Criticisms answered

The assessment of suicidal ideation, although it is a fundamental element of any comprehensive attempt to assess risk, often seems to be taken for granted and not described in any detail even in research literature. The procedure itself has been subjected to criticism and this needs to be answered.

There is no reason to believe that appropriate enquiry concerning the presence of suicidal ideation increases suicide risk. Nor is a person who volunteers to having suicidal ideas necessarily at low risk. Exploration of suicidal ideation is not an isolated exercise. Throughout it must be carried out in full knowledge of the total clinical picture and integrated into the whole therapeutic effort. It is a sensitive empathic reaching out to encourage sharing of distressing and often frightening ideas, as dynamic as any other form of intervention can be. It should not be a questionnaire-type Yes/No interrogation. It should involve progressive exploration of suicidal ideas themselves, their intrusiveness, frequency and acceptance as reasonable, and also of feelings of loss of control, perhaps part of delusional thinking, or of seeing suicide as the only way out of an impossible life situation. The role of ambivalence and personality traits tending towards the impulsive should never be ignored. All this requires gentle patience and warm concern. It is not intended to be a matter of focusing on a patient's weaknesses rather than strengths: it is simply a guide that can help us make things safe. At all points we need to remember that it is for the benefit of the patient and not, for example, just a way primarily of coping with our own anxieties concerning adverse sequelae if things go wrong. A confident, realistically encouraging attitude is essential: an ambivalent patient will soon pick up doubts on our part and the whole exercise can then become counterproductive. It is never appropriate as a carer to identify with a patient's despair. It needs to be recognised that aggressive behaviour can occur in parallel with significant suicide risk. Clarify any risk to others. A brusque hurried approach can leave the patient feeling stigmatised and undervalued, as well as encouraging denial of intent or even encouraging suicidal motivation. Discussing together any reasons why denial may occur and how to mitigate them, being available and offering urgent help at any time should suicidal thoughts recur, both represent a mark of respect for the patient's intrinsic worth.

Although all of this is a complex and difficult clinical task it would be a mistake to lose sight of the important positive part that discussing suicidal ideation can have in caring for suicidal individuals. To carry it out properly is comparable to having responsibility for dispensing medicines: the results of maladministration of either could be dire. So, it is reasonable to expect that each psychiatric hospital ward should ensure that all clinical staff have been well trained in assessing suicidal ideation. Appropriate privacy is important if this sensitive and complex task is to be carried out reliably.

Allowing for variability in severity of intent

To base assessment of ongoing risk on the individual’s mental state during a single interview is clearly likely to be highly unreliable. It can mislead not because suicidal ideation is intrinsically valueless in prediction and should be discarded, but rather because such an approach makes no allowance for the marked short-lasting and ongoing variation in severity of intent that is a common feature of the suicidal state of mind. The following case series and studies illustrate this in suicidal individuals during the weeks before death.

Over a 10-year period (1968–1978), 12 deaths by suicide occurred among patients who were under my sole consultant care, either while they were resident in a single psychiatric hospital ward or within 2 months of discharge. They had all expressed suicidal ideation. I knew each of them well, day in, day out, usually over several weeks, yet in spite of my close continuing contact with them and the majority reassuring me about their safety, they ended their lives. In an attempt to understand all this further, I put all their case records together and looked at them as a whole.^{Reference Morgan8} A half showed marked variability in severity of suicidal intent, in some even over the course of a single day, with short episodes of corresponding misleading clinical improvement in which suicidal ideas were denied. Variable contact and unresolved stress factors and impulse seemed relevant. For example, one patient happily organised a charity walk from the hospital only to take her own life the following weekend when, still apparently well, she went home on leave to be confronted with an unresolved domestic crisis.

Two later case series consisted of all psychiatric in-patients in the City of Bristol during two separate periods (1982–1984 and 1991–1993) who died by suicide either from hospital wards or within 2 months of discharge.^{Reference Morgan and Priest9,Reference Morgan and Stanton10} In the two studies, 52 and 61% respectively showed episodes of misleading clinical improvement. Stress was a common background to relapse: 44 and 50% respectively killed themselves when clinically improved but with stress factors unresolved.

Although the number of cases in these three studies was small (12, 27 and 18 respectively) and they bridged a time span of 30 years, the findings were derived from detailed observations at ward level by clinicians responsible for the patients’ clinical care and were remarkably consistent. They confirmed that short-term variability of intent exists in a considerable proportion of suicidal patients. This reinforces the view that to judge assessment of suicidal ideation as a predictor of ongoing suicide risk merely on its efficacy when recorded in a single interview ignores a common clinical feature of the suicidal state of mind, namely its variability in severity of intent. To put it bluntly, it asks the impossible of it.

The following discussion focuses particularly on out-patient work but its principles apply to the in-patient situation too. We need to ask: In the face of such practical difficulties what can we do to make prediction of suicide more reliable so that we can rely on it over the short term until the next out-patient interview?

Out-patient care of suicidal individuals: guidance on assessment of short-term ongoing risk

Remember that predicting ongoing risk of suicide merely on the basis of the patient's mental state at the time of a single interview can be very unreliable. Evaluation of suicidal ideation and intent should be more broadly based and cover the whole of the follow-up period. Remember that ongoing day-to-day variation in severity of intent can be marked. Always give serious consideration to any evidence of suicidal intent, expressed or otherwise. Keep in mind the possible reasons specific to each case why relapse might occur, as well as the overall risk assessment picture. An important trigger for relapse is stress, particularly stress that has previously been present and has not been resolved. Try to assess the likelihood that difficulties, stress related or otherwise, will recur before the next appointment and whether suicidal ideation will complicate them. Overall evaluation should invariably take such anticipated risk into account. Clarify what urgent help could be made available in such a forthcoming crisis, ensure that it would be acceptable to the patient, work through any hesitation expressed regarding seeking that help and review the appropriateness as well as any security provision related to prescribed medication. All this should be integral to the assessment process.

Implementing the suggested guidance

This guidance is put forward in the hope that it might improve our ability both to predict suicide risk in the short term and to target more effectively the ongoing treatment we offer. It is no more than a care plan, but one that not only focuses on the present: it also insists on searching for potential risk at some point in the future. According to it, no assessment of ongoing suicide risk would be regarded as complete without such a ‘two-take’ approach covering the follow-up period and aimed at what amounts to a moving target. This allows plans to be laid in advance for an acceptable form of urgent help should problems recur. Predicted suicidal ideation at any level of severity without stratification and particularly ideation that is triggered by a stressful event should be taken as sufficient to indicate significant continuing risk.

As mentioned above, variability in severity of intent has been found to occur in about half of patients with suicidal ideation who proceed to suicide.^{Reference Morgan8,Reference Morgan and Priest9,Reference Morgan and Stanton10} In the remaining patients, environmental stress factors that presumably trigger such variability are presumably either absent or do not cause significant problems for the patient. Does the proposed guidance therefore have value only for half of patients? Not necessarily. It is possible that, in other patients, careful anticipatory enquiry might reveal a need to plan help for problems that otherwise would have remained unexpected.

From what has been presented here it is clear that the process of predicting suicide risk even in the short term is a difficult clinical exercise. Regarding someone whose stress factors have not been fully resolved, a letter to the general practitioner (GP) taking all relevant issues into consideration might read as follows:

Great care should be taken before one is ever tempted to suggest that ongoing risk is totally absent in someone who has experienced suicidal ideation in the recent past. In such a situation a letter to the GP might read along the following lines: