Canterbury Law Review
Two competing interests exist in mental health legislation - the protection of patient rights and the protection of the community's right to be safe from harm. The difficulty is finding the balance between these rights to ensure that one does not dominate at the expense of the other. Attempts are made to do this in New Zealand via the Mental Health (Compulsory Assessment and Treatment) Act 1992 which provides the means by which mentally disordered persons can be subjected to non-consensual assessment and treatment. This Act, like most mental health law, is based upon the assumption that certain mental disorders are linked with violent behaviour. These public perceptions and fears have lead politicians towards the legislation which allows mentally disordered people to be civilly detained to protect both society and themselves from potential harm. Evidence that this assumption is present in New Zealand legislation can be found in the parliamentary intent behind the Mental Health (CAT) Act 1992. It was recognised that there was increased awareness in the community about 'people with serious mental disorders, and often the dividing line between the 'mad' and 'bad' is extremely fine... the [Act] is really more for those who pose a serious danger to the health and safety of others'. Further support for this position is found in the Law Commission's report on Community Safety where it considers relevant provisions in the Mental Health (CAT) Act 1992 and Criminal Justice Act 1985 for the purpose of 'protecting members of the public from substantial risk of harm from individuals whose release in the community would pose that risk'. The belief that there is a relationship between mental illness and violence takes the form of dangerousness in the legal system. Dangerousness is not a medical, scientific, psychological or psychiatric concept. It is purely a legal notion. In the New Zealand context, dangerousness is included in the criteria of the legal definition of 'mental disorder'. This requires the courts to assess the level of risk an individual presents and to determine based upon that rating of risk whether or not the person should be detained for treatment in a psychiatric hospital. These risk assessments are the most controversial aspect of mental health legislation. Our legal system is based upon the mandate that incarceration of individuals occurs only after a crime has been committed, not before. However, in the mental health context, people assessed as high risk can be detained on the basis that they may commit a crime in the future.
The fundamental concepts underlying the incorporation of dangerousness into our mental health law are also controversial. There is an assumption of increased violence among the mentally disordered population from which the public need to be protected. Although research suggests that there is a modest relationship between mental illness and violence, there are a number of other factors which influence whether or not a mentally disordered person will behave violently; the person's specific diagnosis, the type of symptoms the person is suffering, whether or not there is a contemporaneous substance abuse problem, and whether the person has a history of violent behaviour. The second assumption is that mental health professionals can accurately predict the mentally disordered who will behave violently. Accordingly, these people will be placed in psychiatric hospitals to protect both themselves and the public from harm. Research indicates that mental health professionals are able to predict those who will behave violently in the future, but at a cost. For each of the people predicted to behave violently and who subsequently do so, there are twice as many who are predicted to behave violently who do not behave in such a way. This implies that a number of people detained because of the danger that they present would not in fact behave violently if they were still living in the community. The deprivation of liberty based upon incorrect risk assessments raises a number of ethical questions. This paper attempts to discuss these concerns in the context of the Mental Health (CAT) Act 1992 and reviews the literature which leads to these conclusions. While accepting there some individuals who are in genuine need of treatment, it is proposed that the dangerousness criteria be removed from the Act and replaced with a more clinical measure which takes into account the best interests of the patient.
The Mental Health (Compulsory Assessment and Treatment) Act 1992 introduced significant changes in New Zealand mental health law. The Act clarifies the conditions under which a person can be subject to compulsory psychiatric assessment and treatment. It defines the rights of patients and aims to protect those rights. The provisions of the Act aim to ensure that both vulnerable individuals and the public are protected from harm. The procedure by which a person may become subject to the Act is clearly set out. Any person over the age of 18 who believes a person is suffering from a mental disorder can fill out an application asking for an assessment of the person. The application must state the relationship of the applicant with the person and why they believe the person is suffering from a mental disorder. It must also be accompanied by a certificate from a medical practitioner stating that he or she has examined the person and considers there are reasonable grounds for believing that the person may be suffering from a mental disorder. On receiving this application, the Director of Area Mental Health Services nominates a medical practitioner to conduct an assessment of the person. During this initial assessment examination, the medical practitioner must determine whether there are reasonable grounds for believing that the person, now considered a patient under the Act, is mentally disordered and whether or not they should undergo further assessment and treatment. If the medical practitioner believes that there are reasonable grounds for further assessment and treatment, the patient is required to undergo a further period of assessment and treatment. This initial assessment can continue for 5 days, followed by two further periods of 14 days before a judge makes an order.
Following the second period of assessment, if the clinician believes that the patient is fit to be released into the community, then they must be released from compulsory status without delay. However, if the clinician believes that the patient is not fit to be released from care, they must apply to the court for a compulsory treatment order. This type of order requires the patient to undergo treatment for his or her mental disorder either in the community or as an inpatient. Under a community treatment order, the patient is required to attend the patient's place of residence or some other place specified in the order for treatment. Alternatively, if an inpatient order is made, the patient must remain at the hospital specified for the duration of the treatment. If a compulsory order is made, the court will make a community order unless it believes the patient cannot be adequately treated as an outpatient and in that case the court will make an inpatient order.
Where practicable, the application for a compulsory treatment order will be heard and determined by a Family Court Judge, otherwise the application can be heard by a District Court Judge. The judge will arrange to examine the patient within 14 days of the application being filed. During this examination, the judge will identify himself or herself to the patient, explain the purpose of the visit, discuss with the patient their situation, the proposed course of assessment and treatment and the patient's opinion on these matters. The judge will also consult with the responsible clinician and with at least one other health professional involved with the patient, and other persons regarding the patient's condition such as family members, caregivers or nursing staff. If the judge considers the patient is fit to be released from compulsory status then the judge will order the patient to be released immediately. However, if the judge is not satisfied that the patient should be released then the judge will hold a court hearing to determine the clinician's application for a compulsory treatment order. Following the hearing, if the judge believes a compulsory treatment order is appropriate, he or she will make the necessary community or inpatient order. The compulsory order is for 6 months duration and when it expires, if the responsible clinician believes the patient still needs to be treated he or she can apply to the court for an extension of the order. Following this, the order can be extended for one more period of 6 months and at the end of that extension, the court may grant a further extension for an indefinite period. However, there are provisions ensuring a clinical review of a patient's situation after the first 3 months and the rights of patients to access review procedures.
There are three types of patients under the Act. The first is a patient who is being compulsorily treated under either a community treatment order or inpatient order. The procedure outlined above applies to this category of patient. The second group of patients subject to the Act are called special patients. These are patients who have entered the mental health system via the criminal justice system. It is also a disposition option for judges for defendants found not guilty by reason of insanity or unfit to stand trial. The third type of patients was a new creation under the 1992 Act. If a patient subject to an inpatient order presents special difficulties because of the danger they pose to others they can be made a restricted patient. Under a restricted patient order, more restrictions are placed over a patient's leave. This new status is further indication of public perception that mentally disordered people are dangerous. Evidence for this is found in the judgement of Re Tahere  where it was stated that the category of restricted patient was 'a reflection of public concern as to the release of mentally disordered persons who are dangerous to others'.
The definition of mental disorder is central to our mental health law as a person must be found to be mentally disordered to become subject to the Mental Health (CAT) Act 1992. Under the earlier 1969 legislation, a person was considered mentally disordered if they were:
Suffering from a psychiatric or other disorder, whether continuous or episodic, that substantially impair(ed) mental health, so that the person belong (ed) to one of the following classes, namely:
a) Mentally ill - that is, requiring care and treatment for mental illness
b) Mentally infirm - that is, requiring care and treatment by reason of mental infirmity arising from age or deterioration of or injury to the brain:
c) Mentally subnormal - that is, suffering from subnormality of intelligence as a result of arrested or incomplete development of the mind.
However, the 1992 Act introduces the concept of danger into the legislation by defining mental disorder as:
An abnormal state of mind (whether of a continuous or an intermittent nature), characterised by delusions, or by disorders of mood or perception or volition or cognition, of such a degree that it -
(a) poses a serious danger to the health or safety of that person or of others; or
(b) Seriously diminishes the capacity of that person to take care of himself or herself.
The change in the definition of mental disorder reflects a trend away from the paternalistic approach towards one of patient autonomy. This is part of a global trend towards protecting the public from mentally disordered people and also protecting mentally disordered people from self-harm.
The test for mental disorder under the 1992 Act has two limbs. The person must be suffering from an abnormal state of mind; and they must either pose a serious danger to the health and safety of that person or others, or have a seriously diminished capacity for self-care. However, this essay is concerned with the 'serious danger to the health and safety of others' part of the definition.
