Risk – how is that to be managed?

Psychiatry is full of risks. Without delay here is a short incomplete list:

  1. Risk to self
    1. Self-harm
    2. Suicide.
    3. Accidental injury or death.
    4. Self-neglect.
    5. Exploitation of various kinds.
  2. Risk to others
    1. Injury – from violence or aggression.
    2. Sex offences
    3. Homicide
    4. Fire and other property damage.
    5. Terrorism
  3. Risk to health
    1. Physical illness
    2. Mental health

The context of this post is about risks managed in mental health services. Some of the ideas here may inspire deeper thought in other circles. For a sound overview start with Risk Management @ Wikipedia. For quick preview scroll down now to the video (and rewind it if necessary). This article is original thinking based on experience and the background knowledge from other sources.

Basic literacy on using a mouse in a web interface is required to navigate this article. No support provided to luddites.

Risk and management overview

At the outset the mindmap below gives some bare bones ideas about how risk and ‘risk management’ are understood at the coalface. Icons within the body of the mindmap are clickable for more information. This mindmap cannot be broken. Do click, drag and zoom using the zoom button, as much as necessary. If mindmap does not show click this.

Lay concepts

The core issue for lay persons is that they have an attitude or knowledge that ‘risk’ is something they must feel and see. That’s true of most situations in life, such as crossing a busy street – which one does not do with the eyes closed (or otherwise blinded). The feeling one gets when standing next to a large growling dog, is what lay persons normally consider to be risk i.e. of getting one’s leg painfully eaten. There is a problem arising from this ‘attitude’ about risk when lay persons become professionals. In professional environments – with complex issues to be considered – the largest most significant risks are not the ones that can be felt. Professionals – who were once lay persons – may attitudinally approach risk in a pre-professional way, though cognitively they are trained and aware of risk in a different way. The latter is a risk in itself.

The concept of risk is so basic, that it becomes difficult to define when one looks deeply into it from a professional perspective. No one definition may cover all purposes. Naturally, people will search dictionaries and the internet when they want to dig deeper. This is a list of 10 definitions of risk. This post will focus on risk in mental health services scenarios. Having read a number of definitions, the author describes risk as follows: ‘Risk is the probability estimate of adverse outcome.’  This is short and sweet covering most situations. Finer bones can be added to the definition in terms of categories. A finer definition of risk is, “The combination of the probability, or frequency, of occurrence of a defined hazard and the magnitude of the consequences of the occurrence” (Royal Society, 1992).

Hazard –  “a property or situation that in particular circumstances could lead to harm” (Royal Society, 1992)

Risk perception – involves people’s beliefs, attitudes, judgements and feelings, as well as the wider social or cultural values that people adopt towards hazards and their benefits.

Risk assessment is the procedure in which the risks posed by inherent hazards involved in processes or situations are estimated either quantitatively or qualitatively. Risk assessment may include an evaluation of what the risks mean in practice to those effected.

Risk evaluation – is the meaning and context of the risk.

Risk control – actions taken to reduce the probability of the risk materialising.

Limitations of risk assessment

Probability estimates

In mental health circles, it is very difficult to say for example, “If he does not take his medication there is a 51% chance of relapse in 3 months.” Why is this difficult? Medication is a major factor in preventing relapse, but there are often several other protective factors in an individual case. How is any psychiatrist to evaluate the protective effects of good community care and social support (among others), in the absence of medication? What tool would they use to measure or estimate any of that? No such tools exist that might be applied to an individual patient. There are certainly figures that can be found from research evidence which usually is applied to larger populations of patients. However, for the individual with their own set of variables, it is difficult or impossible to measure.

Prediction

Lay persons and some legal professionals often wonder, “What’s the point of risk assessment if it is not to predict something?” Then we have a group of psychiatrists who will say, “Poor compliance with medication predicts relapse.” This is poor use of language. What those psychiatrists mean is that relapse has a high probability when there is poor compliance. But none of those psychiatrists could possibly give a reliable time frame for relapse. Of course probability of relapse increases with prolonged non-compliance. So one could say that there is a 90% chance of relapse after 3 years of non-compliance. But that is not a prediction either. It is an estimate of probability. It leaves a 10% chance that there may be no relapse. Such statements are derived from research information that can give such estimates for a large class of similar patients. Therefore the individual patient is caught in a large statistical group. Hence such an assessment (at 3 years) is more reliable – but prediction it is not.

