide-suicide results from the perpetrator feeling an inability to live with or without the victim of the homicide.? Homicide-suicide psychology may be viewed as including both outward and inward attribution of blame. The victim is blamed then the perpetrator blames himself for the crime. This fits the picture we have in this inquest. Various other theories have also been formulated but they are of less immediate relevance to our understanding of this case.?

Because of the scarcity of general research into homicide-suicide, and because of the general unpredictability of suicide, it is not possible to draw any significant conclusions as to whether persons in custody on domestic violence homicide charges ought be assessed as being at higher risk than other remand prisoners simply on that basis.

Indeed, as for other prisoners, it seems that in a case such as this risk assessment must be undertaken on an individual basis. In such an assessment Mr Singh’s attempted suicide following the killing, while not predictive of suicide, would be taken as an indicator that the prisoner is at higher risk of suicide than people in the general population.

Placement in a disabled cell

Areview of the occupancy and disabled cell locations was undertaken as part of the investigation into the death of Ranjit Mr Singh. There are three pods within the MHSU, Pod 19, Pod 20 and Pod 21. Only one cell, which has extra support railings, per pod is classified as a ‘disabled cell’. In Pod 19 itis cell 561, Pod 20 it is cell 585 and Pod 20 itis cell 590, Pod 19 contains a total of thirteen cells including the disability cell (561), Pod

° Marieke Liem “Homicide followed by suicide: a review” Aguressien arg Lidlent Behevior (154 May 2010 153-1olatpls7.

• Liem (2010) p.154.

§ Liem (2010) p.155.

9 See Liem (2010) p.155

42,

44,

20 contains a total of twelve cells including the disability cell (585) and Pod 21 contains a total of twelve cells including the disability cell (590).

A review of the housing history of the MHSU and specifically cell 585 indicates that all cells were occupied at the time of Mr Singh's transfer into the MHSU. On the housing list it indicated that Mr Singh was transferred on 30 August 2013 at 22:43 hours with an end date/time of 03 September 2013 at 21:45 hours, after his death.

According to this report it shows a second inmate was in cell 585 at the time Mr Singh was housed in there. Further investigation into this indicates an inmate, Mr Lord, was moved out of cell 585 to another block on 5 February 2013 at 16:20 hours. This discrepancy appears to be human error in transferring on the computer system. At the time of his death Mr Singh was housed one out in Cell 585.

The reasons that Mr Singh were placed in the disability cell were that he had been classified as suitable for “normal cell placement” (ie, one-out) and that the disability cell was the only available cell. He was not placed on suicide watch and therefore was not placed in a “protection” cell. Such cells are designed to have eliminate all hanging points.

It is well-known, and indeed was stated by Associate Professor Dean, that placement ina one-out cell is a suicide risk factor. A prisoner alone in a cell is more capable of taking undetected steps to end his or her own life than a person with a cellmate. But placement in a two-out situation is not always protective and can bring small but real risk of even greater harm as well. On occasions, homicide-suicides have occurred in prison cells. | have dealt with a case in which a prisoner who was assessed as a suicide risk killed his own cellmate before committing suicide. That event was impossible to predict but demonstrates that there are no perfect solutions to suicide risk.

Once the assessment was made that Mr Singh was not at immediate risk, the cell placement was reasonable in the circumstances.

CCTV footage

During the course of the investigation it was determined that there was no CCTV camera installed in cell 585 but it was determined that there are CCTV cameras located throughout Pod 20. Upon request for recordings of the CCTV footage from Pod 20 for 3 September 2013, the Systems Security Manager reported that the computer hard disk drives, which recorded all the video footage, had failed, and as such, there was no CCTV footage available.

It is unclear why the hard disk drives failed but it is a matter of considerable concern not only that they failed but that the failure went undetected, it appears, until the request for footage was requested. That poses a significant safety and security issue for inmates and staff and steps should be taken to rectify this problem.

47,

49,

Conclusions

There are few more severe forms of interpersonal violence than homicide-murder cases.

They shock and horrify the immediate family and friends of the people who have died, who frequently suffer enormous and irreparable emotional trauma as a result. They also send aripple effect of alarm and incomprehension throughout the communities in which they take place.

It is difficult to imagine the feelings of members of the Thind and Singh families. { hope, however, that they will accept the sincere and respectful condolences of the coronial team and all the staff of the Coroners Court. Our thoughts and sympathy are especially for Momi who will now grow up in the bosom of a loving family but without parents.

Findings s 81 Coroners Act

I find that Sandeep Kaur Thind died about 2:30pm on 10 August 2013 at McDonalds Family Restaurant, 1 Lasso Road, Gregory Hills, New South Wales as a result of blunt force trauma to the neck inflicted by her husband, Ranjit Singh.

