Lawyers have a proper interest in the quality of risk decisions. Was it proper to discharge that patient then? Should the social workers have returned the child to its abusive parents? These familiar examples require a retrospective analysis. We know what happened, the patient suffered loss, the child was abused, but we want to know whether appropriate decisions, which involved risk, were taken. But legal practice is not (contra Aubert, 1963; Campbell, 1974) necessarily retrospective. A lawyer can argue that a risk should be taken, that a detained patient should be discharged, that a prisoner should be granted parole, that the father—rather than the mother—should be granted custody of the child. The common features are decision-making and the potential for justification.
A key problem, making collaboration between law and psychology, research and practice, unnecessarily difficult, is the tendency to associate, even to define, risk exclusively in terms of the possibility of harm (e.g. Yates, 1992). Risk is regularly associated or defined as the possibility of harm (e.g. Royal Society, 1992). But that makes it difficult to justify taking a risk. A patient with a mental disorder wishes to be discharged from hospital. But he or she is thought to be a risk for suicide. So a risk assessment is undertaken. The likelihood that patients with this individual's characteristics will self-injure is examined. A practical plan is developed for managing the risk of self-harm, if the patient were to be discharged. This is amended in the light of the risk assessment. A decision is taken. The risk of suicide is, or is not, too high.
But should a decision be taken, on this information? Has not risk assessment been misused?
Risk practitioners have to balance and to take into account the case for as well as the case against. The question—ethically, professionally or legally—is not whether the risk of harm is sufficiently low. It is not just that we do not have a test for what would constitute 'sufficiently low'. It is a moral or value judgement rather than a scientific test. The question is whether the case against this particular patient being discharged is outweighed by the case in favour. This is so even in cases of terminal care, or services for elderly mentally ill people where practice has to focus on slowing and reducing, rather than preventing, decline and disability. The question involves balance and choice between alternatives. It is perfectly understandable that researchers should seek to provide us with more and better information, and tools for, predicting harm and loss. That is what we want to understand, in order to avoid. But that does not prevent us from acknowledging that risk decision-making necessarily involves comparing the possibility of benefits with the possibility of harm. We can act to minimise the likelihood and the seriousness of the harms. We can act to maximise the likelihood and the value of the benefits. That is the role of risk management. But the eventual risk decision, perhaps after several alternative risk management plans have been investigated, will require a balancing operation—a valuing of the options.
Lawyers, at least in the United Kingdom, also tend to associate or define risk exclusively in terms of harms. Knowing that there is a risk of injury to another person, and yet acting, has been deemed to constitute 'recklessness' sufficient for a conviction for manslaughter (Adamako  1 AC 171). It is submitted that this cannot, and it is hoped that this will not, stand when a case arises where it is appropriate to balance potential benefits against potential harms. (We should not be guilty of manslaughter just because we foresaw that someone might be injured when we got in our car to go to work and, without more fault, we happened to kill someone in an accident.) The need to balance has been recognised for civil cases of negligence. In Bolitho v. City and Hackney Health Authority ( 3 WLR 1151) the House of Lords, the most senior court in the UK, was invited to change the standard of care by which professionals' decisions, including risk decisions, are measured for the law of negligence. They decided to retain the existing law whereby risk decisions do not break the standard of care if a responsible body of professional opinion would support them. They stated:
In particular in cases involving, as they so often do, the weighing of risks against benefits, the judge before accepting a body of opinion as being responsible, reasonable or respectable, will need to be satisfied that, in forming their views, the experts have directed their minds to the question of comparative risks and benefits and have reached a defensible conclusion on the matter. (p. 1159)
Whilst this demonstrates that the judges consider 'risk' to be limited to possibilities of harm—for otherwise there would be no need to weigh benefits against them—it also emphasises the propriety, indeed the critical role, of balancing.
Although it would simplify matters, and as such it is recommended, it does not really matter whether we define 'risk' as including or excluding reference to the possibility of benefit. What matters is that we acknowledge the necessary role of balancing when it comes to making risk decisions. Those who make risk decisions are entitled to take into account the possibility that each of their reasons (objectives and goals), for taking the proposed risk, may be fulfilled or achieved. The potential value of those benefits may be placed in the balance against the potential seriousness of the harms feared. Indeed, if they fail to consider potential benefits they are not making a proper decision! They are laying themselves open to criticism. Imagine that someone is called upon to justify taking a risk, which led to harm. He or she could list and elaborate upon the value of the potential benefits that were being sought, as well as the potential harms feared. Or he or she could explain simply that the likelihood and/or seriousness of the harms was not considered to be bad enough for them not to take the risk. The first approach is more likely to be successful, not least before judges. It is also perfectly proper practice.
But reconsider the role of risk factors and risk assessment. Because of the tendency to associate risk exclusively with the chance of harm, risk factors and risk assessment schemes invariably only inform us about the chances of harm. They provide us with reasons, or excuses, for not taking a risk. They are inherently conservative, biased towards not taking a risk. A few lists include factors, such as availability and skills of supervising staff (e.g. Gunn, 1990), but they are a minority.
Indeed it is surprising that we have little or no research about the 'protective factors'
...in individuals' lives which serve to reduce risk, whether than be to self or to others.
(Langan, 1999, p. 172, reference omitted)
But even when protective factors exist they are not the same thing as risk factors for success. They may tell us features of people less likely to experience harm; but that is not the same as identifying people who are likely to experience benefits. We can examine risk factors for male violence (e.g. Webster et al., 1994) or spouse assault (Campbell, 1995; Kropp et al., 1995). They will help us to predict the likelihood of suicide or spouse assault. But they will not help us to predict the likelihood that the patient will be motivated or retain his or her job if discharged from hospital. One spouse might hit the other. For that reason we contemplate separating them, amongst other tactics. But there are many other things to be considered when deciding whether to support an order excluding one spouse from the matrimonial home. Remaining in the same house could do good, as well as ill. The factors which make it less likely that a person will commit suicide, for example, are not necessarily the same as will make it likely that he or she will get pleasure and motivation from living in familiar rooms with more self-control and responsibility. And yet the person making a risk decision needs to consider potential benefits, as well as the potential harms and rights involved. It is not just that risk assessment tools, which concentrate on the chance of failure and loss, incompletely serve practitioners but measures, which could identify the likelihood of success, would help them contemplate alternative risk management strategies.
Lawyers could, and it is submitted should, investigate the use of risk assessment tools, say in negligence claims or inquiries into untoward incidents. They should not just include questions about whether the measure chosen was appropriate for the decision in question, or whether it was properly used. They should examine the process, the balancing exercise. For example, was a proper attempt made to identify the possible benefits of taking the risk, and their likelihood? It is normal practice for an assessment tool to list a range of risk factors. But does the tool imply that the different factors are of equal importance? If so that is almost certainly going to involve an error. Some factors are much more important, more predictive, than others are. For example age, gender and number of prior offences are more predictive of violence than nature of delusions (Monahan et al., 2001). The tool may not tell its users, one way or the other, about the relative importance of the different items. The author may not be satisfied that the science is sufficiently developed to allow for such differential rating (Hagell, 1998), but he or she is likely to know that—explicitly or implicitly—treating each item as of equal importance is improper, and thereby unprofessional.
Are the risk factors mutually exclusive? For example one question, in a list designed to identify risk factors, might ask about past violence whilst another asks about past crimes. But past violence is a past crime! The same cause for concern, past violence, is being 'double counted'. That is likely to be inappropriate. The authors of the tool might have done it deliberately, as a means of weighting the importance of that topic. But that may not be the case. Either way unsophisticated or mechanistic users of such a tool are liable to be caught out by a knowing lawyer alive to the possibility of such a problem.
Intuitively it would seem proper to conclude that a more detailed list of risk factors is going to be better, more predictive, than a short list. It will appear better, more thorough. But that simply does not follow. A long list of risk factors may have more to do with demonstrating the author's imagination than identifying the things that, according to research or clinical experience, are the important predictors. A few key risk factors may be sufficient, or get decision-makers as close as they are ever going to get (with current levels of knowledge) to an appropriate risk assessment. More information may add little additional predictive power (McNeil and Binder, 1994). Generally, more information will help risk-takers, but they must not make the basic mistakes of assuming that it is of equal value or quality. If it is a risk decision then, by definition, it is going to have to be made on the basis of incomplete information. Knowledge of, or insight into, the predictive power and accuracy of what is known will regularly be more important than sheer quantity of information.
Risk assessment is the start, rather than the end, of risk-taking. By definition it is an imperfect science. Imagine there is little, or only poor quality, information available for a risk assessment. Unless it is clear that any possible harms will be trivial, and/or exceptionally unlikely, it will difficult to justify taking that decision. But that is where risk management comes into the 'equation'. Because there is only poor quality information available for a risk assessment, it is more important that a high-quality risk management plan is devised. High-quality risk management, for example the employment of more resources and control, can justify action on a risk assessment necessarily based upon poor-quality information. Assessment and management have an iterative relationship.
Was this article helpful?