1 INTRODUCTION

The police frequently fulfil the role of gatekeeper for people experiencing mental health problems. There is some evidence of criminalisation of the mentally ill if this policing role is not performed appropriately (Lamb et al., 2014). Other research has shown that police involvement with individuals suspected of minor offences and flagged as experiencing mental health problems are more likely to be arrested than those who are not. Indeed, for offences of equivalent severity, a person with mental state problems is twice as likely to be arrested as one without (Charette et al., 2014) and are more often charged and spend longer in police custody (Kane et al., 2018). This type of encounter may require officers to have specific training and skills (Lamb et al., 2014) and depend on the availability of hospital and community mental health resources for successful outcomes (Kane et al., 2018). The same individuals who have unresolved needs are likely to have repeated contacts. Occasionally, they may involve volatile situations, risking the safety of all involved (Reuland et al., 2012). In England and Wales, the police are given powers under section 136 (s136) Mental Health Act (MHA) 1983/2007 to remove a person from a public place, when they appear to be suffering from a mental disorder, to a place of safety to facilitate assessment of their health and well-being as well as the safety of other people around them (Thomas & Forrester-Jones, 2019).

Clearly, there are complexities facing police and individuals experiencing mental health problems when the two come into contact, whether the individuals are victims, suspects or witnesses. The fairly extensive related literature, however, has significant gaps and few high-quality studies (Kane et al., 2017). The level of police resource tied up by mental state-related incidents is a topic largely devoid of substantial, objective research. This has been complicated by the often inexact and idiosyncratic categorisation of mental state-related incidents by different police forces and key individuals within them. A mental state-related flag on a police Command and Control System (CCR) could have been generated in ways ranging from an inaccurate or out of date note on the system, by a call handler with little training, the caller himself/herself, or have involved a mental health professional or information from a healthcare record. Consequently, the accuracy of the flag varies a great deal, with a danger that such a mixed and un-systematised categorisation leaves too much room for un-evidenced claims and counter claims around demand, response and resources (Kane et al., 2017).

An important part of understanding the perceptions of the police in relation to individuals with mental state problems is how they construct their view of these problems and those experiencing them. There is an extensive literature on the impact of social construction on official data and its impact on how different groups and communities are perceived by organisations and their members (see Bor et al., 2018; Javaid, 2018; Sampson & Raudenbush, 2004). Social construction assumes not only that people create, make or invent their understandings of the world and the meanings of encounters with others but also that they do this jointly, in coordination with others rather than individually (Leeds-Hurwitz, 2016). How does this work for police in their work encounters with people with mental health problems?

This study was commissioned and carried out as a result of a decision by the National Police Chiefs' Council (NPCC) to develop a better understanding of mental state-related demand on police resources. This decision was prompted by concerns expressed about the impact of incidents relating to people with mental state problems on police resources by Her Majesty's Inspectorate of Constabulary Fire and Rescue (HMICFRS; Picking up the Pieces, 2018). Our study involved several elements (1) a national exercise using a 24-hour, national snapshot census carried out on 12 November 2019 in all forces in England and Wales (the date was chosen by the commissioners of the study); (2) an analysis of 12 months of more in-depth data from three case study forces ranging from large metropolitan, mixed urban and predominantly rural forces; and (3) a series of semi-structured interviews and focus groups with a wide range of warranted officers and civilian staff of all ranks in these and seven other forces.

The study was designed to capture, first, the number of incidents that met the criteria set out in a new NPCC and College of Policing (CoP) definition of what constitutes a mental state-related incident, developed after inspections by HMICFRS had identified that forces were using different definitions. The new definition was published in 2019. The definition states ‘Any police incident thought to relate to someone's mental health where their vulnerability is at the centre of the incident or where the police have had to do something additionally or differently because of it’ (NPCC & CoP, 2020).

The other main aim of the study was to understand the nature and impact of incidents meeting this definition.

2 METHODS

3 RESULTS

4 DISCUSSION

It seems clear from both the snapshot and the extended data from the case studies that we analysed, that actual mental state-related incidents are significantly fewer than the anecdotal evidence suggested. The data from the 24-hour national snapshot, the case studies in three very different forces and the detailed analysis of 320 cases, indicate that while mental state-related incidents do consume police resources, some taking more time than others and can be stressful and complex to deal with, both for the police and for the individuals involved, they do not represent a disproportionate amount of demand on police time and resources compared to non-mental state-related incidents. However, the data we collected does indicate that only about a quarter of what the police dealt with were crimes. This in itself may amplify their feeling that they spend a lot of their time dealing with social issues, including those related to mental state problems.