The test of serious danger is complicated and many different interpretations have been suggested. But as recognised In the matter of T  the correct test to be applied is of enormous consequence to both clinicians and patients: to clinicians because they need to know what the criteria are to commit patients or release them; and to patients because if the threshold for serious danger is too low they may be detained unreasonably, and if it is too high then the people who need treatment may not receive it.
It is accepted that there must be a nexus or connection between the abnormal state of mind and the serious danger. Serious danger was originally defined by Ellis DCJIn the matter ofJK originally defined serious danger in relation to others, as 'danger... involv(ing) a demonstrable risk of no less than serious physical violence to the person of another or others...'.  This approach put emphasis on physical danger and was subsequently criticised for being too restrictive. It had the potential effect of removing from the control of the Act a group of people who should be subject to it; for example, people suffering from Erotomania or experiencing delusions of persecution and consequently engaging in fear inducing behaviours such as stalking. The issue was revisited by Judge Doogue in Re RWD who, in recognising the problems of limiting danger to physical danger, extended it to include psychological and emotional health:
No universal application of the term is possible. At one end of the spectrum will be the obvious case where the patient will, unless restrained, inflict physical injury on himself or others - a clear case of serious danger to health and safety. At the other will be equally obvious case where the delusions or disorders are benign or of mere nuisance value. For the cases in between, a judgment will have to made, but in making that judgement.. the statute requires the court must have regard not just to the possibility of physical harm, but also to the impact of he patient's behaviour on the psychological and emotional well-being of himself and others.
Judge Doogue concluded that the phrase 'serious danger to the health of safety':
must be applied in a commonsense fashion, interpreted to achieve the objects of the statute which include care for the patient and protection of the community, but with due regard for the individual freedoms contained in the bill of rights.
This decision has been affirmed in a number of cases, including a case by Ellis DCJ in which he accepted that his earlier decision Re JKwas decided on its particular facts (Re TAV).
In the matter ofJK, Ellis DCJ reviewed the issue of serious danger and held that 'serious' enhanced the use of danger by introducing components of 'imminence and likely consequences which will have, or could have, important consequences' . Although these comments were deemed obiter in Re T, it had been held by Judge Boshier in Re O that the serious
danger means a danger that is imminent or demonstrable. In Re T the tribunal held that a danger is imminent if it is likely to happen at any moment in the reasonably foreseeable future. Recently Judge McElrea in Re GM observed that imminent danger should not be limited to danger expected to happen in the immediate future:
If serious danger was a real and substantial risk to the lives of other people, or in the sense that they might be killed or suffer serious bodily harm, then the fact that it might not happen tomorrow or next week, or even the next week is irrelevant.
McElrea DCJ continued by stating that the imminent danger must be demonstrable, which means on the balance of probabilities serious danger has been established.
In the criminal division a court must be satisfied beyond reasonable doubt that a defendant is guilty. This high standard of proof reflects the seriousness of the potential outcome of depriving a person of freedom. However, mental health legislation is in the civil domain and accordingly the burden of proof is on the balance of probabilities. In both cases an individual's freedom is at stake but the balance of probabilities used in mental health law is a lower standard than that required in criminal cases The judiciary are well aware of this incongruity:
Clearly by the standard required for satisfaction must be on the balance of probabilities by since there is a question of detention... special regard will be given to that restriction of the liberty of the person and the general desirability that all persons should be free.
It would be inappropriate to require the same standard as in criminal cases as it would undermine the purpose of the Act to protect and care for the mentally ill. However, it is a well-recognised principle when considering evidence that may have serious consequences, the balance of probabilities must be set at a high standard; Pallin v Department of Social Welfare. This was recognised by Ellis DCJ In the matter ofJ K who stated; 'that the evidence...must be measured objectively on the balance of probabilities, but requiring a clear tipping of the scales having regard to the possible loss of liberty involved' .
The Mental Health (CAT) Act 1992 is legislation attempting to redefine the situations in which a person may become subject to a compulsory psychiatric treatment and assessment and it purports to define the rights of such people and provide better protection of those rights. However, the Act also provides the mechanisms by which a person's rights can be severely corroded by the state. For this reason it is important to ensure that decisions are not arbitrary and are made by people with relevant expertise. In particular, a determination of dangerousness is often required for a person to be found mentally disordered under the Act. A judge makes this decision, although these decisions are typically based upon the evidence of psychiatrists and other mental health professionals. However, it is not a requirement by law that the opinion of mental health professionals is sought. This was recognised by Greig J who said in Re M:
Though it is plain that the decision and responsibility for it rests with the judge.. he cannot have the knowledge, training or experience to say whether a person is suffering from a psychiatric disorder or whether that substantially impairs his mental health. In those areas the judge must rely on those who have the knowledge and experience...
Judges must satisfy themselves on the whole of the evidence as to the state of mind of the person in question, taking into account both the public interest and interests of the patient; Re M.
There are guidelines that suggest which factors a judge should consider when determining the dangerousness of an individual under application for a compulsory treatment order or for an extension of an existing order. Brookbanks and Simpson proposed an assessment of dangerousness should take into account the following factors:
1. the nature and magnitude of the harm
2. its imminence
3. its frequency
4. situational circumstances and conditions that effect the likelihood of harm occurring
5. balancing the need between the alleged harm on the one hand and the nature of society's intervention on the other
This approach has been commended by the Law Society for taking into account that some risk factors for danger will change in situations . It reflects the current psychiatric view that dangerousness is not so much caused by the patient himself or herself, but results from the interplay of a number of factors such as specific diagnosis, symptomology and drug or alcohol use. These guidelines for risk assessment by Bookbanks and Simpson are considerably different from the guidelines prepared for mental health professionals. They provide a more detailed analysis of the medical and psychological factors involved in assessing risk. For example, the guidelines for clinical risk assessment and management in mental health provided by the Ministry of Health require a detailed analysis of the following: clinical history; a history of risk behaviour; victim characteristics; early warning signs; interventions; and specific abnormalities of mental state associated with risk. The differences between the factors that judges and mental health professionals take into account are indicative that the legal and psychiatric approaches to dangerousness are quite separate. The legal framework is moralistic and pragmatic and dangerousness is attributed to the person.
The clinical perspective is scientific and empirical which see dangerousness as an outcome of various factors like treatment, specific mental disorder, and social and personal factors.
Dangerousness is a legal concept rather than a psychiatric idea and highlights the conflict between the legal system and psychiatry: 'law presumes men to be rational, hedonistic, responsible; clinical thought provides for the probability that they may often be irrational, destructive and most irresponsible'. The two schools of thought appear incompatible. However, mental health legislation forces the two to co-exist in a situation where the freedom of an individual is at stake. This is in part due to the public perception of a relationship between mental illness and violence which has been steadily increasing in recent years. This has been pinpointed as one of the driving forces behind the inclusion of 'dangerousness' in our mental health legislation. Many critics argue this increased public awareness is in part due to media portrayal of the mentally ill and an increased focus on violent acts committed by the mentally disordered. For example, an American study has found that a character with mental illness was shown in 17% prime time American television programs with 73% of those programs showing the character as violent, as opposed to 40% of the 'normal' characters. 23% of these mentally ill characters were homicidal compared to 10% of the 'normal characters'. Subsequent studies show that there has been no decrease in these figures since the original 1981 study. Another study undertook an analysis of all stories from the United Press International database which dealt with people who had formerly been committed to a psychiatric institution. In over 86% of the stories the focus of the article was violent crime and most commonly murder.
Regardless of why or how the dangerousness concept entered our legal system, a more salient concern is whether it is a valid criterion for deciding that a person is mentally disordered. Its validity can be examined by challenging the underlying assumptions of the concept. The use of dangerousness in mental health legislation is based upon two assumptions. The first is that mentally disordered people present more of a danger to society than non-mentally disordered people. This can be tested by investigating the relationship between mental disorders and violent or fear-inducing behaviour. The second assumption is that certain people can predict accurately those who are likely to engage in violent or fear inducing behaviour. This can also be investigated by looking at studies assessing the ability of mental health professionals to predict violent or fear-inducing behaviour in mentally disordered populations.
