Med Sci Law OnlineFirst, published on June 24, 2015 as doi:10.1177/0025802415590176
Original Article
Appeals against detention in excessive security (outcomes of appeals against detention in conditions of excessive security in Scotland)
Medicine, Science and the Law 0(0) 1–6 ! The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0025802415590176 msl.sagepub.com
Alexander Slater1, Daniel M Bennett2, Gabriele Vojt3 and Lindsay Thomson1,4
Abstract The Mental Health (Care and Treatment) (Scotland) Act 2003 introduced the right for patients in high-security psychiatric care to appeal against detention in conditions of excessive security. A previous study examined the first 100 patients to appeal under this provision. In this study we compare them with the next cohort of 110 patients to lodge an appeal, finding, contrary to expectations, no change in patient characteristics or the outcome of their appeals. The clinical, legal and demographic features of successful and unsuccessful appellants, who made up 38% and 27% of the 110 patients, respectively, were also compared. Those patients with the support of their responsible medical officer and those already included on a transfer list had a significantly better chance of success (p ¼ 0.00). It was also found that a history of excessive alcohol consumption was associated with successful appeals (p ¼ 0.002). A diagnosis of learning disability was associated with unsuccessful appeals (p ¼ 0.018), though the sub-sample was very small. These findings are important given the forthcoming extension of this right of appeal to other levels of security. Keywords Forensic psychiatry, law, medical law, legal system, high security, excessive security
Introduction
the full security provided.4 The Mental Health (Care and Treatment) (Scotland) Act 2003, hereafter referred to as ‘the 2003 Act’, confirmed the principle that care must involve the ‘minimum restriction’ sufficient for the patient. The suggestion that patients were held in unnecessarily high security led to two developments in Scottish forensic psychiatry: the establishment of three medium-secure units and the unique right for patients to appeal against detention in conditions of excessive security (rather than detention per se) under the 2003 Act. The patient can make an application to the Mental Health Tribunal for Scotland, hereafter referred to as ‘the Tribunal’, for a declaration that they are being
High-security psychiatric care for Scotland and Northern Ireland is provided by the State Hospital at Carstairs. Its patients require high-security care due to their ‘‘dangerous, violent or criminal propensities’’.1 High security not only involves environmental measures, such as perimeter fences, but also relational and procedural aspects.2 Relational security refers to ‘‘the knowledge and understanding staff have of a patient and of the environment, and the translation of that information into appropriate responses and care’’.3 Procedural security refers to the protocols in place to control the patients’ environment and keep staff and patients safe; this includes regulating the patients’ possessions. The gathering of information 1 The University of Edinburgh, Edinburgh, UK on the history and progress of individuals in high2 University of Aberdeen, Aberdeen, UK 3 security care also allows a degree of tailoring in Risk Management Authority, Paisley, UK 4 The State Hospitals Board for Scotland and the Forensic Mental their security provision, relevant to their Health Services Managed Care Network, Carstairs, UK requirements. Whilst high-security care is undoubtedly necessary Corresponding author: for some, a study in 1997 suggested that 53% of the Lindsay Thomson, Division of Psychiatry, The University of Edinburgh, State Hospital’s patients, in the view of their Royal Edinburgh Hospital, Edinburgh EH10 5HF, UK. Responsible Medical Officer (RMO), did not require Email: [email protected] Downloaded from msl.sagepub.com at Monash University on November 14, 2015
2 held in conditions of excessive security: it is this process which we refer to as an appeal. If the Tribunal upholds an appeal, they order the health board responsible for the patient to relocate them to a lower-security setting within a specified period of up to three months. The appeal does not concern whether the patient should be detained, but specifically whether they require high-security care at the State Hospital. The Tribunal may reconvene twice more to grant the responsible health board up to seven months in total to relocate the patient. Should the responsible health board fail to meet this obligation, its executive directors may be fined or imprisoned.5 The reasons why the Tribunal makes a declaration of detention in conditions of excessive security are of significant interest to the forensic service. A recent study, hereafter referred to as the ‘Bennett study’, examined the clinical, legal and demographic features of the first 100 patients to make such an appeal.6 This showed that successful appellants more often had their RMO’s support and were more likely already to be included on a transfer list. No other significant differences were found between successful and unsuccessful appellants. It had been thought that the characteristics of patients making an appeal would change following the transfer of the accumulation of patients who would be better cared for in medium-security units. However, amongst the first 100 patients to appeal this assumption only seemed to be true for women, all of whom were successful. The anticipated shift in the characteristics of appellants, if it came to pass at all, may have only occurred after the first 100 appeals. This study aims to investigate the demographic, legal and clinical features of the 110 patients who have made appeals since the Bennett study. This includes 33 patients who had lodged an appeal during the period studied by Bennett et al. Between the two studies, all patients were included who had ever lodged an appeal. The hypothesis of this study is that those who are successful will be more likely to have the support of their RMO and already be on the transfer list. They are expected to be similar in all other characteristics. These 110 patients are also compared with the 100 patients in the Bennett study to look for evidence of a change in the characteristics of patients making appeals. It is postulated that fewer recent appeals will succeed, since the initial group of inappropriately placed patients has moved on.
Medicine, Science and the Law 0(0)
Sample Appeals against detention in conditions of excessive security were concluded for 110 patients between midJuly 2008 and the end of July 2013. The sample included every patient who lodged an appeal since the Bennett study, 33 of whom had lodged an appeal during the period studied by Bennett. The outcomes of these appeals were obtained from a Forensic Mental Health Managed Care Network Database. Mid-July 2008 was specified in order to continue from the date at which the Bennett study concluded. Only a patient’s first appeal during this time period was included in this sample.
Sources for data collection Data were gathered by retrospective case note analysis, using a coding sheet adapted from Bennett et al. The data included: demographic details; psychiatric history; psychiatric diagnoses; drug prescriptions; reason for admission; the legislation under which the patient was detained; and drug and alcohol history. Factors which changed over time, such as prescriptions, were recorded from formal review documents. Information regarding the patient’s appeal was gathered from documents provided by the Tribunal or legal correspondence. Data concerning whether the patient was already on a transfer list were gathered from the transfer list database. Patients were recorded as already being included on a transfer list if they were as such on the date they lodged an appeal. If the RMO’s support for an appeal could not be determined from the notes then, where possible, the RMO was asked to indicate their position from memory. This was the case for 21 of the 61 appeals for which the RMO’s support was ascertained.
Statistical analysis
The data were analysed using the programme SPSS Statistics 21 (IBM, Chicago, Illinois). A range of descriptive techniques were used to summarise the data. To identify any difference between successful and unsuccessful appellants in this study, the data were first put into contingency tables for each of the factors studied. Fisher’s exact test was then used to look for a significant association between each factor studied and the outcome of appeal. Those factors which showed a significant association with the outcome of appeal were included in a binary logistic Methods regression model to determine whether they could Literature search predict the outcome. A two-step approach was used to compare the curA literature search was carried out using the rent and Bennett et al. sample: (i) a non-parametric PsycINFO database. A combination of the search Levene test to look for homogeneity of variance; if the terms ‘‘Forensic psychiatry’’ AND ‘‘Security’’ AND data failed this test it showed a significant difference ‘‘Appeals’’, gave 18 results. Only the Bennett study between the two groups; (ii) if the data were similarly was relevant to this study. Downloaded from msl.sagepub.com at Monash University on November 14, 2015
Slater et al.
3
distributed, a Kruskal–Wallis test was applied to look for a significant difference between the median of the groups.
having controlled for the effect of other predictors in the model.
Statement of principal findings Results Of the 110 patients to make an appeal between midJuly 2008 and July 2013, only one was female. The mean length of their admission to the State Hospital (7.0 years) was similar to that of patients in the Bennett study (7.1 years). The characteristics of successful appellants were compared with those who were unsuccessful (Table 1). Using contingency tables and Fisher’s exact test identified four significant differences. Successful appellants were more likely to have the support of their RMO and already to be included on a transfer list. They were also less likely to have a diagnosis of learning disability – although the subsample was very small – and more likely to have a recorded history of excessive alcohol consumption. Next, the four factors significantly associated with the outcome of appeal were used in a binary logistic regression model, the results of which are shown in Table 2. This was based on the 47 patients for whom there was data available relating to all four domains. Used together, these factors were able to predict 96% of appeals which were upheld and 95% of appeals which were not upheld. Beta weights were not standardised and positive values show an association with failure at appeal (since the code for appeal not upheld was 1 in the model, compared with 0 for appeal upheld). The odds ratio relates to the odds of failure at appeal,
The findings of this study are similar to those of the Bennett study. Having RMO support and already being included on a transfer list were significantly associated with success at appeal. This study also identified two clinical characteristics which showed an association with the outcome of appeals, namely, that having a diagnosis of learning disability was associated with unsuccessful appeals, whilst a history of excessive alcohol consumption was associated with success at appeal. However, the sample sizes were small, especially for learning disability. Used together, these four factors were able to predict the outcome of more than 95% of appeals in a binary logistic regression model.
Strengths and weaknesses of the study One of the strengths of this study is that it includes all the patients who have lodged an appeal since the previous study concluded. Between the two studies, all the patients to have made an appeal against detention in excessive security in Scotland are included, albeit the total is only 210 patients. For some characteristics, such as having a learning disability, the number of patients was as few as seven. This is not sufficient to draw any definitive conclusion and, therefore, the associations of alcohol excess and learning disability with outcome of appeal should be considered as requiring future investigation. Another potential
Table 1. A comparison between successful and unsuccessful appellants. Variable
Successful (N)
Unsuccessful (N)
Fisher’s Exact Test (p-value)
Patient background (demographic and previous psychiatric treatment) – no significant difference between groups Offence, or behaviour leading to admission – no significant difference between groups Record of substance misuse History of excessive alcohol consumption 28 9 0.002 Illicit drug use 33 21 0.289 Recorded history of IVDU 7 2 0.185 Diagnoses Schizophrenia 29 21 0.571 Learning disability 1 6 0.018 Personality disorder 11 9 0.462 Depression 2 1 0.626 Primary diagnosis (schizophrenia vs. other) (Schizophrenia) 29 20 0.515 Prescribed medication – no significant differences between groups Circumstances of the appeal Appeal supported by RMO (i) 26 1 0.000 Patient on transfer list 31 6 0.000 Had lodged previous appeal 13 8 0.450 (i) Only 47 of the 110 patients had data in this category Downloaded from msl.sagepub.com at Monash University on November 14, 2015
4
Medicine, Science and the Law 0(0)
Table 2. Logistic regression using patient characteristics to predict the outcome of their appeal. Independent variable
Beta weighti
History of excessive alcohol consumption 16.554 Learning disability 20.017 Appeal supported by RMO 36.707 Patient on transfer list 35.321 Model 2 56.333 p ¼ 0.000 Pseudo R2 (Nagelkerke method) 0.938 (scale is 0–1) N ¼ 72
SE
Wald
Sig.
