Coping with suicide: a perspective from Scotland

Summary

Procedures following suicide differ in Scotland from elsewhere in the UK and we describe the investigation of deaths by procurators fiscal and fatal accident inquiries that may ensue. Higher Scottish suicide rates, and possible reasons for these, are mentioned. Suicide risk cannot be accurately quantified in individual patients but psychiatrists should take the view that good management can collectively reduce the risk among all patients. We comment on practical and emotional issues for clinicians who are coping with the suicide of a patient.

Within the UK, there are some significant differences between countries in official procedures relating to suicide, and this article is intended to complement that of St John-Smith and colleagues (2008, this issue).

National differences

Investigation of deaths in Scotland

In Scotland, procurators fiscal (‘fiscals’) are responsible for investigation of deaths. Investigations are conducted for a number of purposes (Crown Office 1998). These include the need to diminish the risk of undetected homicide or other crime, to eradicate dangers to life and health in the public interest, to allay public anxiety, and to ensure that full and accurate statistics can be compiled. When the investigation is complete, the fiscal sends a report to the Crown Office, where it is decided whether a ‘fatal accident inquiry’ should be held.

Deaths in certain categories must be reported to the fiscal. Box 1 lists those that may arise in connection with psychiatric practice.

Box 1 Categories of reportable death that might be met in psychiatry

1. • Any death due to violent, suspicious or unexplained cause

2. • Any death involving fault or neglect on the part of another

3. • Possible or suspected suicide

4. • Any death by drowning

5. • Any death by burning or scalding, or as a result of a fire or explosion

6. • Any death resulting from medical mishap

7. • Any death where a complaint is received suggesting that medical treatment or the absence of treatment may have been a contributory factor

8. • Any death in legal custody

9. • Any death where a doctor has been unable to certify a cause

Incidents may also be reported to the police, who will then carry out an investigation and submit a report to the fiscal. The fiscal can order a police investigation if this has not already taken place. In the case of in-patient suicides, the police should be immediately informed. As part of their investigation, the police can interview relevant parties such as clinicians.

The investigation by the fiscal will begin by ascertaining the cause of death. The fiscal has control of the disposal of the body until enquiries are complete and has the power to instruct an autopsy. Statements (also known as ‘precognitions’) will be taken from relevant witnesses, who may of course include clinicians. These statements can be taken by a procurator fiscal, procurator fiscal depute or a precognition officer. In the case of suicides by patients under psychiatric care at the time of death, case notes will be obtained as part of the investigation. A police or sheriff officer is sent to obtain these, sometimes with little advance warning. It is therefore important to make a photocopy of all records as soon as possible after the death.

Fiscals normally interview the spouse or next of kin of the deceased. They will be asked if they wish a fatal accident inquiry to be held and this view will be made known to the Crown Office.

A psychiatrist can also be involved in this investigation as an expert witness. A procurator fiscal can commission an expert report into the care and treatment received by the patient as part of their investigation of a death.

The fiscal will usually request reports of internal inquiries carried out by psychiatric services following the deaths of psychiatric patients. Many services are now applying the methodology of root cause analysis to this process. These reviews are sometimes seen as threatening by clinicians, who fear that they may contain information or opinions that could be incriminatory. Our experience is that fiscals usually look for evidence that the service has undertaken a serious scrutiny of the death, that the relevant lessons have been learned and any necessary changes instituted. If the fiscal is satisfied on these points, they may feel that the public interest has been met and there is less need for a fatal accident inquiry. Fiscals sometimes come under pressure from bereaved families to recommend inquiries. An internal review can often give families an opportunity to express concerns and to receive an account from clinicians of the events leading up to the death. This may provide reassurance and diminish the likelihood that families will press for a fatal accident inquiry.

When the investigation by the fiscal is complete, a report is sent to the Crown Office in Edinburgh. The final decision about whether to hold a fatal accident inquiry rests with the Lord Advocate or the Crown counsel acting on his or her behalf. Expert witness reports will make an important contribution to this decision.

Fatal accident inquiries

These are statutory public inquiries and are held in a sheriff court. They often attract detailed coverage in the press. Most suicides will not result in fatal accident inquiries, especially if an expert report has raised no cause for concern: ‘Isolated incidents involving errors of judgment, for example, one-off medical errors … will not normally provide sufficient justification for discretionary FAIs [fatal accident inquiries]’ (Crown Office 2007: p. 16).

