Memorandum submitted by Dr George Szmukler (MH 9)
This submission comprises comments on two issues of relevance to the Bill - (i) an impaired decision making criterion, and (ii) risk assessment.
(i) A 'capacity' or 'impaired decision-making' criterion in a Mental Health Act
A number of questions are commonly raised concerning an 'impaired decision-making criterion.
1 Many who argue against a capacity or 'impaired decision-making' (IDM) test for persons with 'mental disorders' ignore the fact that capacity tests are commonplace in medicine. All patients with 'physical disorders' (and patients with 'mental disorders' who require treatment for a 'physical disorder') who reject a proffered treatment in what appears to be an imprudent manner, become the subject of a capacity assessment to determine whether treatment should be given involuntarily in the patient's best interests. A capacity assessment is nothing new in medicine. Such assessments have been conducted under the common law and will continue in a similar manner under the Mental Capacity Act 2005.
2 Are capacity assessments different, in principle, for patients with 'physical' and 'mental' disorders? The almost identical definitions of mental difficulties in the MCA and the MHA Amendment Bill indicate that similar impairments are being considered. Research at the Institute of Psychiatry, involving both patients on the medical wards at King's College Hospital, on the one hand, and 112 patients on the psychiatric wards at the Maudsley Hospital, on the other, found similar percentages of patients with impaired capacity (around 40%), and did not find significant differences in the reliability of capacity assessments (Raymont et al, 2004; Cairns et al, 2005). The inter-rater agreements were high for both. Other research also supports the reliability of assessments.
3 On the other hand, there is no more research evidence concerning the inter-rater reliability of decisions to detain patients under the current, arguably highly subjective criteria of the Mental Health Act. (e.g. 'is suffering from mental disorder of a nature or degree which warrants the detention of the patient in a hospital for assessment ......; and ought to be so detained in the interests of his own health or safety or with a view to the protection of other persons'). It is likely that significant variations in practice occur, between clinicians, hospitals and regions (Peay, 2003; Perkins, 2003).
4 A criticism has been made that 'capacity' assessments as currently conceived are too 'cognitive' and do not take into sufficient account, 'affective' or emotional elements. This is an important issue which requires further clarification, but at a practical level in the research described above, this did not seem to present a major problem. It must be borne in mind here that 'emotional' elements are expressed in language, which is in a important sense 'cognitive' - for example a depressed patient will describe feeling sad, hopeless about the future, guilty about being burden on others, being unable to find pleasure in previously enjoyed activities, and so on. It can be seen that differentiating what is cognitive from what is affective is not at all straightforward.
5 A way forward in the particular case of the Mental Health Act is to use an IDM criterion which more clearly includes affective or volitional impairments within its scope.
6 All doctors are expected to have a knowledge about capacity assessment. This has always been the case, and it is even more clearly so with the MCA.
7 The potential dilemma posed by patients with fluctuating mental conditions who temporarily regain their capacity, after medication, and again refuse necessary treatment, has often been raised.
8 The extent which capacity fluctuates in patients with mental disorders has not been much researched. Fluctuations on a daily basis may be seen in patients with acute organic brain syndromes (such as delirium due to infections, or drug toxicity or drug withdrawal). Patients with a psychosis are unlikely to show such degrees of fluctuation.
9 Fluctuating capacity did not emerge as an
issue in the Institute of Psychiatry research mentioned above. In fact the ratings by two clinicians on
assessments carried out on separate occasions 1 to 7 days apart were highly
10 In any case, as now, common sense needs to be exercised in deciding when a patient should be discharged from an order. Where the patient has been treated involuntarily with a positive response, and a sustained course of treatment is again considered necessary, sustained resumption of capacity on the part of the patient might be appropriate for their refusal to be honoured. That might be the kind of test applied now by tribunals when considering the discharge of patients from the Mental Health Act who would not be placed under an order at their current level of risk (Peay, 2003; Perkins, 2003).
Would a 'capacity' or IDM criterion result in fewer or more patients being treated under compulsion?
11 The answer to this question is that we do not know. The best evidence we have comes from the research study mentioned above. Of those patients admitted to the Maudsley on section, 85% lacked capacity. The overlap is thus likely to be very high.
12 Although some have assumed that a capacity criterion would reduce the numbers under compulsory treatment, there is no evidence for this. It is equally likely, in my view, that the numbers might indeed increase. Clinicians would no longer have to wait for a clearly mentally ill patient, who is clearly behaving in a way that is against his or her best interests (and thus likely to lack capacity), to do something seriously risky before intervening with a section. Earlier intervention, thus preventing a deterioration of the illness and increased risk, would become possible.
13 Some have argued that a capacity or IDM criterion would exclude from compulsory treatment some patients for whom this would be appropriate. The problem here is that it is unclear how these authorities 'know' who should be thus treated. Their criteria are not usually stated, or at least not with the kind of precision that is demanded of a capacity-type criterion, and is presumably based on a judgement of risk of some kind. The inaccuracies of even the best risk assessment instruments currently available, particularly their high false-positive rates (predicting someone will be violent who turns out not to be), raise doubts about such judgements (Szmukler, 2003).
14 In any case, principles should antecede our decisions as to who should be involuntarily treated, rather than deciding on uncertain bases who we believe need to be thus treated and then trying to fix the rules or criteria to give us the answer we believe to be right. As both the Richardson and Bamford (Northern Ireland) Committees concluded, on ethical and anti-discriminatory grounds, it is hard to argue against a capacity criterion (or an impaired decision-making criterion).
15 There are also concerns about the implications of a capacity criterion for forensic patients. However, most authorities who have considered this question agree that the protection of the public needs to be given prominent attention, especially in the case of those who have already committed a serious offence (see for example, Dawson and Szmukler, 2006).
