Mental health inpatient settings: Creating conditions for learning from deaths in mental health inpatient services and when patients die within 30 days of discharge

This is one of a series of HSSIB investigations on the theme of patient safety in mental health inpatient settings. The investigation examined how providers conduct timely and effective investigations into deaths of patients receiving care in inpatient units or within 30 days of discharge. This included a review of local, regional and national oversight frameworks, as well as data collection mechanisms.

The aim of the investigation was to understand how providers learn from deaths, and how they use that learning to improve. The investigation sought to understand the impact on individuals involved in the care of patients who died while in mental health inpatient care or shortly after discharge. It reflects the experiences of families, carers and staff.

The investigation recognises the complex nature of mental health inpatient care and discharge. The findings present opportunities to improve systems and practices in mental health services, with potential relevance to other healthcare settings in England.

The investigation identified significant challenges in maintaining safety, conducting effective investigations, managing data on deaths, and ensuring system-wide learning. These findings reiterate findings from other reports on inconsistencies in data reporting, lack of consistent terminology, and difficulty in cross-provider comparisons. The investigation found gaps in discharge planning, crisis service accessibility, and access to community therapy that were potentially contributing to poor patient outcomes, including deaths.

The investigation highlighted system-level issues in service commissioning, patient flow, integrated working and accountability, compounded by a lack of system-level learning and application. The investigation found there is a culture of blame in which individuals and organisations are afraid about safety investigation processes. The report emphasises the need for a systemic approach to safety investigations and learning with a focus on collaboration, transparency, and oversight, with a shift from procedural practices to a culture rooted in empathy, person-centred care and active involvement of families. The findings are grouped under the terms of reference for the investigation:

Understanding how providers ensure timely and effective investigations

• Investigations into patient safety incidents in mental health do not always take a system-wide perspective, limiting the ability to capture the full complexity of care.
• The current national framework for incident response faces implementation barriers in mental health settings, due to the differences in care requirements compared to acute physical health settings.
• Training for the implementation of the Patient Safety Incident Response Framework includes developing knowledge of systems thinking and system-based approaches to learning from patient safety incidents. However, some organisations described their training focused on acute physical health contexts and does not sufficiently account for the mental health care context.
• There is no national system to track and ensure the implementation of investigation recommendations, resulting in limited strategic oversight of patient safety investigations and a lack of structured learning for improvement.
• Many families feel marginalised and excluded from the investigation process, experiencing investigation processes as a ‘tick box’ exercise and without a culture of transparency, learning, and accountability.
• Processes for learning from deaths are feared by families, staff and organisations because of a reported focus on blame which does not align with the stated goals of an effective safety culture that is orientated around learning to support systemic change.
• Families often feel excluded from care processes, with their concerns about safety planning and risk mitigation often overlooked, which complicates their ability to help keep their family member safe.
• Legal processes within organisations may unintentionally shut down opportunities for learning, fostering a culture of defensiveness rather than reflection.
• Staff lack the time, permission and safe spaces to support open, reflective conversations about patient safety incidents, which are essential for learning and improvement.
• Patient safety incident investigations, and other associated investigation processes if a death occurs, often do not consider the emotional distress experienced by all affected. This results in compounded harm.

The investigation also became aware of areas of mental health inpatient care where investigations had not effectively addressed ongoing concerns about inpatient mental health care:

• Gaps were identified in discharge planning, crisis service accessibility, and community therapy provision, and staff skilled in mental health, resulting in people being left in unsafe situations where they may self-harm.
• There is significant variability in therapeutic engagement and a lack of personalised care which has left some patients feeling hopeless and disconnected.
• The term ‘therapeutic engagement’ may be interpreted differently across mental health services. This has resulted in some approaches becoming clinically focused rather than person focused.
• Providers told the investigation that incidents of people using items of clothing to ligature resulting in catastrophic self-harm was increasing. However, the investigation did not identify specific guidance on how to reduce and respond to non-anchored ligature risks, or on managing access to known ligature risk items.
• Staff face ongoing challenges in balancing ‘least restrictive approach’ policies and the therapeutic benefit of decisions about care, with the need to ensure patient safety, often creating tension in care delivery.

Examining national, regional, and local oversight and accountability frameworks for deaths in mental health inpatient services

• Many previous national recommendations to improve the care of patients with mental health needs have not been taken forward to date, leaving no clear plan for implementing the recommendations.
• Some integrated care boards do not have full oversight of patient safety risks across all the services they oversee. Instead of having a clear, formal structure for accountability in patient safety investigations, they often rely on informal relationships or collaborations between providers.
• This lack of complete oversight can make it difficult to ensure that patient safety investigations are thorough and standardised across different organisations, leading to potential inconsistencies in addressing safety issues.
• Some integrated care boards and regional teams struggle with gathering and analysing data on patient safety due to resource and reporting limitations. This means they do not always have a complete picture of the risks at a system-wide level.
• Data gaps limit the ability of integrated care boards to identify co-morbidities (when patients have multiple health conditions) and understand health inequalities, especially among people with serious mental illness. This creates challenges for integrated care boards to address broader health trends and inequities within their regions effectively.
• The involvement of patient safety partners and people with lived experience in safety meetings is variable, limiting their ability to contribute.
• Some non-executive directors with responsibility for safety and quality struggle to scrutinise and interpret complex data sets on patient safety and deaths due to the volume and lack of triangulation of information presented. This limits meaningful oversight and learning.

Examining the mechanisms that capture data on deaths (and near misses) across the mental health provider landscape, including up to 30 days after discharge

• There is inconsistency in data reporting. Mental health providers report deaths and near misses in varied ways, using different definitions and methods. This inconsistency makes it difficult to compare data across providers and understand overall trends in patient safety.
• There is not a standardised national system requiring providers to report deaths in the same way. This means that each provider’s reports may look different, which reduces the reliability of data for understanding patient safety across the board.
• There is not a single, comprehensive database that includes all deaths and near misses within mental health services, including those occurring within 30 days after a patient’s discharge. This makes it hard to see the full picture of patient safety outcomes and identify patterns or risks.
• There is not a centralised organisation or process effectively overseeing and co-ordinating data on deaths. This lack of oversight limits the ability to identify systemic issues, reduce duplicated efforts, and drive consistent improvements across mental health services.
• There is currently limited co-ordinated effort among organisations that produce data relating to deaths, and individual providers may be collecting and analysing similar data on their own. This leads to duplicated work, wasting time and resources that could be better used if there was improved collaboration. It also results in considerable variation in the data being presented.

HSSIB makes the following safety recommendations

HSSIB makes the following safety observations

Local-level learning

HSSIB investigation reports include local-level learning where this may help organisations and staff identify and think about how to respond to specific patient safety concerns at the local level.

1.1.1 A person’s mental wellbeing/health influences how they feel, what they think and how they behave (World Health Organization, 2022). Around a quarter of the population of England will experience a ‘mental health problem’ each year (Mind, 2024). A mental health problem is a change to a person’s mental wellbeing that impairs their ability to function as they would do normally. Mental health is determined by a combination of biological (for example genetics and physical health), psychological (for example beliefs, perceptions and previous traumas) and social (for example relationships, culture and life circumstances) factors (Mental Health Foundation, 2024).

1.1.2 Most people experiencing a mental health problem are cared for outside of hospital in the community. For some people admission to hospital on a voluntary or compulsory basis is needed. The Mental Health Act 1983 is legislation that covers the assessment, treatment and rights of people where a person is admitted to hospital on a compulsory basis. In this circumstance, they may be described as ‘detained’ under the Mental Health Act. The Act is split into different sections which contain information about being detained, treatment while detained and the allowance of ‘leave’ from hospital for an agreed purpose and period (this may be referred to as ‘Section 17 leave’). The ‘Mental Health Act 1983: code of practice’ sets out how the Mental Health Act should be implemented in practice (Department of Health, 2015). The code of practice provides statutory guidance to health and social care authorities and staff on how they should proceed when undertaking duties under the Act. It is prepared and published by the Secretary of State. The Mental Health Act was amended in 2007 and at the time of writing further reform via the Mental Health Bill 2024 (Department of Health and Social Care, 2025) were being considered by Parliament.

1.1.3 In addition to the requirements of the Mental Health Act, relevant professionals (particularly those involved in discharging or treating patients in the community) should also consider the general responsibilities of local authorities under Part 1 of the Care Act 2014. This applies to the care and support arranged or provided by local authorities to patients in the community, such as patients subject to community treatment orders (CTOs), guardianship or leave from hospital. The Care Act 2014 requires local authorities, NHS commissioners and providers, and housing services to work together to provide truly person-centred care and support. It places a particular emphasis on managing people’s needs to prevent risks increasing. These duties are particularly important for people with mental illness, as they often require coordinated support from multiple agencies to promote their recovery and enable participation in society following hospital discharge, or while on a CTO, guardianship or on leave.

Mental health inpatient care

1.1.4 In England, there are various mental health inpatient services. The demand on mental health inpatient services in England is high and has been increasing. Between 2016 and 2023 there was a 24% increase in the number of patients in hospital (The King’s Fund, 2024). The Royal College of Psychiatrists recommends a maximum bed occupancy of 85% (Royal College of Psychiatrists, n.d.a). Bed occupancy has consistently been above the recommended maximum of 85% (except during the COVID-19 pandemic) since 2010/11 (Mental Health Watch, 2024).

National strategies to improve mental health

1.1.5 Over the past decade, England has introduced several major strategies to improve mental health services. Examples include the Five Year Forward View for Mental Health (NHS England, 2016), the NHS Long Term Plan (NHS England, 2019a), the NHS Mental Health implementation plan (NHS England, 2019b), and a suicide prevention strategy (Department of Health and Social Care, 2023a).

1.2.1 The Learning from Deaths framework (NHS England, 2017) requires all trusts to carry out mortality reviews (reviews of the deaths of patients in their care) and to publish a quarterly dashboard reporting their data on deaths, including data on preventable deaths and reports on their actions to learn and improve.

1.2.2 To support national work regarding deaths in hospital, the National Mortality Care Record Review Programme was launched in 2016 to retrospectively review the quality of care for patients who had died, from hospital admission to death. A validated Structured Judgement Review tool for case notes of patients who had died was implemented. There can be two stages to the review process. A second-stage review is recommended where care problems have been identified during the first stage review, and where the potential to avoid a death through a different care approach may have been identified (Royal College of Physicians, 2018). The Royal College of Psychiatrists developed a Care Review tool which is based on the Structured Judgement Review methodology, originally developed by the Royal College of Physicians (Royal College of Psychiatrists, n.d.b).

1.2.3 NHS England (2024b) has designed a single national NHS system for recording patient safety events called Learn from Patient Safety Events (LFPSE). LFPSE data can be used to inform the need for safety investigations and highlight where existing safety risk controls may not be working as intended.

1.2.4 In August 2022 NHS England published the Patient Safety Incident Response Framework (PSIRF) as a replacement for the Serious Incident framework (NHS England, 2015). The PSIRF sets out the NHS’s approach to developing and maintaining effective systems and processes for responding to patient safety incidents for the purpose of learning and improving patient safety. The PSIRF has four key aims:

• ‘compassionate engagement and involvement of those affected by patient safety incidents
• application of a range of system-based approaches to learning from patient safety incidents
• considered and proportionate responses to patient safety incidents
• supportive oversight focused on strengthening response system functioning and improvement’ (NHS England, 2022a).

1.2.5 The PSIRF applies to the delivery of healthcare services. Therefore, social care organisations are not required to adopt PSIRF but may wish to do so.

1.2.6 Under PSIRF there is no classification or threshold for a ‘Serious Incident’ and level of harm is not the driver for an investigation to identify learning. A response might involve a Patient Safety Incident Investigation (PSII) but there are other types of learning response available which may include an after action review (AAR) or a multidisciplinary (MDT) review. The focus is on learning and responding proportionately.

1.2.7 The PSIRF outlines guide timelines for patient safety learning responses and asks for them to begin as soon as possible after an incident is identified. The response methodology outlined in the framework asks that responses are conducted for the sole purpose of learning and identifying improvements that reduce risk and/or prevent or significantly reduce recurrence and that they do not seek to determine liability and blame.

Independent investigations/public inquiries

1.2.8 There are differences for example between a Department of Health and Social Care (DHSC) inquiry and an NHS England independent patient safety investigation. The differences often relate to their scope, purpose, oversight, and process. Public inquiries are sometimes launched in cases with large numbers of patient deaths or where there is evidence of system-wide issues. These are commissioned by the Department of Health and Social Care and may involve external experts, legal counsel, and independent chairs. The current Lampard Inquiry (n.d.) is an independent statutory inquiry investigating the deaths of mental health inpatients in Essex.

1.2.9 The National Independent Patient Safety Investigation Framework (NIPSIF) (NHS England, 2023a) is an internal framework for NHS England that guides standardisation of independent investigations, from commission to publication, including the crafting of recommendations.

1.2.10 Previous published independent reviews into care in mental health services include the review of deaths at Southern Health (Mazars, 2015), Independent Review of Greater Manchester Mental Health NHS Foundation Trust (NHS England, 2024c) and a special review of mental health services at Nottinghamshire Healthcare NHS Foundation Trust (Care Quality Commission, 2024a).

1.2.11 NHS England has previously published an annual report which provided an overview of independent investigations commissioned by regional independent investigation teams (NHS England, 2021a). These primarily related to homicides committed by patients under the care of mental health services.