Odds ratio
4541 15707 6992 6992
0 0 0 0
0.997 0.999 0.996 0.996
0 493445089 0 0
i
Beta weights were not standardised and positive values show an association with failure at appeal (since the code for appeal not upheld was 1 in the model, compared with 0 for appeal upheld). The odds ratio relates to the odds of failure at appeal, having controlled for the effect of other predictors in the model.
been subject to recall bias. Finally, this study does not include any qualitative analysis of the decisionmaking process of the Tribunal. Therefore, the mechanism of any association can only be described speculatively.
Strengths and weaknesses in relation to the Bennett study
Figure 1. A comparison of the outcome of appeals in the Bennett study and the current study. Bennett outcomes (above). Current outcomes (below). Key: Green¼upheld; Red¼not upheld; Blue¼withdrawn/cancelled; Yellow¼adjourned.
This study used a similar approach to data collection and analysis as did the Bennett study. Instead of using a Kruskal–Wallis test to compare successful with unsuccessful appellants, a Fisher’s Exact Test and binary logistic regression were used to determine whether the outcome could be predicted, based on patient characteristics. It was hoped that this would reinforce the associations identified in the Bennett study. The Bennett study had the advantage of a complete data set regarding whether an appeal had RMO support; a factor which was only known for 47 of the 110 patients in this study. The cohort of patients in this study had similar rates of success to those included in the Bennett study (Figure 1). The characteristics of the 110 patients were compared with those in the Bennett study to examine change over time (Table 3). No change was evident, whether the three adjourned appeals from the Bennett study were included or excluded on account of the fact that this study only considered appeals which had reached a conclusion.
Meaning of the study: possible mechanisms and implications for clinicians and policy makers
The association of RMO support and already being included on a transfer list with success at appeal is explicable. As discussed in the Bennett study, these factors are linked because the opinion of the RMO, weakness in the statistical comparison is that 33 of the as part of the clinical team, advises whether the 110 patients had been included in the previous study. patient should be included on a transfer list. The supIt is possible that this could have contributed to the port of the RMO may be influenced by many factors, similar results found in both studies. Regarding the 21 both current and historical, including the patient’s patients for whose appeals RMO support was clinical condition, perceived risk and insight. In this obtained from memory, this couldDownloaded possibly have from msl.sagepub.com at Monash University on November 14, 2015
Slater et al.
5
Table 3. A comparison between patients in the Bennett study (N ¼ 100) and those in this study (N ¼ 110). Variable Patient background Gender (Male) Marital status (Married) Responsible Health Board (Greater Glasgow and Clyde) (Rest of Scotland) (Northern Ireland) Previous treatment (In-patient) Legal provision (Restricted patient) Source of admission (Hospitals) (Legal system) Diagnosis (not necessarily primary) Schizophrenia Learning disability Personality disorder Depression Record of substance misuse History of excessive alcohol consumption Circumstances of the appeal Appeal supported by RMOi Patient on transfer list Outcome of hearingii (Successful appeal)
Bennett
% (N)
Current % (N)
Test
P
KWT
0.076
KWT
0.669
KWT
0.484
KWT
0.303
KWT
0.780
KWT
0.408
95 (95)
99 (109)
3 (3)
1 (1)
46 (46) 50 (50) 4 (4)
46 (50) 48 (53) 5 (5)
78 (78)
66 (72)
63 (63)
63 (69)
50 (50) 50 (50)
45 (49) 51 (56)
62 (62) 20 (20) 42 (42) 5 (5)
67 (74) 14 (15) 31 (34) 3 (3)
KWT KWT KWT KWT
0.425 0.218 0.096 0.391
53 (53)
54 (59)
KWT
0.927
59 (58) 59 (59)
52 (32) 49 (54)
KWT KWT KWT
0.326 0.151 0.358
43 (43)
38 (42)
KWT¼Kruskal–Wallis test i Based on 164 cases for which RMO support was known (99 from Bennett study and 65 from this study). ii Based on 97 cases from Bennett study (excludes those adjourned) and 110 new cases.
