When the Procedure Is Followed but Justice Is Missed
Why decision tools cannot replace context, professional judgement and the right to be heard
The following is a constructed educational scenario. It is not based on a real person, organisation or incident.
Consider a distribution centre where employees use handheld scanners to track parcels moving through the building.
One afternoon, a worker bypasses the scanner twice and manually enters two parcel numbers. One parcel is subsequently sent to the wrong depot.
During the review, the questions appear straightforward:
Did the worker know the correct procedure? Yes.
Was the departure from the procedure deliberate? Yes.
Had the worker previously been reminded to use the scanner? Yes.
The emerging conclusion is that the worker knowingly disregarded an established process.
On paper, the reasoning appears consistent and evidence-based.
But then more information becomes available.
The scanners had repeatedly frozen during that shift. Parcels were accumulating on the conveyor. Employees were being measured against hourly throughput targets. Several workers had previously used manual entry when the system failed, and supervisors knew that this informal workaround was occurring.
The worker still departed from the written procedure. That fact has not changed.
What has changed is the meaning attached to it.
Was this reckless non-compliance?
Was it a normalised workaround?
Was it an attempt to meet competing organisational expectations?
Or was it some combination of individual choice and system pressure?
That is where the real work of justice begins.
The problem with an apparently objective process
In their 2013 paper, Sidney Dekker and James Nyce questioned whether organisational justice can genuinely be achieved through an algorithmic assessment of intention, choice and repeated behaviour.
Their argument was not that evidence should be abandoned, nor that individuals should never be held accountable.
Their concern was that evidence does not select or interpret itself.
People decide:
what information should be collected;
which contextual factors are relevant;
whose explanation should be believed;
whether conduct is described as adaptation or disobedience;
and where responsibility should ultimately rest.
Those decisions take place within organisational hierarchies. A structured tool may therefore make a conclusion appear neutral even though human judgement and institutional power remain embedded throughout the process.
The decision tree does not remove judgement.
It can merely make the judgement less visible.
Accountability is not the same as blame
A fair culture does not require organisations to overlook deliberate harm, wilful neglect, dishonesty or genuinely reckless behaviour.
Accountability remains necessary.
The problem arises when an organisation begins with the individual and works backwards, rather than first understanding how the work was actually performed.
A procedure may say one thing while the operational system rewards another.
An employee may be told to follow every step while simultaneously being criticised for working too slowly.
A workaround may be formally prohibited but informally tolerated because it keeps the service functioning.
When something eventually goes wrong, the written rule is rediscovered and applied to the last person in the chain.
The question is no longer simply:
Did the worker depart from the procedure?
It becomes:
What conditions made that departure possible, normal or apparently necessary at the time?
Substantive, procedural and restorative justice
Dekker and Hugh Breakey later argued that a credible just culture should consider more than whether punishment is justified.
It should also examine:
substantive justice — whether the outcome is genuinely fair;
procedural justice — whether the method of reaching the outcome is fair;
restorative justice — what affected people need and what relationships, trust or systems require repair.
This matters because an organisation can follow a procedure accurately while still relying on incomplete evidence, unequal credibility or an overly narrow understanding of responsibility.
A process is not fair merely because every box has been completed.
It must also be possible to explain:
why particular evidence was selected;
how conflicting accounts were assessed;
how system conditions were considered;
whether comparable people would have been treated similarly;
and whether the affected worker had a meaningful opportunity to respond.
How NHS thinking has developed
Although the constructed scenario above is outside healthcare, the same governance principle is now visible in NHS patient-safety policy.
NHS England’s current Being Fair Tool states that patient-safety incidents are usually signs of wider systemic issues and that action singling out one individual is rarely appropriate. The tool should not be used routinely. It should be considered only where a systems-based learning response has already begun and raises a genuine concern about an individual’s conduct or fitness to practise.
The tool also asks a form of substitution question:
Would another person with comparable experience and qualifications have acted similarly in the same circumstances?
a real That is an important shift.
