The existence of a complaints procedure is not evidence of accountability. What matters is whether an institution is capable of being changed by what people tell it.
Britain has an extraordinarily elaborate architecture of complaint. Hospitals, universities, local authorities and other public bodies provide formal routes through which dissatisfied patients, employees, students and service users can raise concerns. Complaints can be acknowledged, investigated, reviewed and escalated. Beyond individual organisations sit regulators and ombudsmen, including the Parliamentary and Health Service Ombudsman (PHSO). Viewed administratively, this is an impressive infrastructure. It gives the appearance of a society in which institutions expect to be scrutinised and citizens are provided with mechanisms through which failures can be identified and corrected.
My experience of these systems, however, has increasingly made me question the assumption on which this architecture rests. The existence of a complaints procedure does not demonstrate the existence of accountability. Nor does the fact that an organisation has responded to a complaint tell us whether it has learned anything from it. An institution can acknowledge correspondence, investigate itself, consult its employees, cite its policies, produce a detailed response and close a case without changing anything about the conditions that generated the complaint. Indeed, the procedural sophistication of a complaints system can obscure a more troubling reality: the machinery ostensibly established to identify institutional failure can become a mechanism for containing criticism of the institution itself.
This is what I call the complaint paradox. Institutions create complaints systems because they formally acknowledge that institutions and the people within them can fail. Yet when an individual actually uses that system to allege failure, the institutional response can become organised around demonstrating that no meaningful failure occurred. The complainant approaches the organisation believing that they are providing information from which it might learn. The organisation can receive the same information as an allegation against which it must defend itself. What was intended as feedback consequently becomes adversarial.
This distinction matters because the purpose of complaint should extend far beyond determining whether a complainant has “won” or whether an institution has been found formally at fault. A complaint is also a form of knowledge. It tells an institution something about itself from a position that those managing it may never occupy. A patient knows something about the experience of a ward that a hospital director cannot know from performance reports. A disabled person encounters dimensions of a service that an able-bodied policy writer may never experience. A Black patient or employee may perceive racialised interactions that are invisible within an organisation’s equality strategy. A bereaved family may recognise patterns of dismissal, communication failure or professional arrogance that are not visible in clinical statistics.
This does not mean that every allegation must automatically be accepted as fact. Complaints require fair investigation. Employees are entitled to fairness, accounts may conflict, memories may differ and some allegations will not be substantiated. But an institution genuinely committed to learning should be capable of distinguishing between the evidential question—can we establish exactly what happened?—and the organisational question—is there nevertheless something here from which we should learn? A complaint can fail to meet the threshold required for a formal finding against an individual and still reveal a weakness in communication, supervision, safeguarding, staffing, accessibility or organisational culture.
The inability to make this distinction produces what I describe as defensive accountability: the performance of accountability procedures without the corresponding vulnerability to institutional change.
When the institution becomes the object that must be protected
Institutional defensiveness is not unique to healthcare. It resembles a phenomenon visible in debates about national history. When the history of British slavery or colonialism is criticised, for example, some responses move almost immediately towards comparative defence: Britain was not as brutal as another empire; slavery existed elsewhere; another colonial power behaved worse; Britain subsequently participated in abolition. Historical criticism becomes interpreted as an attack upon national identity, and the purpose of argument shifts from understanding what happened to preserving a morally acceptable account of Britain.
There is an important epistemological problem here. Once identity becomes implicated in criticism, evidence is no longer encountered neutrally. It is filtered through the need to protect the thing with which one identifies.
Institutions can behave similarly.
The NHS occupies a particularly complicated position because it is not merely a healthcare system. It is also a cherished British institution, deeply embedded within narratives of national identity, solidarity and public provision. Criticism of the NHS can therefore be interpreted as hostility towards the NHS. Concerns about waiting times may be answered by comparisons with the United States. Poor experiences can be contextualised through staff shortages. Criticism of treatment may be met with reminders that healthcare remains free at the point of use. Accounts of inappropriate staff behaviour may be explained through exhaustion and extraordinary workload.
These contextual factors are real and important. Chronic workforce shortages, inadequate resources, high demand and occupational exhaustion have consequences for workers and patients alike. But explanation cannot be allowed to become exoneration. A patient who is humiliated, neglected or treated without dignity is not less harmed because the member of staff responsible was exhausted. Nor is inadequate care rendered acceptable because another country’s healthcare system may be worse.
The appropriate question for a public institution is not whether it can identify somewhere performing more poorly. It is whether it can become better than it currently is.