If an inquiry is held, it need hardly be said that this can be very stressful for clinicians who have been involved with the patient prior to the death. The process of adjustment to the death is prolonged and made more difficult. Inquiries are often held as long as a year after a death. In the time leading up to the inquiry, the psychiatrist will have to give statements to the fiscal and instruct solicitors acting in their defence and those representing their employing organisation. The fact that an inquiry is being held at all indicates a degree of concern on the part of legal authorities about the circumstances of the death. The clinician faces the prospect of their management being closely scrutinised in a public arena.

The organisation of fatal accident inquiries is not such as to promote equanimity in witnesses. One can expect to be kept waiting to give testimony for many hours or even spend an entire day without being called. The hearing is formal and is presided over by a sheriff. The case is led by the fiscal, who leads evidence from witnesses cited by him or her. These witnesses can be cross-examined by other parties or solicitors and counsel representing them. Relatives of the deceased are entitled to question witnesses and can engage legal representation. The fiscal can then re-examine these witnesses to clear up any uncertainties and ambiguities. Other parties can call additional witnesses and the same process of examination, cross-examination by other parties (including the fiscal) and reexamination takes place. Professional and expert witnesses may therefore be in the stand for several hours at a time.

The employing organisation, such as a National Health Service trust or board, will usually have a solicitor or advocate in court to represent its interests. It is vital that the psychiatrist also has individual representation, which is normally arranged by one of the medical defence unions. Psychiatric care is generally provided by multi-disciplinary teams and a patient will have been in contact with a number of professionals prior to the death. Each of these professionals will usually have legal representation. The role of these lawyers is not to contribute to a consensus or to a dispassionate search for truth. It is, quite explicitly, to protect the interests of their clients.

The main source of professional evidence is the patient's case record (Box 2). A record that is comprehensive, accurate and legible is essential if clinicians are to defend their assessments and actions. An incomplete record can place the psychiatrist in a vulnerable position and healthcare professionals should never forget the adage ‘if it ain't in the case record, it didn't happen’. A full account of the events leading up to a suicide should be recorded in the patient's case notes as soon as possible after the event.

Box 2 Suggested medicolegal actions following a patient's suicide

1. • Make a detailed record in the patient's case notes of events leading up to the death

2. • Arrange for full photocopying of case notes

3. • Inform line manager (e.g. clinical director) and employer's legal advisor

4. • Obtain independent legal advice from defence union or other organisation if there is any possibility of a fatal accident inquiry or litigation

If a psychiatrist has submitted a report as part of the process of investigation, they can expect to be summoned to court as an expert witness and the report will be submitted in evidence. At the stage of preparing a report, the psychiatrist should anticipate that every statement might be subjected to detailed questioning. If they are in any way critical of another professional, they should expect searching and perhaps even hostile cross-examination. The expert should bear in mind the usual rules about stating nothing that cannot be justified by the facts and keeping within one's area of expertise.

After the conclusion of evidence, the fiscal will make a submission to the sheriff covering the circumstances of the death and the evidence that has been heard and will invite the latter to make a determination, which can include the following components:

1. • where and when the death took place

2. • the cause(s) of death

3. • the reasonable precautions, if any, whereby the death may have been avoided

4. • the defects in any system of working that contributed to the death.

The sheriff has no power to find fault with individuals or to apportion blame between persons whose acts or omissions may have contributed to the death.

One focus of the article by St John-Smith and colleagues is the fear that a psychiatrist at an inquest will be faced with the unfair expectation that suicides can be predicted or prevented. They make the very important point that there is no rating scale that is capable of providing useful prediction of suicide. It is unlikely that any psychiatrist appearing at a fatal accident inquiry will face examination along these lines. In our experience, most fiscals and sheriffs are realistic about what can be expected of clinicians and about the limitations of even the best psychiatric practice. When a fatal accident inquiry is ordered this is usually because significant concerns have been raised by family members or in expert reports. Such concerns are nearly always ones that would be shared by conscientious and competent psychiatrists.

