The belief that mental illness is associated with violence is common and cuts across cultures [1] . This is not different in Nigeria; Gureje et al. [2] in a study among the Yorubas of Western Nigeria reported that belief in the association between mental disorders and violence is one of the reasons for the stigma associated with the mentally ill. By providing for the detention/compulsory admission of the mentally ill who are likely to pose a danger to others in the law, the notion that the mentally ill are dangerous is further validated [3] .

Generally it is thought that schizophrenia and other psychoses predispose subjects to homicidal behaviour [4] [5] . However, other empirical studies have suggested that violence and mental illnesses are only weakly associated and that apparent relationship may be coincidental [6] . Notwithstanding such suggestion, there are some support in the literature for the assumption that there is a moderate but reliable association between mental disorders and violence [7] [8] . This increased risk of violent behaviour is clearly evident for specific psychiatric diagnoses and symptom constellations. Schizophrenia and other psychoses are associated with violence and violent offending, particularly homicide [9] . There is a positive and significant correlation between schizophrenic syndromes and increased rate of antisocial and violent behaviour [10] [11] ; this relationship reportedly has both clinical and social significance [12] . Reports have also confirmed the association between psychoactive substance use and violent behaviour [13] [14] . However, research has also shown that the vast majority of violent people are not mentally ill [15] .

Most studies done to assess psychiatric morbidity among criminals in Nigeria were conducted on sentenced criminals who are already in prison, including both violent and non-violent offenders [16] . Though violent crimes have been associated with mental disorders, the prison environment itself may precipitate a psychiatric disorder, thus serving as a confounding factor [17] . This study therefore assessed arrested offenders still in police custody, thus removing the effect of the prison environment as a confounding factor [18] .

The study was conducted in two Local Government Areas (LGAs) of Osun State (Ife Central LGA, and Ife East LGA) in the South-Western part of Nigeria, about 220 km North-East of Lagos. Ife Central LGA has one police station; a Divisional Headquarters for all the police stations in District Zonal Police Command Area, Ife East LGA has two police stations: a “B” Division, and a “C” Division. These three police stations serve as the first major points of call by complainants reporting cases of violence or other crimes.

Ethical clearance was obtained from the Ethical and Research Committee of Obafemi Awolowo University Teaching Hospital Complex. Permission to conduct the study was also obtained from the Commissioner of Police, State Police Command Headquarters, Osogbo, Osun State. The approval was subsequently relayed to the Divisional Police Officers concerned.

The subjects were predominantly males, 367 (91.7%). The age range was from 18 to 54 years; but about two thirds of them, 271 (67.8%), were below 32 years. The mean age was 29.9 years (SD ± 7.3). Other socio-demo- graphic variables of the subjects are shown in Table 1. Assault (simple and aggravated) was the most common complaint against respondents (51.2%), also 235 (58.8%) of respondents agreed to currently using psychoactive substances, many of these 131 (55.8) admitting to using multiple substances (alcohol, tobacco, cannabis) while 41.2% deny the use of any psychoactive substance. Among respondents who agreed to the use of at least one psychoactive substance 74% had a specific psychiatric diagnosis (Table 2).

Out of all the respondents, 36 (9%) screened positive (scored 5 or more) for probable psychopathology on the GHQ. However using PSE to interview all 36% and 10% of those who screened negative, 34 respondents were diagnosed as having a specific psychiatric disorder giving a prevalence of 8.5% for psychiatric morbidity among the offenders. Among offenders with specific psychiatric diagnosis, paranoid schizophrenia was the commonest psychiatric disorder (41.2%), followed by bipolar affective disorder (manic episode) (23.5%), persistent delusional disorder was next (14.7%), unspecified psychotic disorder (11.8%) and schizoaffective disorder (5.9%), while 2.9% of the subjects had depressive episode (Table 3). Figure 1 shows the distribution of psychopathol-

ogy among respondents by their sex.

Table 4 shows the comparison of crimes committed by the subjects with PSE diagnosis with those committed by the subjects without PSE diagnosis. Among the respondents with psychiatric diagnosis, 20.6% committed homicidal offences compared with 7.9% of the apparently mentally healthy subjects. This difference was statistically significant (χ² = 4.77, p = 0.023). Also from Table 4, risk analysis showed that there is a Risk Ratio of 20.6/7.9 (2.6). This means that the risk of committing a homicidal offence by a mentally ill subject is more than double the risk by an apparently mentally healthy subject.