In the 1970s the perception that there was a relationship between mental illness and violent or fear-inducing behaviour was questioned by a number of researchers. After reviewing the available literature it was concluded that there was no causal relationship between mental disorders and violent behaviour. In fact, it was thought that the mentally disordered population presented less of a risk than the non-disordered population as a whole. However, in recent years, research has presented new evidence of an association, albeit small, between mental illness and violence. After reviewing current literature on the topic, Mulvey concluded that 'contrary to findings of earlier research, an association does appear to exist between mental illness and the likelihood of being involved in violent incidents'. It has also lead to the retraction by Monahan, a leading expert in the area, of earlier statements that supported findings of no relationship and who now thinks that to continue 'to deny that mental disorder and violence are in any way associated is disingenuous and ultimately counterproductive'. Monahan proposes two ways in which to determine whether a relationship exists between mental disorder and violent behaviour, and if it does, to estimate the strength of that relationship. If being mentally disordered increases the probability that an individual will commit a violent act, then there must be a higher prevalence of violence among mentally disordered persons than the non-disordered populations. Accordingly, it would be logical to assess the prevalence of violent behaviour among mentally disordered persons and compare it to the prevalence of violence among non-disordered people. Furthermore, if mental disorders contribute to the occurrence of violence, the rate of mental disorder should be higher among people who commit violent acts than among those who do not. Therefore it is suggested that the rates of mental disorder among criminal populations will be higher than non-criminal populations. 
Three types of studies can be used to analyse violence among the mentally disordered population: the prevalence of violent incidents before admission to hospitals; the prevalence of violent episodes while in hospital care; and the prevalence of violent behaviour once released into the community. Comparative studies are then necessary in non-disordered populations to determine whether there is an increased risk of violent behaviour among the mentally disordered. This summary of research attempts to look at all these types of studies.
One study by Johnston et al looked at the disturbed and violent behaviour among patients with schizophrenia in the month prior to admission to a psychiatric institution. 19% of this group had behaved in way that put the safety of others at risk and over a third had damaged property. Although this is not necessarily a physical assault on the person, intimidating behaviour suffices as danger to the health and safety of others in our definition of mental disorder. The police also assisted in the admission to psychiatric assessment for over 20% of patients in this study. An American study reported higher rates of violence in the six months prior admission. Over 26% had engaged in an assault and a further 36% behaved in a manner likely to produce fear in other people. Monahan summarised three reviews reporting on 11 studies and found that in the average period of two weeks before admission, approximately 10% to 40% (an average of 15%) of psychiatric admittees committed a physical assault. These reviews also provided data on violence among inpatients ranging from the first day of admission up to a year. Violent behaviour also ranged from 10% to 40% of inpatients with an average of 25% of inpatients acting violently. Walker and Seifert undertook a prospective study of physical assaults for six months in a new psychiatric intensive care unit. There were 48 patients admitted and 37 assaults occurred, three of which were against other patients and 34 were against staff. Results indicated that a third of patients were involved in assaults. However, two patients were responsible for over 40% of all assaults. This suggests that a few patients were responsible for a large percentage of all assaults. In addition, it was found that the features most highly correlated with violent behaviour were a criminal record and a history of drug abuse. A study by Fottrel into the violent behaviour of psychiatric inpatients reported similar findings to the previous study. There were many incidents of petty violence and incidents of serious violence were rare. It was found that only a small percentage of the total patient population behaved violently (from 3-10%) and that a small number of patients were responsible for a large number of the incidents. In one hospital 3% of patients were involved in 70% of incidents and in another 4% of patients were involved in 59% of all incidents. Interestingly enough, it was found that schizophrenia was the predominant diagnosis of patients involved in violent incidences.
It must be noted that there are other plausible explanations for violence among inpatients. Acting out behaviour could result from a number of external factors such as the realities of confinement and provocation from other patients and visitors. An interesting study looked at the events that preceded violent incidences in psychiatric institution. The authors studied over 1000 such incidences in three hospitals over 13 months. It was found that patients who were agitated or disturbed were more likely to act violently (29%); restrictions by the hospital management were responsible for 19.2% of the incidences; and provocation by other patients, relatives or visitors was responsible for 17.5%. Only 0.7% incidents arose from staff contact with patients. Again, 8% of the patients were responsible for over 40% of all incidents.
In one follow up study of 304 high risk male patients with at least one violent incident in their past, it was found that approximately 25% — 30% of the patients were violent in their first year of release from the hospital; the mean time was with 4 months. However, this is a very selective population of high risk individuals and the findings should not be taken as applying to all released psychiatric patients. A study by Steadman and Felson compared the violent behaviour of released patients with a community sample. It was found that the 22.3 % of released patients reported at least one dispute using a weapon compared to 15.1% of the community sample. 8.1% of the disordered population reported using a weapon in disputes. This is compared to 1.6% of the community sample. However, these differences were not considered significant. Swanson et al conducted a highly reported study comparing disordered and non-disordered people. After reanalysing data from the Epidemiological Catchment Study (ECA) of 10,059 subjects it was found that violent acts were reported by 2.4% of the population. For those with a diagnosis of schizophrenia it was 12.7%; for major depression and bipolar it was 11%; for those with alcohol problems it was to 25%; and for drug abuse and dependence it was 34.7%. It must be noted that the co-occurrence of mental illness and substance abuse accounted for much of these higher rates of violent acts among the mentally disordered. The authors concluded that 1) the prevalence of violence is more than five times higher among people who meet criteria for a DSM-III R diagnosis than those who received no diagnosis; and 2) the prevalence of violence among people who received a diagnosis of alcoholism was 12 times that of people with no diagnosis, and the figure increased to 16 for those who had a diagnosis of drug abuse and dependence. One other study by Link et al compared 400 adults from the Washington Heights, New York City, with community sample of 375 patients from the same neighbourhood who had been or were in current contact with psychiatric services. The study found that the patient group was about 2 to 3 times more likely to be violent than the community group, particularly when experiencing psychotic symptoms which may include hallucinations and delusions.
Another way to analyse the relationship between mental disorders and violent behaviour is to look at the prevalence of mental disorders amongst prison populations, for which we have recent data for New Zealand prison population. However, the large numbers of substance abuse and personality disorders often complicate this method. Substance abuse problems, although a formal diagnosis in the DSM-IV R, are usually under the ambit of the Alcohol and Drug Addiction Act 1966 and not the Mental Health (CAT) Act 1992. A personality disorder is a very common diagnosis in prison populations and is considered a somewhat less severe diagnosis than other mental disorders. For example, about 60% of New Zealand prisoners had a diagnosis of a personality disorder. In 1999 the New Zealand Department of Corrections carried out a study investigating the lifetime prevalence of mental illness in New Zealand prisoners compared to community sample. All female inmates, all male remand inmates, and 15% of all sentenced male inmates, were interviewed. This was considered representative of the prison population according to the New Zealand census of 1996. The prevalence of mental illness was compared to a previous study by Wells & Oakley-Browne et al. Results indicated that 89.4 % of the total population had had a lifetime prevalence of a substance abuse disorder; and approximately one third of these prisoners had suffered from another disorder. Over 25% had experienced a major depressive disorder, and at the time of the study 10% were suffering from a current episode of major depression. For major depressive disorder lifetime prevalence was 31.9% in women, 22.3% in male remands and 20.6 % in sentenced men. It is estimated that between 6 and 8% of the prison population have suffered from either schizophrenia or a related disorder such as schizophreniform, delusional and schizoaffective disorders. Approximately 2.8% were suffering from schizophrenia or a related disorder at the time of the study. The lifetime prevalence of bipolar disorder was 1.2% in women, 2.3% in male remands and 2.2% in sentenced men. The lifetime prevalence of alcohol abuse and dependence was 69.1% in women, 74.7% in male remands and 76.0% in sentenced men, and for cannabis abuse or dependence was 43.3% in women, 53.7% in male remands and 55.7% in sentenced men. When the results were compared to the earlier community Christchurch study there were a number of mental disorders that were significantly elevated in the prison population. These were major depressive disorder; bipolar disorder especially with a current episode of mania; schizophrenia and related disorders; substance abuse and dependence, (particularly so in women); post traumatic stress disorder; obsessive compulsive disorder, and extremely high rates of personality disorders. In addition, it was estimated that at the time of the study there were 135 inmates with bipolar, schizophrenia or related disorders that required hospital admission.
These results are consistent with the frequency of mental illness in other western countries. Fazel & Danesh  reviewed 62 surveys of mental illness in prison populations in 12 western countries. It was found that one in seven prisoners in western countries have a psychotic illness or major depression. More specifically, a study of a representative sample of 1,365 male and 272 female prisoners in England and Wales found the prevalence of schizophrenia in males at 1.5% and females at 1.1%. This is significantly higher than the general population where prevalence ranges from 0.2% to 0.5%. Alcohol abuse was found in 8.6% of male prisoners and 4.4% female prisoners. Drug abuse was found in 10.1% of men and 24.2% female.
