Suicide prevention as biopolitical surveillance: A critical analysis of UK suicide prevention policies

Data

We examined a total of eight government produced suicide prevention policy documents, two from each of the four UK nations, in use between 2009–2019 in England, Northern Ireland, Scotland, and Wales.

Analysis

Our analysis of the eight policy documents was guided by Bacchi's WPR post-structural policy analysis framework which aims to ‘question the taken-for-granted view prevalent among many … that policy problems are self-evident, and that subjects, objects, and places simply exist' (2016: 108). We have used WPR in an earlier examination of suicide prevention policy (Marzetti et al., 2022) and the current article expands upon our previous analysis, presenting a theoretically informed and focused consideration of the implications of a surveillance driven approach to suicide prevention. WPR challenges the notion that policies simply address problems that are exogenous, and which are ‘discovered’ by and independent of governmental practices. Instead, influenced by Foucault, this approach recognises that policies produce problems in particular ways (Bacchi and Goodwin, 2016: 6) and the ways in which problems are represented within policy influences how they are understood and responded to by government, experts, services, and publics.

The WPR framework provides researchers with six analytical questions (Bacchi and Goodwin, 2016: 20). Our overall research project was broadly guided by all six questions; however, the specific analysis of surveillance within this paper is driven by two of the six questions: 1) ‘How has this representation of the problem come about?' and 2) ‘What effects are produced by this representation of the ‘problem?' Focusing on these two questions enabled us to 1) consider and defamiliarize the underlying surveillance rationality that was present within many of the suicide prevention policies that we examined and 2) consider the implications of a surveillance driven approach to conceptions of suicide.

Using the WPR approach as a guiding framework, the first stage of analysis involved exploratory, in-depth readings of the eight policy documents and open coding of each document supported by NVivo 12 qualitative analysis software. During this initial coding process, there was ongoing dialogue among the whole research team to identify and agree potential themes of interest After the first stage of coding, Authors A and D produced reports on themes (including ‘surveillance’), which were discussed amongst the team, with Authors B and C providing additional analysis. This collaborative and ongoing process enabled us to refine and articulate our interpretations

A surveillance centred suicide prevention regime

Surveillance practices and technologies played a central part in the suicide prevention policies. There was significant interest across many of the individual suicide prevention policy documents in strengthening and enhancing surveillance capacities and in maximising ‘data’ in terms of both gathering more data on certain selected population groups and monitoring population trends to manage the overall risk of suicide. This was emphasised via key objectives or commitments to enhance general population surveillance and monitoring capacities. For example:

Objective 6: Continue to promote and support learning, information and monitoring systems and research to improve our understanding of suicide and self-harm in Wales and guide action (Welsh Government, 2015: 17).

Across the documents there was clear consensus on the importance of surveillance technology within suicide prevention and a commitment to the role of ‘Reliable, timely and accurate suicide statistics…[as] the cornerstone of any suicide prevention strategy and of tremendous public health importance' (HM Government, 2012: 47). In keeping with WHO guidance (2016), contemporary UK suicide prevention relies heavily upon techniques and instruments of surveillance. An explicit desire was expressed in the policies to continue to enhance and strengthen these mechanisms, a typical commitment being to ‘expand and improve the systematic collection of and access to data on suicides' (HM Government, 2012: 47). To enable this population surveillance, specific tools and technologies were being developed to collect, categorise and cascade data. In the 2013 Scotland prevention policy there was significant mention of the Scottish Suicide Information Database (ScotSID) which was used to profile suicide risk and to shape policy directions. In England, the importance of ScotSID as a surveillance tool for profiling the population was highlighted (HM Government, 2012: 48) and by 2017 there was enthusiasm within English policy regarding the identificatory capabilities of a new surveillance instrument:

Public Health England has been developing further its Suicide Prevention Profile – often referred to as the Suicide Prevention Fingertips tool- which provides a range of data at a local level to support local authorities and the NHS to understand suicide trends in their areas. This tool provides valuable thematic data on local demographics which enables local areas to identify high risk groups and locations (HM Government, 2017: 34, emphasis added).

There was sometimes a tension within documents between the urge to encourage ever more surveillance, to generate ever more data, and the need to effectively utilise this data to construct knowledge. In England, there appeared to be frustration that:

A wealth of data is already collected… but only limited information is collected centrally or easily accessible and available to researchers or public health specialists (HM Government, 2012: 47, emphasis added).

Scotland, presumably because of the advanced development of ScotSID, was more confident about the use of surveillance data for knowledge generation, asserting that:

We have good evidence now about suicide, highlighting risk factors and providing an evidence base of effective interventions (Scottish Government, 2018: 18).