The relationship between violence and mental disorder has interested scholars (psychiatrists inclusive) for centuries. Particular interest has been attached to mentally ill persons committing very violent crimes, especially

^*Multiple: 2 or more drugs (i.e. tobacco, cannabis and alcohol).

^*Fishers’ exact probability test used.

homicide. This has contributed to widespread belief in the society that mentally ill individuals are dangerous, which in turn has contributed to the stigma associated with mental disorders in the society [20] [25] . Nigeria is no exception, Gureje et al. [26] and other researchers have shown that fear of violent behavior by the mentally ill contribute to the stigma associated with mental disorders in the country [27] .

Most of the studies on violent behavior and the mentally ill in Nigeria have been among prison inmates [16] [28] however this study is among offenders yet to be convicted, thus the effect of imprisonment and prison condition as a confounder in the occurrence of mental illness was eliminated [18] .

The finding in this study that majority of the subjects were young adult males (in their 20s and early 30s) is in keeping with earlier findings both in Nigeria and in other parts of the world [15] [28] [29] . We also found that, males predominate among those arrested for violent crimes, this finding agrees with the notion that violent crime is a predominantly male activity, particularly the younger males [29] .

It was observed that assaults, (simple and aggravated) were the predominant violent acts the subjects were alleged to have committed. Abiodun [30] and Ogunlesi et al. [28] reported murder as the predominant crime anomg convicted prisoners. This difference may be explained on the basis of the different setting of these studies, Abiodun [30] and Ogunlesi et al. [28] studied convicted and sentenced offenders, while this study assessed offenders who are basically accused but yet to be convicted or sentenced.

Reports have shown that psychoactive substance use is positively correlated with violent crime [9] [11] [12] . In this study majority (58.8%) of the subjects admitted to currently using a psychoactive substance and more than half of them currently use a combination of psychoactive substances mainly alcohol and cannabis. This finding attests to the association between violent crime and use of psychoactive substance which has been found in many studies [31] [32] .

Among respondents, 8.5% had diagnosable psychiatric conditions while the rest were basically normal and apparently in good mental health. Compared to studies conducted among prison inmates this prevalence rate is low [16] [18] . The rate is also lower than the 12% prevalence of mental disorders in the general population in Nigeria reported by Gureje et al. [2] , the difference may be due to the instrument used which did identify psychoactive substance use disorder.

Another observation was that, among the mentally ill subjects, the most common diagnosis was schizophrenia (41.2%). This agrees with finding in other studies [15] [33] [34] . This study showed that the violent offences most likely to be committed by mentally ill offenders, just like apparently mentally healthy offenders, were non-homicidal offences, especially assaults. But, the proportion of the mentally ill subjects that committed homicidal offences was significantly higher than that of the apparently mentally healthy subjects. Most studies that found a relationship between mental illness and violence, found that, that relationship is more likely with serious violent crimes like murder [28] [35] .

It was found that the risk of committing a homicidal offence by a mentally ill individual is between two to three times that of an apparently normal individual. This suggests that, truly, an association between mental disorder and violence exists, but this association may not be as strong as it is generally believed in the society. This view was also expressed by Mitchell [36] , who after reviewing several studies on the relationship between mental disorder and violent offence, concluded that mental disorders’ contribution to violent crime in the community is generally agreed to be small. Similarly Stuart, et al. [15] concluded in their study that people with mental disorders are not major contributors to police identified criminal violence and that public perceptions of mentally ill persons as criminally dangerous appear to be greatly exaggerated.

The study was conducted in two Local Government Areas in South-Western part of Nigeria. Considering the vast nature of this country, it might be difficult to generalise the findings to the entire country. Also the subjects were interviewed within two days of their arrest, thus ascertaining the sincerity with which they answered the questionnaires may be difficult due to their state of mind at the time. Thirdly, an association between mental disorders and violence was established in this study, however, a causal relationship could not be assumed since the study was not controlled. Another limitation is that the main diagnostic questionnaire (PSE) used in this study cannot diagnose psychoactive substance use disorders, mental retardation and personality disorders which may be common among offenders, thus the list of psychiatric diagnosis is not exhaustive.

Majority of offenders accused of violent crime in this study were found to be mentally healthy and committing non-homicidal offences, especially assaults. Less than 10% of the violent offenders were mentally ill with majority having paranoid schizophrenia. The mentally ill offenders were more likely to commit homicidal offences, especially murder, compared to the mentally healthy offenders. It appears that there exists a weak relationship between mental disorder and violence with the risk of danger posed by mentally ill persons to the society being relative rather than absolute.