1.2.12 Deaths of patients detained under the Mental Health Act or where the Mental Capacity Act applies, and where a death may be linked to problems in care, can have an internal local-level PSII following the PSIRF methodology. This contrasts with detained people in non-healthcare organisations such as the prison service or police custody where there is an automatic, external investigation by an independent national body. For example, the Independent Office for Police Conduct investigates mental health related deaths of detained people in police custody. These bodies publish investigation reports, have oversight of all deaths and policy issues, and share and publicise thematic reports.

Coroners’ inquests

1.2.13 When a death occurs unexpectedly or under concerning circumstances, a coroner holds an inquest to determine the cause and whether there were any preventable factors.

1.2.14 Coroners investigate deaths suspected to have been violent or unnatural, where the cause of death is unknown, or where the deceased died in prison or some other form of state detention. When a death is reported to a coroner, they may decide that a post-mortem and/or an inquest is required (Courts and Tribunals Judiciary, 2024a).

1.2.15 As per the Coroners and Justice Act 2009, if a coroner considers there is a risk of future deaths and that action could be taken to prevent or reduce the risk, a coroner will issue a Regulation 28 report (a formal instruction to protect life) to an individual, organisations, local authorities or government departments and their agencies. These are also referred to as prevention of future deaths reports and can be accessed on the Courts and Tribunals Judiciary website (2024b).

Other types of reviews

1.2.16 Deaths of people with a learning disability and autistic people can be notified to the Learning from lives and deaths – People with a learning disability and autistic people (LeDeR) process (NHS England, 2019c). This review process is not restricted to the last episode of care before the person’s death but looks at key episodes of health and social care the person received that may have been relevant to their overall health outcomes. LeDeR reviews take account of any mortality review that may have taken place following a person’s death (NHS England, 2019c; 2021b).

1.2.17 The Royal College of Psychiatrists’ Safety Incident Response Accreditation Network (SIRAN) standards (Royal College of Psychiatrists, 2024) were written to support member organisations to improve the standards and quality of patient safety and incident response processes. These standards are for service providers and commissioners of mental health services to help them ensure they carry out high-quality safety incident responses and reviews.

1.3.1 When people are in mental health crisis, they need timely access to support that is compassionate and meets their needs. The ‘Mental Health Act 1983: code of practice’ states that:

‘Local authorities, NHS commissioners, hospitals, police forces and ambulance services should have local partnership arrangements in place to deal with people experiencing mental health crises. The objective of local partnership arrangements is to ensure that people experiencing mental health crises receive the right medical care from the most appropriate health agencies as soon as possible.’ (Department of Health, 2015)

The Mental Health Crisis Care Concordat was launched in 2014. This is a national agreement which sets out how organisations will work together to make sure people in crisis get the right help (HM Government, 2014).

1.3.2 Crisis resolution and home treatment teams provide intensive support for people experiencing an acute or ‘crisis’ episode during their mental illness. Also known as ‘hospital at home’ teams, this service is available 24 hours a day, 365 days a year. The teams consist of psychiatrists, clinical psychologists, registered mental health nurses, occupational therapists and support workers, who work with patients to treat them outside of hospital.

1.3.3 Liaison psychiatry services in acute physical health hospitals ‘address the mental health needs of people being treated primarily for physical health problems and symptoms’ (Royal College of Psychiatrists, 2013). In England, liaison psychiatry services are typically commissioned, managed and delivered as part of mental health services rather than acute physical hospital services. This means these staff are employed by the mental health provider but are based in the acute hospital, where they work collaboratively with the physical healthcare staff.

1.3.4 Right care, right person (RCRP) is an approach designed to ensure that people of all ages, who have health and/or social care needs, are responded to by the right person, with the right skills, training, and experience to best meet their needs. The centre of the RCRP approach is a threshold to assist police in making decisions about when it is appropriate for them to respond to incidents, including those which relate to people with mental health needs. The threshold for a police response to a mental health-related incident is: ‘to investigate a crime that has occurred or is occurring; or to protect people, when there is a real and immediate risk to the life of a person, or of a person being subject to or at risk of serious harm’ (Department of Health and Social Care, 2024a). RCRP is underpinned by an agreement between policing, health and other partner agencies which aims to make sure people in mental health crisis get the care they need (Department of Health and Social Care, 2024a; Independent Office for Police Conduct, 2024; Royal College of Emergency Medicine, 2024).

Complexity of mental health care

2.1.1 Section 1.2 describes the national approaches to investigation and reviews of patient safety incidents in mental health settings. Healthcare is provided within a complex system characterised by uncertainty and unpredictability (Braithwaite et al, 2018). Harm has numerous causes, and its wide-ranging ripple effects and negative impacts on human wellbeing and relationships have been documented (Lamiani et al, 2017; Wailling et al, 2022).

2.1.2 Serious incident investigations have become a prominent component of both national and local healthcare safety management, reporting and governance systems worldwide (Kok et al, 2022; Leistikow et al, 2017). However, healthcare governing bodies and organisations have grappled with incident reporting and investigation systems that promote transparency and openness to learning about safety, and to more broadly embrace a participatory approach to investigations (Kok et al, 2022; Macrae, 2016).

2.1.3 The World Health Organization (2021) Global Patient Safety Action Plan strives to eliminate avoidable harm in healthcare’. Contained within this action plan are seven guiding principles:

• engage patients and families as partners in safe care
• achieve results through collaborative working
• analyse and share data to generate learning
• translate evidence into actionable and measurable improvement
• base policies and action on the nature of the care setting
• use both scientific expertise and patient experience to improve safety
• instil a safety culture in the design and delivery of healthcare.

The investigation process when a death occurs – local investigation response

2.1.4 When this investigation began, mental health providers were in the process of transitioning from the former Serious Incident (SI) Framework (NHS England, 2015) to the Patient Safety Incident Response Framework (PSIRF) (NHS England, 2022a). The investigation visited mental health settings during the spring and summer of 2024. At that time, some organisations were still working to the older framework and those that had introduced PSIRF were in the early stages of implementation.

2.1.5 Historically the timeliness of investigations was always described as being a real challenge. All organisations referred to backlogs of old Serious Incident framework investigations. Reasons given for these backlogs included the volume of incidents that met the former SI framework criteria and resources required to undertake investigations. All organisation described a ‘hope’ that PSIRF would improve the investigation process and timeliness of investigations as it is considered less prescriptive than the former SI framework, with the aim of focusing on learning. Due to the timing of the HSSIB investigation and PSIRF implementation being in its infancy in mental health, it was not possible to assess the impact and effectiveness of PSIRF or the timeliness of such investigations.

2.1.6 The investigation identified a divergence in scope and investigation methodology between PSIRF and the Learning from Deaths (LfD) guidance. PSIRF promotes a comprehensive, system-wide perspective on patient safety incidents and focuses resources on investigating incidents where there is the greatest potential for learning and improvement. This means not all deaths are reviewed under PSIRF, as not all deaths necessarily highlight system failures or opportunities for systemic improvement. The Learning from Deaths guidance is more specialised, concentrating on mortality reviews and the quality of care leading up to it.

2.1.7 The investigation was told there is a tension between these frameworks which reflects the broader challenge of balancing systemic learning with individual case review and accountability. This can lead to inconsistencies in how organisations approach patient safety investigations into deaths. This may result in fragmented learning processes and potential overlaps, or gaps, for safety investigations. The investigation was told by various people that the learning from deaths framework needs “urgent review” and that “clearer guidance might be needed to ensure these frameworks complement rather than conflict with each other”. PSIRF guidance states, ‘Some patient safety incidents, […] and deaths thought more likely than not due to problems in care (that is, those meeting the Learning from Deaths criteria for investigation) all require a PSII to learn and improve’ (NHS England, 2022a).

2.1.8 The investigation was told there “is a challenge with Structured Judgement Reviews (SJR) [see 1.2.2] and the interface with PSIRF … there is a lot of duplication”. It was described that some organisations use the SJR process as a triage tool for PSIRF and the numbers of SJRs “is huge” and the demand on time to complete them is significant. There was variability in how organisations identify learning from SJRs at stage 1 and stage 2 of their review process.

2.1.9 One organisation told the investigation that for deaths that are suspected suicides, they use a ‘suicide cultural review tool’. This is because they felt this was a gap in PSIRF and ensures all reviews are done in a systematic way. Other organisations described that they would be investigating patient deaths in the same way as the former SI framework (NHS England, 2015) “but with a different badge of PSIRF”. When questioned what was meant by this, they described “doing the same investigations as they did under the serious incident framework”.

2.1.10 The investigation sought to understand how organisations were developing their understanding of their patient safety incident profile, their ongoing safety actions in response to investigation recommendations, and how these linked with established programmes of improvement. During visits to organisations, the investigation was told about how organisations were implementing PSIRF with examples of provider policies that have considered how they have identified their PSIRF priorities. The decision to carry out a patient safety investigation was invariably based on whether the incident was linked to a national priority or linked to the organisation’s agreed PSIRF priorities.

2.1.11 Examples of how PSIRF priorities were identified included through analysis of an organisation’s patient safety insights data, and thematic analysis of reported incidents. Organisations described that their patient safety priorities would form the foundation for how they would decide to conduct Patient Safety Incident Investigations (PSIIs), patient safety reviews and thematic reviews. Examples of when investigations would be commissioned included reducing restrictive practices, unexpected deaths after unrecognised deterioration in patients’ physical health, or the unexpected death of a patient within mental health services including suspected suicide.

2.1.12 Some organisations told the investigation they felt PSIRF guidance and training was primarily designed for the acute physical health sector, with limited applicability to mental health settings. In addition, the investigation was told that allied health professionals are not all being trained and included in PSIRF ‘in the way they need to optimise benefits’. The intention of PSIRF is to help organisations conduct investigations relevant to their context and the populations they serve. However, organisations described that they were grappling with what was meant by a ‘proportionate response’ in mental health. In addition, they described challenges of ‘involving people’ and that more specific guidance and training was needed. A doctor in academia told the investigation that there is less understanding of “systems approach to investigations in mental health” and that this is reflected in staff not being supported to investigate across organisational boundaries.

2.1.13 PSIRF training is not centrally provided by NHS England. Training is available from HSSIB and can also be procured via suppliers on NHS England’s training procurement framework. Suppliers on the framework have been assessed against criteria to ensure they have relevant skills and experience and deliver content in line with PSIRF requirements (NHS England, 2019d). The investigation was told by NHS England that the procurement framework has recently been suspended as part of a review of training.

2.1.14 Reviews conducted by healthcare organisations suggest that the quality of incident analysis (investigation) can be highly variable and is often poor. Many of the analyses that are conducted do not lead to effective actions or improvements (Peerally et al, 2017).

2.1.15 The investigation undertook a review of several trust-level investigation reports. Given the timing of the HSSIB investigation, the investigations reviewed were all within the former SI framework (NHS England, 2015), they were not PSII’S under PSIRF. Domestic homicide reviews were excluded. Many of the reports described ‘highly complex patients’ who had been under the care of mental health services for many years. The investigation considered several factors including the quality of the local investigation, and reviewed the findings. Most reports had been carried out in line with the investigation policy that had been in place at the time. Most reviewed reports evidenced engagement with staff and suggested they had responded to questions from the families. However, while most of the investigations had responded to questions from the families there was no evidence that families had been involved to any further extent during the investigation.

2.1.16 The methodology of the local investigations generally referred to reviews of patient/clinical records and considered local contributory factors. The recommendations were locally focused but underlying factors often described links to known national issues.

2.1.17 Many local investigations highlighted that they had incomplete information to inform the report findings and subsequent recommendations. Examples included previous incidents not being recorded, previous clinical history not being available, and incomplete risk assessments. One report described a patient as ‘calm, engaging, chatting, compliant with medication and routine, describing not having suicidal ideation’; the patient was later found to have catastrophically self-harmed within 17 minutes of being with staff and patients in a communal area. The local investigation concluded that the patient’s risk status was often incorrectly described. However, it is possible the patient was not showing suicidal ideation (thoughts of suicide) 17 minutes before, or did not tell staff their true feelings. The risk statement was not necessarily wrong, but the local investigation may have assumed it was without considering other factors. Research evidence shows that a change from not having suicidal ideation to suicidal ideation and action can be 10 minutes or less (Paashaus et al, 2021).

2.1.18 The local investigations identified that workforce challenges were a significant contributory factor in care delivery issues, including longstanding vacancies for core and specialist posts, increased reliance on agency or bank staff and often requiring additional staff from other wards to complete a task. All of these were considered and reported in HSSIB’s investigation report ‘Mental health inpatient settings: creating conditions for the delivery of safe and therapeutic care to adults’ (Health Services Safety Investigations Body, 2024a).

2.1.19 All of the investigations reviewed focused on a single mental health provider and did not cut across organisational care boundaries, even when a patient’s care had involved multiple contact points. The reports looked at issues solely from a mental health hospital perspective but not from the perspective of the emergency department or liaison psychiatry. Several reports did not identify any service delivery issues relating to the death of a patient that could be used to inform improvements.

2.1.20 During visits to service providers, the investigation explored how current patient safety investigations were carried out at the time of the visit. Staff said that a safety investigation would involve reviewing healthcare records, reviewing any policies and then asking staff to give their account of the event, either by a statement or by interview.

2.1.21 The investigation’s review of local patient safety investigations showed that the methodology used did not include observations of clinical work to understand how care was delivered in practice. There was no rationale provided as to why observational work was not completed although the investigation heard repeatedly of the resource constraints involved in undertaking patient safety investigations. In addition, PSIRF was at an early stage of implementation at the time of the HSSIB visits. Including observations of normal work in investigations is considered best practice in the safety science literature (Havinga et al, 2017). The benefits are that the investigators gain a deep understanding of the normal stresses on healthcare professionals which then inform the actions proposed to improve patient care. The PSIRF learning response toolkit includes a tool to support organisations in how to conduct observational work (NHS England, 2022a).