While these studies indicate that there is a large percentage of prison populations suffering from mental illness, information was not provided regarding the nature of the offences. It is therefore unknown what percentages of these mentally disordered offenders were convicted of violent or fear-inducing offences. One well-known study looked at the type of offence and psychiatric status of 1,241 men remanded to Brixton prison in London before trial. It was found that 9% of the remandees who were subsequently convicted of a non-fatal violent offence had schizophrenia, and 11% of those convicted of a fatal violent offence had schizophrenia. Additionally it was found that 112 out of the 121 schizophrenic offenders had positive symptoms at the time the offence was committed. Follow up studies look at the number of mentally disordered people who violently offend after release from psychiatric institutions. One such study in Stockholm traced conviction records of over 644 patients diagnosed with schizophrenia over 14 years. It was found that the crime rate for males with schizophrenia was the almost the same as that of the general non-schizophrenic male population whereas for females with schizophrenia it was twice as high as the non-schizophrenic general female population. However, combining male and females with schizophrenia, it was found overall they were more than 4 times likely than the general population to commit violent offences. The majority of these violent offences were minor, the most violent being a charge of aggravated assault. There were some methodological weaknesses in the study as it failed to look at relevant factors such as the length of time spent in psychiatric institutions during the 14 years. Another study looking at violence by people with schizophrenia that had similar results was carried out in the United Kingdom. Five hundred and thirty eight people with schizophrenia were drawn from a community registrar and their criminal records over 20 years were examined. It was found that females with schizophrenia had an increased conviction rates compared with other mental disorders, but the conviction rates of men with schizophrenia did not differ. Consistent with the previous Stockholm results, overall people with schizophrenia were more than three times more likely than a non-schizophrenic control group to have a conviction for assault or serious violence. These studies suggest that violent offending and criminal convictions are related to mental disorders, particularly schizophrenia. However there are a number of other reasons that may contribute to this relationship. For example, some mentally disordered people in the community are known by the police and are kept under careful observation. Dietz identified five patterns that are frequently observed among mentally disordered offenders according to the relationship between their mental disorder and criminality. The first is crime in response to psychotic symptoms in which people respond to delusions in a criminal way, particularly to persecutory delusions. For example, if a person believes others are out to 'get them' then they may act violently to protect himself or herself. The second pattern is crime to gratify compulsive desires where mental disorders may provide the motive, but they do not impair an offender's knowledge of what he is doing. The types of factors that influence this pattern are character, mood, intoxication, recent stress, likelihood and severity of punishment. These factors will determine whether a crime will be committed regardless of whether the desire stems from a mental disorder. The third pattern is crime reflecting personality disorder which often explains reckless and impulsive behaviour. The fourth pattern is coincidental crime and mental disorder in which the combination of mental disorder and criminal behaviour is by chance alone. The fifth pattern is true or feigned mental disorder in relation to crime. This includes disorders that develop in response to the crime committed such as post-traumatic stress disorder and malingering so as to achieve benefits such as not guilty by reason of insanity or placement in a safer environment. Dietz emphasises that these patterns are not exhaustive of the possibilities in explaining mental illness and crime, nor does he propose a typology. This approach focused on the basic issues of cause and effect between crime and mental illness.
From the above studies, it is obvious that there are some factors which increase the probability of violence among mentally disordered people. Substance abuse combined with a mental disorder significantly increases the possibility of violence with estimates as high as 17 times more likely to be violent. The diagnosis of schizophrenia or other major affective disorder is associated with a fourfold increase in the odds of violence in one year. Another study found a significant relationship between psychotic symptoms and rates of violence. Some personality disorders have recently been identified as a risk factor for increased violence among people with mental disorders, in particular cluster B personality disorders which are narcissistic, borderline, histrionic and antisocial personality disorders. These disorders are often associated with impulse control problems, dysregualtion of emotion, behaving in an over dramatic manner and severe interpersonal deficits. Furthermore, antisocial personality disorder is a prerequisite for a diagnosis of psychopathy (other names include sociopath, psychopath and dissocial personality disorder). Psychopathy is associated with significant increases in violent behaviour, substance abuse and criminal offences. Given these factors, one must question how accurate mental health professionals are in predicting mentally disordered people who will behave violently or commit violent offences.
It would be an ideal situation if we could identify exactly which mentally disordered people were going to behave violently to the degree needed for preventative detention. Unfortunately this is not the case. The notion dangerousness requires a prediction of how a person is likely to behave. Accordingly, risk prediction is necessary to determine if a person is dangerous and what level of danger they present. As mentioned earlier the judiciary rely upon the advice of people who have expertise and knowledge in this area. A medical practitioner must perform the original assessment to determine whether a person is suffering from a mental disorder; however, a responsible clinician may be involved in further assessment. A responsible clinician is a person who holds a professional qualification relevant to the assessment, treatment and care of patient with a mental disorder and who is in charge of treatment for that patient. Although the responsible clinician will often be a psychiatrist, this may be extended to include other mental health specialists like psychologists. For this reason, the term 'mental health professionals' is preferred. When earlier researchers concluded there was no relationship between mental disorder and violence, it was correspondingly thought that mental health professionals were unable to predict dangerousness among the mentally ill with any reliability. A three-year prospective study by Cocozza and Steadman looked at the accuracy of psychiatric assessments in defendants found unfit to stand trial. From the patients evaluated as dangerous only 14% of those released into the community were subsequently arrested for a violent crime (13 out of 96). However, among the patients assessed as non-dangerous, 16 % who were released into the community were arrested for a violent crime (11 out of 70). The authors concluded that there was no difference of dangerousness between the two groups and psychiatric predictions were not at all accurate. There has been some criticism of this study as it provided no evidence of what the psychiatrists took into consideration when making assessments or what level of community support was available as a buffer against violent behaviour. The disagreement over mental health professionals' expertise in predicting dangerous behaviour has caused some professional bodies to distance themselves from the job of risk prediction. Following atask force investigation The American Psychiatric Association concluded that 'psychiatric expertise in the prediction of dangerousness is not established'. The American Psychological Association came to a similar resolution:
the validity of psychological predictions of violent behaviour, at least in the sentencing and release situations we are considering, is extremely poor, so poor that one could oppose their use on the strictly empirical grounds that psychologists are not professionally competent to make such judgements.
This is a harsh view because additional studies have suggested that mental health professionals can predict future perpetrators of violent behaviour. Unfortunately in doing so, they over predict who will act violently. Steadman and Cocozza conducted another well-known study regarding violence predictions after the case Baxstom v Herold in the United States. A patient who was serving a sentence in prison was certified insane and was transferred to a psychiatric institution where he was detained there beyond the length of his sentence. The Supreme Court held that this was illegal and following this decision the authorities released nearly 1000 patients from New York's two forensic hospitals to civil hospitals and to the community. Steadman and Cocozza attempted to predict from this population who would reoffend. The percentage of people who were dangerous in the community was 14%. Using clinical variables and the criminal records of the patients they were able to accurately predict 80% of these patients who exhibited dangerous behaviour within four years of their release. This was only one third of the patients identified as likely to exhibit violent behaviour which means that two thirds of the patients identified as high risk did not engage in dangerous behaviour. This study suggests that mental health professionals may be able to predict accurately, but it is at the expense of those labelled high risk who will not behave violently. These findings were supported in other studies and it was concluded by Monahan in a review of the five major studies available in 1981 that 'psychiatrists and psychologists are accurate in no more than one out of three predictions of violent behaviour over a several year period among institutionalised populations that had committed violence in the past ... and who were diagnosed as mentally ill'.
One study that offered more positive results was by Lidz et al. It looked at subjects who were assessed in the psychiatric emergency room of a large hospital. Psychiatrists and nurses were asked to assess patients for potential violence over the next six months. 53% of patients that caused concern regarding future violence were in fact violent after their release. This is significantly more than the 36% who did not raise concerns but subsequently were violent. Overall the results suggest that although accuracy of predications regarding male patients were higher than chance predictions, those regarding women patients did not differ from chance. There is also emerging evidence that well trained clinicians should be able to provide accurate short-term predictions on the violence potential of a patient i.e. a few days to a week. Although this is encouraging, it is not at all practical. Under the Mental Health (CAT) Act 1992, compulsory treatment orders are made for six months at a time with clinical reviews every three months. To take advantage of this ability to predict short-term behaviour orders would need to be made on a weekly basis. This would over extend both the judges and mental health professionals involved in the patient's case.
