In the setting up of a multi-agency suicide prevention task force in northern England (further details available from the authors) we wanted to know the likely incidence and type of contact with the police prior to suicide. In one of the few studies addressing this topic, Murphy et al (Reference Murphy, Clendenin and Darvish1971) acknowledged that roughly 80% of a police officer's activity is devoted to social service, including dealing with suicidal individuals, but that few police officers receive training in dealing with such people, although this contact provides significant opportunities for intervention.

METHOD

One of us (N.J.) identified all residents of County Durham and Darlington who had died between 1 January 1999 and 31 December 2001, with a suicide, open or accidental death verdict. This information was collected from two sources and cross-referenced to reduce the possibility of cases being missed. First, a systematic search was made of the mortality register. Second, information was collected directly from the local coroner's office. All cases in which a suicide verdict was recorded by the coroner were included in this study. Cases with open or accidental verdicts were scrutinised by two of us and a decision made on whether these deaths should be included in this study as ‘probable’ suicides; see Linsley et al (Reference Linsley, Schapira and Kelly2001) for a discussion of this problem.

A long time can elapse between death and the issuing of a coroner's verdict. Therefore records for the 24 months following December 2001 were scrutinised to ascertain ‘late’ verdicts where the death had occurred between 1 January 1999 and 31 December 2001. Coroner's inquest reports were analysed and information collected in a semi-structured format following a tool used in previous research (Reference Linsley, Schapira and KellyLinsley et al, 2001; Reference Schapira, Linsley and LinsleySchapira et al, 2001). Cases were then cross-referenced with the local police force computer records for the area, concentrating on individuals seen within 3 months of death (90 days or less), considered to be a realistic time frame for intervention. This police force covers the same geographical area as the coroner's records. Evidence of contact was recorded in a semi-structured format. All data were entered into a Statistical Package for the Social Sciences (SPSS), version 14 for Windows database alongside existing data detailing contact with other agencies.

RESULTS

A total of 133 suicide verdicts were recorded in the 3-year period; 43 open and 29 accidental death verdicts were included as ‘probable’ suicides. This gave a total of 205 probable suicides within the 3-year period.

Twenty-four individuals (12%) had had contact with the police within 3 months of death as victims of crime and 24 individuals (12%) had been arrested as alleged perpetrators of crime (Table 1). Seven individuals had been both a victim of crime and an alleged perpetrator in the 3-month period, leaving an actual total of 41 people (20%) who had been in contact with the police. A wide range of violent and non-violent crime was evident (further information available from the authors).

Table 1 Contact with police during the 3 months prior to suicide

	All suicides	Reported victim of crime within 3 months of suicide	Arrested for alleged offence within 3 months of suicide	Impending court appearance
n (%)	205 (100)	24 (12)	24 (12)	21 (10)
Male, n (%)	152 (74)	16 (67)	21 (88)	18 (86)
Mean age, years	42.1	36.0	35.9	38.8
Median contact time within the 3 months before suicide, days	NA	30	17	NA
GP contact within last 3 months, n (%)1	98 (48)	11 (46)	9 (38)	8 (38)
Contact with mental health services within last year, n (%)2	61 (30)	11 (46)	7 (29)	11 (52)
Contact with A&E department within last 3 months, n (%)3	24 (12)	3 (12)	5 (21)	3 (14)

Among the 41 cases with police contact 17 (41%) had also seen their general practitioner (GP) within the same period (although we were unable to access general practice records in 6 cases). Four cases had no diagnosis recorded at the last consultation, but all four patients had received either antidepressant or benzodiazepine medication. The main diagnoses in the remaining 13 cases were all related to physical health. No particular condition predominated.

Six (15%) individuals had attended a local accident and emergency department in the same period: 3 for self-harm and 3 for other reasons. Almost a third (32%; n=13) had a history of local mental health service contact in the year prior to suicide. Diagnoses included depression (n=6), personality disorder (n=3), alcohol problem, adjustment disorder and anxiety disorder (all n=1). One case had no diagnosis.

In addition to the main findings, 21 cases had impending court appearances (for criminal matters), making it likely that the individual had ongoing police contact. Of these, 14 had been arrested within the last 3 months, and of these 14 people, 6 had also reported crime within that period. Contact with health agencies is summarised in Table 1.

DISCUSSION

The results show a high rate of contact with police for either arrest or reporting crime. Seven cases additionally had impending court appearances. Thus, in nearly 25% of all cases of suicide the person had had a criminal justice contact within 3 months of their death. The National Confidential Inquiry into suicides in the UK found that in just 24% of suicides and open verdicts the person had been in contact with mental health services within 1 year of death (Reference Appleby, Shaw and SherrattAppleby et al, 2001): thus, as many people see a police officer within 3 months of their suicide as see a mental health professional within 12 months. This makes it imperative that suicide prevention addresses this area.

We accept that the police see a huge number of people and therefore the rate of suicide per contact would be low. However, the same could be said of contact with accident and emergency, mental health and primary care services. This should not deter agencies from trying to prevent suicide.

The median time of contact with police (for those with contact in their last 3 months) was 17 days for alleged perpetrators and 30 days for victims of crime (Table 1). Further analysis of police contact up to 1 year before suicide identified 34 additional contacts (14 reporting crime, 20 arrested) – that is, over a 9-month period. This suggests that the frequency of police contact increases nearer to the time of suicide and warrants further research, as does the nature of the contact and how this relates to health service interaction.

Perhaps most surprising was the more or less equal distribution between victims of crime and alleged perpetrators. It is noteworthy that a comparatively greater proportion of females reported crime than were alleged perpetrators, but nevertheless two-thirds of those reporting crime were male. Services have been developed jointly by criminal justice agencies and mental health services but these have not focused on victims of crime. This study found that a greater percentage of victims than alleged perpetrators had been in contact with mental health services (46% v. 29%), suggesting victims may warrant a greater degree of liaison between police and mental health services.

Of equal importance is the medical and psychiatric contact these individuals had. Cases of suicide with police contact recorded a higher rate of contact with mental health and accident and emergency services compared with cases of suicide without police contact, and in around 40% of cases the person had seen a GP in the same period. This suggests it is important for police to have some understanding about health contacts, and conversely for health services to be aware of police involvement.

Further thought and work are needed before these findings can be applied generally in practice. However, we believe there is a need to help the police to identify vulnerable individuals, ascertain level of risk and obtain guidance on whom they can contact if concerns are raised. Local protocols between health agencies and police might help. These should include a number of levels of action that apply to victims as well as alleged offenders. This has to be backed up by training programmes for police on recognising suicide risk factors and dealing with suicidal individuals effectively. Furthermore, it should be acknowledged that contact with the criminal justice system itself can have a negative impact on already vulnerable individuals. Whether this contact is the final stressor for some or part of being under existing stress, it is important that all criminal justice agencies – including court diversion schemes, prison authorities and agencies dealing with victims of crime – are aware of this negative impact and the increased risk of suicide.

Finally, we should remember that police have contact with individuals at risk of suicide not included in our research: for instance, individuals who have self-harmed, those expressing suicidal thoughts (e.g. people saved from jumping off bridges) and those detained via police powers under section 136 of the Mental Health Act 1983. Policies should cover these eventualities as well, in particular to ensure that such individuals receive appropriate assessments and do not ‘fall through the net’.