2.1.22 NHS England commissioned independent patient safety investigation or DHSC commissioned public inquiries are sometimes launched in cases where there are large numbers of patient deaths or evidence of system-wide issues (see 1.2.8 to 1.2.10). These are independent to the care provided and often carried out by individuals or organisations external to a provider.

2.1.23 NHS England told the investigation that to support the process for commissioning independent investigations, it has an established procurement framework of independent investigation providers (NHS England, 2019d). The investigation was told this was due for review in 2024/25 and is being updated to reflect the approved National Independent Patient Safety Investigation Framework, which builds on the national approach to patient safety incident management set out in the PSIRF.

Reporting and investigation of patient safety incidents across organisational boundaries

2.1.24 The PSIRF oversight roles and responsibilities specification (NHS England, 2022b) states that ‘where a response involving multiple providers and/or services across a care pathway is too complex for a single provider to manage, ICBs should support the co-ordination of cross-system response’. This is also reflected in the ‘Patient safety incident response standards’ (NHS England, 2024d).

2.1.25 People with a serious mental health illness will have care interventions that span many aspects of the health and care system. However, the investigation identified that there are not system-wide investigations that capture the full complexity of mental health care provision. The investigation heard from staff that patient safety incident investigations still “feel very local”. The investigation considered this may be due to PSIRF being in its early stages of implementation at the time of the HSSIB investigation.

2.1.26 An NHS regional lead told the investigation that ideally these should be integrated care board (ICB) level coordinated investigations to allow for a full system understanding; however, they described that the ICB is often unaware of investigations that might need system-level oversight. In addition, ICBs were not always aware of safety risks in their areas, in part because there are different PSIRF learning responses that do not trigger a specific investigation report. Referring specifically to mental health incidents, the regional lead said that “many incidents will have an investigation that involves more than one provider, and the oversight of cross-boundary investigations is very much dependent on relationships and people working effectively together”. NHS England told the investigation that under PSIRF ICBs have wider access to safety intelligence, through involvement in developing Patient Safety Incident Reporting Plans (PSIRPs) rather than only seeing serious patient safety events. In addition, the investigation was told by NHS England that “the Learning from Patient Safety Events team (LfPSE) have considered this a few times with regional leads, and provisions were in place to ensure nothing was “lost” from their existing processes when we migrated over. We have discussed, supported and sought assurance since that all regions have appropriate measures in place to work around”.

2.1.27 The investigation was told by staff that reporting LFPSE (see 1.2.3) is time consuming and “it is as if the incidents go into the ether”. NHS England told the investigation they publish regular case studies to show the direct action taken in response to patient safety events recorded by organisations, staff and the public, and how their actions support the NHS to protect patients from harm. At the time of writing, the investigation could not identify any specific published learning for mental health specialty (NHS England, 2024g).

2.1.28 The investigation was told by NHS regional teams that the former Strategic Executive Information System (StEIS) reporting system was still being used which meant organisations were dual reporting. This was because LFPSE did not enable regional teams to have sight of significant safety events in their area. NHS England told the investigation “this is on an interim basis until organisations' local risk management systems transfer [to the next version]”.

2.1.29 The issue of co-ordinating responses to patient safety events will be described in the HSSIB investigation report on safety management systems due to be published imminently and therefore will not be repeated. This report covers many of the issues heard during this investigation.

Narratives from families who have experienced the investigation process

2.1.30 Involving patients and families who are affected by patient safety events is vital for learning. The experiences that patients and families shared with the investigation were incredibly informative but showed repeatedly that families often did not feel involved in a full or meaningful way during investigations. Families described wanting to know that organisations cared about their family member and want to learn; they wanted the death of their family member “to mean something”. This is not a new finding and is reflected in research and reports.

2.1.31 Some organisations were described by bereaved families as “gaslighting, bullying and had toxic environments”. Bereaved families described having to fight to be involved in investigations with some describing the investigation process as “worse than the actual death because they were reliving the death [of their family member] over and over again”. In relation to guidance about an open and transparent process in investigations, they described “tick box responses to being open”. As reported in literature, the feeling of not being heard or valued can destroy opportunities for reparation, and intensify grief, harm, isolation and anger for all involved (Kok et al, 2022; Ramsey et al, 2022, Wailling et al, 2022). One family told the investigation:

2.1.32 The investigation was told by families that they had “lost confidence and trust” in their local hospital’s investigations. Families described the investigation being done to them rather than involving them. A representative of a family group told the investigation that involvement had been very “tokenistic” and “sometimes it feels like people want to parade your grief for learning”, sharing that involvement only happened when it mattered to the organisation.

2.1.33 When poor engagement and involvement happened, the risk of reputational damage to organisations was high as some families described escalating their concerns to the media, their local MP and establishing campaign groups. McHugh et al (2024) reported that:

‘… inclusion of all stakeholders because it is the ‘right’ thing to do is important, but an unintended consequence of involvement based on this alone is compounded harm, where patients and families feel there is little, or no value assigned to their experience.’

2.1.34 Bereaved families and carers described that investigations were not independent in their opinion and some families described having to go to extreme measures to try and get a more independent investigation.

2.1.35 In a recently published paper relating to organisational policies, ‘involvement’ is described as a passive process of providing information to families, rather than inviting them into the investigation as ‘experts’ or partners in the process, able to contribute to organisational learning (McHugh et al, 2024).

2.1.36 The patient safety charity Action against Medical Accidents (AvMA), in collaboration with the Harmed Patients Alliance, has designed a harmed patient pathway (Action against Medical Accidents, 2024) which is intended to minimise the compounded harm that arises when a patient safety incident occurs. This pathway is intended to encourage providers to recognise harmed patients as suffering a particular form of trauma for which there should be a pathway that seeks to optimise recovery. The pathway is also intended as an obligation for providers to do what is possible to ease suffering and avoid causing further distress. The investigation was told that AvMA is also committed to publishing follow-on guidance to help healthcare staff to think about how to apply the commitments in practice. AvMA told the investigation they "hope to pilot the pathway with some Trusts in 2025 as a way to help develop the ‘How To’ guidance we will draw up to support it".

2.1.37 Working well with families in mental health crises has been considered and ‘Life Beyond the Cubicle’ has been co-produced with patients, family carers and clinicians in a partnership between Making Families Count and Oxford Health NHS Foundation Trust (NHS England, 2025). These resources have been designed to help clinicians work well with the family and friends of people who have mental health crises in order to improve care and patient safety.

2.1.38 The investigation is aware that some organisations have employed Family Liaison Officers (FLO). The FLO is a professional designated to support and communicate with families and loved ones, particularly following patient safety events. The investigation spoke to several FLO’s and there was variability in how their role was resourced. Some FLO’s described very positive resources and support to undertake their role and others described that they were just one person and struggled to meet the needs of so many families.

Narratives from staff involved in safety investigations

2.1.39 Healthcare workers involved in patient safety events involving death or serious harm to a patient can suffer lasting psychological harm and deep-seated feelings of guilt and self-criticism (World Health Organization, 2021).

2.1.40 Staff said that in their experience of being involved in investigations, the ones that went well were those where people felt safe to talk openly and honestly. This included involving everyone in the conversation and reviewing the incident through a safety lens to identify learning. However, the investigation heard from many staff in mental health services a perception that “someone needs to be held accountable” for inpatient deaths by catastrophic self-harm or suicide.

2.1.41 The investigation was told by a consultant psychiatrist:

“… healthcare professionals working in mental health are trained to hold anxiety, however when very tragic deaths occur there is an observable blame shifting and of pushing responsibility in every other direction but ourselves – pushing responsibility towards peers, towards seniors – whilst holding a toxic positivity that everything is OK around here.”

2.1.42 The same senior nurse shared the importance of tackling difficult conversations and “facing up to when I got it wrong”. They told the investigation about a death of a young person in their inpatient unit and their actions on completion of the local investigation report.

2.1.43 A doctor told the investigation that there is a danger in how language is used in investigations. They described that:

2.1.44 Staff told the investigation that “serious incident review processes and inquests get it wrong because they ask the wrong questions which focus in on individual human decision, but in the vast majority of cases you cannot point to a person or a factor and say ‘if that person had done something differently that person wouldn't have ended their lives’”. This cause-and-effect language was described as “unrealistic” and staff said that “when things go wrong, of course we should learn from them but there is a need to look at the systemic factors”. It was described that senior leaders are rarely asked to give evidence at an inquest or speak to a serious incident review about organisational pressures and sometimes there was a feeling of being exposed and having “to make excuses for organisational and systemic factors”.

Narratives from people involved in coronial inquests

2.1.45 The intention of an inquest is to establish who died, when they died, where they died and how their death came about. A coroner told the investigation that an inquest “is not intended to be an overly complex process and it should not morph into a surrogate public inquiry … the key words are it’s inquisitorial which is different to adversarial”.

2.1.46 An area coroner told the investigation that an inquest is about learning and about protecting the public. However, they described situations in which a barrister for a trust will come along to an inquest and “they’ve got two things in their brief – the first is avoid a neglect [decision] … The second is avoid a PFD [prevention of future death report]”.

2.1.47 Some bereaved families and carers described their experiences of coroner’s courts as “poor” and observed that trusts appeared defensive and not wanting to be open. Bereaved families told the investigation they felt excluded from the process and what they expected from the inquest was not the reality. The Voicing Loss project (2024) examined the role of bereaved individuals in coroners’ investigations and inquests in England and Wales; its findings mirrored what this investigation was told by families (Jacobson et al, 2024a). One area of concern for families was linked to when coroners relied on poorly conducted patient safety investigations for their evidence and how this affected coronial outcomes. Families described “point scoring” by legal representatives of the trust and proceedings that were far from non-adversarial, which they were told they would be.

2.1.48 Specifically, families told the investigation that they thought the coroner had a responsibility for making recommendations that would be acted upon, but were left very dissatisfied, leading to feelings of frustration and disappointment. The Voicing Loss research described long-lasting emotional and psychological effects on the bereaved resulting from the overall inquest process “what were perceived to be missed opportunities for learning and prevention of future deaths caused particular anger and engendered mistrust in the coronial system and state authorities more widely” (Jacobson et al, 2024b).

2.1.49 Inquests were described by some staff as “scary”, “adversarial”, and that they felt a “sense of impending doom when faced with a request to attend an inquest”. Research on consultant psychiatrists losing patients to suicide revealed the inquiry and coroner’s inquest as a potentially difficult part in the aftermath of a death (Tamworth et al, 2022). Common feelings included:

2.1.50 A consultant psychiatrist told the investigation that their experience at a coroner’s court “was the most frightening experience of my life … it was to do with the state of my own mind and when I was very traumatised – in retrospect I know that I had a post-traumatic stress disorder”. They went on to describe how they ruminated over the deaths fearing that they were their responsibility, and they were “being tried in the coroner's court”.

2.1.51 Research (Gibbons, 2024a; 2024b; Royal College of Psychiatrists, 2022b) suggests that following the death by suicide of a patient, 80% of clinicians were significantly affected emotionally both in their personal and clinical life, both impacting their clinical work. The research reported that some healthcare professionals changed careers and left mental health care as a result.

2.1.52 Staff said that investigations, including inquests, have:

“…two different tasks – the primary conscious task of the who, where, when and how someone died and another very important and powerful emotional task to do with helping and supporting the grieving process. Not only are the family and friends grieving of course but also the clinicians because they are bereaved as well … and other stakeholders … there could be a very powerful emotional experience where you’re caught in the maelstrom of the most intense grief.”

2.1.53 The Royal College of Psychiatrists has a working group on the effect of suicide and homicide on healthcare staff and has produced guidance to support staff following the death of a patient by suicide (Royal College of Psychiatrists, 2022b). In addition, other guidance for the workforce is also available (Health and Safety Executive, 2019; Tamworth et al, 2025).

The primary purpose of learning and the investigation process

2.1.54 The HSSIB investigation has considered learning from deaths which covers multiple processes; including the Learning from Deaths Framework, coroner's inquests, and investigations following patient safety events. The primary purpose of any investigation – whether conducted locally by a provider or by an independent organisation – is to facilitate learning. However, research indicates that there is a lack of clarity about what ‘learning’ truly entails (Brummell et al, 2021; 2023a; 2023b; McHugh et al, 2024; Jorg et al, 2007).

2.1.55 PSIRF aims to foster a more flexible, transparent, and compassionate approach to investigations and learning. It focuses on understanding the range of factors that contribute to incidents to ensure that organisations can learn from them, promoting a range of system based approaches for learning from patient safety incidents (see 1.2.4). However, PSIRF does not explicitly define what is meant by ‘learning’.

2.1.56 During a learning event at the Royal College of Psychiatrists in September 2024, healthcare professionals discussed the concept of ‘learning’ within the context of safety investigations. Many participants viewed “findings of the investigation” as the primary learning outcome. This aligns with research by McHugh et al (2024), which suggests that investigations often focus on producing recommendations or actions rather than fostering a collaborative process of learning. In this process, knowledge is deconstructed and reconstructed from diverse perspectives, making learning a central, ongoing aspect of safety improvement.

2.1.57 By reviewing investigation reports and speaking with those involved in investigations, this investigation found that although local findings and some actions often follow patient safety incident investigations, there is limited evidence of ongoing learning through structured quality and safety improvement initiatives.