The importance placed on identification in solving the ‘problem’ of suicide thus ran through the policies. The Northern Ireland Government policies even defined suicide prevention as ‘identifying and reducing the impact of risk factors associated with suicidal behaviour…' (Northern Ireland Government 2012: 5, emphasis added). In other policies many actions and intentions relate back to identificatory objectives; for example, the Northern Ireland Government's 2016 document emphasises the importance of destigmatising suicide in part because a reduction in stigma ‘will increase the likelihood of early recognition of suicidal behaviour and suicide risk' (9). Policies which sought to advance cross government and cross sector collaboration and which emphasised broader social and economic contexts still retained a surveillance logic. For example, in a Welsh policy, there is recognition that preventing suicide requires a broad collaborative approach that moves beyond a narrow health focus on suicide but there is also an emphasis on the importance of ‘the identification and management of depression and other mental disorders' (Welsh Government, 2015: 14). The ministerial foreword of Scotland's 2018 policy document expresses optimism ‘that by working together across sectors, organisations and society, we can better identify and support people in distress, strengthen communities, and save lives' (Scottish Government, 2018: 2). This optimism seemed to be based in part upon the surveillance data generated by the ScotSID risk profiling surveillance system, with the Scottish policy seeking to strengthen this surveillance capacity, stressing that ‘it is essential that appropriate data about individuals is able to be shared' (Scottish Government, 2018: 19) through ScotSID. As such, even where policies highlight the role of support and collaboration, surveillance remained prominent. Surveillance is thus a central logic within suicide prevention policy.

This emphasis on surveillance existed in conjunction with a focus on ‘risk factors’. Indeed, surveillance enables ‘risk factors’ to become a central part of the definition of suicide prevention, with Northern Ireland defining the activity of suicide prevention as ‘identifying and reducing the impact of risk factors associated with suicidal behaviour' (Northern Ireland Government, 2012: 5, emphasis added). There is a nexus of surveillance and risk management:

Some groups of people …[are] at higher risk of suicide than the general population. We have been able to identify these groups from research and can monitor numbers from the routine data collected. In this way we identified:

Risk profiles are used to demonstrate the ability and viability of predictive health surveillance, for example, in the key learning section of one policy document, findings were shared on suicide risk:

Specific risk profiles identified included:

In addition to the specific uses of surveillance technology and commitments to the enhancement of surveillance and data generation which were prominent throughout the prevention policies, there also existed a more subtle surveillance discourse. Prevention policy documents in Northern Ireland, Scotland and Wales all used similar phrasing, informing readers that suicide prevention was ‘everyone's business'. For example, the ministerial foreword to the Northern Ireland Government's 2012 document ends with the words ‘Suicide prevention is everyone‘s business' (4); likewise, a ministerial foreword from Scotland claims ‘that suicide is preventable, that it is everyone‘s business' (Scottish Government, 2013: 1). A focus on identificatory practices within suicide prevention regimes is discussed in the Australian context by East et al., who analysed documents published by both government and charities. East et al. found that these documents exhorted the public to play an active role in identifying individuals who are experiencing suicidal thoughts (2021: 184). Such messages may be read as positive and action affirming, emphasising that we all have agency to prevent suicide. However, in linking prevention with identification, they serve to reinforce a construction of suicide as inherently visible, framing the primary role of prevention policy as the extension of this visibility.

Suicide as visible

There is a politics of visibility at the heart of the surveillance prevention approach which frames suicide as a ‘surveillable’ object, amenable to observation. Those who are experiencing suicidal distress are thereby framed as unproblematically identifiable to a societal gaze. Ceyhan has argued that surveillance generally takes the ‘visibility of the surveilled … as seriously as the process of observing, classifying and studying' (2012: 41).

Consequently, in this context, there is a conscious effort to train people to be able to spot suicide. For example, as part of a specific local intervention action plan, there is a commitment to:

developing suicide awareness and education or training programmes to teach people how to recognise and respond to the warning signs for suicide in themselves or in others (HM Government, 2012: 29, emphasis added).