There is something profoundly dangerous about comparative institutional reassurance. At least we are better than X establishes mediocrity as an acceptable destination. Once criticism can always be answered by locating something worse, improvement becomes optional. The institution becomes protected precisely from the dissatisfaction that might have forced it to change.
To criticise the NHS is therefore not necessarily to oppose or devalue it. The opposite may be true. There is little value in protecting an institution from criticism while the conditions producing that criticism deteriorate. Institutions are not preserved by insisting upon their goodness. They are preserved by identifying weaknesses early enough to correct them.
Complaints are often made for people we will never meet
One of the assumptions embedded within defensive complaint handling is that the complainant is primarily seeking something for themselves. Sometimes they are. A person may require compensation, correction of a record, disciplinary action, treatment, an apology or another form of redress. There is nothing illegitimate about seeking remedy after harm.
Yet this understanding misses one of the most striking features of serious complaints. Again and again, complainants explain that they are pursuing the matter because they do not want the same thing to happen to somebody else.
This motivation deserves far more attention than it receives.
Complaining is labour. A serious institutional complaint can require somebody to reconstruct events, locate records, identify dates, remember conversations, write statements, respond to questions, challenge inaccurate accounts and repeat painful experiences through several stages of review. For somebody living with illness, disability, trauma or bereavement, the administrative labour can itself become exhausting. Most people have more desirable uses for their finite time. They could be resting, working, reading, watching football, spending time with their families or simply trying to rebuild an ordinary life.
The fact that somebody nevertheless spends weeks or months constructing a complaint should therefore generate institutional curiosity. What happened that made this individual decide that documenting the experience was worth so much of their time?
Often the answer is remarkably simple: I do not want this to happen to somebody else.
A parent whose baby has died cannot obtain their child back through a complaints procedure. A patient who has experienced degrading treatment cannot make the encounter unhappen. A person who has left an institution may derive little direct benefit from a policy change implemented years later. Yet people continue to complain because complaint can be prospective even when the harm itself is retrospective. They are asking an institution to use their experience to protect the next person.
Seen in this way, complaint is not simply dissatisfaction. It can constitute a form of civic labour: an individual expends their own time and emotional resources to provide information from which a public institution and its future users may benefit.
It is particularly troubling when such an act is encountered primarily as a threat to institutional reputation.
The temporal violence of administrative accountability
The question of time reveals another fundamental weakness within complaints systems.
The PHSO ordinarily operates a 12-month time limit, although it has discretion to consider complaints brought later where there are good reasons for the delay. The existence of discretion is important and should not be ignored. Yet the underlying structure still exposes a profound tension between administrative time and lived time.
Twelve months is straightforward to an administrative system. It can be calculated, recorded and applied.
Human life is not organised in this way.
Consider the year following a traumatic hospital admission, the death of a child, a serious diagnosis or a period of psychiatric illness. A person may leave hospital still unwell. Recovery may take months. Another illness may intervene. A family member may become sick. Employment may be lost. Housing circumstances may change. A person may have caring responsibilities. They may experience bereavement. They may spend a considerable period simply attempting to survive what has already happened.
The capacity to transform experience into a coherent administrative complaint cannot be assumed to appear immediately after harm.
Indeed, the seriousness of the event may itself explain the delay. The person is effectively placed within a paradox: the more destabilising the experience, the less able they may initially be to undertake the bureaucratic labour necessary to challenge it.
I describe this as temporal exclusion: the unequal distribution of access to accountability according to people’s capacity to translate harm into administrative action within institutionally preferred time.
This is not an argument that time limits can never exist. Evidence deteriorates. Staff move. memories fade. Records may become harder to reconstruct. Administrative bodies require workable procedures. But these practical considerations should not lead us to treat the expiry of an investigatory period as though it extinguishes the informational value of the complaint itself.
Suppose a former psychiatric patient reports, outside the usual complaint period, that they experienced degrading or aggressive treatment from staff on a particular ward. Perhaps too much time has elapsed to reach a fair disciplinary finding against an identifiable employee. That is one question.
But another remains.
Has anybody else reported similar treatment?
Do contemporary complaints indicate a pattern?
Does current supervision adequately protect vulnerable patients?
Should staff practice be reviewed?
Does this historical account provide information relevant to present safeguarding?
The inability to adjudicate an old allegation does not require an institution to become incapable of learning from it.
This distinction is essential wherever human safety is involved.
What maternity scandals teach us about the “isolated” complaint
The importance of early complaints becomes painfully apparent when institutional failure is examined retrospectively.
