Our main findings were that both the numbers of assessments and of detentions increased in the first two years after implementation of the MHA 2007 but fell in the third year. The rates of detention, as a proportion of assessments, remained about the same. There were no assessments or detentions of patients with sexual deviance. There was a significant increase in the proportions of patients assessed under the MHA who had mental retardation or who could be labelled as having ‘any other disorder or disability of the mind’ after implementation of the 2007 Act. There was no significant change in the proportions of patients detained with PD, LD or disorders usually occurring in childhood and adolescence. These findings indicate that the underlying aim of facilitating the compulsory care of patients with PDs, sexual deviance and LDs was not achieved in the first three years post-implementation of the MHA 2007.
Many socio-political and service provision-related factors can act to change the way the MHA is used or revised, complicating the interpretation of data of this nature [
10]. For example, although rates of detention have been rising for a number of years, the overall rate of admission for mental disorders has fallen as community-based alternatives have been provided; reduction of in-patient beds within the NHS has, however, been closely associated with an increasing rate of detention to NHS hospitals since the 1980s, especially when analysis includes a time lag of one year between bed provision and subsequent involuntary admissions [
11]. Public and government concern about limiting risks posed by dangerous individuals have even produced new categories of mental disorder, such as Dangerous Severe Personality Disorder, without any other theoretical or scientific pedigree [
12,
13].
Legal and ethical issues can influence research methodology. For example, the data on multiple assessments for the same patient can not be collated when a study such as AMEND is on MHA assessments and not individuals. This is because ethics and data protection law-related requirements prevent social care and mental health provider organisations passing patient identifiable details to research teams. A secondary analysis of assessments of the same patient is therefore precluded by research teams’ inability to cross check data from the two types of service (social care collecting data on assessments and mental health trusts on patients and their care).
Strengths and limitations of the study
The sample size calculation for this study was robust. Expected numbers of assessments per day were consistent with published data [
14].
There can be variation between geographical areas and service provision. Analysis, however, of anonymised records of over 1.2 million people for 2010/11 held in the Mental Health Minimum Data Set (the mandatory return from providers of National Health Service-funded mental health care in England), Census data and data from commissioning and provider healthcare organisations indicated that 85% of the variance in detention rates occurred between individuals; only around 7% of statistically significant variance occurred between places (Census areas) and around 7% again between provider healthcare organisations [
15]. There was also heterogeneity in the socio-economic and ethnic make up of the catchment populations and service delivery across the three sites, making the findings of AMEND nationally generalisable.
The AMEND study only covered a limited period between 2008 and 2011. We do not know how representative the 2008 data was of assessments undertaken in previous years.
Although multiple sources for data collection were used and cross-checked, it is possible that some MHA assessments were missed. Given that not all assessments had all of their specific subcategories of diagnosis recorded, some groups (e.g. personality disorders) are only approximations of categories of interest. Approximately 70 cases were excluded from the analysis as the immediate outcome of the assessment was not clear from the notations from the database. Given the small numbers of forensic detentions, all of which occurred after the 2007 Act was implemented, conclusions could not be drawn about whether there was shift between forensic and civil sections.
Comparison with other studies
Other post-legislative reports on the MHA 2007 do not present detailed data on assessments or diagnoses of assessed or detained patients [
16‐
18]. Government data from the years leading up to implementation of the 2007 Act is available on formal admissions and detentions subsequent to admission under the MHA. Data includes detentions after being subject to a place of safety order, detentions in NHS facilities and independent hospitals. In 2006–07, 2007–08 and 2008–09 numbers detained were 46,539, 47,610 and 47,725 respectively [
19].
The pattern in the AMEND data of an increase in assessments and detentions from 2008 to 2010 and the subsequent decrease in 2011 is difficult to explain. When we compare AMEND data with Government data, we need to remember that the AMEND data was collected July to October in the years of 2008–2011 and Governmental data is collected for years running from 1st April to 31st March for each data year. The latter also does not give data on the numbers of assessments, collect data from Child and Adolescent Mental Health Services (CAMHS) or Learning Disability Services or provide information on diagnoses but did indicate that the total number of detentions (excluding Guardianship) increased by 13% over the five years from 2008 to 9 to 2012–13 [
20]. An interesting observation is that in the third year post implementation of the MHA 1983 there was also a slight dip in rates of detention in the context of year by year increases in numbers of detention from from 1984 to 1991 [
10].
Data from the Mental Health Minimum Data Set is difficult to compare with AMEND data as the former provides information on the number of adults of working age and people over the age of 65 years who spent time formally detained in an NHS hospital under the MHA [
21]. Each person is counted only once and in the category with the most restrictive legal status that applied to them in the year. The focus on detained patients makes invisible those who are assessed but not detained. The only other large-scale study of MHA assessments in England was published a quarter of a century ago and reported data collected in 1984 by the Social Services Research Group to monitor the implementation of the MHA 1983 [
22].
Implications for research and clinical practice
Despite the importance, good quality research and interpretation of data in this area is difficult because of the large number of explanatory, confounding and biasing variables. This in turn means that adequately powered studies, collecting data on the appropriate range of individual, socio-political, professional, service- and practice-related variables will be required. Governmental monitoring of the use of the MHA, e.g. through the Care Quality Commission, should also make visible the needs and experiences of those who are assessed but not detained, as their health and rights are also affected by implementation of the Act.
AMEND, unlike parallel Government data, did not collect information from private sector (independent) hospitals. Government data [
20] shows that although NHS hospital detentions in England increased from 2008 to 9 (with the exception of 2011–12) to 2012–13 (25,185, 27,475, 26,937, 27,855 and 28,779 respectively), there was a sharper rise in detentions to independent hospitals from 2011 to 12 (2761, 2712, 2620, 3045 and 3445 respectively). By the data year 2012–13, independent providers were responsible for the care of just over a quarter of people detained under the MHA. Some of the main independent providers are not however submitting all records pertaining to detentions and hence the significance of private provision may be rising even faster [
20]. Future research should be just as assiduous in collecting data from independent as NHS providers.
It will be important to enquire why the MHA 2007 may not have had the hoped for impact. Perhaps it reflects a lack of knowledge of the MHA amongst relevant professionals [
23] or the individual policies and practices of relevant agencies [
22]. Perhaps the role that capacity plays in decisions about whether to treat people against their will needs further enquiry [
24] as although people with personality disorder may experience high levels of mental distress, loss of capacity is rare [
25] and the majority of mental health practitioners have been shown to be concerned about detaining people with personality disorder against their will [
26]. Another possible explanation is a consequence of the very common association between serious personality disorder and severe mental illness, a combination very likely to be associated with compulsory admission [
27]; In practice clinicians may ignore the personality disorder when making a diagnosis of mental illness. The diagnosis of personality disorder then remains covert even though it is one of the main drivers of coercion when detaining a patient under the Act. It is likely that qualitative studies, such as AMEND’s parallel enquiry into mental health professionals’ understanding and interpretation of the Appropriate Treatment Test, will be needed to add to our understanding of the reasons why changes to the 1983 Act appear to have had such little impact on clinical practice.
Future research in this area could also benefit from the additional use of validated, structured clinical interview schedules to establish the presence or absence of specific diagnoses rather than using approximations of diagnostic categories. This would help in any comparison of practice in differing jurisdictions, however, we acknowledge that there would be significant ethical, practical procedural challenges that would make such research very difficult to execute and complete.