Incident learning and knowledge retention

Approx. 75–110 minutes

Independent SHAREhub process-safety awareness. Selected learning maps to a numbered process-safety syllabus topic; this is not the NEBOSH qualification, approved provision, specialist design instruction or work authorisation. GB guidance is distinguished from corresponding NI requirements; establish applicability for the actual jurisdiction and establishment.

This reader supplies the complete teaching. Use the interactive LMS course for assessed activities, saved progress and completion.

LESSON 1

Protect before investigating

Separate emergency response from evidence collection.

Safety team reviewing photographed evidence and an isolated work area
Existing SHAREhub photographic illustration. The written fictional case supplies the condition and decision evidence.

Protect before investigating

An investigation begins after the immediate response protects people and prevents further harm. Follow the site’s emergency arrangements, preserve restrictions and obtain competent support. Do not enter a suspected release area or disturb an isolation to collect a better photograph.

Evidence matters, but collecting it must not create another exposure. Note what is known safely, identify records that may be overwritten and arrange controlled preservation through the responsible roles. Investigators should understand which conditions changed during the response so later observations are not mistaken for the original event.

What should happen immediately after an incident?

The first priorities are life safety, emergency response, scene control, evidence preservation and escalation.

Investigation must never delay first aid, emergency response or making the area safe. Once people are protected, the scene should be preserved as far as reasonably possible. Equipment settings, isolation status, permits, alarms, photographs, closed-circuit television (CCTV), witness names and environmental conditions can disappear quickly.

The organisation should also decide whether the event needs Reporting of Injuries, Diseases and Dangerous Occurrences Regulations (RIDDOR) review, regulatory notification, insurer notification, specialist technical support, occupational health, human resources (HR), legal privilege consideration or family/victim liaison.

At the evidence-collection decision

Preserve necessary functional medical/support facts through authorised restricted arrangements. Do not copy unrestricted clinical records into the general investigation folder; decide lawful access, purpose and retention before collection.

Applied case

Fictional exercise: A leak has been reported and the area remains restricted; a supervisor wants photographs from beside the flange.

Response

Protect people and manage escalation first.

Preservation

Record what changed during response as well as what was found.

Test your decision

What takes priority?

  1. Follow the response restrictions and arrange safe evidence preservation.
  2. Enter briefly because photographs are important.
  3. Reopen the line to recreate the leak.
Compare the explanation

Follow the response restrictions and arrange safe evidence preservation.

Investigation cannot justify an uncontrolled exposure.

Open this interactive lesson →

LESSON 2

Judge potential as well as outcome

Explain why a no-injury event may merit substantial investigation.

Safety team reviewing photographed evidence and an isolated work area. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Avoid / redesign: Remove the need to open the line by redesigning sampling or maintenance access if technically valid. Control the task: Specify positive or otherwise justified proved isolation, with a safe vent/drain destination. Verify before release: Engineering validates process function; stop if the redesign leaves an untested hazard.

Judge potential as well as outcome

Actual harm is one consideration when choosing investigation depth. A small release stopped by chance may share a pathway with a much more serious event. Consider the credible potential, affected barriers, recurrence and uncertainty rather than using the injury count alone.

Do not invent a worst-case consequence without a basis. State the scenario and information needed to establish its significance. The responsible organisation selects a proportionate investigation with appropriate competence; the learner’s contribution is a defensible explanation of why the event should not be dismissed as harmless.

How should investigation levels be selected?

Investigation depth should match actual harm, potential harm, complexity and learning value.

A low-consequence event with clear facts may need a simple supervisor review. A high-potential near miss, major injury, dangerous occurrence, repeated event or process safety deviation needs deeper investigation, technical support and management oversight.

Potential severity matters. A dropped spanner that misses a person, a failed isolation, a gas detector alarm or a vehicle near miss may deserve serious investigation even where no injury occurred.

This page’s escalation scheme

Level 1/2/3 and the process-safety review category are an illustrative SHARE/company escalation scheme. HSG245 uses minimal, low, medium and high investigation levels. Resolve overlap by escalating for worst credible consequences, recurrence likelihood, complexity and specialist needs even where the actual injury was minor. The seven-stage workflow is an operational expansion of HSG245’s four steps.

Applied case

Fictional exercise: A wrong-route transfer is discovered before material flows. The route-selection control was bypassed.