A number of these studies are American and because of the litigious climate and fear of being sued many mental health professionals are over cautious in risk predictions. This is following the famous case of Tarasoff v Regents of the University of California in which a student at the university's hospital where he was receiving voluntary treatment as an outpatient admitted to the psychologist that he wanted to kill a girl with whom he had an emotional attachment. He was briefly detained but released when he later appeared rational. The head psychiatrist directed no further action be taken despite concluding he was a risk to the specific girl in question. Consequently he murdered the girl and her parents bought an action against the mental health professionals on the basis that they were negligent in not warning a third party of potential danger. The judgment was for the plaintiff and opened the way for people to bring negligent actions against mental health professionals. The court also stated that where there was danger to undifferentiated people in the community, the appropriate response would be to initiate civil committal proceedings to remove the dangerous person from society. Although in New Zealand we have an accident compensation system which limits the liability of mental health professionals, the consequences of serious violence by a mentally disordered person often reflects poorly upon those involved with the person's assessment and treatment. This increases pressure for professionals involved in the care of mentally disordered people to take extra care in risk assessments. Overall the research questions the accuracy of mental health professionals making judgements about the risk of violence in mentally disordered patients. It is also generally accepted that about one out of three predictions will be accurate. The Law Commission has recognised that overseas studies... indicate that at least 50% and as many as 66% of judgments that a given person is dangerous will be inaccurate. More recent consideration of the data from these studies suggests that 50% may be the more appropriate assessment of the accuracy rate. Furthermore, it appears that mental health professionals may not be comfortable in the role of risk assessment due to a lack of confidence in such procedures. Unfortunately, the criminal justice system requires these predictions to be made and the responsibility usually falls upon these mental health professionals.
Over-prediction seems to be the only way to catch mentally disordered people likely to behave violently. One author suggests that it is because psychiatrists are being over cautious. He explains that caution is inherent in risk assessment because of the climate that has been established which results in hysterical response every time a mentally disordered offender commits a serious offence. Intertwined with the cautiousness exercised by mental health professionals are problems of low base rates and false positives. These are explained by the concepts of specificity and sensitivity. In this type of situation, sensitivity refers to the mentally ill who are violent and who are correctly identified by the risk assessment. Specificity is the number of non-violent mentally ill who are correctly identified by the assessment. The ideal assessment is one where there is high sensitivity and high specificity. If there is low sensitivity, it suggests that too many people are labelled as non-violent who are in fact violent. Conversely, if the assessment has low specificity it implies that it was unable to correctly identify those mentally ill who are non-violent. The concepts used to help explain sensitivity and specificity are true positive, false positive, false negative, and true negative. True Positive is where the person receives a positive risk assessment (i.e. it is predicted that the person is likely to behave violently) and he or she do behave violently. False Positive is where the person received a positive risk assessment but did not behave violently. False Negative is where a person is given a negative risk assessment (i.e. it is predicted that they will not behave violently) but he or she does behave violently. True Negative is where the person has a negative risk assessment and does not behave violently (see diagram 1).
Diagram 1: Assessing violence in the mentally disordered
Violent mentally disordered
Non-violent mentally disordered
Risk Assessment Positive
Risk Assessment Negative
For example, using a test of 90% sensitivity and 94% specificity we can try and identify 10 violent people from a population of 100 mentally disordered people. The positive predictive value is 64% and negative predictive value is 99%. If we rely on this test, 64% of those labelled high risk and detained in a hospital would actually be violent. This means that 36% of people labelled high risk and detained would not be violent. Consequently, of the 86 mentally ill people released into the community, just over 1% will be violent. This is illustrated in diagram 2.
Diagram 2: Assessment with a prevalence of violence at 10%
Positive Risk (Detained)
Negative Risk (Released into community)
These predictive values indicate 5 people will be detained who would not be violent to protect society from those who would be violent. This may be an acceptable cost for society. But when the population is increased and we try to identify 10 violent people from a population of 1000 mentally ill people the task becomes more difficult. The cost of accurately identifying these violent people means that 59 people will be detained who would be violent, that is 87% of all mentally ill people detained would not be violent (diagram 3).
Diagram 3: Assessment with a prevalence of violence at 1%
Positive Risk (Detained)
Negative Risk (Released into community)
This second example is more appropriate because it highlights the problems of low base rates. The behaviour, which is the occurrence of violent behaviour, is the base rate measured and occurs at a low frequency. A problem arises with attempts to predict low frequency behaviours because it results in high numbers of false positives. High numbers of false positives equate to high numbers of people given positive risk assessments that did not or will not act violently. This suggests that there are significant numbers of patients detained under compulsory treatment orders who would not be violent if they were released into the community. It has been argued that these conservative clinical predictions have led to a 'public protectionist' model with a lot of false positives and fewer false negatives. These false positive errors are not seen as they are hidden away in psychiatric institutions. However, there would be much more public outcry if clinical predictions caused large numbers of false negatives as more violent acts by mentally disordered people would take place in the community. The question is how many false positives - those labelled high risk and detained but who would not be violent - is society prepared to accept to achieve the goals of mental health legislation. Based upon studies, mental health professionals can only accurately predict from 33% to 50% people who will behave violently. Evidence suggests that psychiatrists over predict by about two-thirds, those who will be violent. For example, from a population of 100 mentally disordered people, there are 10 who are likely to behave violently or in a fear-inducing manner and who would be considered serious enough for compulsory detention. Mental health professionals will predict 30 people, including the 10 who will act violently, who should be detained. This leaves 20 people who will be treated and detained in a psychiatric institution but who present no danger to others. The loss of the rights of 20 people may seem an acceptable loss to society for protection against potential harm, but what if there were 20 people from the 100 who would behave violently. This would mean that 60 people would be detained on the basis that 20 will behave violently. Although this is a hypothetical question which can cause endless philosophical, sociological and ethical debate, it is one that needs to be asked. How prepared is our society, which professes to be dedicated to the protection of people's rights, to be seen to encroach on individual rights so that we can be free from potential harm?
The vague references to the rights of the individual and those of society are not arbitrary constructs. In New Zealand these rights are found in statutes or through ratification of international documents. There are also a number of international documents which although not binding are highly recommended. The consequences of a determination of dangerousness are severe. A person can be assessed, detained and treated without consent. This conflicts with a number of individual rights. New Zealand has ratified the International Covenant on Civil and Political Rights 1966 (ICCPR) which recognises the right of the individual to liberty and security of the person. Section 11 of the New Zealand Bill of Rights Act 1990 states everyone has the right to refuse to undergo any medial treatment and s 22 affirms the right not to be arbitrarily arrested or detained. Detention is arbitrary if it is capricious, unreasoned or without reasonable cause: R v Goodwin. It could be argued that a judicial determination of dangerousness, particularly if it is based upon inaccurate risk assessments, is unreasoned. Furthermore, detaining a person based upon an uncertain prediction could be considered arbitrary. Similarity the Bill of Rights also asserts that everyone has the right not to be arbitrarily arrested or detained on the ground of disability. It is possible that a person with a mental disorder will be considered to be under a disability and it could be argued that this is arbitrary.
It is not only individual rights that are in danger of being infringed, but also the rights of the mentally disordered population. There are a number of international instruments which encourage the least restrictive provisions for people with mental disorders. The Declaration of the Rights of Disabled Persons defines a disabled person as 'any person unable to ensure by himself or herself, wholly or partly, the necessities of a normal individual and/or social life, as a result of deficiency, whether congenital or not, in his or her physical or mental capacities'. According to the Human Rights and Equal Opportunity Commission of Australia in the Burdekin Report this definition would include many people with a mental disorder. Principle 9 of the Declaration recognises 'the right not to be subjected to more restrictive conditions than necessary'. Similarly, the United Nations principle for the Protection of Persons with Mental Illness and for the Improvement of Mental Health Care states that discrimination on the basis of mental illness is not permitted (principle 1.4) and every person with a mental illness has the right to live and work, as far as possible, in the community (principle 3). The judiciary are aware of the potential abuse of human rights. As stated by Callaghan DCJ in determining an application for a compulsory treatment order:
. .the Mental Health Act is a complicated piece of legislation and it makes substantial inroads into a patients rights and the Court must be ever vigilant to ensure that the patient's rights to remain in the community are not encroached upon unless there is satisfactory evidence.
Unfortunately few, if any, human rights are absolute. At any time two or more human rights are in conflict. This is the case under the Mental Health (CAT) Act 1992 between patient rights and societal interests. As Henry Prins noted one of the 'most crucial dilemmas' facing those who make risk predictions is to 'balance the need to act in the interests of the community as agents of control and custody on the one hand, and to serve the interests of the individual on the other'. The right of society to live free from potential harm or violence is in competition with the rights of a mentally disordered person to not be detained. The Bill of Rights states that no one shall be deprived of life and this is affirmed by Article 6 of the ICCPR. The ICCPR also provides that everyone has the right to liberty and security of the person. In cases where fear-inducing behaviour such as stalking or property destruction may be the problem, Article 17 of the ICCPR states that no on shall be subjected to arbitrary or unlawful interference with his privacy, family, home or correspondence, or to unlawful attacks on his honour and reputation.
