Mental Health Risk Assessment: Traps to Avoid in Day-To-Day Clinical Work
The emphasis on risk assessment in psychiatry has increased over the past four decades as a consequence of complex socio-political factors, including media pressure after a small number of catastrophic events. [1]
Given risk assessment is a quotidian task, it is important that psychiatrists have an awareness of the limitations of risk assessment in psychiatry. In this article, we outline the limitations of risk assessment and propose five key areas to consider when undertaking risk work in psychiatry: the evidence; values; legal constraints; operational constraints; and documentation.
PRACTICAL PROBLEMS WITH RISK ASSESSMENT AND MANAGEMENT
Use of invalid or insufficient information
Structured professional judgement, which combines validated risk assessment tools and professional judgement, is the current risk assessment gold standard. [2]
Structured professional judgement (SPJ) requires the psychiatrist to synthesise a variety of independent risk factors, weigh these factors, and make a judgement. Such a task requires time.
When time is sparse, as is often the case in publicly funded mental health services, most psychiatrists utilise a pragmatic, intuitive, quasi-Bayesian approach to risk assessment.
With this approach, an initial opinion about risk – usually documented as ‘low’, ‘medium’ or ‘high’ by a mental health clinician or registrar – is updated and modified as more objective information becomes available.
This is variably supplemented with the use of risk assessment documents or ‘tools’ (generally site-specific checklists rather than the evidence-based SPJ guidelines), depending on service policies.
In busy mental health services, risk documents tend to be completed using ‘System 1’ thinking. ‘System 1’ thinking, as described by psychologist and Nobel laureate Daniel Kahneman, is rapid, instinctive and emotional, as opposed to ‘System 2’ thinking, which is more deliberative and logical. [3] Learn more about System 1 and System 2 thinking in this video by Prof Gordon Parker.
Time constraints inevitably mean that clinicians rely on ‘cognitive short-cuts’ of the kind summarised in Table 1. [4], [5], [6]
This can lead to busy psychiatrists making judgements that have ‘face validity’ but that, on closer scrutiny, are based upon misleading heuristics. (Table 1)
Table 1: Common heuristics and cognitive errors in risk assessment (3, 4, 6)
Misconstruing risk assessment as a predictive task
The anxiety sometimes felt by psychiatrists completing risk assessment proformas is generally underpinned by the sense that the clinical task is one of prediction.
The terms ‘low’, ‘moderate’ and ‘high’ are generally assumed to refer to the predicted likelihood of a serious adverse event. A subsequent harmful event in a patient assessed as low risk could be perceived to be evidence of an error on the part of the clinician.
This ‘crystal ball’ model of risk assessment persisted despite a clear statement by leading forensic mental health experts several years ago that:
We are not now, and probably never will be, in a position to be able to determine with certainty who will or will not engage in a violent act. Relying on a range of empirically supported risk factors, though, we can make a reasoned determination of the extent to which those we are assessing share the factors that have been found in others to relate to an increased level of risk. [8]
There is a similar problem with the low positive predictive value of suicide risk assessments. [9]
Additionally, there are problems with categorising patients as low/moderate/high risk because there is a lack of consensus about what likelihoods these terms pertain to. [10]
A PRAGMATIC APPROACH TO RISK UNCERTAINTY
Despite the problems with risk assessment, we cannot eschew dealing with risk. Judgement in the face of uncertainty is an inescapable task for all psychiatrists.
Our view is that the clinical task of ‘risk work’, which simultaneously addresses a range of potentials for a range of harms, is a complex, iterative and circular process requiring consideration of how to best augment protective factors and diminish dynamic risk factors.
In making decisions about risk, we encourage doctors to consider the following:
The evidence
Academic research has yielded valuable information with respect to risk and protective factors and has shown the limits of unstructured ‘clinical intuition’. The first element when making judgements under uncertainty is, therefore, familiarity with the available evidence and consideration of how it applies to the clinical risk decision at hand.
This can be assisted by evidence-based checklists to act as an aide-memoire – a reminder to the psychiatrist to methodically consider a set of predefined items garnered from the research literature.
This reduces the likelihood of mistakes being made because of failure to ask relevant questions or obtain relevant collateral information about risk.