Actual

No release occurred in this case.

Potential

The failed route control can still warrant examination.

Test your decision

Why investigate?

  1. No release means there is nothing to learn.
  2. It proves the most severe imagined accident was inevitable.
  3. The event reveals a potentially important control weakness despite no release.
Compare the explanation

The event reveals a potentially important control weakness despite no release.

Potential and control failure matter without asserting a certain outcome.

Open this interactive lesson →

LESSON 3

Build an honest timeline

Separate observations, records and hypotheses.

Safety team reviewing photographed evidence and an isolated work area. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Change at source: Remove the need to open the line by redesigning sampling or maintenance access if technically valid. Make it observable: Engineering validates process function; stop if the redesign leaves an untested hazard. Keep useful evidence: Retain the evidence log, causal map, revised documents, field verification and later effectiveness check.

Build an honest timeline

A timeline connects events without pretending every entry is equally certain. Distinguish a time-stamped record from recollection and an inference from an observation. Preserve conflicting accounts until they can be examined; agreement is not created by deleting an inconvenient entry.

Record the source and limits of each item. An alarm log may show a signal but not prove how it was understood. A witness may explain the work context without recalling the exact time. Combining these sources can reveal a sequence while uncertainty remains explicit.

What evidence should be gathered?

Evidence should show what happened, what controls existed and how the work was actually done.

Evidence is broader than witness statements. It can include photographs, plant status, alarms, permits, maintenance history, training records, risk assessments, inspections, CCTV, environmental conditions, drawings, samples, isolation logs, competence records, contractor documents and previous incidents.

A good evidence log records what was collected, who collected it, where it came from, when it was collected and any limitations.

Illustrative evidence-register row

Fictional record E-04: original control-system event log; source system and export location; collector and collection time; known clock offset; restricted read-only original; named working-copy location; access/custody history; limitation “alarm clock not yet reconciled with CCTV”. Preserve CCTV and control data before routine overwrite through lawful access and proportionate retention.

Collect only necessary functional medical facts for the investigation. Keep clinical/support records in the authorised restricted process and link the necessary finding rather than copying unrestricted health records into the investigation folder.

ICO: restricted handling of workers’ health information

Applied case

Fictional exercise: The alarm log records a warning; one witness recalls a phone call but cannot place it before or after the warning.

Recorded

A log establishes its own recorded event and timestamp.

Recalled

A witness account can be useful without exact timing.

Test your decision

How should the call appear?

  1. Nowhere, because recollection is never useful.
  2. As a reported event with timing unconfirmed.
  3. As definitely before the alarm for a tidy sequence.
Compare the explanation

As a reported event with timing unconfirmed.

Preserve useful evidence with its uncertainty.

Open this interactive lesson →

LESSON 4

Ask without supplying the answer

Use open questions to understand work as performed.

Safety team reviewing photographed evidence and an isolated work area. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Setting: Maintenance line breaking at a fuel facility after a small release from a wrongly opened drain. Act: Make safe, preserve the scene, secure permit, drawings, tag history and prior near-miss actions. Hold / review: Hold restart if isolation cannot be independently verified or the same deficient permit remains in use.

Ask without supplying the answer

Begin with an open account of what happened and what information was available. Ask how the task normally works, what made it difficult and how decisions were understood. A leading question can replace the witness’s explanation with the investigator’s assumption.

Check understanding and distinguish direct observation from what the person later heard. Follow the organisation’s fair process and confidentiality arrangements without promising secrecy that cannot be delivered. Avoid asking people to agree with a pre-written cause before equipment, instructions and working conditions have been examined.

How should interviews be carried out?

Interviews should be fair, calm, structured and designed to understand facts rather than force admissions.

People remember events differently, especially after shock. Interviewers should explain the purpose, avoid blame language, use open questions first, ask one question at a time, let the witness draw or describe the task, and check understanding. Leading questions can contaminate evidence.

Interviews should explore task, equipment, environment, instructions, training, supervision, time pressure, communication, previous warnings and what would have made the safe action easier.

Applied case

Fictional exercise: An interviewer begins, “You ignored the alarm because you were careless, correct?”

Open

Invite an account before testing particular details.

Leading

An assumed motive can distort the evidence.

Test your decision

What is the better opening?