Suicide is here constructed as both knowable and visible, resulting in a related imagining within prevention policy of a certain type of idealised and visibilised suicidal subject being imagined. This subject emits ‘warning signs', signals which we can all be trained to identify. This representation resonates with Morris’ critique of suicide prevention training in which he argues that there is a knowledge/power relation which constructs suicide as visible through the recognition and interpretation of specific warning signs which claim to ‘make the suicidal subject knowable and recognizable' (Morris, 2016: 85). The surveillance approach is predicated upon framing the identification of a suicidal subject as unproblematic, assuring us that we can, and we will, be able to know and recognise the suicidal subject through certain knowable warning signs. For example, ‘The majority of suicides have been preceded by warning signs…it is important to understand what the warning signs are and look out for them' (Northern Ireland Government, 2016: 16, emphasis added). The policies imply that, through training combined with the collection of data, suicide can be made visible and prevented; professionals and the wider public can be trained to identify suicidal behaviours/warning signs. For example, ‘educational providers (both at schools and colleges/universities) have an important role in identifying and supporting at-risk young people' (Scottish Government, 2018: 16, emphasis added). This is reinforced by the continued and consistent use of the phrase ‘suicide prevention is everyone's business' (see above), which encourages a form of disciplinary surveillance, an ever vigilant, ever watchful population. Suicide is the object of this gaze, with communities encouraged to participate in their own self-surveillance and self-monitoring (Ajana, 2007: 2).

There is also a broader politics of visibility within the surveillance approach to prevention, one which is explicitly stated in the foreword to the England 2012 policy document by Professor Louis Appleby, Chair of the National Suicide Prevention Strategy Advisory Group, who wrote that:

In identifying the high-risk groups who are priorities for prevention (action 1), we have selected only those whose suicide rates can be monitored – this is essential if we are to report on what the strategy achieves (HM Government, 2012: 4, emphasis added).

This statement demonstrates how groups are only prioritised for prevention if they are deemed amenable to monitoring. The groups which are prioritised for specialist interventions are selected based on their perceived visibility, on how effectively the state and its agencies believe that they can gather data, categorise, and monitor. Indeed, because of this surveillance driven risk management approach there comes the temptation to collect data which is available and collectable rather than to search for data that is actually relevant or useful (Power, 2004: 30). Some groups ‘whose risk is hard to measure or monitor, such as minority ethnic communities' (HM Government, 2012: 4, emphasis added) were largely placed outside of categorisations, and thus, outside of, and peripheral to, the policy gaze. Whilst there is some acknowledgement that provisions for minority ethnic communities are inadequate (Northern Ireland Government, 2012: 74), within the documents we examined there appeared to be little attempt to strategize interventions or increase visibility of these groups. Indeed, in the period we studied, ethnicity was not recorded on death certificates and thus ethnicity and suicide could not be made visible, unlike male suicide for example (Yue, 2021: 2). Indeed, it has only been since 2021 that the ONS has released information on ethnicity and suicide, despite gaps in knowledge being recognised as far back as the England 2012 policy document.

Whilst some of the policies acknowledge contextual factors such as deprivation as being relevant to suicide, frequently the policies pull back from specifying actions or commitments to actual economic and social changes, instead favouring ever more and enhanced surveillance, seemingly as an end in itself. This can clearly be seen in discussions of training, which are prominent within the documents and build directly upon the surveillance approach. In a Scottish policy document, there is a focus on enhancing training for staff of various frontline services:

training for social security staff will enable them to recognise signs of distress, and to signpost people to appropriate support (Scottish Government, 2018: 10, emphasis added).

English prevention policy also highlighted the area of social security as a site of surveillance, specifically suggesting that jobcentres would be good locations in which to identify ‘vulnerable people who may be at risk of suicide' (HM Government, 2012: 29). Jobcentres, as locations that people in search of employment and in receipt of state benefits must attend, are key disciplinary and surveillance sites (Fletcher and Wright, 2018) for economically marginalised groups. However, within the document there was no policy consideration or awareness as to why a Jobcentre may be a site that was inhabited by vulnerable individuals. As Mills has argued the construction of benefits claimants ‘as a burden, through policies designed to dehumanise and impoverish … [them], are key in creating the conditions that make people's lives unliveable, making many people feel suicidal' (Mills, 2022: 53). The current benefits system produces vulnerability and can, indeed, make certain suicide deaths more likely. Whilst this is tacitly accepted in the policies there is a failure to acknowledge, let alone address, the political contexts of injustice that produce suicide deaths by benefit claimants. Instead, the Jobcentre was framed as an apolitical site for opportunistic identification of those individuals deemed ‘at risk', rather than as part of a complex systemic process of injustice. Whilst the Scottish policy did acknowledge that interactions with social security staff could come at a time of ‘great difficulty' (Scottish Government, 2018: 10) for individuals, the emphasis remains on using these interactions as population surveillance opportunities rather than encouraging a deeper consideration of the structural contextual factors that make individuals in these situations vulnerable to suicide and which make their lives less liveable. Consequently, there is a minimalization of the importance of ‘unequal concentrations of primary goods (substantive rights, social connections, resources, opportunities, etc.)' (Button, 2016: 270) that can lead to suicide deaths (see also, Mills, 2018).