to potential destabilisers in a lower-security environrespect, RMO support could be considered as a proxy ment as ‘testing out’. It is possible that prompt idenmeasure to the patient’s condition and the wider clintification and treatment of patients with a history of ical team’s view. excessive alcohol consumption might expedite their If learning disability is associated with failed transfer to a lower-security setting. appeals, the explanation may be that it is less amenThere is no evidence to suggest that these appeals able to treatment than other psychiatric conditions. have changed the characteristics of the patients Without a change in the underlying clinical condition appealing against their detention at the State which led to admission, it is less likely that there will Hospital or that more recent appeals are less likely be a change in the level of security provision required to be successful (see Figure 1). This is interesting by a patient. The risk that they pose to themselves or since the appeals included in the Bennett study others is not so easily decreased by treatments. might be expected to have had a higher success rate, Conversely, excessive alcohol consumption is relaon account of the many patients thought to have been tively responsive to treatment. Since the State held in conditions of excessive security before the 2003 Hospital is a maximum security environment, access Act came into effect. to alcohol should be impossible and abstinence is The Supreme Court’s ruling in the case of RM vs. enforced. The State Hospital also has a programme The Scottish Ministers [2012] UKSC 58 obliges focusing on alcohol and drug abuse, in which each Scottish Ministers to bring forward regulations definpatient is assessed and provided with psychological ing the terms under which excessive security appeals treatment. It may, therefore, be easier for patients will be extended beyond the State Hospital.7 These with alcohol-related problems to demonstrate that they have remained abstinent and engaged with treatnew appeals are legislated for by section 268 of the ment. The Tribunal may be persuaded to expose them 2003 Act. Scottish Ministers need to define Downloaded from msl.sagepub.com at Monash University on November 14, 2015
6 ‘‘qualifying patients’’ and ‘‘qualifying hospitals’’ for these new appeals. The current consultation on the Mental Health (Scotland) Bill includes the proposal to limit extending the appeal against excessive security only to mediumsecure units.8 Appeals against excessive security were introduced in the knowledge that there was a significant cohort of patients unable to move on from high security. This provision has led to a significantly changed forensic estate with a much reduced State Hospital (from 240 to 140 beds) and the development of two new medium-secure units in addition to a pre-existing medium-secure unit. This has considerably alleviated the problem of entrapped patients within high security. The proposed extension of the appeal against excessive security to medium security may result in the development of the low-secure estate. However, these proceedings are stressful to patients, divert clinical time and resources away from the direct care and treatment of patients and have a financial cost associated. It remains to be seen how this will apply to those treated within medium-secure units, where independent access to the community is entirely possible and is often a clinical aim. The Bill also proposes the repeal of Section 266, such that a Health Board would have a maximum period of six months to find a suitable alternative placement for a patient declared to be held in conditions of excessive security.9 In order to implement section 268 appeals a good understanding of the function of section 264 appeals is essential. This study offers some indication of the likely outcome of appeals. Moreover, having found strong predictive factors of success suggests that a sifting step may be possible to operate as part of the appeal process. Of course, the RMO cannot act as the gatekeeper for appeals, but evidence of engaging with treatment and progress may provide useful information. Such a step would be important in minimising the waiting time for an appeal to be heard, as well as being cost effective. It would also address the clinical concern that extending the provision of appeals may lead to conflict between patients, their RMOs and clinical teams, as well as diverting much time and resources to the legal process, without providing significant benefit to the patient.
Unanswered questions and future research
Medicine, Science and the Law 0(0) the mechanism of the associations identified by this study. Finally, a formal cost-benefit analysis may be helpful in describing whether a sifting step for appeals would be effective and just. Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Declaration of conflict of interest None declared.
Ethical approval An ethical assessment form was submitted to the University of Edinburgh for this project and, based on the methodology, no further action was required.
Statement of guarantor Professor Lindsay Thomson is the guarantor for this work.
Statement of contributors Alexander Slater collected and analysed the data. Lindsay Thomson supervised and advised on all aspects of the project. The above drafted and edited the manuscript with the help of Daniel Bennett, on whose previous work this study was based. Gabriele Vojt advised on the method for data analysis.
References 1. National Health Service (Scotland) Act 1978 S.102. http://www.legislation.gov.uk/ukpga/1978/29/section/ 102 (accessed 7 June 2015). 2. Forensic Network. Definition of security levels in psychiatric inpatient facilities in Scotland, http://www.forensicnetwork.scot.nhs.uk/documents/previous_reports/ LevelsofSecurityReport.pdf (accessed 27 January 2014). 3. Department of Health Secure Services Policy Team. Your guide to relational security, http://www.rcpsych. ac.uk/pdf/Relational%20Security%20Handbook.pdf (accessed 27 January 2014). 4. Thomson LDG, Bogue J, Humphreys M, et al. The State Hospital Survey: A description of psychiatric patients in conditions of special security in Scotland. J Forensic Psychiatry 1997; 8: 263–284. 5. Thomson LDG. The mental health (care and treatment) (Scotland) Act 2003: Legislation for mentally disordered offenders. Psychiatr Bull 2006; 30: 423–429. 6. Bennett D, Skilling G, Brown K, et al. Appeals against detention in conditions of excessive security in Scotland. J Forensic Psychiatry Psychol 2013; 24: 386–402. 7. RM v The Scottish Ministers UKSC 58 (2012), https:// www.supremecourt.uk/decided-cases/docs/ UKSC_2011_0212_Judgment.pdf (accessed 7 June 2015). 8. Scottish Parliament. Mental Health (Scotland) Bill Policy Memorandum, http://www.scottish.parliament.uk/S4_ Bills/Mental%20Health%20(Scotland)%20Bill/b53s4introd-pm-bookmarked.pdf (accessed 16 October 2014). 9. Scottish Parliament. Mental Health (Scotland) Bill [as introduced], http://www.scottish.parliament.uk/S4_ Bills/Mental%20Health%20(Scotland)%20Bill/b53s4introd-bookmarked.pdf (accessed 16 October 2014).