2.1.58 In April 2022, the Royal College of Psychiatrists held two workshops on the serious incident (SI) process, bringing together carers, patients, and clinicians. Participants emphasised that ‘learning’ in SI investigations often lacked a clear definition and structured process. True learning was seen as more than simply producing recommendations; it was discussed that it may be intrinsically linked to a mourning process, requiring reflection, open communication among all parties, and time for psychological digestion. This approach to learning was viewed as an ongoing, reparative process that fosters shared understanding and healing for everyone affected by an incident. Rather than focusing solely on corrective actions, learning should include emotional and relational dimensions to enable meaningful and lasting improvements. This investigation findings would support that the desired approach to learning from investigations requires further work.

2.1.59 The investigation has evidenced that there continues to be a blame culture in the process of investigating deaths in mental health inpatient settings and where deaths occur shortly after discharge. The workshops held by the Royal College of Psychiatrists in 2022 acknowledged that blame is a natural part of the grieving journey but should not be the endpoint – “when blame dominates, it hinders healing and prevents meaningful learning”.

2.1.60 The investigation has referred to the creation of a psychologically safe work environment, within which health workers can speak up regarding patient safety and other concerns without fear of negative consequences, as being fundamental to learning. The investigation found evidence to suggest this is not happening.

2.1.61 Research implies that learning is a social and participative process that involves people deconstructing and reconstructing their understanding of shared knowledge of different aspects of the healthcare system (Macrae, 2016). Creating space for multiple perspectives, without prioritising specific voices, may provide opportunity for effective and meaningful organisational learning, as well as facilitating reparation and trust-building (Wailling et al, 2022; Wu, 2000). The investigation was told by many people that providing a psychologically safe space will allow teams and organisations the time for collaborative scrutiny and reconstruction of systems and processes to first understand how care is delivered and experienced, and then to try to re-organise it in ways that help to prevent recurrence of specific incidents and/or patient safety risks.

2.1.62 Developing and sustaining a learning culture requires strong leadership at all levels. This includes from the Department of Health and Social Care, regulators, integrated care boards, providers and clinical teams. The power of space for reflective conversations is well recognised (NHS England, 2023b). However, the reluctance to engage in reflective learning, particularly after patient safety events and deaths, is a common organisational defence and the investigation found this is not happening in a consistent way.

2.1.63 A subject matter advisor told the investigation that “true learning takes time, freedom from persecution, and space for genuine reflection”. The importance of a relational approach to learning, for example communicating with others that embodies respect, inclusiveness, honesty, compassion, cooperation and humility and a ‘just culture’ is required to mitigate the risk of compounded harm and maximise opportunities for healing, learning and improvement (The National Collaborative for Restorative Initiatives in Health, 2023).

2.1.64 In healthcare, learning is viewed as essential for refining system processes and improving patient outcomes (NHS England, 2023b). However, in practice, ‘learning’ is described by families and staff as a tick-box exercise, more focused on rapid data gathering or assigning responsibility than fostering meaningful insight. This approach is particularly evident after the review of patient safety events, where the focus shifts toward deriving immediate lessons rather than achieving a deep understanding – ultimately serving to contain emotional impact rather than truly learning from it. Such processes often lack the psychological safety and time necessary for honest reflection, leading to conclusions that address surface symptoms rather than core issues.

2.1.65 Research (Macrae, 2016; McHugh et al, 2024) argues that to demonstrate commitment to organisational learning at a policy level, the linguistic use of the term ‘learning’ must be transformed, with an explicit definition or outline of what learning means in the context of incident investigations. Rather than learning being represented by delivery of a set of recommendations to be disseminated, learning should be repositioned as a social deconstruction and reconstruction of shared knowledge.

2.1.66 A subject matter advisor told the investigation that the challenge of creating a culture of learning:

“…is especially acute following traumatic events, such as patient suicides, which leave a profound impact on all involved. Attempting to analyse and draw lessons in the immediate aftermath of these events is problematic; the emotional toll disrupts clear thought and hinders meaningful processing. A culture of true learning requires space for mourning and reflection, enabling healthcare providers, families, and the organisation to engage constructively. This transcends isolated incidents, fostering a resilient system capable of learning from adversity without bypassing the essential healing process. Embracing this humane, reflective approach would cultivate a culture that prioritises long-term safety and genuine improvement over procedural compliance.”

2.1.67 A move towards true learning and enabling a participative process would support eliciting ‘everyone involved’ perspectives without judgement or value based on hierarchy, power or emotional connections. The investigation found that there remains a lack of clarity about the meaning of ‘learning’ despite it being the intended key outcome of an investigation.

Summary

2.1.68 Patient safety investigations aim for transparency and learning within the context of mental health which is complex and unpredictable. Principles include patient engagement, data analysis, collaborative working, and fostering a safety culture. However, the investigation has found variable quality in investigations and incident analysis, often failing to lead to effective improvements. Some organisations have struggled with applying PSIRF in mental health contexts and in cross-boundary patient safety investigations involving multiple organisations. Integrated care boards face challenges co-ordinating cross-boundary investigations. The investigation notes that PSIRF implementation in mental health organisations was in the early stages of implementation.

2.1.69 The HSSIB investigation has considered learning from deaths which covers multiple processes; including the Learning from Deaths Framework, coroner's inquests, and investigations following patient safety events. The purpose of these different processes is to understand and learn from avoidable harm however the approaches are not aligned. The impact and consequences of legal processes and how these interact with the objectives of patient safety incident investigation within the PSIRF framework for learning has highlighted some of the tensions and challenges for healthcare staff and organisations. In addition, the meaning of ‘learning’ in patient safety events and investigations has yet to be defined and understood.

2.1.70 Conversations about patient safety events are difficult, especially following unexpected deaths. There remains a cultural and emotional impact from investigations which significantly affect staff and families, often leading to feelings of blame and trauma. Families feel excluded and mistrust investigation and inquest processes. Staff experience emotional strain, fear of blame, and feelings of abandonment. An emphasis on fairness, transparency, and support for both families and staff is needed. The importance of learning and accountability, rather than blame, is stressed for improving patient safety.

HSSIB makes the following safety recommendation

HSSIB makes the following safety observation

Investigations should aim to learn from the care provided with a view to improving future services. However, the investigation was told that there were key areas within mental health care where organisations did not feel they were learning from deaths. These included:

• person-centred care
• self-strangulation without a ligature anchor point
• right place of care.

The investigation explored care in these settings to understand what learning had been embedded from patient safety investigations and learning from deaths reports. This section includes insights that could potentially reduce the risk of deaths in the future, acknowledging the inherent uncertainty associated with catastrophic self-harm and suicide.

Person-centred care – what matters to me

2.2.1 The investigation spoke with several families and carers of people who had died while receiving inpatient care for their mental health problem. Many families described that their family member’s care was not person centred and gave examples where a lack of personalised therapeutic care had left their family member feeling “hopeless, causing them unnecessary distress”. Some families described their family member’s isolating themselves in their bedrooms and not being engaged in activities and no evidence of staff seeking to engage with them in that setting to identify the barriers to leaving their room. This is not in keeping with the Mental Health Act code of practice, which states that ‘providers should encourage patients to avoid staying in their bedrooms for prolonged periods during the daytime’ (Department of Health, 2015).

2.2.2 Person-centred care is fundamental to the core principles in the NHS Constitution (Department of Health and Social Care, 2023b). It is emphasised in guidance by the National Institute for Health and Care Excellence (2016; 2022) and forms regulation 9 of the Care Quality Commission (2024b) standards. Research emphasises that when the focus is on the person, considering their holistic needs, there is improved engagement with therapy and satisfaction and communication between the patient and the healthcare professional (Boardman and Dave, 2020). Upholding these principles is crucial for delivering patient-centred, ethical and effective mental health services (Khosravi et al, 2024).

2.2.3 The investigation met with an expert by lived experience who said:

2.2.4 The investigation was told, and observed during visits, that activities such as music, art or physical activity, for example a supervised gym session or dance session, are important as they give people a sense of purpose, and structure to the day. The investigation observed examples of occupational therapists (OTs) supporting the wider healthcare team to deliver therapeutic interventions. OTs described the satisfaction they felt when working with the team to support engagement of patients in activities they enjoyed. Examples of patients being actively engaged by staff, and patients included activities like dance, pottery and cooking evenings which patients said, “prevented them getting bored” and “reduced their anxiety and likelihood of self-harm”. However, the investigation was told about concerns related to staffing issues which resulted in a lack of these therapeutic activities and lack of one-to-one sessions with staff. The HSSIB investigation report ‘Mental health inpatient settings: creating conditions for the delivery of safe and therapeutic care to adults’ (Health Services Safety Investigations Body, 2024a) described that many factors contribute to the therapeutic environment and the development of therapeutic relationships.

2.2.5 The investigation observed variability in how staff carried out one-to-one observations for patients. In one organisation the investigation was told about staff wearing “bum bags” when undertaking one-to-one observations which contained toys, cards and sweets that could be used during one-to-one engagement activity. The investigation heard the term “given toxic waste as per their request” during a handover meeting. The investigation explored this and was told “toxic waste” is a very sour sweet which once in the mouth tastes “pretty disgusting”. Patients described how this helped them manage their desire to self-harm as “they couldn’t think of anything else”. Other interventions that focused on keeping the patient safe and meeting their needs were the use of ice buckets that some patients liked to put their heads in as a form of distraction, and being able to throw red dye at a wall. Patients described that these interventions “really helped take their minds off wanting to seriously hurt themselves” and were examples of how staff were trying to work with their patients.

2.2.6 The investigation saw examples during one-to-one observations where staff appeared very distant from the patient, with very little engagement, and sitting with arms folded watching or following the patient around with no conversation. This has been reported in a previous HSSIB investigation, ‘Patients at risk of self-harm: continuous observation’ (Health Services Safety Investigations Body, 2024b).

2.2.7 The ‘Mental Health Act 1983: code of practice’ (Department of Health, 2015) is clear that access to fresh air and leave is important for people’s recovery, and that decisions about people’s ability to take leave should consider the benefits and any risks to the patient’s health and safety of granting or refusing leave and the establishment of appropriate conditions to ensure the safety and well-being of both the patient and the public. Families of young people told us that access to outside space was restricted and that leave could be restricted if it did not fit in with staffing needs.

2.2.8 The investigation was told by some families that their family member’s admission was the start of a worsening condition. Bereaved families told the investigation that their family member’s self-harming patterns worsened while they were in inpatient care as they learned new ways of self-harming when witnessing other patients self-harm. In addition, bereaved families described the trauma their family member’s experienced at witnessing and listening to other patients being restrained – “this was frightening at times when they were at their most vulnerable”.

2.2.9 There was an understanding by families that their family member would start having interventions to manage their mental health once admitted but were often told they were “too unwell for therapy”. Some families were told that their family member “would not engage in therapy” and that no further offers were then made.

2.2.10 The investigation was told that psychological therapies and forms of medical treatments which, to be effective, require the patient’s co-operation are not automatically inappropriate simply because a patient does not currently wish to engage with them. Such treatments can potentially remain appropriate and available as long as it continues to be clinically suitable to offer them and they would be provided if the patient agreed to engage, as outlined in guidance by NHS England (2024e). A Consultant Psychiatrist told the investigation "…understanding why patients don’t wish to try [psychological therapies] it and talking though other options (true even if detained and receiving interventions without their consent), you keep working with them to find some way forward […] if they say no to all reasonable options then we have to enforce something so not just let them drift and get worse with more harms".

2.2.11 The investigation heard from families that visits were prevented or restricted and when visits were allowed, they were not in an environment that supported the family to hold conversations. One family described visits being in “a goldfish bowl” with no privacy. All parents the investigation engaged with described being overly restricted in terms of where they were allowed to spend time with their family member.

The investigation was told by a Subject Matter Advisor that any restrictions imposed must be reasonable, proportionate, documented and explained.

2.2.12 The Mental Health Act code of practice (Department of Health, 2015), supports that patients have the right to maintain contact with, and be visited by, anyone they wish to see, subject to carefully limited exceptions. The value of visits in maintaining links with family and community networks is recognised as a key element in a patient’s care, treatment and recovery. Article 8 of the European Convention on Human Rights (European Court of Human Rights, 2024) protects the right to a family life. In particular, every effort should be made to support parents to support their children. Patients should be able to see all their visitors in private, including in their own bedroom if the patient wishes (Department of Health, 2015).

2.2.13 Many families told us they felt excluded while their family member was an inpatient. This was particularly evident when their family member had been repeatedly self-harming, with families describing either not being told or not being involved in any opportunity to help with care or by being involved in their safety plan. One bereaved family described a “whacking great barrier” between hospital staff and patients. The family described how their daughter became fearful of telling them how she was feeling because of the repercussions from staff caring for her:

2.2.14 The investigation findings were consistent with a listening day held by INQUEST (2023) in which it was reported that:

‘… several participants spoke about their anger and frustration at the inadequacy of systems and policies on information sharing prior to their relatives’ death. Most commonly, families wanted to discuss medical needs, changes in health and well-being or broader concerns around their relatives’ treatment. Some participants expressed guilt and remorse, suggesting they could have done more, but in fact faced an administrative system that was hostile to family involvement. Many tried to inform medical professionals about inappropriate treatment, deterioration in their relatives’ mood and concerns about behaviour that they knew to be indicative of unhappiness and isolation.’

Evidence supports that families and carers should have as much involvement as possible in the assessment process, including the opportunity to express their views on potential risk (The National Confidential Inquiry into Suicide and Safety in Mental Health, 2018).

2.2.15 The Triangle of Care (Carers Trust, 2013) is a therapeutic alliance between a service user, staff member and carer (family) that promotes safety, supports recovery and sustains wellbeing. Carers are key partners with health and care services and local authorities in providing care, especially for relatives and friends who have mental disorders (Care Act, 2014). Empowerment and involvement are guiding principles of the Mental Health Act code of practice (Department of Health, 2015), which states clearly that patients should be fully involved in decisions about care, support and treatment. However, the investigation found that patients and their families are not always involved.