Suicide as predictable

The suicide prevention documents are replete with discussions of risk, with risk factors deeply entwined with the surveillance approach because risk factors are simultaneously created through surveillance generated data whilst also providing a specific target for surveillance capabilities. By creating categories of ‘high risk’ groups and through statistical monitoring and surveillance the documents construct suicide as an object amenable to prediction. Underlying this policy is a logic that sees population level behaviour as categorically patterned and foreseeable. According to Welsh policy ‘suicide and self harm are largely preventable, if risk factors at the individual, group or population level are effectively addressed' (Welsh Government, 2015: 6, emphasis added). Such a statement could provide an opportunity to make interventions centred on structural contextual factors such as poverty and deprivation, but these factors are not reflected upon in the policy's suggested prevention actions (Welsh Government, 2015: 15–17). Whilst the policies may recognise that poverty and deprivation are risks, they do not explicitly intervene in the conditions producing poverty, instead focusing on poverty as a risk predictor rather than a contingent context that could be actively addressed and acted upon through policy intervention.

Within the policies we see the biopolitical necessity of increasing data and surveillance to facilitate prediction and manage risk. A suicide policy document from Northern Ireland sets this out clearly when it plans to ‘manage risk’ through:

Screening and improved understanding of clinical, psychological, sociological and biological risk to help identify high risk individuals in health care settings, and subsequent management of risk (Northern Ireland Government, 2016: 106, emphasis added).

Crucially, such an approach does not encourage better understandings of the individual, but rather of the ‘risk’ factors, for the risk factors are linked to the ability to identify, predict and thus manage suicide. Individuals thus become a constellation of risk factors. Castel has argued that preventative strategies no longer focus upon a concrete individual but instead seek to address risk through ‘abstract factors which render more or less probable the occurrence of undesirable modes of behaviour' (1991: 287).

There is a belief that through categorisation future behaviour can be predicted – we can ‘know’ who is most likely to die by suicide. Importantly, a strategy of focusing on predictive risk often only occurs within delineated settings, such as health care. This is because through risk assessments there is a belief that clinicians can categorise patients into risk groups (Mulder et al., 2016: 271) and thus predict future outcomes. Such an approach seeks to enhance the perception of security by minimising the appearance of contingency and maximising the perception of the knowable certainty of trends and behaviour. For instance, the policies have a consistent focus on mental ill health, which is framed as an unproblematic, identifiable, predictive population level risk factor for suicide:

Depression is one of the most important risk factors for suicide. The early identification and prompt, effective treatment of depression has a major role to play in preventing suicide across the whole population (HM Government, 2012: 6, emphasis added).

The biopolitical urge to portray suicides in this way expands, leading to ever more ‘at risk’ groups and ‘risk factors’ being sought out to demonstrate some form of predictability in the face of contingency. For example, the desire to identify and monitor suicide leads Scottish policy to set out a specific action point to consider ‘how risk of suicide is elevated for groups within the broader general population, and identify specific action to address this' (Scottish Government, 2018: 14). In England the ‘high risk’ groups encompass a startlingly large proportion of the population:

The National Strategy identified the following high risk groups:

As Douglas has argued, the process of defining what exactly risk is, and therefore who is at risk, is a political process (Douglas, 1992). As with the politics of visibility, the politics of risk is selective, and categories and classifications are related to wider political and social anxieties and presumptions. This is explicitly seen in Northern Ireland policy which considers ‘rural dwellers’ as potentially being an ‘at risk’ group. This consideration seems, in some respects, linked to perceptions and assumptions regarding the behaviour and culture of such ‘rural dwellers’ which is explicitly described in the policy document, the document noting that ‘The rural culture of self-reliance and stoicism works against help-seeking' (Northern Ireland Government, 2012: 31). However, whilst the policy acknowledges that service provision is seen as poor within rural communities (see also The Samaritans, 2018: 4) it remains vague as to how to increase rural support services, and there is more focus on the existing behaviours and attitudes of people in one rural region than on ways of increasing service and support provision:

Mental health as a term is most generally associated with mental illness within the rural reaches of the South West Action for Rural Development (SWARD) region; seen as something from which an individual will not recover and will have for life, and is generally referred to in negative and stereotyping terms such as ‘psycho’, ‘schizo’, ‘mad’ and ‘nutter’ etc (Northern Ireland Government, 2012: 74).

This demonstrates how concerns and anxieties about certain communities can drive risk management and highlights that the creation of groups as visible and ‘at risk’ is not a neutral process.