National suicide rates

St John-Smith and colleagues quote from the National Confidential Inquiry into Suicide and Homicide in England and Wales (Appleby 2006) the figure of 10.2 suicides per year per 100 000 population during 2000–2004. This document also reports a reduction in the England and Wales suicide rate from a figure of 12 per 100 000 in 1997. The equivalent report for Scotland (which covers January 2000 to December 2005) describes an annual suicide rate in Scotland of 18.7 per 100 000 population. A fall in general population suicides in Scotland will be reported from 19.7 per 100 000 in 1998 to 17 per 100 000 in 2005 (Appleby 2008: pp. 32–35).

Suicide rates in Scotland, especially among men, diverged from those south of the border in the 1970s (Crombie 1990) and have remained higher since then. However, given different national methods of identifying suicides and undetermined deaths, as mentioned above, considerable caution is merited in comparing national suicide rates (Neeleman 1997a). O'Donnell & Farmer (1995) examined 242 deaths, all known to be self-inflicted, on the London underground. Not only was there a significant underestimation of suicides (54 with open verdicts and 33 deemed to have been accidental) but there was striking variation in the frequency of suicide verdicts from one coroner's court to another. More recently, Gosney & Hawton (2007) found that of 14 deaths by hanging among young people, all of which were agreed to be suicides by a panel of psychiatrists, coroners in West Yorkshire classified 3 as suicides, 5 as open verdicts and 6 as deaths by misadventure. Especially when there has been a recent increase in male suicides by hanging and strangulation in England and Wales (McClure 2000) and an even more striking increase in Scotland (Stark 2004), how such deaths are classified is clearly highly important in assessing whether or by how much Scottish suicide rates are higher. That being said, we are unaware of any equivalent studies on the verdicts of procurators fiscal, and thus we should probably continue to assume that Scottish rates are indeed higher. It does seem possible, interestingly, that when Scots move to London they take with them an increased likelihood of dying by suicide (Neeleman 1997b).

Differences between England and Scotland

Crombie (1990) discussed possible reasons for divergent suicide patterns on either side of the Scotland/England border. He ascribed the differences to a ‘complex but unknown set of social changes’, and the situation has probably become no clearer in the intervening years. Drinking patterns may well be of relevance, and Brady (2006) has ably reviewed the relationship (from a personal to a national level) between alcohol misuse and suicide. Certainly, as a proxy indicator of rates of alcohol misuse, rates of liver cirrhosis have risen more steeply over recent years in Scotland than they have in England and Wales (Leon 2006). However, teenagers in Scotland may be bucking national stereotypes by drinking rather less than their contemporaries in England and Wales (Crawley 1997). Social deprivation is strongly associated with suicide, and in Scotland suicide rates cluster with deprivation to the extent that suicide can be regarded as ‘geographically contagious’ (Exeter 2007). The highest rates of suicide in Scotland, especially for men, are found in remote rural areas (Stark 2004) and remote rural residence is more common in Scotland than in the more densely populated remainder of the UK. Whatever the reasons might be for national differences in suicide rates within the UK, they certainly seem to merit further study since important factors relevant to suicide prevention may emerge.

Suicide risk and its management

Certainly in the context of giving evidence at an inquest into a death by suicide, St John-Smith and colleagues appropriately highlight the great difficulties of predicting suicide among psychiatric patients. Although there may be ways of slightly improving assessments of risk (Owens 2005; Cooper 2006), their calculations put the matter into perspective. To present their figures in a slightly different way, it would be necessary (if entirely impracticable) to admit 266 people, perhaps each for as long as a year, in order to prevent one death by suicide after applying Cooper and colleagues' (2006) risk assessment strategy. Risk assessment scales also have the potential disadvantage that staff may cease to consider the possibility of suicide in those deemed, through this fallible process, to be at low risk.

The authors highlight the conclusions of Appleby and colleagues regarding ‘preventable’ suicides and note the lack of scientific rigour in the process of determining ‘preventability’ (Appleby 2006). For example, Appleby and his team concluded that deaths of patients by suicide following hospital discharge but before follow-up were necessarily preventable although, among all suicides during the 3-month post-discharge period, 58% had been seen by a psychiatric professional in the preceding week. The questionnaire sent to consultants by the National Confidential Inquiry following a patient's suicide contains the question: ‘In your opinion, could the suicide have been prevented?’, to which there is only a ‘Yes/No’ response option. This false and simplistic dichotomy, particularly when it is amplified in the subsequent report, does a disservice to psychiatrists. If an influential report written by colleagues implies that the suicides of patients under the care of psychiatric services can be categorised straightforwardly as preventable/avoidable, it renders conveying the subtle aspects of suicide risk and prediction very difficult, notably when attempting to explain the complexities to a coroner or at a fatal accident inquiry. In this context, it is worth noting that, even among psychiatrists, the perception of risk of suicide or homicide is increased by ‘hindsight bias’ if rated when thinking that an adverse outcome has occurred (LeBourgeois 2007).