Major investigations into maternity services in England have repeatedly identified failures to listen to women and families, inadequate investigation of concerns, defensive organisational cultures and failures to learn. Once such cases are assembled retrospectively, the existence of a pattern appears obvious. Multiple deaths, injuries, complaints and missed opportunities can be placed beside one another and understood as manifestations of systemic failure.
But institutions do not encounter systemic failure in this form.
They encounter the first person.
Then perhaps a second.
The first complaint does not arrive with a label announcing that it is the beginning of a future public inquiry. It arrives as one mother’s concern, one family’s account, one patient disputing what professionals have told them, one person insisting that something is wrong.
This is the central epistemological challenge of institutional learning: a pattern is always composed of cases that once appeared individual.
If institutions require a visible pattern before taking individual testimony seriously, then the pattern can only become visible after enough people have been harmed.
This does not mean treating every complaint as evidence of impending catastrophe. It means preserving institutional curiosity even where an allegation cannot be conclusively established. A complaint that is not upheld should not become organisationally meaningless. It should remain available as intelligence that can be considered alongside future complaints, incidents, safeguarding concerns and workforce information.
Otherwise, the first complainants become useful only retrospectively. Years later, an inquiry discovers that they had been warning the institution all along.
Workforce pressure cannot become an alibi for poor care
The question of staff exhaustion requires similar nuance.
Healthcare workers can be extraordinarily overworked. Long shifts, inadequate staffing, overtime, bank work, secondary employment and insufficient recovery can affect judgement, patience, emotional regulation and interpersonal behaviour. These conditions deserve serious organisational attention.
But this should lead to investigation, not excuse-making.
If an employee’s conduct towards a vulnerable patient is being affected by exhaustion, the institutional question should not end with our staff are under pressure. It should begin there.
Why is this worker exhausted?
What hours have they worked?
Are rostering arrangements safe?
Are staff moving between substantive employment and additional bank shifts without adequate recovery?
Are economic pressures encouraging excessive hours?
Does the organisation know enough about fatigue risk?
Are staffing shortages creating environments in which otherwise competent professionals are becoming irritable, detached or unsafe?
And most importantly, what does this mean for the patient who is dependent upon them?
To invoke overwork simply to excuse inappropriate behaviour is to fail both worker and patient. The worker remains within potentially harmful conditions; the patient remains exposed to their consequences.
Protecting employees should not mean protecting every behaviour performed by an employee.
There is an important distinction between protecting a worker from unfair or malicious allegations and protecting an organisational culture from scrutiny.
A responsible employer can defend the procedural rights of an employee while simultaneously investigating whether working conditions, supervision, conduct or culture require intervention. These obligations are not mutually exclusive.
The fiction of the training certificate
Institutions also frequently invoke training as evidence that risk has been addressed.
The employee completed equality training. Safeguarding is mandatory. Staff have undertaken dignity training. An online module has been completed and a certificate generated.
But completion is not transformation.
Anyone who has undertaken mandatory workplace e-learning knows the weakness of treating completion statistics as proxies for behavioural change. People learn how to complete training. They click through slides, remember the answers required for the assessment, consult notes, repeat a quiz if necessary, obtain the required score and upload the certificate.
I have done versions of this myself.
The certificate proves that an administrative requirement was satisfied. It does not necessarily establish that a belief was reconsidered, that behaviour changed or that learning became embedded in professional practice.
This distinction is particularly important where work involves vulnerable human beings.
An organisation that responds to a complaint about dignity, racism, disability or safeguarding by stating that staff have received training has answered the wrong question.
The relevant question is not:
Was training delivered?
It is:
Did practice change?
This is the difference between compliance evidence and behavioural evidence.
Institutions are often rich in the former and surprisingly poor in the latter.
When the complainant becomes the person on trial
Defensive accountability can also produce what I call epistemic reversal.
The individual begins by asking the institution to account for what happened. Gradually, scrutiny moves towards the complainant.
Did you misunderstand?
Why did you not complain sooner?
Could you have interpreted the interaction incorrectly?
Was it really racism?
Are you certain those words were used?
Could your health have affected your recollection?
Individually, some of these questions may be legitimate components of an investigation. The problem arises when credibility testing becomes asymmetrical: institutional records and professional accounts are treated as presumptively reliable while the complainant’s account requires continual authentication.
This becomes especially troubling within mental-health services.
I know this problem not merely as a researcher but through experience. During my own inpatient health care, I observed how interactions could differ according to who was present. A relative’s presence could produce one kind of professional interaction; their absence could produce another.