The New Zealand Bill of Rights Act 1990 recognises that there are situations of conflicts and has built in mechanisms which allow some rights to be overridden. Section 5 states that the rights and freedoms contained in the Bill of Rights may be subject only to such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society. It could plausibly be argued that the detention of dangerous mentally ill persons is a reasonable limit on individual rights for the benefit of society. However, because of the inaccuracy of risk prediction it is unknown exactly which mentally disordered persons present a serious danger. Section 6 provides that wherever an enactment can be given a meaning that is consistent with the Bill of Rights, that meaning shall be preferred to any other meaning. It is a reasonable contention that the Mental Health (CAT) Act 1992 could be seen to be consistent with the Bill of Rights. This would require mentally disordered people not to be arbitrarily detained on the basis of dangerousness unless they are accurately pinpointed as dangerous, which is not possible with the current methods of risk prediction.
An obvious suggestion to the problem is to try and increase accuracy in risk assessments. A recurring theme in the area is to use actuarial methods of prediction over the clinical method. Actuarial systems are statistical methods which are based on empirically determined frequencies and set out exactly what kinds of data are to be considered in a risk prediction. This is opposed to clinical methods which are based upon impressions gained from interviews and the choice of data varies from case to case. Support for actuarial methods has been increasing since evidence suggests that the same predictors for violently re-offending among the mentally disordered are the same among the general population. Despite this, the use of actuarial methods has been criticised for being unethical. It uses information gained from a population and applies it non-discriminately to an individual. For example, between 40% and 50% of violent crime in America is committed by black people, and 80% — 90% of crime by young black men. Consequently, high percentages of young black men may be labelled as likely dangerous offenders according to actuarial risk assessments and 'judging a person by race or gender is not a practice that is easily justified on ethical grounds. From the perspective of the individual it clearly represents racial and sexual discrimination with consequent violations of personal rights'. 
In practice, actuarial methods would be subject to the same claims of right infringements as the current clinical methods.
There are other suggestions on how to improve psychiatric predictions of risk. They include taking special notice of particular risk factors such as criminal history, clinical and legal education, types of information about the patient, and increased communication between clinicians in charge of a patient. However, this essay is not a critique of the clinical methods of predicting dangerousness. It is an analysis into the validity of the dangerousness concept which is included into our mental health legislation. In summary, evidence indicates that there is a slight increase of violent behaviour among the mentally ill population. However, from research one has to add the following qualifications. People with a severe mental illness such as schizophrenia or bipolar disorder may present an increased risk to others, particularly due to psychotic symptomology. Those people suffering from severe mental illness who have active (psychotic) symptoms and who also misuse drugs or alcohol may present a seriously increased risk to others. In addition, a diagnosis of psychopathy may present an elevated risk to others. Also, a small number of mentally ill people appear to be responsible for a large percentage of violent incidents. This suggests that the great majority of mentally ill people present no increased risk to others.
With regard to predictions of violent behaviour or future offending, evidence suggests that the best indications of future offending among mentally disordered people are the same as those for the rest of the population. These are previous offending, criminality in the family, deviant peers and antisocial attitudes. Professional risk predictions are accurate in approximately one third of cases and short-term risk predictions are more accurate than long-term predictions. Psychiatrists tend to be over-cautious with risk predictions which suggests that a number of people involuntarily detained are so detained on the basis of inaccurate predictions of dangerousness.
Many philosophers and rights activists believe that state intervention on the prediction of dangerousness violates the most fundamental rights in a purportedly democratic society. Szasz provides an apt example of this attitude:
The ghost of the 'dangerous mental patient' will not be laid to rest until it is recognised that the institution to which the co-called mental patient is committed is not a hospital but a prison.
This position contends that patient rights are being unfairly eroded by the state and provokes much criticism from both the judiciary and the mental health professionals involved in patient care. As stated by Pfhol:
the contemporary concern with identifying and isolating dangerous people appears tied to liberal of reformist movements within mental health and criminal justice circles, with efforts to secure rights for involuntarily institutionalised through depopulating mental hospitals and prison of all but those who are truly dangerous.
But dangerousness is not the only reason people are detained under compulsory treatment orders. Mental illness causes a great deal of distress and disturbance in a person's life and treatment and professional support are appropriate in a large number of cases. This is consistent with the idea of the Mental Health (CAT) Act as a piece of social care legislation, dedicated to improving services and care forpeople suffering mental illness. As noted by the authors of Butterworth Family Law Journal 'the overall purpose of the statute is protective and therapeutic, for the care and treatment of the mentally ill, not the prosecution and punishment of criminal offenders'. This perspective has been supported in a number of cases. In Re E and Re C it was held that the nature of proceedings for a compulsory treatment order is best regarded as inquisitorial rather than adversarial. Judge Doogue in The matter of D held under the Mental Health (CAT) Act 1992 'that the Court's role is to protect the welfare of the patient and to protect the public interest' and that if the procedure was adversarial then 'the responsible clinician could become the patient's adversary. The very person who is responsible for the health and welfare of the patient would be in competition with the patient and that would make a mockery of true intent of the statute' . A therapeutic relationship is built on trust that the patient holds for his or her responsible clinician and compromising this relationship would be detrimental to the patient. The Act also advocates the least restrictive alternative in treating patients: directing that orders for compulsory treatment must be community treatment orders unless adequate treatment can only be provided through an inpatient order. The intent of the Act is to clarify circumstances and conditions under which people may be subjected to compulsory assessment and treatment, and to define and protect the rights of these people. It is the author's opinion that the concept of dangerousness does not help achieve these goals and in particular it fails to protect patient rights. Because of the differences relating to the assessment of dangerousness and taking cognisance of mental health patient's rights, it may be an advantageous move for the legislature to delete the dangerousness aspect from the definition of mental disorder. It is suggested that the criteria by which a person may be subjected to compulsory assessment should reflect the purpose of the Act as a piece of legislation dedicated to the care and treatment of mentally ill people in the community. It should also incorporate the reality that the judiciary does not have the relevant knowledge or training to make a judgment regarding the mental health of an individual. Accordingly, it would be preferable to assess 'abnormal state of mind' on a more clinical basis which would remove the absolute need to prove dangerousness. The need for compulsory assessment and treatment will be decided from a perspective taking into account the best interests of the patient. This would serve to increase awareness of the fact that risk prediction is not an exact science and decisions are made in the best interests of the patient. This approach would still incorporate the concept of dangerousness because it is in the patient's best interests not to commit violent offences and suffer the consequences of the criminal justice system and the public pressure for preventative detention. This type of legislative change may also have the benefit of alleviating stereotypes that mental illness is synonymous with violent and fear-inducing behaviour. While there will always be cases involving likelihood of harm or injury to others, the inaccurate and often 'guess work' approach of assessing dangerousness can be put to one side. This will preserve both the true intent of the Act and the integrity of patients.
[*] The author is an undergraduate student in the Law Faculty, University of Canterbury.
This paper was written as part of the undergraduate Honours programme.
 J Monahan, ' Mental disorder and violent behaviour - perceptions and evidence' (1992) 47 American Psychologist 511.
 P E Mullen, 'Reassessment of the link between mental disorder and violent behaviour, and its implications for clinical practice' (1997) 31 Australian and New Zealand Journal of Psychiatry 3, 9.
 New Zealand, Parliamentary Debates, Hansard (1992) vol 522, 6479, 6860 (Hon Katherine O'Regan).
 Law Commission, Community Safety: Mental Health and Criminal Justice Issues, Report No 30 (1994).
 Ibid para 54.
 R J Menzies, C D Webster & D S Sepejak, 'Hitting the forensic sound barrier: predictions of dangerousness in a pretrial psychiatric clinic' in C D Webster, M H Ben Aron & S J Huvker (eds), Dangerousness (1985) 137.
 G Amer, 'Mental Disorder: "Serious Danger" - a minor faux pas or a serious mistake?' (1996) 2 Mental Health and the Law 40, 42.
 Under the Mental Health (CAT) Act 1992 (NZ) s 14 a person can be made the subject of a ccompulsory treatment orders.
 The Long Title of the Mental Health (CAT) Act 1992.
 Law Commission Report, above n 4, para 5.
 Mental Health (CAT) Act 1992 (NZ) s 8.
 Mental Health (CAT) Act 1992 (NZ) s 9 a written notice to be issued requiring the person to attend the assessment examination and detailing the place and time.
 Mental Health (CAT) Act 1992 (NZ) s 10.