Concerns regarding information sharing were identified in the interim report ‘Creating conditions for learning from deaths and near misses in inpatient and community mental health services: Assessment of suicide risk and safety planning’ (Health Services Safety Investigations Body, 2024c).

Person-centred care – autism and neurodevelopmental conditions

2.2.16 Research looking at autistic adults’ experiences of support and treatment for mental health difficulties has reported that they are at high risk of mental health problems, self-injury and suicidality (Camm-Crosbie et al, 2019). The investigation heard many examples from patients, carers and staff where inpatient mental health environments did not always cater for people with autism and neurodevelopmental conditions.

2.2.17 Bereaved families gave examples of environments being too noisy, unpredictable and chaotic, which they said contributed to changes in their family member’s behaviour and a pattern of increasing self-harm. This meant their family member would have more restrictive interventions such as being placed on enhanced observations or not being allowed on home leave, which was something they looked forward to. This finding is in keeping with the evidence that autistic people who are detained under the Mental Health Act are more likely than the general population to experience restrictive practice (Health and Social Care Committee, 2021a). The investigation observed examples of wards with sensory rooms and further evidence of this is reported in ‘Mental health inpatient settings: Creating conditions for the delivery of safe and therapeutic care to adults — HSSIB’ (Health Services Safety Investigations Body, 2024a).

2.2.18 The investigation was also told by bereaved parents that their autistic child was able to hide the extent to which they were struggling and were seen as “coping” and “tolerating” their inpatient environment; after a period of leave they would then beg not to be taken back.

2.2.19 NHS England told the investigation that it had produced an ‘autism informed’ supporting document for inclusion in the culture of care standards for mental health inpatient services. This guidance is expected to be approved in the near future (early 2025). Draft guidance includes that an:

‘… individualised approach is necessary whereby practitioners seek to embrace compassionate curiosity to work alongside autistic individuals and their families. It requires recognition of the autistic individual as equal partner and expert in their own experiences and needs, to explore the accommodations and adaptions that are beneficial.’

2.2.20 In addition, guidance currently being produced by NHS England, ‘Staying safe from suicide: Best Practice Guidance for Safety Assessment, Formulation and Management’, may help promote a shift towards person-centred care models that prioritise individualised assessments and holistic approaches over standardised risk assessments, enhancing patient outcomes. This guidance is expected to be published in early 2025.

Self-strangulation without a ligature anchor point (non-suspensory strangulation)

2.2.21 The HSSIB investigation report ‘Mental health inpatient settings: creating conditions for the delivery of safe and therapeutic care to adults’ (Health Services Safety Investigations Body, 2024a) considered where ‘fixtures and fittings were able to be used as ligature anchor points (points where something could be attached for the purposes of self-harm)’.

2.2.22 Ligature anchor points have been associated with multiple patient deaths (National Confidential Inquiry into Suicide and Safety in Mental Health, 2024). During the investigation’s site visits to providers, it observed that ligature audits and risk assessments described the location of anchor points and mitigations to prevent or reduce the likelihood that they could be used for self-harm by hanging. Some hazards could be eliminated (for example removal of non-functioning fixtures) or changed for something less hazardous (for example replacing with an anti-ligature type).

2.2.23 As defined by the Care Quality Commission (2023):

‘A ligature is anything, like a cord or other material, that could be used for the purpose of hanging or strangulation. A ligature anchor point is anything that could be used to attach a ligature. Ligatures do not necessarily need to be attached to a ligature anchor point.’

The term ‘non-anchored ligatures’ refers to items that patients may use to attempt self-harm without needing to fix these items to a physical structure, such as clothing, bedding, or personal belongings.

2.2.24 Between 2011 and 2021, there were 342 deaths by suicide on inpatient mental health wards in the UK. Of these, 80% (272) of patients died by using a ligature; in 8% (18) of these cases, no ligature point was used. This increased to 26% in the under 25 year old age category. In 167 (64%) patient deaths by ligature, patients used an item of clothing or a personal object (National Confidential Inquiry into Suicide and Safety in Mental Health, 2024).

2.2.25 As described by the Care Quality Commission (2023), one of five key factors to reduce harm from ligatures is the built environment:

‘Controlling the built environment reduces opportunities for a patient to use fixtures, fittings, or furniture or their personal items (such as clothing) as ligatures or ligature points to cause harm to themselves or attempt suicide ... clothing and other items commonly available in hospital wards and in the home are the main materials used.’

One of the key considerations is ‘are there mitigations to minimise or eliminate potential areas of risk?’

2.2.26 During engagement with bereaved families, carers, and with patients, the investigation heard about occurrences of self-strangulation without the use of ligature anchor points. Families described how they felt that there were not sufficient mitigations put in place for known and recent behaviours and risks, which could have prevented the death of their family member. In addition, families told the investigation they were not included in developing and reviewing decisions about how to keep their family member safe, sometimes describing not being told about significant events of self-harm. One family described how their daughter died by non-anchored self-strangulation using a specific item of their clothing.

2.2.27 The patient had attempted to self-strangulate using a specific item of their clothing on several occasions during the previous 2 days, and earlier on the day that she died. The clothing was untied, if the specific clothing was removed from the patient’s possession it was documented that the patient would become aggressive as they were a sensory need. The patient was left with their item of clothing. The investigation was told that this was because there is a balance with managing patients in line with the least restrictive practice policy.

2.2.28 On review of the trust’s after action review, and of the follow-up patient safety incident investigation, no learning was identified regarding leaving the patient with items known to be used regularly to ligature.

2.2.29 Another family member told the investigation about a known and ongoing ligature risk of their daughter using a different but specific item of clothing:

2.2.30 On exploration across providers, the investigation heard that extensive audit activity, time, resource, and money are invested into minimising physical ligature anchor points to mitigate ligature opportunities. The safety benefits of this activity are however null and void for patients who self-strangulate without the need of a ligature anchor point and are left with items to carry out that self-harm ligature action. Staff told the investigation that the use of items of clothing and hair accessories was significantly increasing and that “it was difficult to know what to do”. The investigation was told by a Subject Matter Advisor that there is a risk in seeking to remove non – fixed ligatures as a blanket approach and is very unlikely to reduce suicide rate.

2.2.31 Although there is ligature anchor point audit guidance provided across the mental health inpatient system, the investigation did not identify specific guidance on how to manage non-anchored ligature risks, or for managing access to known ligature risk items and maintaining a therapeutic environment.

2.2.32 The investigation was told by a subject matter advisor that non-anchored ligature risks are particularly difficult to address because it is often impractical, counterproductive, or even impossible, to remove all potential items without infringing on a patient’s comfort or autonomy. Effective guidance would shift some of the systemic responsibility for these complex ethical and moral issues from individual staff – who are often working in highly stressful environments – toward a structured, supportive approach that allows for safe and thoughtful management in real time. Another Subject Matter Advisor told the investigation that “any guidance needs to recognise the delicate and dynamic balance between restrictive practice to try to reduce harms and restrictive practice reducing therapeutic effectiveness leading to more harm”. HSSIB has made a safety recommendation relating to this point which can be found at the end of this section.

Right place of care – deaths from catastrophic self-harm or suicide after discharge

2.2.33 The Parliamentary and Health Service Ombudsman (2024) report on discharge and transition between services within mental health care stated that ‘unsafe discharge potentially leads to poorer outcomes for patients and the risk of repeated cycles of readmission: a revolving door in and out of services’. The report also referred to the impact on family and carers caused by poor discharge planning. Walter et al (2019) reported that discharged mental health patients were 32 times more likely to die by suicide than the general population, and stated that fatal drug overdose, irrespective of intent, was more than 90 times more likely among this group. Findings from the PHSO report included those unsafe discharges which had led to serious deteriorations in people’s mental health and wellbeing and their death.

2.2.34 The investigation saw evidence of people being discharged from hospital with very little safety planning or involvement of the wider multidisciplinary team in the process, which had led to deaths or serious harm. This was explored in the interim report ‘Learning from inpatient mental health deaths and near misses: assessment of suicide risk and safety planning’ (Health Services Safety Investigations Body, 2024c). Family members told the investigation they were not involved in discharge and safety planning, as explored in the recent investigation report ‘Mental health inpatient settings: Supporting safe care during transition from inpatient children and young people’s mental health services to adult mental health services’ (Health Services Safety Investigations Body (2024e). The report identified examples where parents of young people were not made aware of their children’s changing patterns of behaviour, or key safety risks, as part of the discharge process.

2.2.35 The investigation also heard from staff and families of patients being discharged as homeless without their housing needs being addressed, or discharged to caravans or bed and breakfast hostels. Prevention of future deaths reports refer to poor risk assessment, care planning and consideration of Section 117 aftercare needs.

2.2.36 The investigation heard one mother say:

2.2.37 The investigation heard about challenges to getting patients discharged, with patients describing themselves as “stuck” or “not belonging”. The investigation heard about patients who had been ready for discharge from acute inpatient mental health care but who had died by suicide while awaiting an onward place of care with appropriate support. Staff told the investigation that when there are delays to discharging people from inpatient mental health care, they have witnessed increasing self-harming behaviours which have resulted in catastrophic harm.

2.2.38 The investigation was told by a national stakeholder that using the wider multidisciplinary team is important to support safe discharge. They stated, “OTs are ideally positioned to support discharge and accommodation discussions. They also provide rehabilitation to help individuals retain or regain skills that may be at risk during acute episodes or extended stays in ward settings". The investigation observed examples of multidisciplinary involvement in discharge planning.

2.2.39 Statutory guidance (Department of Health and Social Care, 2024e) sets out best practice on how NHS organisations and local authorities should work closely together to support the discharge process and ensure the right support in the community, and provides clarity in relation to responsibilities patient and carer involvement in discharge planning. The investigation was told by the Department of Health and Social Care that it has been exploring the barriers and enablers to discharge from mental health settings to better understand the issues. The investigation was told the deep dives focused on discharge from mental health settings and built on the best practice principles within the statutory guidance. The investigation engaged with the leads of this work to understand their findings.

2.2.40 The findings of this investigation, in relation to the barriers to discharge, were aligned with the Department of Health and Social Care’s findings, which included the complexity of patient needs, particularly those with co-existing neurodiversity and autism, forensic history or substance misuse, and the systemic barriers such as lack of housing, provision of supported living accommodation, and a culture of risk aversion among community providers. The Department’s early findings suggested discharges are often delayed because of funding constraints, workforce shortages, and inter-agency communication issues. This is echoed by findings in the HSSIB investigation reports ‘Harm caused by mental health out of area placements’ and ‘Mental health inpatient settings: Supporting safe care during transition from inpatient children and young people’s mental health services to adult mental health services’ (Health Services Safety Investigations Body 2024d; 2024e). The departments findings included enablers for better discharge practices which included having dedicated housing resources within discharge teams, early identification of barriers to discharge, risk-tolerant approaches to patient management, and collaboration with the voluntary sector and community organisations. The importance of a single integrated data system and a clear well-embedded governance structure for discharges was emphasised.

Patients presenting in crisis

2.2.41 When people are in mental health crisis, they need timely access to support that is compassionate and meets their needs. A key aim of Integrated Care Boards (ICBs) and MH services and systems, is to reduce the numbers of people reaching MH crisis point (NHS England, 2024h). The investigation heard about gaps in community mental health care which were adding to pressure across the mental health system, with people being cared for in inappropriate environments (Care Quality Commission, 2024c). The investigation did not explore in detail cross-provider working on discharge of patients to the community, or cross-agency working such as with the police, local authorities or social care. However, the investigation heard about and considered some of the barriers to effective discharge and people presenting to other services in crisis.

2.2.42 Referring to a patient who had presented many times in “crisis” and trying to coordinate their care and treatment to avoid crisis situations, a Consultant Psychiatrist told the investigation:

2.2.43 The HSSIB investigation report ‘Mental health inpatient settings: creating conditions for the delivery of safe and therapeutic care to adults’ (Health Services Safety Investigations Body, 2024a) referred to limited availability of crisis resolution and home treatment teams, and limited availability/access to community mental health services, outpatient psychiatric care and social care (Rethink Mental Illness, n.d.; The King’s Fund, 2024).

2.2.44 The investigation heard many examples of concerns regarding accessing the right place of care (see 1.3). The investigation was told of challenges with the introduction of the ‘right care, right person’ approach (see 1.3.4) which left gaps in service provision for vulnerable people with mental health problems.

Section 136 suites

2.2.45 The investigation witnessed examples of Section 136 suites being used as ‘holding areas’ while inpatient beds were found. Section 136 suites are described as being suitable only for a temporary placement for those suffering an immediate mental health crisis. As self-contained units, they may provide suitable accommodation for people, provided that appropriately trained staff are available, but using them in this way removes this limited resource from their intended use. The investigation saw examples where suites were occupied for several days. The investigation spoke to a patient who was being cared for in a 136 suite and described being “frightened” and having to request to have her door locked as “the other people here are very scary and loud … I don’t feel safe but there’s no bed”.

2.2.46 The investigation reviewed prevention of future deaths reports (PFDs) relating to incidents where a patient died while being detained in a 136 suite for an extended period.

Emergency departments

2.2.47 The investigation was told by staff working in some emergency departments (EDs) that the number of people who have been discharged from mental health inpatient settings and then come to their ED is increasing. A review of evidence on the Urgent and Emergency Care (UEC) MH dashboard shows no significant change in total numbers of percentage of MH attends to UEC since 2019. However, it does show a significant rise in people with mental health problems waiting over 12 hours (FutureNHS, n.d.).