The results of this study suggest that a history of excessive alcohol consumption and a diagnosed learning disability may be associated with the outcome of appeals against detention in conditions of excessive security. This should prompt further investigation with a larger sample size. Given the nature of the field, this may have to wait until more patients have lodged appeals. Moreover, when such appeals become available in conditions of lower security this will merit investigation in these new settings. A qualitative study of the decision-making by the Tribunal may elucidate Downloaded from msl.sagepub.com at Monash University on November 14, 2015
Original Article
Appeals against detention in excessive security (outcomes of appeals against detention in conditions of excessive security in Scotland)
Medicine, Science and the Law 0(0) 1–6 ! The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0025802415590176 msl.sagepub.com
Alexander Slater1, Daniel M Bennett2, Gabriele Vojt3 and Lindsay Thomson1,4
Abstract The Mental Health (Care and Treatment) (Scotland) Act 2003 introduced the right for patients in high-security psychiatric care to appeal against detention in conditions of excessive security. A previous study examined the first 100 patients to appeal under this provision. In this study we compare them with the next cohort of 110 patients to lodge an appeal, finding, contrary to expectations, no change in patient characteristics or the outcome of their appeals. The clinical, legal and demographic features of successful and unsuccessful appellants, who made up 38% and 27% of the 110 patients, respectively, were also compared. Those patients with the support of their responsible medical officer and those already included on a transfer list had a significantly better chance of success (p ¼ 0.00). It was also found that a history of excessive alcohol consumption was associated with successful appeals (p ¼ 0.002). A diagnosis of learning disability was associated with unsuccessful appeals (p ¼ 0.018), though the sub-sample was very small. These findings are important given the forthcoming extension of this right of appeal to other levels of security. Keywords Forensic psychiatry, law, medical law, legal system, high security, excessive security
Introduction
the full security provided.4 The Mental Health (Care and Treatment) (Scotland) Act 2003, hereafter referred to as ‘the 2003 Act’, confirmed the principle that care must involve the ‘minimum restriction’ sufficient for the patient. The suggestion that patients were held in unnecessarily high security led to two developments in Scottish forensic psychiatry: the establishment of three medium-secure units and the unique right for patients to appeal against detention in conditions of excessive security (rather than detention per se) under the 2003 Act. The patient can make an application to the Mental Health Tribunal for Scotland, hereafter referred to as ‘the Tribunal’, for a declaration that they are being
High-security psychiatric care for Scotland and Northern Ireland is provided by the State Hospital at Carstairs. Its patients require high-security care due to their ‘‘dangerous, violent or criminal propensities’’.1 High security not only involves environmental measures, such as perimeter fences, but also relational and procedural aspects.2 Relational security refers to ‘‘the knowledge and understanding staff have of a patient and of the environment, and the translation of that information into appropriate responses and care’’.3 Procedural security refers to the protocols in place to control the patients’ environment and keep staff and patients safe; this includes regulating the patients’ possessions. The gathering of information 1 The University of Edinburgh, Edinburgh, UK on the history and progress of individuals in high2 University of Aberdeen, Aberdeen, UK 3 security care also allows a degree of tailoring in Risk Management Authority, Paisley, UK 4 The State Hospitals Board for Scotland and the Forensic Mental their security provision, relevant to their Health Services Managed Care Network, Carstairs, UK requirements. Whilst high-security care is undoubtedly necessary Corresponding author: for some, a study in 1997 suggested that 53% of the Lindsay Thomson, Division of Psychiatry, The University of Edinburgh, State Hospital’s patients, in the view of their Royal Edinburgh Hospital, Edinburgh EH10 5HF, UK. Responsible Medical Officer (RMO), did not require Email: [email protected] Downloaded from msl.sagepub.com at Monash University on November 14, 2015
2 held in conditions of excessive security: it is this process which we refer to as an appeal. If the Tribunal upholds an appeal, they order the health board responsible for the patient to relocate them to a lower-security setting within a specified period of up to three months. The appeal does not concern whether the patient should be detained, but specifically whether they require high-security care at the State Hospital. The Tribunal may reconvene twice more to grant the responsible health board up to seven months in total to relocate the patient. Should the responsible health board fail to meet this obligation, its executive directors may be fined or imprisoned.5 The reasons why the Tribunal makes a declaration of detention in conditions of excessive security are of significant interest to the forensic service. A recent study, hereafter referred to as the ‘Bennett study’, examined the clinical, legal and demographic features of the first 100 patients to make such an appeal.6 This showed that successful appellants more often had their RMO’s support and were more likely already to be included on a transfer list. No other significant differences were found between successful and unsuccessful appellants. It had been thought that the characteristics of patients making an appeal would change following the transfer of the accumulation of patients who would be better cared for in medium-security units. However, amongst the first 100 patients to appeal this assumption only seemed to be true for women, all of whom were successful. The anticipated shift in the characteristics of appellants, if it came to pass at all, may have only occurred after the first 100 appeals. This study aims to investigate the demographic, legal and clinical features of the 110 patients who have made appeals since the Bennett study. This includes 33 patients who had lodged an appeal during the period studied by Bennett et al. Between the two studies, all patients were included who had ever lodged an appeal. The hypothesis of this study is that those who are successful will be more likely to have the support of their RMO and already be on the transfer list. They are expected to be similar in all other characteristics. These 110 patients are also compared with the 100 patients in the Bennett study to look for evidence of a change in the characteristics of patients making appeals. It is postulated that fewer recent appeals will succeed, since the initial group of inappropriately placed patients has moved on.
Medicine, Science and the Law 0(0)
Sample Appeals against detention in conditions of excessive security were concluded for 110 patients between midJuly 2008 and the end of July 2013. The sample included every patient who lodged an appeal since the Bennett study, 33 of whom had lodged an appeal during the period studied by Bennett. The outcomes of these appeals were obtained from a Forensic Mental Health Managed Care Network Database. Mid-July 2008 was specified in order to continue from the date at which the Bennett study concluded. Only a patient’s first appeal during this time period was included in this sample.