Proper use of these checklists requires a formulation-based understanding of the meaning, relevance and implications of particular ‘items’ in the guide rather than a perfunctory tick-box approach. [11]
For example, if the box ‘past episodes of self-harm’ is ticked, proper assessment requires further exploration of the nature and circumstances of those past episodes to help to properly determine what steps might now be required to minimise future risk.
Structured assessments based on validated guides can assist with minimising reliance on potentially misleading cues such as ‘guarantees of safety’ and ‘settled behaviour’ (Table 2).
Table 2: Commonly encountered examples of faulty assumptions
VALUES
Risk judgements made by psychiatrists involve value-laden questions such as balancing the risk of harm against overriding patient autonomy and balancing current safety against long-term recovery.
Contemporary risk work should, therefore, consider the values of the patient in the clinical-decision making process.
Relevant information and values-based concerns can be shared between the patient (expert by experience of illness) and the psychiatrist (expert by training).
Ideally, the patient is empowered as an active and responsible agent in their risk management, resulting in what has been termed the ‘co-production of safety’. [14]
With this approach, the psychiatrist’s job becomes more akin to a coach than an agent of control.
LEGAL CONSTRAINTS
The powers of psychiatrists to manage risk are appropriately constrained by law. Dilemmas may arise when the law constrains what appear to be common sense control measures to reduce risk.
This element of practical risk management becomes particularly salient in situations where harm of some kind is foreseeable, but the legislative threshold for control to mitigate the risk (by means, for example, of involuntary detention or a Guardianship Order) is not met.
Clearly, optimal risk management requires the psychiatrist to be familiar with the relevant legislation and its limits.
OPERATIONAL CONSTRAINTS
By ‘operational constraints’, we refer to those real-world factors that limit the provision of ideal care and optimal risk management. All health care takes place in the context of finite economic resources.
In psychiatry, this can lead to dilemmas for the psychiatrist such as determining when to discharge a ‘difficult to help’ patient to enable the service to provide resources for a more responsive patient.
Concerns about staff safety also constrain the possibility of optimal service delivery.
Understandably, staff may be reluctant to provide certain services such as home visits, or any services at all, to patients with a propensity for aggression.
In combination with resource limitations, these constraints can lead to situations whereby the most ‘complex’, high-risk patients receive a less intensive service than required or no service at all
The documentation of operational factors when making high-stakes decisions can enhance service quality (by clarifying where service demand exceeds capacity) and may be of medicolegal relevance in the event of future litigation if harms do eventuate.
RISK ASSESSMENT DOCUMENTS
Risk assessment should be part of an integrated and seamless process of clinical work, instead of an additional, separate task. Psychiatrists should record awareness of the potential for adverse outcomes as they arise during the clinical process. [15], [16]
Risk factors should be noted from the history and mental state, and hypothesised pathways to adverse events included in the formulation. This synthesis can inform the relevant risk management strategies in the management plan.
Instead of focusing on ‘likelihood risk categories’ (high/medium/low), psychiatrists could more usefully adopt action terms that indicate what steps should follow, such as ‘routine care’, ‘enhanced monitoring’, or ‘specific management interventions required’, with specific risk mitigation strategies documented.
CONCLUSION
Serious adverse outcomes in psychiatry are improbable but inevitable. Psychiatrists must therefore grapple with risk, but an over-emphasis on risk perpetuated by clinician anxiety and community fear does little to serve the interests of patients, clinicians or the community.
Risk assessment is best framed as a professional judgement under conditions of uncertainty rather than an exercise in probabilistic prediction.
An awareness of the shortcomings of current risk assessments, and a reminder to incorporate discussion of contemporary values and service delivery realities in risk decisions, may lead to a more sanguine approach to risk management in psychiatric practice.
QUIZ
References
Mullen PE, Ogloff JRP. Assessing and managing the risks of violence towards others. In: Gelder MG, Andreasen NC, Lopez-Ibor JJ, Geddes JR, editors. New Oxford Textbook of Psychiatry. 2. Oxford: Oxford University Press; 2009. p. 1991-2002.
Scurich N. The case against categorical risk estimates. Behavioral sciences & the law.0(0)
Douglas KS, Hart SD, Webster CD, Belfrage H. HCR-20: Assessing Risk for Violence, Version 3 (HCR-20 V3). Lutz, Florida: PAR, Inc.; 2013.
VogeL; VD, Ruiter; CD, Bouman; Y, Robbe MDV. SAPROF Structured assessment of PROtective factors of violence risk Utrecht: Forum Educatief; 2009