2.2.48 The challenges of people presenting to ED in mental health crisis and subsequent long waits, were described by staff as “significant”. Reasons given included having people at high risk of harming themselves or others in the ED without any of the powers, or skills, that mental health services have.

2.2.49 The investigation reviewed PFDs which highlighted issues including a patient being placed in an unsuitable environment and exiting through a window, inexperienced staff caring for unwell patients and patients absconding from an ED.

2.2.50 One of the responses to this PFD advised that:

‘all systems … must ensure that there are clear pathways for mental health patients who are accessing care via EDs and who need to remain in acute hospital settings until their care can be transferred. This should be supported by access to 24/7 mental health liaison teams (or other age-appropriate equivalents for children and young people), both in Accident & Emergency settings, and on the wards.’

However, the investigation was told that in some cases this does not happen and that liaison services are very stretched, and care is not safe. This was considered in a previous HSSIB investigation report, ‘Provision of mental health care to patients presenting at the emergency department’ (Healthcare Safety Investigation Branch, 2018).

2.2.51 There is a UK-wide consensus statement (Royal College of Psychiatry et al, 2020) on working together to help patients with mental health needs in acute hospitals. However, there is ongoing evidence that there is a lack of joint ownership for the safe care of people with mental health needs in the ED.

2.2.52 The investigation was told of examples of people taking overdoses and then refusing treatment. In those circumstances physical healthcare staff need on-site mental health professional presence and support to act in patients’ best interests, which may go against their wishes. ED staff said: “Often we do not have the skilled mental health expertise in the department and people can be waiting days for an assessment.” It was also described that when alcohol is involved “no one is interested in seeing patients who then have to stay in ED until they are sober or withdrawing from a toxic substance”.

2.2.53 The investigation was told of recent examples where the language of “we’ve got one of yours here” was used (referring to a mental health patient in the ED) and examples of patients being considered “not fit for assessment” and “requiring medical clearance” before being seen by liaison psychiatry. The investigation was told that very poorly patients in mental health crisis were being cared for by health care support workers until an inpatient bed in a mental health service could be made available.

2.2.54 Evidence was submitted to the House of Commons Select Committee in January 2023 following scrutiny on the draft Mental Health Bill and criminal justice system in November 2022 (UK Parliament, 2022; Department of Health and Social Care, 2022b). The recommendations requested greater powers for ED clinicians; however, the recommendation to give these powers has not been implemented with the situation described as “stuck”.

Crisis resolution and home treatment teams

2.2.55 The investigation was told by many people who try to access crisis services that they are not “heard or believed”. One patient told the investigation that she was told “to have a cup of tea and warm bath” when she telephoned a crisis service feeling very suicidal.

2.2.56 SANE (n.d.) reported that:

‘… far too many people in crisis are being turned away because no local beds are available or deprived of their liberty under section as the only way to receive treatment. They may find themselves locked in police cells, shunted around the country or placed under the care of overstretched crisis resolution teams, where they are now three times more likely than in-patients to take their own lives.’

Media reports have also referred to concerns in mental health crisis care and community provision (BBC News, 2024).

2.2.57 The Royal College of Psychiatrists (2022a) sets out guidance in the form of ‘Practice guidelines for crisis line response and crisis resolution and home treatment teams’, created as part of its quality network that looks at improvements and standard setting. The purpose of the practice guidance is to help services to improve their support for patients in crisis and their carers/families, including contributing towards improving the quality of care and service delivery within crisis services. The document outlines a number of points that could be followed by crisis services so that they can ensure they are working to provide effective, timely and appropriate crisis care to individuals who need it. These include inclusion and exclusion criteria, referral pathways, safety planning, onward referral, least restrictive options and compassionate and psychologically informed care.

Additional services – local authority/social care/safeguarding

2.2.58 The investigation found concerns relating to deaths, particularly of young people, due to problems with multi-agency working between health trusts and relevant agencies with regard to safeguarding and keeping people safe.

2.2.59 Prevention of future death reports refer to lack of multi-agency staff to respond to requests for patient assessments or attend discharge planning meetings, not allocating social workers to people in need prior to discharge and a lack of assessment for any Section 117 needs to facilitate a safe discharge. This has resulted in people being discharged to “wholly unsuitable” living environments, or even being left homeless, and their subsequent death.

Community provision and therapies

2.2.60 The investigation was told of significant gaps in community provision of therapies beyond traditional talking therapies. A GP told the investigation that they had written numerous letters to mental health services concerning a very vulnerable patient who had been in and out of inpatient settings and discharged back to community mental health teams with no support. They described “a stream of rejected referrals”:

2.2.61 A consultant psychiatrist told the investigation that pressure to maintain patient flow is a significant issue. They described challenges regarding waiting for prescribed medications to take effect (generally 6 weeks) against an expectation of bed stays of 20 days or less. They tearfully told the investigation: “We are trying to do things with a degree of hope and flying by the seat of our pants.” Another member of staff at a different organisation told the investigation: “Akin to water boarding, is bed management.” These challenges were described in the HSSIB investigation report ‘Harm caused by mental health out of area placements’ (Health Services Safety Investigations Body, 2024d).

2.2.62 The investigation has been made aware of concerns regarding mental health service signposting and that “community hubs are merely a ‘one stop shop’ whereby interventions consist of chats with Peer Support Workers, are time limited, and inappropriate for their needs”. Staff at the sites visited by the investigation referred to the use of peer support workers working outside of their scope of practice and that “they can’t replace clinical expertise”. In addition, the investigation was told that some services offered to people in mental health crisis have no clinical expertise.

2.2.63 SANE’s experience is that far too many people in crisis are turned away from help when they are at their most vulnerable and told they must make do with returning home to await a phone call from a crisis resolution/home treatment team.

2.2.64 The investigation heard about different approaches to managing care pathways for people with mental health problems and how services are being transformed to provide better care. One organisation shared that its intention was to move to a ‘place based’ model of care. The investigation was told that that this revised structure would enable consistency of clinical services in each of the localities with ‘strengthened clinical leadership’. Specialised services would be supported through the adoption of ‘hosting’ arrangements within each locality group, ensuring a maintained concentration of clinical expertise and ‘economies of scale’.

2.2.65 The investigation heard positive examples of using the voluntary sector but also examples where this was not meeting the needs of people. Some bereaved families described services not meeting their family member’s needs, with the parent of a young person describing that their child “was simply bouncing between services and they never had a sense of belonging”.

2.2.66 NHS England’s aim was to transform community mental health care to improve patient care and address long-term system pressures in mental health pathways by 2024. NHS England told the investigation about its new pilot model. Six new Neighbourhood Mental Health Centres have launched offering 24/7 community support for individuals with serious mental illness. These centres integrate crisis intervention, community support, and open access beds to facilitate extra support, tailored to local needs. These centres are rooted in local neighbourhoods. Individuals can visit without a referral to receive help from a range of professionals including psychiatrists, social workers and peer support workers, and support such as psychological therapies, medication support and assistance with related issues such as housing or employment. Each centre, led by an NHS provider, will work in partnership with people with lived experience, as well as voluntary, charity, faith and social enterprise organisations.

2.2.67 With a strong focus on open access, continuity of care and fostering trusted therapeutic relationships, the centres will provide support closer to home, reducing the need for out-of-area hospital inpatient treatment, and ensuring people can maintain a sense of citizenship and belonging in their community while accessing the service.

2.2.68 The investigation was not able to fully explore the community and voluntary sector support offers for people discharged from inpatient mental health services across England due to the investigation’s focus on mental health inpatient settings. However, given the feedback from patients, and from staff who had been involved in the care of a patient who had died within 30 days of discharge, it was considered important to highlight the concerns expressed, particularly around the described lack of meaningful support.

2.2.69 Where community mental health support and provision does not appropriately meet the needs of people discharged from inpatient mental health settings, risks to their ongoing recovery, and in turn their safety, are increased. Prevention of future deaths reports have shown that this has led to deaths.

2.2.70 A Subject Matter Advisor told the investigation that continuity of care is important and that up to date patient care records, including what may happen if a person is to relapse and their potential behaviour pattern if they do relapse, are fundamental to this. This is because it is likely that people with a SMI are likely to relapse and may need help in a crisis. Having patient specific information readily available for healthcare professionals and for the person picking up the phone on a crisis line, will better help safe decision making.

2.2.71 The HSSIB investigation report ‘Mental health inpatient settings: Supporting safe care during transition from inpatient children and young people’s mental health services to adult mental health services’ (Health Services Safety Investigations Body, 2024e) made a safety recommendation to NHS England regarding communication of essential safety and risk mitigation information when patients transition from inpatient mental health services to community mental health services.

Summary

2.2.72 This section emphasises the importance of person-centred care in mental health services, highlighting that focusing on individual needs can enhance patient engagement, satisfaction and recovery. Patients and families shared that activities such as music and exercise helped them cope, but staff shortages often limited these options, leaving some feeling isolated and disconnected and subsequently catastrophically self-harming. Families voiced concerns over a lack of personalised care and restricted involvement in care decisions, which affected their ability to support their family member, leading to guilt and anger that they could not stop their family member dying.

2.2.73 Although there is ligature anchor point audit guidance provided across the mental health inpatient system, the investigation did not identify specific guidance on how to manage non-anchored ligature risks, or on managing access to known ligature risk items whilst maintaining a therapeutic environment. There is recognition that there are issues in discharge planning and community crisis support, with many patients unable to access timely and compassionate care, leading to increased presentations in emergency departments, worsening their mental health crises. NHS England is developing guidelines to improve personalised care and is piloting Neighbourhood Mental Health Centres to strengthen community support and continuity of care.

HSSIB makes the following safety recommendations

HSSIB makes the following safety observation

National oversight

3.1.1 The accountability and oversight of mental health inpatient deaths involves multiple organisations and mechanisms intended to ensure the safety of patients and provide assurance that care standards are upheld (NHS England, 2024f). Figure 1 is a summary of the oversight mechanism (Healthcare Financial Management Association, 2023).

Figure 1 Summary of the oversight mechanism for deaths of patients in mental health inpatient settings

3.1.2 The Care Quality Commission (CQC) is the independent regulatory body responsible for monitoring and inspecting health and social care services in England, including mental health facilities (Care Quality Commission, 2024d). The CQC also has a duty under the Mental Health Act 1983 to monitor how services exercise their powers and discharge their duties when patients are detained in hospital or are subject to community treatment orders or guardianship. The CQC requires mental health facilities to promptly notify them of any deaths, especially if the death was unexpected or occurred under unusual circumstances. This enables the CQC to assess whether the facility met safety and care standards in the lead-up to the incident. Recent CQC reports have highlighted safety concerns leading to demands for immediate improvements (Care Quality Commission, 2024a).

3.1.3 The expectations in relation to reporting, monitoring and board oversight of incidents involving patient deaths are set out in the NHS England National Quality Board’s Learning from Deaths (LfD) guidance (National Quality Board, 2017; NHS England, 2017). The LfD framework places particular responsibility on trust boards to ensure their trust has robust systems for recognising, reporting and reviewing or investigating deaths where appropriate. The LfD states that ‘the aim of this process is to ensure that all deaths of people under the Trusts’ care are reviewed at the appropriate level and organisational learning occurs’. This framework specifically targets the deaths of individuals with serious mental illness who are under the care of inpatient, outpatient and community mental health services, aiming to identify systemic issues and improve care.

3.1.4 Trust boards are accountable for ensuring compliance with the LfD framework and working towards achieving the highest standards in mortality governance. NHS England has published guidelines for implementing the LfD framework for trust boards (NHS, England, 2017).

System-level oversight

3.1.5 In 2022, Parliament passed the Health and Care Act 2022, which aimed to make it easier for services to work together to provide joined-up care for patients. This formalised the work of integrated care systems (ICSs) which are partnerships, consisting of NHS services, social care, and other organisations, which together provide care in defined geographical areas. Each ICS has an integrated care board (ICB), which determines what care is needed and how funding will be allocated to the various bodies in the ICS, including mental health trusts.

3.1.6 The government (Department of Health and Social Care, 2024b) describes that:

‘DHSC [the Department of Health and Social Care] and NHS England will work with ICS leaders to highlight the importance of ICSs taking a leadership role in ensuring the quality of data, to improve patient safety and therapeutic care in inpatient settings and reduce the time needed for frontline staff to input data.’

3.1.7 In all sites visited during this investigation, people frequently described relationships with ICBs as “challenged” and that integrated working is “worse than before [the establishment of ICBs]”. One mental health provider told the investigation that because of the nature of its services and the geographical spread, it had to report to three separate ICBs, “all of whom operated differently”. The investigation was told “relationships across the system outside of [their location] are very challenging – they [ICBs from other areas] have their own priorities and we can’t find common ground at the moment”.

3.1.8 NHS England (2022c) describes that ‘commissioners, providers and other relevant organisations should establish effective relationships to ensure efficient working with accountability defined through joint governance arrangements’. However, the investigation was told of multiple examples where working relationships are strained and that “he who blinks first takes responsibility”.

3.1.9 The investigation was told that there was sometimes a disconnect between health and social care. This was described in the HSSIB investigation reports ‘Harm caused by mental health out of area placements’ and ‘Mental health inpatient settings: Supporting safe care during transition from inpatient children and young people’s mental health services to adult mental health services’ (Health Services Safety Investigations Body, 2024d; 2024e).

