The assessment of potential for suicide is a large part of a psychiatrist's work. Within the past decade these footnotes have popped up even in algorithms that are designed to guide decisions about psychopharmacology. They have always been present in treatment guidelines for most major psychiatric disorders. They are a major cause of anxiety for practitioners, because we all know that our predictive capacity is low, but more importantly we know that unlike Internists and Surgeons we have limited access to the resources necessary to address the problem.
Considering for a moment a typical outpatient crisis, for a person known in the practice with depression who is now clinically changed in an office assessment with suicidal thoughts, the options are very limited. In the case of an assessment of extreme risk, inpatient treatment may be offered. If the patient has any inpatient experience at all, he or she knows that inpatient units are generally miserable places where very little active care happens and where they are enclosed with a number of very ill patients. They may also know that there are an arbitrary number of hoops that must be jumped through in order to be discharged and that as a result they may be in that environment much longer than they need to be. They may also have had a typical experience of the inpatient psychiatrist not talking with their outpatient psychiatrist and making a number of abrupt medication changes that are neither necessary or indicated based on their brief familiarity with the case. For those reasons and also because most people are averse to sitting in hospitals - people will balk at the suggestion of inpatient care.
The suggestion of inpatient care also assumes there is the availability of that option in the community. Most hospitals in any given state do not offer inpatient psychiatric care. That level of care has been discriminated against at a political and financial level for 30 years and as a result hospital services and inpatient psychiatric beds have contracted in an expected manner. Patients are often transferred hundreds of miles within states to reach these beds. A related issue is the availability of electroconvulsive therapy (ECT) for severe depressions. In the case of high risk depression it may be the only effective option. Many states have no availability of this option for patients who need it.
The suggestion of emergency department (ED) care is an even bigger dead end. The vast majority of ED care is provided by mental health professionals who are not psychiatrists and who are making triage decisions that ED physicians can sign off on. The wait is hours and if a high risk determination is made it might be days in the ED before any disposition can be made. Patients are often discharged on the basis of whether their suicidal ideation is chronic or not and whether they are saying that they have a suicidal thought and an intent to harm themselves
right at the time of the assessment.
All of the above factors generally place the burden of care back on the original treating psychiatrist, even when the risk is higher that he or she would want. Most psychiatrists recognize that if they are treating very ill patients, there needs to be an element of
acceptable risk in order to provide treatment and the hope of recovery. Psychiatrists realize that resources are severely rationed, that their patient needs acute treatment, that the patient will only accept certain treatment, and that there is a societal expectation of medical paternalism if the patient in not able to remain safe. The psychiatrist and the patient are frequently operating in this zone of acceptable risk that is perceived very differently by others. Family members are the clearest case in point. Like society in general, many family members have their biases when it comes to psychiatry. Many have been instrumental in discouraging their family member from getting treatment. In some cases they have interfered with treatment and suggested that the family member discontinue treatment or throw away any medications that they have been taking. At the same time, family members generally favor a
zero risk treatment environment. They would prefer that the patient's suicidal thinking resolve completely so that there is no risk that they will attempt suicide. They see suicidal thoughts as controllable and the product of a series of correctable decisions. They don't understand why the thoughts just can't be turned off by the patient, their psychiatrist, or in some cases - the medication the patient is taking. In extreme cases, they may threaten litigation if the patient suicides or makes a suicide attempt implying a volitional and controllable basis for suicidal thinking.
An understanding of human consciousness provides a way to analyze this situation and the misperceptions about suicidal thinking and behavior. The predominant model of risk assessment for both suicidal ideation and aggressive potential is risk factor analysis. It generally proceeds from an elaboration of the specific thoughts to past history of attempts, availability of lethal means, diagnostic risk factors, past history and analysis of attempts, and specific demographic risk factors associated with suicide attempts. Many texts like the
Harvard Medical School Guide To Suicide Assessment and Intervention have detailed approaches to the problem and further conceptualizations like
proximate and
distal risk factors. In an
earlier post, I discovered a checklist of risk factors that looked at the issue of
Increased Reasons or
Decreased Barriers to suicide called the Convergent Functional Information for Suicide Scale (
CFI-S). Many institutions these days prefer the Columbia Suicide Severity Rating Scale (
C-SSRS). All of these methods are essentially based on risk factor analysis. Some are more elaborate than others. There all estimate risk to one degree or another and in some cases factors that mitigate risk. I won't debate the merits of these methods here. All that I want to say about them is that after the risk has been estimated, the psychiatrist may still be working with a high risk patient who is unpredictable in both an inpatient and an outpatient setting. Interventions can be initiated to reduce the risk, but there is no assurance that they will be effective fast enough to prevent a suicide attempt. In many clinics where a standardized approach like this is used with an electronic health record and a cutoff score is used to determine risk, a psychiatrist may find the patient visits being flagged for months or longer based on these numbers.
