top of page
Gradient Bar.jpg
Homepage

Compassion in the NHS: Why Listening Is a Patient Safety Intervention

Writer: Steve Franklin
Steve Franklin
Oct 12
14 min read

What Poppy Russell’s story asks of compassion, candour and learning in the NHS



A baby, a family and a question for the NHS

Poppy Russell lived for just 12 hours. She died at the Princess Royal Hospital in Telford in April 2021 after suffering a hypoxic brain injury. In 2023, the coroner concluded that failures in fetal heart-rate monitoring led to her preventable death and that, with effective monitoring, signs of distress should have been identified for more than an hour (Old Square Chambers, 2023).


The clinical failure was devastating. What followed compounded the harm. Poppy’s mother, Katie, was blamed on the basis that she had refused fetal monitoring. The inquest found that claim was untrue. Shrewsbury and Telford Hospital NHS Trust later acknowledged failings in Poppy’s care and apologised (Patient Safety Learning, 2026).


Neil and Katie Russell could reasonably have walked away. Instead, they have worked with the Trust to develop Poppy’s Promise, founded on compassion, acknowledgement, respect and empowerment. That response should make us pause. A family harmed by care has offered the NHS a route towards better care. The first responsibility of the system is to listen.


Poppy’s story is not simply about being kind. It is about whether compassion is treated as a safety control, whether families are recognised as holders of vital information, and whether organisations can remain open when the information is painful, inconvenient or reputationally difficult.


Compassion is not the soft part of patient safety

Compassionate leadership is often described through four connected actions: attending, understanding, empathising and helping. The sequence matters. We cannot understand without paying attention. Empathy without action may feel sincere, but it does not remove a barrier, correct an unsafe condition or repair harm.


In patient safety, compassion improves the quality of information available to the organisation. A patient may notice a change that is not visible in a set of observations. A family may recognise that a person is “not themselves”. A member of staff may detect a weak signal before it becomes a measurable trend. Listening allows that intelligence into the system.


The familiar SOAP structure makes the point in miniature. It begins with the Subjective: what the patient feels, notices and experiences, expressed in their own account. Only then does it move to objective findings, assessment and plan. That ordering is not accidental. Patient experience is not an optional narrative added after the clinical facts; it is often the first source of information from which observation, interpretation and action should develop. If we fail to establish the subjective account, we risk building a technically ordered response on an incomplete understanding.


This is why compassion and value are connected. Value is not created only through activity, productivity or cost reduction. It is created when the organisation uses the knowledge available to it to make better decisions, reduce avoidable harm and design care around what matters to people. Ignored information has no organisational value, however important it might have been.


PSIRF makes compassionate engagement part of the method

The Patient Safety Incident Response Framework places compassionate engagement and involvement alongside system-based learning, proportionate responses and supportive oversight. NHS England’s guidance expects organisations to work to answer the questions of patients, families and staff, and to involve those affected in patient safety investigations (NHS England, 2022a, 2022b).


That changes the framing. Compassionate engagement is not a meeting added to the end of an investigation. It begins before first contact, continues throughout the response and includes decisions about communication, accessibility, emotional support, involvement and feedback. It is both relational and practical.


The quality of engagement also affects the quality of the learning response. Families may hold information that is absent from the clinical record. Staff may explain how work was actually done under pressure. Patients may identify where a pathway felt fragmented, dismissive or unsafe. A response undertaken without those perspectives may be technically complete but informationally weak.


Involvement should be appropriate and proportionate to the learning opportunity, but proportionality must not become a reason to exclude. There will be occasions when patients, families and carers can and should contribute to a rapid learning response, including a SWARM or other facilitated team review. Their involvement may be direct, through questions or attendance where that is safe and useful, or indirect, through an engagement lead who brings their account and questions into the discussion. The method should flex around the people affected rather than requiring them to fit a process designed only for professionals.


Compassionate engagement must also shape the terms of reference, not merely the communications plan. The questions patients, families, carers and staff need answered should influence the scope, evidence sought, people involved, language used and way findings are shared. Agreeing that scope openly helps prevent a learning response from becoming technically neat but disconnected from the harm experienced. It also makes clear that the purpose is learning and improvement, while maintaining fair accountability, rather than constructing the narrowest defensible account.