This creates an obvious problem of evidence. Poor treatment does not necessarily occur while managers, relatives or potential witnesses are standing nearby. Yet subsequent investigation may privilege precisely those forms of evidence controlled by the institution: clinical notes, staff accounts and incident records.
What happens when the professional writes the record and the psychiatric patient disputes it?
What happens when an inappropriate interaction was never documented because the person whose behaviour is being questioned was one of the people responsible for documentation?
What happens when race, disability, gender and psychiatric diagnosis already affect assumptions about whose account appears rational, reliable and credible?
The answer cannot be to assume automatically that either patient or professional is truthful.
But neither can institutional records be treated as though they constitute a neutral God’s-eye account of events. Records are created by people. They contain decisions about what is worth recording, how an encounter is characterised and whose interpretation becomes institutional fact.
Complaints systems need sufficient epistemic humility to recognise this.
What would a learning institution look like?
A genuinely learning institution would still investigate complaints fairly and reject allegations where evidence requires it to do so. But its analysis would not stop at the binary categories of upheld and not upheld.
It would ask what complaints collectively reveal.
It would examine recurring themes across cases, including complaints that individually failed to meet formal evidential thresholds. It would preserve information from complaints that could not be investigated because of time limitations. It would examine whether particular wards, teams, services or groups repeatedly appear within concerns. It would distinguish training completion from behavioural outcomes. It would examine workforce fatigue rather than invoke it rhetorically. It would ask whether race, disability, gender or other inequalities shape how apparently similar service users experience the institution.
Most importantly, it would ask a question that is surprisingly absent from much complaint handling:
What changed because this person complained?
Not merely: was a response sent?
Not: was the case closed?
Not: did the institution successfully defend itself?
But: what did we know after this complaint that we did not know before it, and what did we do with that knowledge?
An organisation should be able to improve without treating improvement as an admission of legal liability. It should be possible to remind staff about dignity without first proving misconduct against a named employee. It should be possible to review racial inequalities without formally declaring an individual racist. It should be possible to strengthen safeguarding after an inconclusive complaint. It should be possible to learn from an old complaint even where disciplinary adjudication is no longer feasible.
Learning is not confession.
Improvement is not liability.
Institutional self-examination is not institutional betrayal.
Protecting institutions by refusing to protect them from criticism
I remain struck by how frequently institutional criticism is treated as though the critic wishes the institution to fail.
Often the opposite is true.
There is no meaningful sense in which the NHS can be protected by protecting it from knowledge of its own weaknesses. A healthcare system does not become safer because complaints are successfully rebutted. A hospital does not become better because its director concludes that no further correspondence is necessary. A ward does not become safer because an old complaint can no longer proceed through a particular administrative route.
Institutions can survive criticism.
What they may not survive indefinitely is the accumulation of failures they refused to confront.
We have seen this principle repeatedly, not only in healthcare but across governments, universities, corporations, churches, police forces and other powerful institutions. Organisational decline can coexist for remarkably long periods with confident institutional narratives about competence. Procedures continue. Reports are written. Policies exist. Training is completed. Leaders reassure themselves that systems are robust.
Then something breaks sufficiently publicly that the distance between institutional representation and lived reality can no longer be maintained.
At that point, the language changes.
There were missed opportunities.
Lessons were not learned.
Concerns were not escalated.
People were not listened to.
The culture had become defensive.
But cultures do not become defensive overnight. They are produced incrementally, every time an organisation learns that the safest response to criticism is rebuttal rather than curiosity.
Perhaps, then, the most important measure of institutional integrity is not the absence of complaint.
It is the capacity to receive complaint without experiencing it as an existential threat.
The NHS does not become weaker because a patient says that something went wrong. A university does not become weaker because an employee identifies racism. A public authority does not become weaker because a disabled person describes an inaccessible or degrading service.
Institutions become weaker when maintaining the belief that they are good becomes more important than discovering where they are failing.
The question at the conclusion of a complaint should therefore never be merely:
Was it upheld?
It should also be:
What did this person allow us to see?
Could it happen again?
What are we doing differently because we now know this?
If those questions are absent, we are left with something that resembles accountability administratively while failing to perform its most important function socially.
The citizen speaks. The institution responds. The correspondence is filed. The complaint is closed.
And somewhere, perhaps on the same ward, in the same department or within the same organisational culture, the conditions that produced the complaint remain intact.
Then somebody else encounters them.
And somebody else complains.
Eventually, when enough individual stories accumulate to become a scandal, we ask why nobody acted sooner.
Perhaps the answer, sometimes, is that somebody did.
They complained.
We simply built a system better equipped to answer them than to hear what they were saying.