There is a relatively commonly held view that mental state tags/flags in police systems under-record mental state incidents; however, the in-depth work we did as part of this project indicates that even when a sample of 320 mental state-related incident cases was analysed in detail, no significant level of under-recording was found. In fact, in one of the case studies, over-recording seemed to have been the issue. This is in line with other published quantitative research that has shown that in fact mental state tags, particularly those in CCRs, tend to over identify mental state-related incidents (Kane et al., 2018). Studies that have, for example, used text mining algorithms can also over record because they often search on any mention of mental state/disorder or related terms regardless of the accuracy of their application, currency or, now, the likelihood that the new NPCC definition is being applied. It is also probable that the application of the new definition, where it occurred, has produced a more accurate reflection of actual demand rather the reliance on historic police systems flags, outdated local records and misattribution of mental state problems to an individual caller, for example. We know from other research that it appears that it is the application of the mental state flag, accurate or not, that triggers a differential response from police staff such as, for example, blue light responses, specialist team deployments and influences the on-street decision-making by officers (Kane et al., 2018). Improving the accuracy and application of this mental state flag in line with the NPCC & CoP definition is an area that could help manage demand more effectively than at present. Some forces already audit their mental state incidents in detail on a frequent basis and have been successful in weeding out some of the historic and inaccurate flags being applied. The new NPCC & CoP definition is also a step in the right direction if used across forces and rigorously backed up by clear, supportive force policies.

It is clear from our data that police staff, warranted and civilian, are dealing with a significant and potentially increasing number of vulnerable individuals; however, the data we have analysed in this study do not appear to support the contention that this demand is chiefly associated with mental state-related incidents. It seems likely, from the data we collected, that a significant number of the cases often being flagged as mental state-related incidents are in fact connected to other vulnerabilities rather than the individual's mental state. These most commonly include homelessness and drug and alcohol abuse. It is important if progress is to be made on understanding demand on policing created by all vulnerabilities, that the debate is broadened to include a wider range of national and local partners beyond just the Department of Health, National Health Service England and local health commissioners and service deliverers. In some areas, Crisis Concordat arrangements have helped develop more robust partnerships but in others partnership working around vulnerable individuals, including those experiencing mental state-related problems, appears weak. The Mental Health Crisis Care Concordat is a national agreement between services and agencies involved in the care and support of people in crisis. It sets out how organisations will work together better to make sure that people get the help they need when they are having a mental health crisis. Each area in England is required to action these arrangements locally (HM Government, 2014).

What we could not do in this project, and may be an avenue for future research, is to explore in detail whether there is any co-relation between the actual demand on policing and the performance and resources of other local public and third sector services, as well as mental health providers and commissioners. We did carry out some initial scoping of the health service aspects of this using open-source data and regression analysis, but the data were difficult to match at a sufficiently granular level and consequently the results were not clear and hence not reported here. More granular local data are needed to do justice to this area. Interestingly, it was clear from our interviews with police staff and other partner agency staff that demand was not perceived as high or as problematic where there were good relationships and trusted inter-agency ways of working, even though the data showed their demand was within the national profile and therefore broadly the same as other areas. Where these relationships and inter-agency working were not mature, demand was usually seen as high, even when falling firmly inside the national profile range. It is likely that many of the vulnerable individuals who are placing demands on police resources are in the remit of housing, social care, drug and alcohol services and related agencies, and it is partnership with them as well as health that will lead to reduced police demand and improved services for individuals.

It should also be borne in mind that given the general incidence of expressed mental state-related problems (around 1:6 individuals say they have experienced some mental state issues in the previous week (McManus et al., 2016) there will inevitably be contact between police and those who feel they are experiencing mental state-related problems. It is also important to recognise that the fact an individual is experiencing such problems does not of itself mean that there should be a differential police response unless their mental state has led to a particularly enhanced level of vulnerability or risk, for example.

5 LIMITATIONS

All data we present in this report are from police forces in England and Wales. We did find some data issues. Many of the forces found it a challenge to share complete data on mental state incidents, with consistency in some areas an issue. For example, when forces were asked about their data quality, some of the issues they reported included: 79% had mental state-related data recorded on a stand-alone system (as such the data were not integrated with the forces main systems and available to all who needed it), forces tended to think the data available to measure the time spent on mental state-related incidents were not accurate – 36% said they were very or fairly accurate, and 46% said the data were not very or not at all accurate, only two thirds of forces could provide data on the use of MHA Act 1983 section 136.

Data were gathered from a 24-hour snapshot and may not have been a typical profile for any given force. This is mitigated by the fact that when we looked at the three in-depth forces their 12-months data reflected the same or lower levels of mental state-related incidents.

We are aware that the use of the new definition was not consistent among forces, and this may also have had an effect on the numbers of incidents included by some forces, but should have helped the data from the forces using it to provide a more accurate reflection of demand. Our checks of mental state-related incidents in the case study forces did not suggest substantial over or under-recording of mental state-related incidents.

CONFLICT OF INTEREST

There were no conflicts of interest. The research team carried out the work independently and the findings reflect our views.