It asks decision-makers to examine the environment before treating an individual’s conduct as exceptional.
The Patient Safety Incident Response Framework similarly promotes systems-based learning rather than simplistic, linear searches for a single cause. It also warns organisations not to combine a patient-safety learning response with a separate process designed to determine blame, culpability or professional fitness. Different processes may all be necessary, but their purposes should remain distinct.
Five questions for fair decision-making
Before an organisation classifies a worker’s conduct, leaders should ask:
1. What evidence has not yet been considered?
Written procedures may show what should happen. They do not necessarily show how the work is routinely performed.
Workload, staffing, equipment, training, supervision, conflicting targets and accepted workarounds may all affect the meaning of an individual action.
2. Who controls the description of the event?
Terms such as resistant, careless, reckless, difficult or non-compliant are conclusions, not neutral observations.
A defensible process should separate what was directly observed from the interpretation later attached to it.
3. Would another worker have acted similarly?
Where comparable colleagues would probably have made the same decision, the strongest explanation may sit within the system rather than within one person’s character.
4. Was the person genuinely heard?
Being invited to a meeting is not the same as being meaningfully heard.
The person should understand the concern, see the material being relied upon, explain relevant context and challenge factual inaccuracies before conclusions become final.
5. What outcome would improve the system?
Discipline may sometimes be justified.
But other responses may be more effective:
redesigning the process;
repairing faulty equipment;
clarifying conflicting expectations;
improving supervision;
providing training;
supporting those affected;
or addressing an unsafe organisational norm.
The necessary limitation
Systems thinking must not become a method of avoiding every form of individual responsibility.
Some actions may still warrant remediation, restriction of practice, disciplinary action or external referral.
The point is not that the individual never matters.
The point is that the organisation must not decide that the individual is the problem before it has understood the system in which the decision was made.
A 2026 scoping review found that just-culture and restorative-just-culture approaches are increasingly recognised in healthcare, while also noting that evidence about interventions, implementation and outcomes remains limited. This is another reason to avoid treating any single framework as an automatic producer of justice.
A process should structure judgement—not hide it
Decision tools can be useful. They can prompt consistency, identify missing questions and discourage impulsive punishment.
But they cannot carry moral responsibility.
The people using the tool remain responsible for:
the evidence they selected;
the context they excluded;
the credibility they assigned;
the conclusion they reached;
and the consequences imposed.
The final governance question is therefore not simply:
Was the procedure followed?
It is:
Did the procedure help the organisation understand what happened—or did it merely give institutional authority to the first explanation available?
A fair process should do more than classify the worker.
It should examine the conditions that shaped the work.
Educational boundary
This article provides general educational commentary on organisational justice and safety governance. The scenario is entirely constructed and is not based on any real worker, employer or incident. The article does not provide legal advice or determine the appropriate outcome in any individual employment or professional matter.
Sources
Dekker SWA, Nyce JM. Just culture: “Evidence”, power and algorithms.Journal of Hospital Administration. 2013;2(3):73–78. DOI: 10.5430/jha.v2n3p73.
Dekker SWA, Breakey H. ‘Just culture’: Improving safety by achieving substantive, procedural and restorative justice.Safety Science. 2016;85:187–193. DOI: 10.1016/j.ssci.2016.01.018.
NHS England. Being Fair Tool: Supporting staff following a patient safety incident. Published 9 May 2025. Official guidance.
NHS England. Patient Safety Incident Response Framework. Published 16 August 2022; updated 23 September 2025. Official framework.
Brown K, Danby G, D’Souza N, et al. Just culture and restorative just culture in healthcare settings: a scoping review of interventions, activities, factors and outcomes.BMC Health Services Research. 2026;26:285. DOI: 10.1186/s12913-026-14095-z.
Source review date: 3 August 2026.
Subscribe to The Workplace Compass for evidence-led explanations of workplace systems, governance and organisational justice.a real a real