I find that Ranjit Singh took his own life by way of hanging on Tuesday 03 September 2013 in cell 585 at the Metropolitan Remand and Reception Centre Silverwater, New South Wales.

Recommendations s 82 Coroners Act

I make the following recommendations to the Minister for Corrective Services and Minister for Health:

00] Areview of the Mental Health Screening Unit be conducted to identify and, if reasonably practicable, remove hanging points and any other identifiable hazards to both staff and inmates.

(ii) An audit be carried out all CCTV equipment within the Mental Health Screening Unit area and a system of daily back-up of CCTV footage be installed.

Jr the

Magistrate Hugh Dillon Deputy State Coroner

POPE NRITK "1"

Research statement

lL brrep diterucure Pevicw ta response Co qglestiais GPEstilg

from (he Coronial faquest fate the death af Mr Ranjit SINGH

Prepared by: A/Prof Kimberlie Dean”?

‘Justice Health and Forensic Mental Health Network, NSW

School of Psychiatry, University of New South Wales

May 2016

This review is focused on providing a brief overview statement on the peer-reviewed and other literature relevant to a number of specific issues raised by the Coroner in the inquest into the death of Mr Ranjit SINGH.

The epidemiclogy of homicide-suicide

A small proportion of homicide perpetrators commit suicide following the homicide, a phenomenon labelled ‘homicide-suicide’ or H-S. These events are rare and there has been limited research devoted to understanding their occurrence. When they do occur the perpetrators are most commonly male and their victims are most commonly their female partners or ex-partners, sometimes also their children. There is a literature focused on the extent to which H-S cases resemble homicide-only cases or suicide-only cases. In a published systematic and quantitative review of 49 studies, the authors concluded that H-S appears epidemiologically closer to hamicide than ta suicide in regions with high homicide rates such as the US, while the opposite may be true for regions without high homicide rates (Large, Smith, & Nielssen, 2009}. H-S cases in this systematic review were defined as those where the homicide perpetrator committed suicide prior to conviction while some individual studies require the suicide to have occurred in close temporal proximity to the hemicide; most suicides in H-S cases are known to occur within 24 hours of the homicide. In a recent Australian study of H-S, homicide-only and suicide-only cases, the characteristics of the H-S cases were found to more closely resemble the suicide-only cases than the homicide-only cases (McPhedran et al., 2015). For example, both the H-S and suicide-only cases were more likely to share a history of mental itl health and to have had alcohol use problems. All three groups were likely to have a history of past suicide attempts.

Other acts of co-occurring aggression to self and others

Homicide-suicide or H-S events may be usefully considered as at the extreme or lethal end of a continuum of aggression which is directed both towards others and towards the self {Hillbrand, 2001), a continuum which would include attempted H-S events (attempted homicide followed by suicide and homicide followed by attempted suicide). The lethal intent to commit an act of H-S may be present in such cases but factors potentially outside the individual’s control may have intervened to limit the seriousness of at least one of the outcomes. The extent to which such attempted H-5 events occur and to what extent they differ in character from completed H-S events is not well understood. One Dutch study examined a sample of men who had attempted or completed an act of intimate partner homicide and found that those who also demonstrated suicidal behaviour were more likely to be: unemployed, motivated by a fear of abandonment rather than narcissistic injury, depressed, and to have previously expressed suicidal threats (Liem & Roberts, 2009), The risk of subsequent completed suicide following an attempted H-S is also not well studied.

One study of mortality among 176 homicide offenders in Sweden reported 13% died by suicide, a rate which contributed significantly to the three-fold increase in mortality for this group (Lindqvist, Leifman, & Eriksson, 2007).

Given the overlap in risk factors for aggression aimed at the self and at others, it is perhaps not surprising that such behaviour can co-occur (O’Donnell, House, & Waterman, 2015).

Although approaches to the assessment and management of suicide and violence are often considered separately, they share an important overall goal of reducing the level of identified risk factors for aggression (Hillbrand, 2001). Some have extended this argument to suggest that those individuals presenting to mental health clinicians with a history of violence should be considered a potential risk of self-harm and vice versa, and it is to some extent on this basis that structure professional risk assessment tools such as the START (Short-Term Assessment of Risk and Treatability), which considered multiple potential risk outcomes, have been developed (Webster, Nicholls, Martin, Desmarais, & Brink, 2006). In practice however, particularly in settings over which the clinician has limited control such as prison, this approach can present difficulties. The decision to place an individual with risks of aggression in a single cell represents an example of such a difficulty. It is clear that further research is needed to guide policy and practice in this area.