3.1.10 One provider described a “fractured relationship with the ICB” in relation to the provision of crisis services given to a local authority to manage; however, there were concerns about the quality and safety of the new arrangements. The concerns were described as having been escalated to the ICB, but this had resulted in the provider that escalated the concerns being “ejected from meetings as they were not seen as collaborative”. The impact of the change to service provision meant that other teams were picking up crisis concerns, which in turn meant they were not able to fulfil the requirements of their usual roles and responsibilities. The provider told the investigation that the ICB was making decisions on implementation of services but was not involved in the delivery of services. The impact on provider staff when there was a serious system-level safety concern created fear, with a staff member saying, “they [ICBs] are not held to account … I am the one in the coroners' courts”.

Local oversight

3.1.11 As part of the LfD programme (see 1.2.1), healthcare providers (including mental health inpatient services) are legally required to annually report information about patient deaths. The intention of the programme was that healthcare care providers would share this learning and take measurable action to prevent future deaths. Research by Lalani and Hogan (2021) noted that some providers have taken on the LfD programme as a ‘tick-box’ exercise and their research refers to the need for formal evaluation when new policies are implemented.

3.1.12 The investigation reviewed the quality accounts related to patient death information across 10 mental health trusts and found significant variability in the data presented. This aligns with research findings that engagement with the LfD programme across healthcare providers varies significantly (Brummell et al, 2023a). There was a wide variation in the proportion of case record reviews or patient safety investigations that were carried out into patient deaths (for example, one trust reported 2% compared to 42% at another). The investigation acknowledged that this may be due to unclear reporting rather than the actual investigations that take place.

3.1.13 The investigation’s findings are consistent with research that suggests wide variation in reporting in quality accounts, demonstrating that some healthcare providers have engaged fully with LfD, while others appear to have disengaged with the programme (Brummell et al, 2021; 2023b). It was reported (Brummell et al, 2023a) that some providers did not believe the guidance was written for or applied to them, and that effective learning about patient safety needs to be defined and agreed on (Brummell et al, 2023b). This was more prevalent across mental health and community providers: ‘… there was no nationally agreed definition for mental health services or community health services with regard to what constitutes a death from problems in care and therefore this data was not reported.’

3.1.14 The investigation was told by some providers and ICBs that trying to report and respond to mortality data was incredibly time consuming with very little benefit to organisational learning, and with little support in terms of resource. The investigation was told that as part of current improvement work, NHS England intends to work alongside CQC to better understand the causes of data burden in frontline services, and work together on principles, case studies and permission-giving messaging to reduce the data burden where it does not clearly drive quality improvement.

3.1.15 The investigation reviewed organisational policies in relation to learning from deaths. The policies referred to the national guidance on the Learning from Deaths framework (NHS England, 2017), recommendations made by the Mazars (2015) investigation into Southern Health, and the Learning Disability Mortality Review (LeDeR) process managed by NHS England. The policies included multiple people having similar accountability and responsibility, a similar finding to recent research on policies by McHugh et al (2024).

3.1.16 Within the implementation of the Learning from Deaths framework (NHS England, 2017), it is intended that healthcare providers ensure quality improvement remains key by championing and supporting learning that leads to meaningful and effective actions that continually improve patient safety and experience and supports cultural change. Examples of how this might be achieved include spending time developing board thinking; ensuring a corporate understanding of the key issues around the deaths of patients; and ensuring that sufficient priority and resource is available for improvement work.

3.1.17 Most organisations told the investigation they had a learning from deaths lead who chaired a learning from deaths group with a non-executive director and clinicians. The frequency of meetings was variable. The learning from deaths meeting would feed into trust-wide patient safety oversight groups (PSOGs). Some organisations described how their PSOG would link with their organisation's quality improvement workstream and described newsletters and bulletins to share learning.

3.1.18 Some organisations included patient safety partners or people with lived experience in their governance forums; however, there was variability in their involvement. The investigation observed meetings which included patient safety partners. However, in the meetings observed there was more of a focus on telling the attendees what was happening rather than actively involving them.

3.1.19 There was also feedback from people with lived experience that, when done in a psychologically safe environment, holding a conversation with senior leaders regarding “difficult issues on death and the impact of this when followed by a poor investigation” can inform thinking and subsequently improvement. Experts by lived experience told the investigation that while it is great they are being increasingly used in patient-facing areas, they consider their experiences and insights would also have significant impact with senior leaders in how organisations can learn from deaths and serious patient safety incident investigations.

3.1.20 The National Confidential Inquiry into Suicide and Safety in Mental Health produced a toolkit for specialist mental health services and primary care ‘10 ways to safer services’ (National Confidential Inquiry into Suicide and Safety in Mental Health, 2022). A nurse consultant told the investigation “this toolkit is an example of how we can learn from previous investigations and recommendations and a method of focusing in on themes which are repeatedly found […] It would be a good place to start if organisations implemented learning from 20 years of data on deaths of MH patients”.

3.1.21 The investigation spoke with some non-executive directors (NEDs) who have responsibility for chairing their local quality committee. They described the enablers and barriers to creating a culture that prioritises patient safety, compassion, and genuine engagement with staff and patients. The investigation was told that there is variability in the way quality committees operate and there are known difficulties in interpreting complex data sets. It was described that this is made more difficult when safety data and learning is not triangulated, necessitating the reading of multiple separate reports related to patient safety incidents, complaints, claims, mortality, and investigation reports without the “so what?” question being answered. The NEDs stressed the importance of moving away from a tick-box mentality towards meaningful oversight and learning, particularly in relation to Structured Judgement Reviews, patient safety incidents and data triangulation. One NED highlighted the role of “tone” and “values” in ensuring quality and safe care, which is more about human interaction than formal processes.

3.1.22 The discussion explored the challenges NEDs face, particularly those from non-clinical backgrounds, in understanding and effectively contributing to the oversight of complex healthcare environments. The investigation identified that having a clinical background enabled greater confidence to challenge data and ask the “so what?” of how the data could support improvements. NEDs emphasised the significance of supportive challenge, asking critical questions, and creating an environment where staff feel comfortable discussing and addressing patient safety concerns. They reflected on the need for better induction and training for NEDs, focusing not just on technical aspects but on human factors, experiential learning and effective challenge.

3.1.23 Guidance is available to help boards to consider their approach to handling and acting on the information they receive (NHS England, 2024i; NHS England 2024j). The guidance and competency framework considers the leadership behaviours and culture of the board and how these can affect the information it receives and the actions it takes, as well as metrics that can support the board to better understand the organisation’s performance. NHS England patient safety syllabus level 1 training has a module for Boards and senior leaders (NHS England, n.d.). In addition, NHS England told the investigation that in response to feedback from Boards, in particular NEDs and Executives with non-clinical background, a new resource was developed to enable patient safety specialists to facilitate a group session either face to face or virtual. NHS providers told the investigation that a training module on quality of care and patient safety specifically for NEDs was introduced from November 2024. Therefore, the investigation has not made a safety recommendation in this area although it would be helpful for the training to be evaluated.

HSSIB makes the following safety observations

Independent investigations

3.2.1 The investigation reviewed a small sample of NHS England commissioned independent investigation reports for people who died while an inpatient in mental health services. The reports contained recommendations, and sometimes an action plan. However, there was limited oversight of the next steps in terms of action implementation. Some actions referred to ‘reviewing’ or ‘reflecting on’ an incident and were ‘RAG [red amber green] rated’ green as achieved, but it was not possible to see the outcome or how learning had been embedded. The investigation was also provided with independent reports from companies commissioned to undertake investigations; however, there was no evidence that recommendations had been implemented. The investigation saw examples of comprehensive follow-up letters from bereaved families to organisations requesting updates on the recommendations, to which the families were still awaiting responses.

3.2.2 NHS England told the investigation that it has an Independent Investigations Committee chaired by the National Patient Safety Director, with representation from all NHS England regions. The business of the committee covers mental health homicide as well as other independent investigations. With the support of the National Patient Safety Independent Investigations team it has oversight of any recommendations generated for NHS England, or arm’s length body, to ensure there is ownership at a national level where required. In addition, NHS England told the investigation that it is working to enhance mechanisms for translation of insight from mental health homicide investigations into actionable policy via the national mental health team.

3.2.3 There is national oversight of nationally led recommendations. However, NHS England told the investigation that these recommendations may not always be implemented at the local or regional level. As a result, many regional and local recommendations have not been actioned. This was considered in the HSSIB investigation reports ‘Mental health inpatient settings: Supporting safe care during transition from inpatient children and young people’s mental health services to adult mental health services’ and ‘Recommendations but no action: improving the effectiveness of quality and safety recommendations in healthcare’ (Health Services Safety Investigations Body, 2024e; 2024f).

Prevention of future death reports and learning

3.2.4 The Chief Coroner produced guidance for coroners on ‘reports to prevent future deaths’ (Courts and Tribunals Judiciary, 2016). It states that ‘A prevention of future deaths report raises issues and is a recommendation that action should be taken, but not what that action should be’. A coroner told the investigation that “when a coroner writes a prevention of future deaths (PFD) report they are not making recommendations, they are raising a concern”. Other coroners told the investigation the same.

3.2.5 The investigation heard many times the term “recommendations made by the coroner”. Many people, including families, perceive that a coroner has follow-up powers. However, this is not the case. Section 28 of the Coroners (Investigations) Regulations 2013 imposes a requirement on recipients of PFDs to send a response within 56 days. If a recipient of a PFD fails to respond this will be evident on the Chief Coroner’s website (Courts and Tribunals Judiciary, 2024b). The investigation was told that some coroners copy their PFDs to the CQC and the Director of Public Health for their region, but this was not mandated or consistent.

3.2.6 The investigation was told about different mechanisms for capturing themed learning from PFDs. PFD reports and responses are available via the Courts and Tribunals Judiciary website. There is also a ‘Preventable Deaths Tracker’ platform that provides a database of all published coroner reports in England and Wales since 2013; however, whilst basic information on the website is available for free, more detailed reports require a payment subscription. Recent figures show that nationally, between 2013 and 2024, a total of 501 PFDs relating to mental health related deaths have been issued (Richards, 2024).

3.2.7 The investigation was told by some ICB leaders that they report every PFD in their system to their Board and monitor improvements. Some also feed this into “System Mortality” groups. Regional teams told the investigation that NHS England has a mechanism for collating PFDs that NHS England is cited on and the regional teams meet quarterly to review them. In response to a specific PFD, the NHS England National Medical Director stated:

‘I would also like to provide further assurances on national NHS England work taking place around the Reports to Prevent Future Deaths. All reports received are discussed by the Regulation 28 Working Group, comprising Regional Medical Directors, and other clinical and quality colleagues from across the regions. This ensures that key learnings and insights around preventable deaths are shared across the NHS at both a national and regional level and helps us pay close attention to any emerging trends that may require further review and action.’ (NHS England, 2023c)

3.2.8 The Independent Advisory Panel on Deaths in Custody (2023) report ‘More than just a paper exercise’ emphasised the need for meaningful oversight and accountability in preventing deaths in custody. It highlights that investigations and reviews often lack depth and fail to address systemic issues, leading to repeated mistakes and missed opportunities to save lives. The report calls for a stronger commitment to transparency, better support for families, and a proactive approach to implementing recommendations to prevent future deaths.

Oversight of recommendation implementation

3.2.9 At present there is no mechanism for tracking implementation of recommendations after an inquiry or investigation is complete to ensure they are implemented. This links to the findings of the HSSIB investigation ‘Recommendations but no action: improving the effectiveness of quality and safety recommendations in healthcare’ (Health Services Safety Investigations Body, 2024f). This investigation did not find evidence of a framework for, or responses required from, public bodies to ensure inquest and investigation outcomes feed into effective learning through the implementation of appropriate actions. The Thirlwall Inquiry (2024) completed a review of previous recommendations by inquiries relating to events which took place in hospitals and other healthcare settings. The review report included a table of recommendations from over 30 inquiries that were coded to indicate whether there was evidence to suggest that they had been implemented; many had not. This was a clear example of how despite recommendations being made, action has not been taken.

3.2.10 Within the HSSIB report ‘Recommendations but no action: improving the effectiveness of quality and safety recommendations in healthcare’ (Health Services Safety Investigations Body, 2024f), there was a proposal for a monitoring system that supports the co-ordination, prioritisation and oversight of safety across the system, although no specific safety recommendation was made. Other industries often have a ‘state safety board’ with officials, regulators and ministers to perform this role.

3.2.11 As part of the No More Deaths' campaign INQUEST (n.d.a) has repeatedly stated that there is a pressing need for a new independent public body with singular responsibility for collating, analysing and following up on recommendations arising from inquests, inquiries, official reviews and investigations into state-related deaths.

3.3.1 The NHS faces significant challenges with system visibility and oversight, meaning the challenges associated with responding in a timely way to feedback and how the oversight of NHS trusts, foundation trusts and integrated care boards operates, due to its complex, fragmented structure. This investigation and others highlight organisations working in silos, leading to duplication of effort, miscommunication, and ambiguity in understanding patient outcomes (Health Services Safety Investigations Body, 2024f). The investigation identified a lack of insight across departments in the same organisations regarding areas of focus. This appeared to create duplication of effort across teams and created confusion due to mixed messages.

3.3.2 The Department of Health and Social Care (2024c) has commissioned an independent review of patient safety across the health and care landscape. The current phase aims to assess whether the range and combination of patient safety organisations deliver effective leadership, listening, learning and regulation within the health and care system. The findings from this review are expected to shape the government’s forthcoming 10-year plan for the NHS. The final report is anticipated to be published in early 2025 to support a streamlined, patient-centred model for patient safety. The investigation has not explored this further given the ongoing work, and to avoid further duplication.

Summary

3.3.3 This section highlights the oversight and accountability mechanisms for mental health services and the challenges in implementing meaningful changes following investigations and inquiries. Trust boards are responsible for ensuring compliance with the Learning from Deaths framework, but variability in engagement and reporting persists. The Health and Care Act 2022 formalised integrated care systems to improve collaboration, though relationships between health and social care remain strained.