Why a family liaison role matters

A well-designed family liaison or compassionate engagement role provides continuity at a time when the organisation can otherwise feel fragmented. The role can act as the principal point of contact, agree how and when communication will happen, explain unfamiliar processes, ensure questions reach the learning team, and help the family understand what is known, what remains uncertain and what will happen next.


The role must not become a protective layer between the family and the truth. Nor should it be confused with counselling, advocacy or investigation. Its value lies in skilled, reliable liaison: maintaining the human relationship while enabling meaningful involvement and accurate information exchange.


Healthcare has something to learn from policing. Police family liaison officers are specially trained investigators who provide the primary link between an investigation team and a bereaved family in homicide and fatal-collision investigations. They provide updates, gather information, understand communication needs and signpost specialist support. Their deployment is managed within a clear investigative strategy (Foreign, Commonwealth & Development Office, 2026; West Yorkshire Combined Authority, 2026).


The systems are different and the model should not be copied uncritically. However, the principle transfers well: after serious harm, continuity, role clarity and skilled communication should be designed into the response rather than left to chance.


Candour is a practice, not a letter

The statutory organisational duty of candour has applied to NHS trusts and foundation trusts in England since November 2014. It was extended to other Care Quality Commission-registered health and social care providers in April 2015. Regulation 20 requires openness and transparency and, for notifiable safety incidents, includes informing the person affected, providing reasonable support, giving truthful information and offering an apology (Care Quality Commission, 2025a; Department of Health and Social Care, 2024).


An apology is important, but candour cannot be reduced to a template letter or a completed field. It is the quality and continuity of honesty after harm. Are difficult facts volunteered, or released only after repeated challenge? Are uncertainties explained? Are records corrected? Are families kept informed when timescales change? Does the organisation remain open when its first account proves wrong?

The CQC is explicit that saying sorry is not an admission of liability. Yet organisations can still behave defensively, treating openness as a legal risk rather than a foundation for trust (Care Quality Commission, 2025a). That defensive reflex can create a second harm: after the original event, the family must fight the organisation for recognition, information and truth.


Robbie Powell and the long road to a legal duty

The case of 10-year-old Robbie Powell exposed the absence of a legal duty requiring healthcare providers to share information with people harmed by care or their families. Robbie died in 1990 from untreated Addison’s disease after earlier suspicion of the condition had not been communicated to his family. His parents’ long pursuit of answers became part of the background to the statutory duty of candour (Care Quality Commission, 2025b).


Robbie’s case matters because it shows that the harm caused by clinical failure can be deepened by failures of transparency and accountability. A family should not require unusual persistence, legal knowledge or decades of campaigning to obtain an honest account. If truth depends on a family’s stamina or resources, candour is not functioning as an organisational duty.


The statutory duty was an important change, but the continuing review of its operation underlines a wider truth: regulation can create a requirement, but only culture, leadership and governance can make openness routine (Department of Health and Social Care, 2024).


Different inquiries, the same signal

The NHS does not lack reports. It lacks consistent action on what they repeatedly tell us.


The Francis Inquiry concluded that the Mid Staffordshire board did not listen sufficiently to patients or staff and did not ensure that deficiencies brought to its attention were corrected (Francis, 2013). The Berwick review urged the NHS to engage, empower and hear patients and carers at all times, embrace transparency and support staff to improve the systems in which they work (National Advisory Group on the Safety of Patients in England, 2013).


The Ockenden review of maternity services at Shrewsbury and Telford identified failures to listen to families, failures to learn from incidents and a culture in which mothers were sometimes blamed for poor outcomes. It said boards must listen to and hear both local families and their own staff (Ockenden, 2022). Kirkup’s East Kent investigation called for compassionate care, better teamwork and an honest response to challenge (Kirkup, 2022).