The lay notion of event B following event A often gives lay persons an idea that A predicts B. Hence stopping medication (total non-compliance) is seen as event A. Relapse is event B.

The problem runs deeper. Prediction leads to a ‘silent binary‘ i.e. either something happens or it does not. For example, either he is run over by a car if crossing a busy street with his eyes closed, or not. Or, either he relapses or not. When the adverse outcome happens people being human, are likely to think, “There we are – I predicted this!” When it does not happen, they remain silent. The occurrence of the adverse outcome is not prediction.

Nobody values a prediction such as “99% of all people born today will be dead in 100 years.” Why? Because it is a statistical fact – not a prediction. Birth becomes event A and death becomes event B. Does birth predict death? This sort of thinking becomes silly. Therefore a statistical fact for a large population of people when applied to an individual who is part of that population, is no prediction for the individual.

Risk assessment is more about honing in on scenarios relevant to an individual. Evaluation is about making probability estimates about relapse or other adverse consequences. In the scenario where a patient has had several relapses , one per year for the last 10 years, due to non-compliance, and relapses have been closely associated with brandishing knives – it is reasonable to say that, “If this patient is non-compliant with medication for 3 months there is a high probability of relapse and the probability of a homicide.” But nobody – except a charlatan – could say that a homicide will happen at 3 months plus 1 day. If the patient moves into a statistical class of similar individuals, it may be possible to make a better estimate at say 1 year or 3 years.

Risk management – differentiation from other things.

This is a complex task. It is best done by an expert team working closely together. Many health services approach the topic of risk as ‘risk assessment’. The mindmap shows that risk assessment is a smaller aspect of the overall process. It is not to be confused or conflated with risk management. Some believe that risk assessment is risk evaluation. The author departs from such beliefs – and thinks that risk evaluation is a rather separate discipline requiring a high degree of knowledge, skill and experience.

Risk control – is often neglected in the management of patients. So – doing the risk assessment and an evaluation of scenarios is not enough. Managing risk means controlling factors that may lead to an adverse outcome.

Risk reviews – are often taken to mean updating ‘the risk assessment’. It is not. Reviews require a high degree of concentration, reasonably accurate information and adaptation to changing circumstances in the bio-psychosocial and legal domains of care.

There is a big problem for risk management in mental health services. It is like this: Risk is not a thing that can be held in the hands. Compare this with driving a car. Everyday millions of good drivers manage risk quite successfully by using  their steering wheels, accelerators and brake pedals. When drivers do this they are close to the risk. They can sense their speed (or see it on a speedometer). They can see surrounding traffic and make what might seem to be automatic adjustments.  They do all this in real time. See video below and compare to how risk management works in mental health services.

Can adverse outcomes be prevented? Yes – but nothing in this business is 100% or near that. Note carefully the ‘silent binary’ appearing in the question i.e. prevented or not. No individual or team can prevent most bad outcomes. Some spend more time worrying about legal consequences instead of doing the job properly. The law is not going to be punitive, where workers demonstrate the right degrees of diligence. The law will be punitive where workers were unforgivably slack or asleep on the job, and that led to serious adverse outcomes. The correct levels of diligence are to be found in Codes of Conduct and Codes of Practice.

Take away points

  1. Prediction is not truly the business of risk assessment or risk management.
  2. Risk assessment is not risk management.
  3. Risk evaluation should be a well circumscribed activity.
  4. Sound risk management is an adequate skilled team working adaptively on complex issues.
  5. Risk reviews are not simply updating risk assessments.
  6. Risk controls do not have to be anywhere near perfect.
  7. Legal duties of care are best discharged by paying attention to standards of practice, and doing what is required.

Reference:

1 – Chapter 1  Environmental Risk Assessment: Approaches, Experiences and Information Sources 2020