Medication and suicidality

One of the potentially unfortunate results of this dichotomous slant on whether or not suicides are preventable is that psychiatrists may profess impotence in order to avoid blame. This process may be occurring when St John-Smith and colleagues discuss ‘medication and suicidality’. There is surely pretty unequivocal evidence of the benefits of lithium in reducing suicide risk in affective disorders (Cipriani 2005). It is also highly likely that clozapine has a specific antisuicidal effect in people with schizophrenia (Meltzer 2003). The possibility that antidepressants may increase suicidality in some patients cannot be dismissed (Cipriani 2007) but is hopefully now in a much more appropriate perspective (Simon 2007) and should be viewed against the growing evidence of their antisuicidal properties (Tiihonen 2006; Gibbons 2007). Indeed, given that depression is by far the most common condition associated with suicide and that the majority of depressed people who complete suicide are untreated at the time of their deaths (Gibbons 2005), it may behove psychiatrists to proactively encourage the detection of depression and its treatment with antidepressants. In general, as experts in the diagnosis and management of psychiatric disorders, we surely believe that our efforts are of benefit to the patients we see. As a corollary, it is illogical to suppose that patients' suicides are inevitable events that will occur irrespective of the quality of care we provide.

Quantifying risk

It is crucial for psychiatrists to view suicide risk as a continuous and not as a dichotomous variable, and as one that applies to all of their patients. In essence, all of our patients are at increased risk, but we cannot predict, with any likelihood of accuracy, which ones will die by suicide. The pragmatic approach, therefore, to suicide prevention is to aspire to clinical excellence with all patients (Eagles 2001) while practising in a culture of suicide awareness. In so doing, and unfortunately this is impossible to quantify, we shall probably prevent many deaths; at the same time we should be able to mount a logical defence to suggestions that any suicide may or should have been prevented by our service.

Coping with suicide

We would like to add a few comments on more general aspects of how psychiatrists might cope with the suicide of a patient. Other helpful sources of advice on how services and clinicians might respond include Rose (2000), Hodelet & Hughson (2001) and Campbell & Fahy (2002).

We agree about the central importance of a multidisciplinary meeting involving those who have contributed to the patient's care, both for information gathering and for mutual support. However, St John Smith and colleagues suggest that this meeting should be used to consider the appropriateness of the management plan and any changes that might be made to reduce the risk of suicide in other patients. Our experience is that such considerations should be deferred until a later date. In the early stages, clinicians may be struggling with a sense of disbelief and loss of control, and a fear that there will be a repetition of the tragic event. This may be followed by a stage characterised by feelings of anger, guilt, anxiety and loss of confidence (Little 1992). Judgements about how a patient was treated and whether there is a need for change are likely to be more objective when everyone has had a chance to calm down.

Over recent years, there has been an appropriate increase in the attention given to the emotional responses of psychiatrists to suicides among their patients. The largest survey was conducted in Scotland (Alexander 2000; Dewar 2000). In the context of this article, in terms of their personal response to events, it is noteworthy that 15 of the 31 consultants who had attended a fatal accident inquiry reported this to have been either ‘unhelpful’ (for 8) or ‘very unhelpful’ (for 7). By contrast, of the 83 consultants who had experienced an internal critical incident review, this was deemed unhelpful/very unhelpful by only 9 respondents, while 56 (69%) rated the experience as helpful or very helpful (Alexander 2000). It is also reassuring that, when medical defence organisations were involved, this was deemed helpful or very helpful by 76% of respondents (Alexander 2000).

In essence, it is important to appreciate that it is usual for psychiatric professionals to experience emotional trauma following the suicide of a patient, and to expect that support will be required by everyone involved, irrespective of their level of seniority (Alexander 2000; Dewar 2000; Courtenay 2001; Ruskin 2004). All of these papers give useful insights and advice relevant to psychiatrists in coping with the emotional aftermath of a patient's suicide, and in this area the paper by Campbell & Fahy (2002) may be found to be especially helpful.