Cairns, R., Maddock, C., Buchanan A., et al. (2005) Reliability of mental capacity assessment in psychiatric inpatients. British Journal of Psychiatry 187: 372-378
Dawson, J & Szmukler, G. (2006) Fusion of mental health and incapacity legislation. British Journal of Psychiatry 188: 504-509
Peay, J. (2003). Decisions and Dilemmas: Working with Mental Health Law. Oxford: Hart.
Perkins, E (2003). Decision-Making in Mental Health Review Tribunals. London: Policy Studies Institute.
Raymont, V., Bingley, W., Buchanan, A., et al. (2004) Prevalence of mental incapacity in medical inpatients and associated risk factors. Lancet 364: 1421-1427
Szmukler, G. (2003) Risk assessment: 'Numbers' and 'values'. Psychiatric Bulletin 27: 205-207
(ii) Risk assessment in mental health
16 Risk assessment has two components, 'numbers' and 'values'. 'Numbers' refers to the estimation of the likelihood that an adverse event will occur in a stated period of time. The methods are mathematical and statistical. 'Values' refers to the processes of attaching a value to the risk and deciding what should be done about it. Benefits are weighed against costs in what is largely a moral enterprise.
17 Most authorities that there is a need for a scientific approach to establishing best risk assessment instruments - structured, standardised, replicable and tested on appropriate populations. Transparent assessments are preferable to vague clinical judgements. The development of such a risk assessment instrument is a major task, as its precision in each population needs to be separately tested.
18 A huge limitation on the predictive value of a risk assessment instrument is the 'base rate' problem. Rare events are extremely difficult to predict with any accuracy. Risk assessment is of little value when the base rate of violence, especially serious violence, in the population being tested is low (as it is, for example, in general community mental health patients). However, most people fail to appreciate the astonishing effect of base rate on the accuracy of prediction. I illustrate this in Table 1. The 'positive predictive value' of two tests is shown (that is, the proportion of patients the test predicts will be violent who in fact turn out to be violent). For example, using the average precision of all the known tests assessed by Buchanan and Leese (2001), if 5% of the population of interest is violent, the test will be right 8 times out of a hundred and wrong 92 times out of a hundred (PPV is 8%). Serious violence is likely to occur in less than 1% of community care patients over a period of a year. Homicides occur at a rate of 1 in 10,000 patients suffering from a psychosis, per annum.
Table 1. How the positive predictive value varies with the base-rate of violence in the population of interest
Base-rate = frequency of violence in the patient population over a specified period (e.g. one year). Serious violence is less than 1% of community patients.
PPV = proportion of those who are predicted by the test to be violent who turn out to be violent
Specificity = proportion of cases who will not be violent who are accurately predicted; at 0.68, 32% of non-violent persons will be incorrectly predicted to be violent.
Sensitivity indicates the proportion of violent cases that will be accurately predicted; at 0.52, 48% of violent cases will be missed.
1 Based on review of instruments by Buchanan & Leese (2001)
2 Based on data from MacArthur Foundation study (Monahan, 2001) - This was a single study, costing millions of dollars, which needs to be replicated. It represents the best prediction achieved so far and is unlikely to be bettered - but it was in a research, not a 'real-world', setting.
19 An immense effort went into the MacArthur study (Monahan 2001) and I believe it is unlikely that better accuracy can be achieved. After all, unlike the prediction of the release of toxic waste from an industrial plant, we are trying to predict an act committed by a person, an agent with intention who is engaged in ongoing myriad and complex interactions with others. For a relatively simple way of understanding the mathematics see Szmukler (2000)
20 The action to be taken following estimation of the likelihood of the risk is fundamentally a question of values. For example, what false positive rate is acceptable? Who should decide? Of special relevance to the prediction of violence in the mentally disordered are the potentially serious consequences of being wrongly classed as dangerous, and being unnecessarily detained. There are also other more general costs. Risk assessment may: emphasise control and containment at the expense of treatment; divert resources towards those assumed risky and away from the majority of those with mental illness; reinforce stereotypes of the mentally ill as dangerous; and, deter people who would benefit from mental health services, including risky persons, from seeking help for fear of coercive treatments.
21 But there are still more troubling aspects. Risk management decisions weigh up the costs and benefits to society in general, or to particular social groups, of particular courses of action aimed at controlling risk. In the case of mental disorders, we are weighing the benefits to society as a whole, against costs that largely fall on a small segment of the population, the mentally ill. Given society's long history of prejudice against, and social exclusion of, the mentally ill, the threat to this group is very worrying. The costs fall on the marginalised; the perceived benefit accrues to the rest.
22 In conclusion, risk factors for violence are derived from population-level analyses and lead to conclusions framed in terms of probabilities. As a result useful population-level interventions may be suggested. For example, if patients with co-existing psychosis and substance abuse are more likely to be violent than the average, services for this group of patients might be the subject of special investment.
23 Prediction in the individual case is another matter. Unless the base rate of violence is very high in the patient population being assessed (as it might be in forensic groups who have committed a serious offence), the false positive rate is extremely high.
Buchanan A, Leese M. (2001) Detention of people with dangerous severe personality disorders: a systematic review. Lancet 358: 1955-9.
Monahan, J., Steadman, H.J., Silver, E. et al (2001) Rethinking Risk Assessment: The Macarthur Study of Mental Disorder and Violence. Oxford, Oxford University Press.
Szmukler G (2001) Violence risk prediction in practice. British Journal of Psychiatry 178: 84-85.
 Institute of Psychiatry, King's College London