Is there another model that might supplement or improve upon the risk factor analysis models? For about 15 years now, I have been looking at a model that considers the basic question of what happens when a human conscious state shifts from one that would never contemplate suicide to one that does or in the extreme state proceeds rapidly to suicide. The usual psychiatric model considers the development of an illness state like depression, bipolar disorder, borderline personality disorder, or alcoholism as a precursor state. The cognitive changes, like depressogenic thinking seen in the precursor states are seen as the basis for suicidal thinking. The intervention is generally directed at reversing the precursor state, acutely structuring the environment as necessary for safety, and direct verbal interventions to address the suicidal thinking.
It is possible to explore with people the transition of their conscious state from a person without suicidal thinking to a person who develops suicidal thinking and consider a broad array of associated factors. Just being able to recognize that this transition has occurred is an important part of any evaluation and intervention. Some people are so severely depressed that it seems like the suicidal thinking has been there forever. They can barely recognize a time when they felt better or were not suicidal. In many cases they are preoccupied with
existential factors such as meaningfulness of their life, personal freedom, and of course life and death - factors that they were only peripherally focused on during their daily life. In some cases they are important psychodynamic factors such as the death of a family member or friend from suicide. I speculate that many psychiatrists have heard of or been involved in situations like this. These events are also described in some of the psychoanalytic literature but not necessarily the risk factor analysis literature. John Bowlby described some examples in his book
Loss:
"From many examples from Cain and Fast we select two: one eighteen year old girl who drowned herself alone at night in much the same fashion as had her mother many years earlier; the other a thirty-two-year old man who drove his car over the same cliff that his father had driven over twenty-one years earlier. Some of these individuals, it seems, had lived for many years with a deep belief, amounting to a conviction, that they will one day die by suicide. Some quietly resign themselves to their fate. Others seek help." (p. 389)
Of course the complexity of this situation is much greater than Bowlby can capture in his brief explanation. Just at the psychodynamic level there is the issue of identification with the parent and their suicidal actions. Do they believe that they have a deeper understanding of the parent's action and consider them to be logical? Have they incorporated this into their worldview and consider it to be their fate? At the neurobiological end of the spectrum, is it a case of straight genetic vulnerability to suicide or were there epigenetic factors related to a severe disruption of the home environment that the suicide of a parent can cause? Do they remember an event or series of events during childhood when the affected parent seemed to transmit a tendency to anxiety or depression directly to them? All of these are relevant considerations when examining what is going one at the conscious level in an individual who has become suicidal.
Elementary risk factor analysis also benefits from the broader perspective of considering other conscious factors. It allows for an exploration of additional degrees of freedom. For example, the issue of firearms possession and the elaboration of risk often depends on possession and risky behavior with that gun. But what constitutes risky behavior and what needs to be asked? Have you had the gun in your hand when you were thinking about suicide? Was the gun loaded? Did you actually point the loaded gun at yourself? What were you thinking about at that time? The questions and responses cannot be anticipated in a linear risk factor analysis or algorithm.
A nonlinear consciousness approach can also incorporate an informed consent approach to provide active feedback to the patient on the current risk and the limitations of treatment. This often opens a window into the dynamics of how the patient conceptualizes risk and their ability to work with the psychiatrist in minimizing it. A more linear assessment often takes on the structure of the psychiatrist trying to guess whether or not the patient is going to kill themselves and leaves the patient as a relatively passive participant. A consciousness based approach recognizes that the patient has entered at least partially into a conscious state that is foreign to them, less predictable, and represents some degree of risk to them. They need to hear very clearly that they and the psychiatrist need to work together to restore their baseline conscious state and reduce risk in the meantime. The process encourages them to not leave the interview leaving something that is potentially important - unsaid.
George Dawson, MD, DFAPA
References:
1: John Bowlby. Attachment and Loss - Volume III: Loss - Sadness and Depression. Basic Books. New York. Copyright by the Tavistock Institute of Human Relations. 1980, p 389.
2: Douglas G. Jacobs (ed). The Harvard Medical School Guide to Suicide Assessment and Intervention. Jossey-Bass Publishers; San Francisco. Copyright by the President and Fellows of Harvard College. 1999.