More recent reports have not made the signal weaker. The 2026 Ockenden review at Nottingham drew on more than 2,500 family cases and engagement with families and staff. The Amos investigation again placed listening to women, families and staff at the centre of national maternity improvement. The Thirlwall Inquiry found profound failures of management, governance and safeguarding, including a gulf between clinical concerns and hospital leadership and parents being kept in the dark for years (Ockenden Maternity Review, 2026; The King’s Fund, 2026; Thirlwall Inquiry, 2026).


These are different services, organisations and circumstances. They should not be collapsed into one story. Yet the recurring pattern is difficult to ignore: warning signs existed; people raised concerns; information did not travel, did not retain its meaning, or did not produce timely action.


Situational awareness: information must be perceived before it can be used

Endsley’s model of situational awareness provides a useful frame for information that is ignored, diluted or not shared. Level 1 is perception: noticing the signal and ensuring it reaches attention. Level 2 is comprehension: integrating the information and understanding what it means in context. Only then is Level 3, projection, possible: anticipating what may happen next—the “so what?” of the information (Endsley, 1995). If an organisation does not perceive a concern, or receives it without comprehending its meaning, it cannot reliably predict risk or choose a timely response. Information held somewhere in the system is not the same as organisational awareness.


Staff voice is patient safety intelligence

Listening to patients and families cannot be separated from listening to staff. The workforce often sees the system from its sharpest edge. Staff know where workarounds have become normal, where demand repeatedly exceeds capacity, where equipment or digital systems create risk, and where policy does not reflect work as done.


The 2025 NHS Staff Survey illustrates the gap between encouragement and confidence. Although 86.16% of respondents said their organisation encourages reporting of errors, near misses and incidents, 67.30% said action is taken to prevent recurrence and 61.02% said they receive feedback about changes. Only 55.49% were confident their organisation would address a concern about unsafe clinical practice. For broader concerns, 60.29% felt safe to speak up and 47.59% were confident their organisation would address the concern (NHS Staff Survey, 2026).


Those figures do not tell us that staff are unwilling to speak. They suggest that organisations must pay equal attention to what happens after somebody speaks. A speaking-up route without response can become a collection mechanism for organisational disappointment. People learn from experience whether raising a concern leads to understanding and action, or to silence, delay and personal exposure.


Listening must also be equitable

MBRRACE-UK reports continue to document serious inequalities in maternal outcomes, while wider research into confidence in maternity services has described how ethnic minority women’s trust can be affected by whether relationships, communication and care feel safe and responsive. Listening is therefore inseparable from equity. A system cannot claim to listen well if some voices are more readily believed than others or if language, bias, disability, deprivation or power determine whose concern is acted upon (Office for Equality and Opportunity, 2025; MBRRACE-UK, 2026).


This makes compassionate engagement more than an interpersonal skill. Organisations need interpreting, accessible information, culturally safe practice, reasonable adjustments, representative data and routes for challenge. Respect must be designed into the system.


People are not outside the system

Human factors in healthcare gives us a useful corrective to the habit of treating people as the unreliable part of an otherwise sound system. In SEIPS, outcomes emerge from interactions between people, tasks, tools and technology, the physical environment, organisational conditions and the external environment. NHS England’s SEIPS guide deliberately places people at the centre and states that design should support, rather than replace or compensate for, people (NHS England, 2022c).


The SHELL model, developed in aviation, makes a similar point through the interfaces between liveware, software, hardware and environment, including the liveware-to-liveware relationships between people. The value is not in forcing healthcare into an aviation diagram. It is in asking where the interfaces do not fit: the procedure that conflicts with reality, the alarm that does not support attention, the team relationship that inhibits challenge, or the environment that makes dignified communication difficult.


Patients, families and carers are not visitors to this system model. They do work too: noticing, remembering, coordinating, communicating, making decisions and adapting. SEIPS 2.0 explicitly broadened attention to the work done by patients, families and other non-professionals (Holden et al., 2013). Excluding them from analysis removes part of the system we are trying to understand.