Sources for data collection Data were gathered by retrospective case note analysis, using a coding sheet adapted from Bennett et al. The data included: demographic details; psychiatric history; psychiatric diagnoses; drug prescriptions; reason for admission; the legislation under which the patient was detained; and drug and alcohol history. Factors which changed over time, such as prescriptions, were recorded from formal review documents. Information regarding the patient’s appeal was gathered from documents provided by the Tribunal or legal correspondence. Data concerning whether the patient was already on a transfer list were gathered from the transfer list database. Patients were recorded as already being included on a transfer list if they were as such on the date they lodged an appeal. If the RMO’s support for an appeal could not be determined from the notes then, where possible, the RMO was asked to indicate their position from memory. This was the case for 21 of the 61 appeals for which the RMO’s support was ascertained.
Statistical analysis
The data were analysed using the programme SPSS Statistics 21 (IBM, Chicago, Illinois). A range of descriptive techniques were used to summarise the data. To identify any difference between successful and unsuccessful appellants in this study, the data were first put into contingency tables for each of the factors studied. Fisher’s exact test was then used to look for a significant association between each factor studied and the outcome of appeal. Those factors which showed a significant association with the outcome of appeal were included in a binary logistic Methods regression model to determine whether they could Literature search predict the outcome. A two-step approach was used to compare the curA literature search was carried out using the rent and Bennett et al. sample: (i) a non-parametric PsycINFO database. A combination of the search Levene test to look for homogeneity of variance; if the terms ‘‘Forensic psychiatry’’ AND ‘‘Security’’ AND data failed this test it showed a significant difference ‘‘Appeals’’, gave 18 results. Only the Bennett study between the two groups; (ii) if the data were similarly was relevant to this study. Downloaded from msl.sagepub.com at Monash University on November 14, 2015
Slater et al.
3
distributed, a Kruskal–Wallis test was applied to look for a significant difference between the median of the groups.
having controlled for the effect of other predictors in the model.
Statement of principal findings Results Of the 110 patients to make an appeal between midJuly 2008 and July 2013, only one was female. The mean length of their admission to the State Hospital (7.0 years) was similar to that of patients in the Bennett study (7.1 years). The characteristics of successful appellants were compared with those who were unsuccessful (Table 1). Using contingency tables and Fisher’s exact test identified four significant differences. Successful appellants were more likely to have the support of their RMO and already to be included on a transfer list. They were also less likely to have a diagnosis of learning disability – although the subsample was very small – and more likely to have a recorded history of excessive alcohol consumption. Next, the four factors significantly associated with the outcome of appeal were used in a binary logistic regression model, the results of which are shown in Table 2. This was based on the 47 patients for whom there was data available relating to all four domains. Used together, these factors were able to predict 96% of appeals which were upheld and 95% of appeals which were not upheld. Beta weights were not standardised and positive values show an association with failure at appeal (since the code for appeal not upheld was 1 in the model, compared with 0 for appeal upheld). The odds ratio relates to the odds of failure at appeal,
The findings of this study are similar to those of the Bennett study. Having RMO support and already being included on a transfer list were significantly associated with success at appeal. This study also identified two clinical characteristics which showed an association with the outcome of appeals, namely, that having a diagnosis of learning disability was associated with unsuccessful appeals, whilst a history of excessive alcohol consumption was associated with success at appeal. However, the sample sizes were small, especially for learning disability. Used together, these four factors were able to predict the outcome of more than 95% of appeals in a binary logistic regression model.
Strengths and weaknesses of the study One of the strengths of this study is that it includes all the patients who have lodged an appeal since the previous study concluded. Between the two studies, all the patients to have made an appeal against detention in excessive security in Scotland are included, albeit the total is only 210 patients. For some characteristics, such as having a learning disability, the number of patients was as few as seven. This is not sufficient to draw any definitive conclusion and, therefore, the associations of alcohol excess and learning disability with outcome of appeal should be considered as requiring future investigation. Another potential
Table 1. A comparison between successful and unsuccessful appellants. Variable
Successful (N)
Unsuccessful (N)
Fisher’s Exact Test (p-value)
Patient background (demographic and previous psychiatric treatment) – no significant difference between groups Offence, or behaviour leading to admission – no significant difference between groups Record of substance misuse History of excessive alcohol consumption 28 9 0.002 Illicit drug use 33 21 0.289 Recorded history of IVDU 7 2 0.185 Diagnoses Schizophrenia 29 21 0.571 Learning disability 1 6 0.018 Personality disorder 11 9 0.462 Depression 2 1 0.626 Primary diagnosis (schizophrenia vs. other) (Schizophrenia) 29 20 0.515 Prescribed medication – no significant differences between groups Circumstances of the appeal Appeal supported by RMO (i) 26 1 0.000 Patient on transfer list 31 6 0.000 Had lodged previous appeal 13 8 0.450 (i) Only 47 of the 110 patients had data in this category Downloaded from msl.sagepub.com at Monash University on November 14, 2015
4
Medicine, Science and the Law 0(0)
Table 2. Logistic regression using patient characteristics to predict the outcome of their appeal. Independent variable
Beta weighti
History of excessive alcohol consumption 16.554 Learning disability 20.017 Appeal supported by RMO 36.707 Patient on transfer list 35.321 Model 2 56.333 p ¼ 0.000 Pseudo R2 (Nagelkerke method) 0.938 (scale is 0–1) N ¼ 72
SE
Wald
Sig.