 Mental Health (CAT) Act 1992 (NZ) s 12.
 Mental Health (CAT) Act 1992 (NZ) s 11.
 Mental Health (CAT) Act 1992 (NZ) ss 13, 14.
 The responsible clinician must satisfy a number of requirements during this period and the patient is entitled to respect for cultural identity; appropriate medical treatment; an explanation of the expected effects of any treatment offered to the patient; an entitlement to seek a consultation with a psychiatrist of his or her own choice in order to get a second opinion and to request a lawyer to advise the patient on his or her status and rights.
 Mental Health (CAT) Act 1992 (NZ) s 14.
 Mental Health (CAT) Act 1992 (NZ) s 28. It will be an inpatient treatment order or community treatment order.
 Mental Health (CAT) Act 1992 (NZ) s 29.
 Mental Health (CAT) Act 1992 (NZ) s 30. 22 Mental Health (CAT) Act 1992 (NZ) s 17.
 Mental Health (CAT) Act 1992 (NZ) s 18.
 Mental Health (CAT) Act 1992 (NZ) s 20.
 Mental Health (CAT) Act 1992 (NZ) s 24, states that the hearing is not open to the public. The only people who may attend are the judge, officers of the court, parties to the proceedings and their barristers and solicitors and any other person nominated by the patient, witnesses, any other person to whom the certificate of final assessment was sent and any other person whom the judge permits to be present: ss 19, 20, 21, 22 & 23.
 Mental Health (CAT) Act 1992 (NZ) s 28.
 Mental Health (CAT) Act 1992 (NZ) s 33. 28 Mental Health (CAT) Act 1992 (NZ) s 34.
 For more detail on procedure with special patients, see W J Brookbanks & S A Bell, Mental Health Law in New Zealand (1998) 36-71.
 Mental Health (CAT) Act 1992 (NZ) ss 54, 55. 31 Re Tahere  DCR 545, 552.
 It is suggested that not only does the restricted status reflect public concern, but it also adds a political dimension to a clinical concern as the state has more power to intervene into a person's life and deprive them of their personal liberty: W J Brookbanks & A I F Simpson, 'Restricted Patients in New Zealand: A Legal and Clinical Overview' (1996) 3 Journal of Law and Medicine 336, 340.
 Mental Health (CAT) Act 1992 (NZ) s 8.
 Mental Health (CAT) Act 1992 (NZ) s 2.
 Mental Health (CAT) Act 1992 (NZ) s 34.
 Brookbanks and Bell, above n 29, 11.
 In the matter of T  NZFLR 351, 355 (Boshier DCJ).
 In the matter of GBW  NZFLR 428, 430 (MacCormick J).
 In the matter of JK  NZFLR 678, 695.
 S Cooper, 'Mental Disorder and establishing Serious Danger'  2 Mental Health and the Law 5.
 Erotomania is the delusional belief that one is loved, perhaps secretly, by another person. Usually the other person is of higher status than the person. It is commonly associated with stalking behaviours.
 Delusions of persecution are beliefs that one is in danger of being followed or monitored, harassed or conspired against. This may involve government agencies, criminal groups (the mob), neighbours, co-workers and even family. They are often associated with stalking and other intimidating behaviours due to the nature and type of delusions.
 Cooper, above n 40, 3.
 Re RWD (1994) 12 FRNZ 387, 403.
 Re TAV  NZFLR 846, 850.
 In the matter of JK  NZFLR 678, 679.
 Re T  NZFLR 946, 956.
 Re O  NZFLR 545, 546.
 Re T  NZFLR 946, 955.
 Re GM  NZFLR 665, 670-1.
 Re M  M 716/85 (Unreported, High Court of Wellington, Grieg DCJ, 21 April 1986).
 P R H Webb & P J Treadwell, Butterworths Family Law in New Zealand (10th ed, 2001) para 9.53.
  NZLR 266.
 In the matter of JK  NZFLR 964
 Mental Health (CAT) Act 1992 (NZ).
 Webb and Treadwell, above n 53, para 9.51.
 Re M  M 716/85 (Unreported, High Court of Wellington, Grieg DCJ, 21 April 1986).
 Brookbanks and Bell, above n 29, 17.
 S McCarthy & S Simpson, Running a Case Under the Mental Health Act 1992 and Related Legislation (1996) 10.
 Ibid 59.
 Ministry of Health, The guidelines for clinical risk assessment and management in the mental health service (1998).
 S Shah, 'Dangerousness: a paradigm for explaining some issues in the law and psychology' (1978) 36 American Psychologist 224, 226.
 S Dinitz & J P Conrad (1978) in Menzies, Webster & Sepejak, above n 6, 137.
 L Childs & P M J Brinded 'Rehabilitation of the Mentally disordered Offender' (2002) 37 (3) Autralian Psychologist 229-36.
 Monohan, above n 1, 511.
 Ibid 515.
 Mullen, above n 2, 5.
 Ibid 7.
 Monohan, above n 1, 513.
 Case law has established that serious danger can include psychological and emotional harm danger to others. See Re RWD (1994) 12 FRNZ 387, 403.
 J Monahan & H J Steadman 'Toward a rejuvenation of risk assessment research' in J Monahan & H J Steadman (eds), Violence and Mental Disorder; Developments in Risk Assessment (1995).
 Hafner and Boker (1982): 'if we define dangerousness of the mentally abnormal as the relative probability of their committing a violent crime then our findings show that this does not exceed the dangerousness of the legally responsible adult population as a whole' in P E Mullen 'The Dangerousness of the mentally ill and the clinical assessment of risk', in W Brookbanks (ed), Psychiatry and the Law — clinical and legal issues 94.
 Mullen, above n 2, 4.
 Monohan, above n 1, 514.
 E Johnston, T Crow, A Johnston, & F Macmillan, 'The Northwick Park Study of first episodes of schizophrenia 1: Presentation of the illness and problems relating to the admission' (1986) 145 British Journal of Psychiatry 728.
 'Assault' under the Crimes Act 1961 can include a threat: R v Mwai  NZCA 306;  3 NZLR 149, 155.
 'Health' includes psychological and emotional wellbeing: Re RWD (1994) 12 FRNZ 387, 403.
 Monohan, above n 1, 515-8.
 Ibid 514.
 A Walker, R & Seifert, 'Violent incidences in a psychiatric intensive care unit' (1994) 164 British Journal of Psychiatry 826.
 E Fottrell, 'A study of violent behaviour among patients in psychiatric hospitals' (1980) 136 British Journal of Psychaitry 216.
 Schizophrenia is a chronic thought disorder which begins early adulthood and the person never regains premorbid levels of functioning. It is marked by prominent hallucinations or bizarre delusions and the person needs to be ill for at least 6 months before receiving a diagnosis of schizophrenia. See American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, (4th ed, 2000) 297.
 G Powell, W Caan, & M Crowe, 'What events precede violent incidents in psychiatric hospitals' (1994) 165 British Journal of Psychiatry 107.
 Fottrel, above n 84.
 Klassen and O'Connor (1998) in P E Mullen, 'The Dangerousness of the mentally ill and the clinical assessment of risk' in W Brookbanks (ed), Psychiatry and the Law — clinical and legal issues (1996).
 Steadman and Felson in Monahan, above n 1.
 Statistically the differences did not indicate a difference.
 J W Swanson, 'Mental disorder, Substance Abuse, and Community violence: An epidemiological Approach' in J Monahan & H J Steadman, Violence and Mental Disorder (1995) 101.
 The Diagnostic and Statistical Manual of Mental Disorders IV TR (2000) is published by the American Psychiatric Association contains the diagnostic criteria for all mental disorders.
 B G Link & A Stueve, 'Psychotic symptoms and the violent/illegal behaviour of mental patient compared to community controls' in J Monahan & H J Steadman Violence and Mental Disorder (1995) 137-59.
 Hallucinations are sensory perceptions in the absence of an actual external stimulus. They can occur in any sensory system of the body: sight, hearing, taste, smell and touch.
 Delusions are generally bizarre thoughts described as fixed, false beliefs that are not share with the members of an individual's culture or religion. The content of delusions can vary for example delusions of grandiosity are beliefs of special powers, abilities, talent or identify e.g. someone may believe they are God. For more information see American Psychiatric Association, above n 85, 91.
 A I F Simpson, P M J Brinded, T M Laidlaw, N Fairley & F Malcon, The National Study of Psychiatric Morbidity in New Zealand Prisons: An Investigation of the Prevalence of Psychiatric Disorders among New Zealand Inmates Department of Corrections (1990) 1999.
 Swanson, above n 91.
 Alcohol and Drug Addiction Act 1966, s 9 allows a committal order be made for the purposes of detaining a person for treatment of a substance addiction in an institution.