Civility and compassion must work hand in glove

The Civility Saves Lives movement has helped make visible a patient-safety truth: how people treat one another affects attention, communication, teamwork and performance. Incivility can narrow what people are willing or able to say, discourage challenge and consume cognitive capacity that should be available for care. Civility is therefore not etiquette at the edge of safety; it is one of the conditions that allows safety-critical information to travel (Civility Saves Lives, n.d.; NHS England, n.d.).


But civility alone can become a demand for politeness that leaves difficult experience untouched. It must work hand in glove with compassion: respectful behaviour joined to attention, understanding and helpful action. A compassionate team can hold challenge without humiliation; a civil organisation can make it possible to speak; together they create the relational conditions in which weak signals, uncertainty and painful truths can be heard and used.


Psychological safety: speaking is only the first half

Psychological safety at work means people can ask questions, admit uncertainty, report mistakes and challenge decisions without interpersonal fear closing down the conversation. It does not mean comfort, agreement or the absence of accountability. It means that relevant information can surface while there is still time to use it.


For leaders, the test is not whether they say “my door is always open”. It is how they respond when somebody walks through it with information they would rather not hear. Defensive questions, premature reassurance, personalisation of challenge and delayed feedback all teach the system what is safe to say next time.

Compassionate leadership and psychological safety therefore meet at action. Attending and understanding create the conditions for voice. Helping means doing something proportionate with what has been heard, explaining the decision and returning to the person who raised the concern. Accountability remains. The standard of evidence remains. But curiosity comes before conclusion.


A listening organisation is a learning organisation

There is a phrase often used after harm: lessons will be learned. The real question is whether the organisation has changed the conditions that produced the outcome and whether it can show that the change is working.

Listening supplies the raw material for learning. Healthcare governance determines whether that material becomes a decision. Good governance is not the volume of papers reaching a committee. It is the quality of informed decision-making: bringing together patient experience, family insight, staff voice, safety data, operational reality, inequalities, professional evidence and uncertainty; testing what they mean; and deciding what to do.


This requires a conscious organisational choice: learning before protection. When the first instinct is to defend reputation, position or an earlier account, listening becomes restricted to information that feels safe. Questions are narrowed, language hardens and contradictory evidence is treated as threat. A learning organisation does not abandon due process or fair accountability; it creates enough openness to discover what it may have misunderstood. Unless leaders adopt that stance, the organisation will hear only within the limits of what it is prepared to admit.


This is also how listening drives value. It reduces the cost of avoidable repetition, identifies risk earlier, improves the fit between services and the people using them, and directs attention towards changes that matter. Crucially, it treats experience and expertise as assets rather than interruptions.


Questions for boards and leaders

A board cannot listen directly to every conversation, but it can design an organisation in which information travels without being stripped of uncertainty, emotion or meaning. Useful questions include:

  1. Whose voice is missing from this account, and what might they know that we do not?

  2. How do we know that patients, families and staff can raise concerns safely and accessibly?

  3. What happens after a concern is raised, and how quickly does the person receive feedback?

  4. Where have we heard the same issue before, and why has it not produced reliable change?

  5. Do family liaison and engagement roles have training, authority, supervision and continuity?

  6. Are candour and involvement evident in everyday practice, not simply in policy compliance?

  7. What inequalities affect whose account is believed, recorded or acted upon?

  8. How have we tested that safety actions changed work as done and did not create new risk?

  9. What uncomfortable information has influenced a board decision in the last six months?


Successful by design, not luck

The only real failure is the failure to learn from failure. That does not diminish the seriousness of harm or remove the need for accountability. It recognises that repeating a known pattern after people have already tried to warn us is a governance failure as much as a clinical or operational one.


Poppy Russell’s story asks something simple and difficult of the NHS: be human, listen carefully, tell the truth and act on what people tell you. The same message runs through Robbie Powell’s case, Francis, Berwick, Ockenden, Kirkup, Amos, Thirlwall, staff surveys and maternity reports. Different voices have been carrying the same signal for years.


Listening makes us informed. Informed organisations make better decisions. Better decisions create safer care, stronger trust and greater value. That is what good governance should achieve: success by design, not luck.



References



Comments

bottom of page