Odds ratio
4541 15707 6992 6992
0 0 0 0
0.997 0.999 0.996 0.996
0 493445089 0 0
i
Beta weights were not standardised and positive values show an association with failure at appeal (since the code for appeal not upheld was 1 in the model, compared with 0 for appeal upheld). The odds ratio relates to the odds of failure at appeal, having controlled for the effect of other predictors in the model.
been subject to recall bias. Finally, this study does not include any qualitative analysis of the decisionmaking process of the Tribunal. Therefore, the mechanism of any association can only be described speculatively.
Strengths and weaknesses in relation to the Bennett study
Figure 1. A comparison of the outcome of appeals in the Bennett study and the current study. Bennett outcomes (above). Current outcomes (below). Key: Green¼upheld; Red¼not upheld; Blue¼withdrawn/cancelled; Yellow¼adjourned.
This study used a similar approach to data collection and analysis as did the Bennett study. Instead of using a Kruskal–Wallis test to compare successful with unsuccessful appellants, a Fisher’s Exact Test and binary logistic regression were used to determine whether the outcome could be predicted, based on patient characteristics. It was hoped that this would reinforce the associations identified in the Bennett study. The Bennett study had the advantage of a complete data set regarding whether an appeal had RMO support; a factor which was only known for 47 of the 110 patients in this study. The cohort of patients in this study had similar rates of success to those included in the Bennett study (Figure 1). The characteristics of the 110 patients were compared with those in the Bennett study to examine change over time (Table 3). No change was evident, whether the three adjourned appeals from the Bennett study were included or excluded on account of the fact that this study only considered appeals which had reached a conclusion.
Meaning of the study: possible mechanisms and implications for clinicians and policy makers
The association of RMO support and already being included on a transfer list with success at appeal is explicable. As discussed in the Bennett study, these factors are linked because the opinion of the RMO, weakness in the statistical comparison is that 33 of the as part of the clinical team, advises whether the 110 patients had been included in the previous study. patient should be included on a transfer list. The supIt is possible that this could have contributed to the port of the RMO may be influenced by many factors, similar results found in both studies. Regarding the 21 both current and historical, including the patient’s patients for whose appeals RMO support was clinical condition, perceived risk and insight. In this obtained from memory, this couldDownloaded possibly have from msl.sagepub.com at Monash University on November 14, 2015
Slater et al.
5
Table 3. A comparison between patients in the Bennett study (N ¼ 100) and those in this study (N ¼ 110). Variable Patient background Gender (Male) Marital status (Married) Responsible Health Board (Greater Glasgow and Clyde) (Rest of Scotland) (Northern Ireland) Previous treatment (In-patient) Legal provision (Restricted patient) Source of admission (Hospitals) (Legal system) Diagnosis (not necessarily primary) Schizophrenia Learning disability Personality disorder Depression Record of substance misuse History of excessive alcohol consumption Circumstances of the appeal Appeal supported by RMOi Patient on transfer list Outcome of hearingii (Successful appeal)
Bennett
% (N)
Current % (N)
Test
P
KWT
0.076
KWT
0.669
KWT
0.484
KWT
0.303
KWT
0.780
KWT
0.408
95 (95)
99 (109)
3 (3)
1 (1)
46 (46) 50 (50) 4 (4)
46 (50) 48 (53) 5 (5)
78 (78)
66 (72)
63 (63)
63 (69)
50 (50) 50 (50)
45 (49) 51 (56)
62 (62) 20 (20) 42 (42) 5 (5)
67 (74) 14 (15) 31 (34) 3 (3)
KWT KWT KWT KWT
0.425 0.218 0.096 0.391
53 (53)
54 (59)
KWT
0.927
59 (58) 59 (59)
52 (32) 49 (54)
KWT KWT KWT
0.326 0.151 0.358
43 (43)
38 (42)
KWT¼Kruskal–Wallis test i Based on 164 cases for which RMO support was known (99 from Bennett study and 65 from this study). ii Based on 97 cases from Bennett study (excludes those adjourned) and 110 new cases.