 Section 4(d) of the Mental Health (CAT) Act 1992 excludes people who are substance abusers from the assessment and treatment provisions of the Act.
 Personality disorders appear on the second axis in the American Psychiatric Association, above n 85, 91. They are described as enduring pattern of inner experience and behaviour that deviates markedly from the expectation of the individual's culture. They are pervasive and inflexible with onset occurring in adolescence or early adulthood. It is stable over time and leads to distress and impairment. There is a great deal of controversy over whether personality disorders are considered mental disorders. For a more detailed discussion see R F Farmer, 'Issues in the assessment and conceptualization of personality disorders' (2000) 20 Clinical Psychology Review 823.
 Simpson, Brinded, Laidlaw, Fairley & Malcon, above n 96.
 Delusional disorder is usually a chronic illness and symptoms are delusions, although they are not bizarre like schizophrenia, but can occur in real life such as infidelity and illness. Schizophreniform disorder is characterised by delusions, hallucinations and manic or depressive syndrome, although unlike schizophrenia the symptoms only have to last for 1 month instead of 6 months. Schizoaffective disorder has the characteristic features of schizophrenia, but the duration of the mania or depressive syndrome is not relative to the duration of these features. See American Psychiatric Association, above n 85, 297-345.
 Simpson, Brinded, Laidlaw, Fairley & Malcon, above n 96.
 Posttraumatic stress disorder occurs when a person has witnessed, exposed, experienced or been confronted with a traumatic event which involved actual or threatened death or serious injury. The person's response to the event involved intense fear, helplessness or horror and the even is re-experienced in a number of ways such as dreams, flashbacks or physiological reactions. See American Psychiatric Association, above n 85, 463.
 Obsessive compulsive disorder is characterised by recurrent and persistent thoughts, impulses, or images that are experienced, at some time during the disturbance, as intrusive and inappropriate and cause marked anxiety and distress. The person attempts to ignore or suppress such thoughts, impulses, or images with some other thought or action. See American Psychiatric Association, above n 85, 456.
 Farmer, above n 100.
 S Fazel & J Damesh, 'Serious mental disorder in 23,000 prisoners: a systematic review of 62 surveys' (2002) 359 The Lancet 545.
 Psychotic disorders are characterised by episodes of psychosis which are severe mental disturbance indicating gross impairment in reality testing. Manifestations of psychosis include delusions, hallucinations, and looseness of associations, grossly disorganised and bizarre speech or behaviour.
 Taylor & Gunn in Mullen, above n 2.
 The positive symptoms of schizophrenia are delusions, hallucinations, disorganised speech and extremely disorganised behaviour
 P Lindqvist & P Allbeck, 'Schizophrenia and Crime. A longitudinal follow-up of 644 schizophrenics in Stockholm' (1990) 157 British Journal of Psychiatry 345.
 P E Mullen, 'The Dangerousness of the mentally ill and the clinical assessment of risk' in W Brookbanks (ed), Psychiatry and the Law — clinical and legal issues (1996) 97.
 S Wessely, 'The epidemiology of crime, violence and schizophrenia' (1997) 170 British Journal of Psychiatry 8-11
 P E Dietz, 'Mentally disordered offenders: patterns in the relationship between mental disorder and crime' (1992) 15 Clinical Forensic Psychiatry 539.
 This is consistent with the Simpson, Brinded, Laidlaw, Fairley & Malcon, above n 96, finding that 59.6% of New Zealand's prison population had a diagnosable personality disorder.
 Monohan and Steadman, above n 91.
 American Psychiatric Association, above n 85, 297-345.
 Monohan and Steadman, above n 91.
 For more information see R F Farmer, 'Issues in the assessment and conceptualization of personality disorders' 20 (2000) Clinical Psychology Review 823.
 P G Nestor, 'Mental disorder and violence: Personality dimensions and clinical features' (2002) 159 The American Journal of Psychiatry 1973.
 Although psychopathy is not a clinical diagnosis, its construct has a great deal of support. For more information see N J Wilson, 'Adult psychopathic personality disorder: treatment issues and guidelines' (Paper presented at the Annual Conference New Zealand Psychological Society, Wellington, New Zealand, August 31 - September 2, 1998).
 Re M  M 716/85 (Unreported, High Court of Wellington, Grieg DCJ, 21 April 1986).
 Section 8.
 Mental Health (CAT) Act 1992 s 2, definitions of 'responsible clinician' and 'clinician'.
 J J Cocozza & H J Steadman, ‘The failure of psychiatric predictions of dangerousness: clear and convincing evidence' (1976) 29 Rutgers Law Review 1184.
 American Psychatiric Association, 'Clinical aspects of the violent individual task force' (Report 8, 1974) 33, n 1.
 Monahan, above n 1.
 Cocozza and Steadman, above n 129.
 s8s U.S 107 (1966).
 J Monahan & H J Steadman, Violence and Mental Disorder (1995) 5.
 C W Lidz, E P Mulvey, & W Gardner, 'The accuracy of predictions of violence to others' (1993) 269 The Journal of the American Medical Association 1007.
 J Monahan, 'Limiting therapist exposure to Tarasoff liability' (1993) 48 American Psychologist 242.
 Mental Health (CAT) Act 1992 (NZ) s 20.
 131 Cal Rprtr 14, 551 P 2d 344(1976).
 B M Dickons, 'Predictions, professionalism and public policy' in C D Webster, M H Ben Aron & S J Huvker, Dangerousness (1985).
 A Buchanan, 'The investigation of acting on delusions as a tool for risk assessments in the mentally disordered' (1997) 170 British Journal of Psychiatry 12.
 Mullen, above n 2, para 56.
 Brookbanks, above n 114, 103.
 Mullen, above n 2, para 56.
 P Bacon, 'Assessing risk; are we being overcautious?' (1997) 170 British Journal of Psychiatry 30.
 P Duggan, 'Introduction' (1997) 170 British Journal of Psychiatry 1.
 Cocozza and Steadman, above n 129, 1088.
 Webster, Ben Aron & Huvker, above n 6, 116.
 W Bingley, 'Assessing dangerousness: protecting the interests of patients' (1997) 170 British Journal of Psychiatry 28.
 Adopted by New Zealand in The New Zealand Bill of Rights Act 1990 (NZBOR Act 1990).
 International Covenant on Civil and Political Rights 1996 Article 2.
 New Zealand Bill Of Rights Act 1990 s 11.
 New Zealand Bill Of Rights Act 1990 s 22.
 For further discussion on arbitrary detention in relation to mental health see J Mclean, 'Forensic psychiatry and the constitution' in W Brookbanks (ed), Psychiatry and the Law — clinical and legal issues (1996).
  2 NZLR 153. See ibid.
 Law Commission Report No 30, above n 4, para 57.
 The Declaration of the Rights of the Disabled Persons was adopted by the United Nations General Assesmbly in 1975.
 The Human Rights and Equal Opportunity Commission Annual Report 2001 - 2002.
 Adopted by the United Nations Commission on Human Rights in 1991.
 In the Matter of PH  Ca & T 19/95 (Unreported, Invercargill District Court, 31 January 1996).
 H Prins, 'The clinical assessment of dangerousness' in R Bluglass & P Bowden (ed), Principles and Practice of Forensic Psychiatry (1990) 492.
 New Zealand Bill Of Rights Act 1990 s 8
 International Covenant on Civil and Political Rights 1996 Article 9.
 New Zealand Bill Of Rights Act 1990 s 5.
 J Monahan, Predicting Violent Behaviour (1981) 96.
 J Reed, 'Risk Assessment and Clinical Risk Management: the lessons from recent inquiries' (1997) 170 British Journal of Psychiatry 4.
 N Pollock & C Webster, 'The clinical assessment of dangerousness' in R Bluglass & P Bowden (eds), Principles and Practice of Forensic Psychiatry (1990) 489-97.
 Bluglass and Bowden, above n 160, 505.
 Monahan, 'Limiting therapist exposure to Tarasoff liability', above n 136.
 Reed, above n 165.
 D A Andrews & J Bonta, The Psychology of Criminal Conduct (2nd ed, 1998).
 Monahan, Predicting Violent Behaviour, above n 164, 29.
 Webster, Ben Aron & Huvker, above n 6, 116.
 Webb & Treadwell, above n 53.
 (1994) 11 FRNZ 354; (1993) 10 FRNZ 545, 547.
 NZFLR 28.
 Ibid, 33.
 Mental Health (CAT) Act 1992 s 28(3).
 The Long Title of the Mental Health (CAT) Act 1992.
 Wessely, above n 115.