3.3.4 There is limited follow-up on recommendations from inquests and patient safety investigations, with a noted lack of national mechanisms to ensure implementation. The investigation has highlighted concerns around data visibility management. The investigation supports previous suggestions for a dedicated body to oversee and enforce the implementation of recommendations, emphasising the importance of systemic learning and sustained improvement in patient safety.

HSSIB makes the following safety recommendation

Mental health inpatient provider perspective on mortality data

4.1.1 The investigation was told by providers that it is not possible to compare data on deaths, and that defining deaths is variable. A subject matter advisor told the investigation that there are different causes of death in mental health services which need to be thought about differently. Examples of complexities include the difficulty of classifying deaths in a timely way, as many deaths are initially treated as suicide even though coroners may later find differently.

4.1.2 A subject matter advisor told the investigation that deaths by suicide need to be separated out from other deaths as the learning pathway is different. Importantly, they stated that the “set of causes” in a death by suicide cannot be determined and is always 100% hypothetical, so to link a death to care provision could be profoundly biased and lacking a broader curiosity (Gibbons, 2023; 2024a). This is because a coroner’s judgement is based on the evidence presented; crucially, they cannot ask the deceased why they died. The set of causes of other deaths may be clear or certainly known to a better degree. However there was no consensus on this amongst other stakeholders.

4.1.3 The investigation was told by providers that there are “no agreed definitions for deaths in mental health services … people do not know what they are measuring, and data collected is inconsistent and cannot be compared in a helpful way”. The investigation was told that more work is needed to understand reporting of data on deaths in mental health settings and that organisations cannot be expected to “lift and shift” a physical healthcare data mapping model to record deaths and identify learning.

4.1.4 There was variation in the timeframe within which organisations classify a death within their care. For example, some included deaths that happened up to 6 months after a patient was discharged, while others had a different timeframe. One organisation told the investigation that “it is difficult to benchmark data as there is inconsistency and variability in reporting nationally”. Another organisation said that “there is no national measure, and we can’t compare as there are no standards … we know we need a national benchmark of mortality … it will help us measure like for like”.

4.1.5 Bereaved families campaigning for better oversight of data on deaths shared their thoughts and reflected: “It is different to just numbers, and more a requirement about reassurance on learning … Publish data that is understandable and comparable to everyone”.

4.1.6 Most NHS mental health trusts follow the Mazars Framework (Mazar, 2015) which was written to help trusts to develop a case selection process for Structured Judgement Reviews. Examples of the different categories of death include:

• Unexpected Natural, for example cardiac arrest, stroke, diabetes
• Expected Natural, for example a person receiving end of life care
• Unexpected Unnatural deaths that potentially meet the PSIRF priorities, for example all unexpected inpatient deaths, which is nationally mandated
• Awaiting cause of death/unable to obtain cause of death.

4.1.7 The investigation was told by a healthcare professional that ‘the Mazars categories are hard to apply in practice’. This was reported with the independent review into data on mental health inpatient care which stated, “the use of natural and unnatural in relation to deaths can be unhelpful in that they relate to the way someone has died rather than the cause of death” Department of Health and Social Care (2024d). All deaths can be multifactorial with many contributory factors and other terminology may be more helpful.

4.1.8 The investigation reviewed reports by individuals, campaign groups and some commissioned for a specific purpose. Many reports referred to poor data collection and an inability to understand data on deaths and how it can then be used to influence learning (Aldridge et al, 2023; Care Quality Commission, 2024c; NHS England, 2024c; Tees, Esk and Wear Valleys NHS Foundation Trust, 2023; Thornton, 2023).

Physical health and deaths of people with severe mental illness

4.1.9 People who have a severe mental illness (SMI) have a greater risk of poor physical health and also have a greater risk of dying prematurely compared with the general population (Hert et al, 2011; Public Health England, 2018; Reilly et al, 2015). NHS data reveals that 130,400 adults with SMI in England died prematurely (before the age of 75) between January 2020 and December 2022 (Office for Health Improvement and Disparities, 2023) (see figure 2).

Figure 2 Deaths of adults under the age of 75 with severe mental illness in England

4.1.10 The CQC (2024c) reported that the data on deaths of people detained under the MHA, and for patients on a community treatment order, found the prominent cause of natural death was pneumonia and reported that those with an SMI are at higher risk of dying of respiratory disease. Physical health checks help identify early signs of physical illness in people with SMI. Only 59% of those registered with an SMI received the stipulated full set of six physical health checks in the year to June 2024, leaving 41% at risk of missed early diagnosis and treatment.

4.1.11 In 2021 the All-Party Parliamentary Group on Mental Health highlighted that to deliver on the 5-year forward view commitment to prevent poor physical health outcomes for people with SMI, a national measure for reducing premature mortality with targets to hold services to account was required (Health and Social Care Committee, 2021b). It was recommended that NHS England and Public Health England should ensure mortality data is published as part of their sustainability and transformation plans and local authority levels, and that local plans set out how to meet reduction targets, including rolling out social prescribing in every primary care centre. The investigation was told by ICBs and staff in NHS regional teams that there are challenges associated with having the required data, analytical skills, resources and capacity to achieve this.

4.1.12 Deterioration of patients’ physical health was reported in the HSSIB investigation report ‘Mental health inpatient settings: creating conditions for the delivery of safe and therapeutic care to adults’ (Health Services Safety Investigation Body, 2024a), with a corresponding safety observation and safety recommendation. The investigation saw examples of a failure to recognise deterioration in patients on wards and a lack of understanding of how to respond to emergencies requiring basic and advanced life support. Further information on physical health will feature in a future overarching report on the theme of mental health care.

The national lens on mental health inpatient mortality data

4.1.13 There remains no single, complete and coherent set of data on the number of deaths in mental health settings in England. This has been the subject of previous reports (Independent Advisory Panel on Deaths in Custody, 2024; INQUEST, n.d.b). It is reported that people in state custody are at a significantly elevated risk of death, both from natural and unnatural causes, compared with the general population (Independent Advisory Panel on Deaths in Custody, 2024).

4.1.14 The Mazars (2015) report into mental health and learning disabilities deaths in Southern Health NHS Foundation Trust looked at the deaths of people with a learning disability or mental health problem, including investigations. The Mazars report made 23 recommendations to the Trust, 9 for commissioners, and 7 for national bodies.

4.1.15 The Care Quality Commission (CQC) reports on deaths as part of its monitoring of the Mental Health Act (Care Quality Commission, 2024c). Its latest report, which covers 2022/23, states that it was notified of 318 mental health related deaths (264 detained patients and 54 patients on a community treatment order (CTO)). It should be noted that the reporting of CTO deaths is not compulsory, and for this reason, figures may be underestimated. Of the 318 deaths reported, CQC highlighted:

• 189 were from natural causes (that is, a result of old age or a disease, which can be expected or unexpected)
• 63 were due to unnatural causes (which encompasses death as a result of an intentional cause, that is, harm to self or by another individual, or unintentional cause (an accident)) for example hanging, self-strangulation or suffocation
• 66 deaths were currently still undetermined (the cause of death had not yet been determined by a coroner or CQC did not hold information on cause of death).

4.1.16 In January 2023, the Department of Health and Social Care (2024d) commissioned a rapid review into data on mental health inpatient settings. This was due to concerns that the data and information needed to provide early alerts to identify risks to patient safety in mental health inpatient settings and prevent safety incidents was not available. As part of its data mapping work, the rapid review scoped out the range of organisations that collect and use data on service user deaths in mental health inpatient services and included over eight organisations.

4.1.17 The review report was published on 28 June 2023 and the government’s response was published in March 2024 (Department of Health and Social Care, 2024b; 2024d). The Department of Health and Social Care convened a ministerial led steering group to oversee the implementation of the recommendations, which has met once.

4.1.18 Considering the rapid review, and to prevent duplication of existing work, the investigation did not further examine the mechanisms that capture data on deaths (and near misses) across the mental health provider landscape.

4.1.19 NHS England has a commitment to establish an ‘early warning signs framework’ for all NHS-commissioned mental health, learning disability and autism inpatient settings (in response to recommendation 1 of the rapid review). This is to support providers, commissioners and national bodies in ‘measuring what matters’ for mental health inpatient services, and to ensure they can access the information they need to provide safe, therapeutic care. This work aligns with the wider National Quality Board work on Quality Early Warning Signs (QEWS).

4.1.20 To achieve this, NHS England told the investigation that during November and December 2023 it consulted with over 150 service users, staff, stakeholders and system leaders to co-produce a suite of ‘metrics that matter’ which providers and integrated care boards (ICBs) should prioritise in their regular quality reviews. Nine Early Warning Signs have been proposed, encompassing:

• the experience of patients, families and carers
• experience of staff and the effectiveness of organisations
• inherent risk factors.

4.1.21 These Early Warning Signs are supported by a further 19 secondary measures that help providers, commissioners and quality governance bodies to drill into issues when they emerge. The investigation has been advised that, once agreed, NHS England will:

• embed the full suite of measures in National Quality Board guidance (quarter 1, 2024/25)
• embed Early Warning Signs in the 2024 refresh of the National Oversight Framework (where possible) (quarter 1, 2024/25)
• review and update/align all system-facing dashboards and tools (quarter 1 to quarter 2, 2024/25).

4.1.22 NHS England told the investigation that nationally it is reconfiguring its system-facing dashboards to align to the Early Warning Sign metrics where possible. Some of the data cannot be aligned due to it being qualitative (non-numeric) rather than quantitative (numeric).

4.1.23 Recommendation 4 of the rapid review report set out that “DHSC, in partnership with NHS England and CQC and supported by key experts from across governmental and non-governmental organisations, should convene all the relevant organisations who collect and analyse mortality data to determine what further action is needed to improve the timeliness, quality and availability of data on deaths”. The Department of Health and Social Care accepted this recommendation and, in response, established a Mortality Data Working Group (MDWG) that included organisations that gather, analyse and distribute data on deaths in inpatient mental health services, which met for the first time in May 2024. The MDWG identified initial actions to address gaps and drive improvements in the data on deaths. The investigation was told by the Department of Health and Social Care in December 2024, that “Ministers have agreed that the work of the MDWG should continue under the new government”.

4.1.24 Recommendation 13 of the rapid review stated that:

‘Except where specified, these recommendations should be implemented by all parties within 12 months of the publication of this report. Government ministers, through the Department of Health and Social Care (DHSC), should review progress against these recommendations after 12 months.’ (Department of Health and Social Care, 2024d)

4.1.25 The investigation requested an update on the status and next steps of the rapid review. As of December 2024, the Department of Health and Social Care told the investigation that new ministers have confirmed that the mortality data working group established in relation to recommendation 4 will be reinstated.

4.1.26 The investigation considers that the recommendations of the rapid review, while beneficial and appropriate, are substantial and will require considerable investment and resources at local, regional and national level to achieve meaningful and achievable service improvement and planning.

Summary

4.1.27 The investigation highlights significant challenges in collecting and using mortality data in mental health inpatient settings. Definitions and classifications of deaths vary, complicating comparisons and hindering learning. Deaths by suicide and from catastrophic self-harm in mental health require different investigative approaches from other deaths, yet there are no consistent national benchmarks for mortality data. Physical health issues significantly contribute to premature deaths among those with severe mental illness, with systemic gaps in preventive measures such as physical health checks.

4.1.28 The rapid review by the Department of Health and Social Care identified the need for improved data systems, including the establishment of an early warning signs framework. However, implementation faces delays, highlighting the complexity and resource demands of enhancing data-driven safety measures in mental health care. The investigation has found that there is duplication of effort in some of the diagnostic and improvement work being undertaken in relation to mortality data and learning from deaths. There is potential for increased co-ordination and collaboration to mitigate duplication of effort and increase the impact of the different streams of improvement work.

HSSIB makes the following recommendation

The investigation’s findings were drawn from analysis of available intelligence (serious incident investigation reports, coroners’ prevention of future deaths reports, research and policy literature) and through activities undertaken by HSSIB (observational visits, patient and staff interviews, wider stakeholder interviews and focus groups).

Stakeholder engagement

This is one of a series of HSSIB investigations into patient safety in mental health inpatient settings. This meant it was able to draw on evidence from across the four separate investigations in the series. Specific stakeholders engaged with primarily for this investigation are shown in table A and listed below.

Table A Patients and families, providers and regional stakeholders engaged with primarily for this investigation

The investigation directly engaged with the following national stakeholders and academics as part of the investigation:

• Department of Health and Social Care – various teams
• Professional Standards Authority for Health and Social Care
• NHS England – various teams
• Ministry of Justice
• National Confidential Inquiry into Suicide and Safety in Mental Health
• royal colleges and professional bodies
• service regulators – Care Quality Commission
• charities – VoiceAbility, INQUEST, Action against Medical Accidents, Voicing Loss
• independent sector – Independent Healthcare Provider Network and two large independent sector providers.

Further stakeholders were also engaged with during the consultation phase for this report.

The findings presented in this report were identified following triangulation of various evidence sources and following consultation with stakeholders involved in the investigation. The AcciMap model (Svedung and Rasmussen, 2002) was used to analyse the information gathered and support the direction of the investigation. The analysis focuses on identifying relationships between the different levels of the system, which include government policy and budgeting; regulatory bodies and associations; local area management; physical processes and actor activities (what staff, people, organisations, systems did); and equipment and surroundings. The contributory factors are arranged into a series of levels representing the different parts of the health and care system. The analysis focuses on identifying connections between the different levels of the system (see figure A).

Figure A Representation of the AcciMap method