Role of single cell occupation in terms of risk of in-prison suicide

Single cell occupation is an established risk factor for completed suicide in prison settings.

In a published systematic review of 34 prison suicide studies, the following risk factors were highlighted as the most important factors consistently found to be assaciated with suicide — occupation of a single cell, recent suicidal ideation, a history of attempted suicide, and having a psychiatric diagnosis or history of alcohol use problems (Fazel, Cartwright, NormanNott, & Hawton, 2008). Occupation of a single cell was associated with odds of suicide 9.1 times higher (95% confidence interval 6.1-13.5) than for prisoners not occupying a single cell. The authors of the review comment, however, on the likelihood that single cell occupation as a risk factor for prison suicide is influenced by the presence of mental illness since those with active mental illness are more likely to be placed in a single cell when disturbed behaviour and/or aggression directed towards others is demonstrated. In such circumstances, mental health clinicians and prison officers face a difficult situation where there is a need to balance risk of suicide with other risks including risk of harm to others. It is also clear that while single cell occupation is associated with suicide in prison, such an event is uncommon and the vast majority of individuals placed in single cells do not attempt or complete suicide. This highlights the difficulty in applying findings derived from study groups, particularly when defined by an uncommon event or characteristic, to an individual case.

Role of assessments of suicidal ideation/assurances of safety in terms of suicide risk

The literature regarding the association between expressing thoughts/ideas/plans of selfharm or suicide and the occurrence of subsequent completed suicide is conflicting.

Similarly, statements ‘guaranteeing safety’ or denying suicidal ideation in the context of a recent suicide attempt and the extent to which they can justifiably reassure mental health clinicians about suicide risk is unclear. It is true that the ability of clinicians to make accurate predictions about the risk of an individual attempting suicide at some point in the

future is very difficult indeed and is often further complicated by limitations placed on clinicians to alter the status of important risk factors in individual cases. In the systematic review of prison studies mentioned earlier, recent suicidal ideation was identified as an important risk factor for prison suicide, with odds of suicide being 15.2 times greater (95% confidence interval 8.5-27.2) among those reporting thoughts of suicide compared to those without such ideas (Fazel et al., 2008}. The risk of completed suicide among individuals with a history of suicidal behaviour but recent denial of suicidal ideation in the prison setting has not been established. In the context of mental health settings, for those patients with a history of suicidal behaviour who go on to die by suicide, denial of suicidal ideation prior to the completed suicide is not uncommonly documented (Busch, Fawcett, & Jacobs, 2003).

Overall, mental health clinicians working with those who present a risk of aggression, to themselves and/or others, must assess the significance of denial of suicidal ideation in the context of the circumstances and characteristics of the individual.

Busch, K. A., Fawcett, J., & Jacobs, D. G. (2003). Clinical correlates of inpatient suicide. Journal of Clinical Psychiatry.

Fazel, S., Cartwright, J., Norman-Nott, A., & Hawton, K. (2008). Suicide in prisoners: a systematic review of risk factors. J Clin Psychiatry, 69(11), 1721-1731.

Hillbrand, M. (2001), Homicide—-suicide and other forms of co-occurring aggression against self and against others. Professional Psychology: Research and Practice, 32(6), 626.

Large, M., Smith, G., & Nielssen, O. (2009). The epidemiology of homicide followed by suicide: a systematic and quantitative review. Suicide and life-threatening behavior, 39{3), 294-306.

Liem, M., & Roberts, D. W. (2009). intimate partner homicide by presence or absence of a selfdestructive act. Homicide Studies, 13({4), 339-354.

Lindqvist, P., Leifman, 4., & Eriksson, A. (2007). Mortality among homicide offenders: a retrospective population-based long-term follow-up. Criminal Behaviour and Menial Health, 17(2), 107112. doi:10.1002/cbm.643

McPhedran, S., Eriksson, L., Mazerolle, P., De Leo, D., Johnson, H., & Wortley, R. (2015).

Characteristics of Homicide-Suicide in Australia A Comparison With Homicide-Only and Suicide-Only Cases. Journal of interpersonal violence, 0886260515619172.

O'Donnell, 0., House, A., & Waterman, M. (2015). The co-occurrence of aggression and self-harm: Systematic literature review. Journal of affective disorders, 175, 325-350.

Webster, C. D., Nicholls, T. L., Martin, M. L., Desmarais, 5. L., & Srink, J. (2006). Short Term Assessment of Risk and Treatability (START): the case for a new structured professional judgment scheme. Behavioral sciences & the law, 24(6), 747-766.