to potential destabilisers in a lower-security environrespect, RMO support could be considered as a proxy ment as ‘testing out’. It is possible that prompt idenmeasure to the patient’s condition and the wider clintification and treatment of patients with a history of ical team’s view. excessive alcohol consumption might expedite their If learning disability is associated with failed transfer to a lower-security setting. appeals, the explanation may be that it is less amenThere is no evidence to suggest that these appeals able to treatment than other psychiatric conditions. have changed the characteristics of the patients Without a change in the underlying clinical condition appealing against their detention at the State which led to admission, it is less likely that there will Hospital or that more recent appeals are less likely be a change in the level of security provision required to be successful (see Figure 1). This is interesting by a patient. The risk that they pose to themselves or since the appeals included in the Bennett study others is not so easily decreased by treatments. might be expected to have had a higher success rate, Conversely, excessive alcohol consumption is relaon account of the many patients thought to have been tively responsive to treatment. Since the State held in conditions of excessive security before the 2003 Hospital is a maximum security environment, access Act came into effect. to alcohol should be impossible and abstinence is The Supreme Court’s ruling in the case of RM vs. enforced. The State Hospital also has a programme The Scottish Ministers [2012] UKSC 58 obliges focusing on alcohol and drug abuse, in which each Scottish Ministers to bring forward regulations definpatient is assessed and provided with psychological ing the terms under which excessive security appeals treatment. It may, therefore, be easier for patients will be extended beyond the State Hospital.7 These with alcohol-related problems to demonstrate that they have remained abstinent and engaged with treatnew appeals are legislated for by section 268 of the ment. The Tribunal may be persuaded to expose them 2003 Act. Scottish Ministers need to define Downloaded from msl.sagepub.com at Monash University on November 14, 2015
6 ‘‘qualifying patients’’ and ‘‘qualifying hospitals’’ for these new appeals. The current consultation on the Mental Health (Scotland) Bill includes the proposal to limit extending the appeal against excessive security only to mediumsecure units.8 Appeals against excessive security were introduced in the knowledge that there was a significant cohort of patients unable to move on from high security. This provision has led to a significantly changed forensic estate with a much reduced State Hospital (from 240 to 140 beds) and the development of two new medium-secure units in addition to a pre-existing medium-secure unit. This has considerably alleviated the problem of entrapped patients within high security. The proposed extension of the appeal against excessive security to medium security may result in the development of the low-secure estate. However, these proceedings are stressful to patients, divert clinical time and resources away from the direct care and treatment of patients and have a financial cost associated. It remains to be seen how this will apply to those treated within medium-secure units, where independent access to the community is entirely possible and is often a clinical aim. The Bill also proposes the repeal of Section 266, such that a Health Board would have a maximum period of six months to find a suitable alternative placement for a patient declared to be held in conditions of excessive security.9 In order to implement section 268 appeals a good understanding of the function of section 264 appeals is essential. This study offers some indication of the likely outcome of appeals. Moreover, having found strong predictive factors of success suggests that a sifting step may be possible to operate as part of the appeal process. Of course, the RMO cannot act as the gatekeeper for appeals, but evidence of engaging with treatment and progress may provide useful information. Such a step would be important in minimising the waiting time for an appeal to be heard, as well as being cost effective. It would also address the clinical concern that extending the provision of appeals may lead to conflict between patients, their RMOs and clinical teams, as well as diverting much time and resources to the legal process, without providing significant benefit to the patient.
Unanswered questions and future research
Medicine, Science and the Law 0(0) the mechanism of the associations identified by this study. Finally, a formal cost-benefit analysis may be helpful in describing whether a sifting step for appeals would be effective and just. Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Declaration of conflict of interest None declared.
Ethical approval An ethical assessment form was submitted to the University of Edinburgh for this project and, based on the methodology, no further action was required.
Statement of guarantor Professor Lindsay Thomson is the guarantor for this work.
Statement of contributors Alexander Slater collected and analysed the data. Lindsay Thomson supervised and advised on all aspects of the project. The above drafted and edited the manuscript with the help of Daniel Bennett, on whose previous work this study was based. Gabriele Vojt advised on the method for data analysis.
References 1. National Health Service (Scotland) Act 1978 S.102. http://www.legislation.gov.uk/ukpga/1978/29/section/ 102 (accessed 7 June 2015). 2. Forensic Network. Definition of security levels in psychiatric inpatient facilities in Scotland, http://www.forensicnetwork.scot.nhs.uk/documents/previous_reports/ LevelsofSecurityReport.pdf (accessed 27 January 2014). 3. Department of Health Secure Services Policy Team. Your guide to relational security, http://www.rcpsych. ac.uk/pdf/Relational%20Security%20Handbook.pdf (accessed 27 January 2014). 4. Thomson LDG, Bogue J, Humphreys M, et al. The State Hospital Survey: A description of psychiatric patients in conditions of special security in Scotland. J Forensic Psychiatry 1997; 8: 263–284. 5. Thomson LDG. The mental health (care and treatment) (Scotland) Act 2003: Legislation for mentally disordered offenders. Psychiatr Bull 2006; 30: 423–429. 6. Bennett D, Skilling G, Brown K, et al. Appeals against detention in conditions of excessive security in Scotland. J Forensic Psychiatry Psychol 2013; 24: 386–402. 7. RM v The Scottish Ministers UKSC 58 (2012), https:// www.supremecourt.uk/decided-cases/docs/ UKSC_2011_0212_Judgment.pdf (accessed 7 June 2015). 8. Scottish Parliament. Mental Health (Scotland) Bill Policy Memorandum, http://www.scottish.parliament.uk/S4_ Bills/Mental%20Health%20(Scotland)%20Bill/b53s4introd-pm-bookmarked.pdf (accessed 16 October 2014). 9. Scottish Parliament. Mental Health (Scotland) Bill [as introduced], http://www.scottish.parliament.uk/S4_ Bills/Mental%20Health%20(Scotland)%20Bill/b53s4introd-bookmarked.pdf (accessed 16 October 2014).
The results of this study suggest that a history of excessive alcohol consumption and a diagnosed learning disability may be associated with the outcome of appeals against detention in conditions of excessive security. This should prompt further investigation with a larger sample size. Given the nature of the field, this may have to wait until more patients have lodged appeals. Moreover, when such appeals become available in conditions of lower security this will merit investigation in these new settings. A qualitative study of the decision-making by the Tribunal may elucidate Downloaded from msl.sagepub.com at Monash University on November 14, 2015