  1. Ask what the person saw, understood and did when the signal occurred.
  2. Ask them to sign the assumed cause immediately.
  3. Tell them which answer will close the investigation fastest.
Compare the explanation

Ask what the person saw, understood and did when the signal occurred.

Open questions preserve the person’s evidence and decision context.

Open this interactive lesson →

LESSON 5

Look beyond the last action

Connect immediate and underlying conditions.

2D teaching diagram: Change the evidence before restarting
Existing SHAREhub 2D teaching schematic. Use alongside the explanation and fictional case in this lesson.

Look beyond the last action

“Operator error” describes a label, not a complete explanation. Examine the task, information, interface, equipment condition, workload and supervision that shaped the action. Ask why the relevant barriers did not prevent or reveal the error.

This does not mean assuming every system failed or removing personal accountability. Evidence may support some contributing factors and leave others unconfirmed. A useful analysis explains a causal connection and tests alternatives. Repeating “why” without evidence can turn a plausible story into an unsupported root-cause claim.

Use Five Whys as a branching evidence enquiry

A causal question is a hypothesis until evidence supports it. Several paths may contribute to the same event; five answers are not a required stopping point.

Fictional learning example

Separate the evidence from the decision
Part of the pictureWhat the example establishes
KnownOil was found on a walkway and a slip was reported.
Mechanical pathA seal defect may explain the release; inspect the relevant records and condition.
Organisational pathA reported defect may not have reached the right decision owner; test the communication and response history.

Pause and decide: Should “the operator did not notice” become the root cause?

Compare the reasoning

Only if supported, and it would still leave other mechanisms to investigate. Preserve the timeline, test competing explanations and look for missing or failed controls. Choose actions that address supported causes, then verify their practical effect and retain the reasoning for future users.

What is root cause analysis in practice?

Root cause analysis identifies the deeper system conditions that allowed the event pathway to exist.

A root cause is not always a single cause. Incidents usually have a network of causes: a technical weakness, a procedure weakness, a supervision gap, a competence issue, a maintenance backlog, a poor design decision and a culture signal may all combine.

If the only root cause is written as 'operator error', the investigation is probably unfinished. The better question is why the error was possible, likely, unrecovered or consequential.

Applied case

Fictional exercise: The wrong valve was selected; similar labels and an interrupted handover are reported but not yet examined.

Immediate

The selected valve is part of the event sequence.

Underlying

Conditions influencing selection require evidence, not assumption.

Test your decision

What should analysis do?

  1. Close with “carelessness” alone.
  2. Declare all management systems failed.
  3. Test those factors and the relevant barriers against evidence.
Compare the explanation

Test those factors and the relevant barriers against evidence.

Analysis needs supported causal relationships, not a blame label or universal conclusion.

Open this interactive lesson →

Explore causes with Fishbone and Five Whys

A fishbone, also called a cause-and-effect or Ishikawa diagram, places the event at its head. The main bones group possible contributors; smaller branches explore what led to each one. Categories are prompts to investigate, not findings.

Compare the worked examples

The two BuildOut reports examine the same invented excavator-arm/scaffold incident: 18 explanations and six actions. Their evidence references are fictional teaching records.

LESSON 6

Design an action that changes the mechanism

Compare a reminder with a more effective control response.

2D teaching diagram: Plan for people inside and beyond the site
Existing SHAREhub 2D teaching schematic. Use alongside the explanation and fictional case in this lesson.

Design an action that changes the mechanism

Choose actions that address the established mechanism and consider stronger risk reduction before relying on another reminder. If ambiguous identification contributed, merely telling people to concentrate leaves the ambiguity in place. The appropriate solution still requires task-specific review and authorised implementation.

Specify the intended outcome, owner, implementation evidence and effectiveness check. Avoid prescribing an engineering modification from awareness training. Different actions may address immediate exposure and longer-term improvement; their status should not be merged into a single reassuring tick.

What makes investigation actions effective?

Actions should be specific, proportionate, risk-based, owned and verified.

Weak actions sound like reminders: 'retrain staff', 'be more careful', 'communicate better'. Strong actions change the conditions that made the event possible: redesign access, change maintenance priority, rewrite a permit hold point, add supervisor verification, improve procurement or fix a monitoring gap.

Actions should be ranked by risk and checked after completion. A photo of a sign is not evidence that behaviour or control reliability has changed.

Applied case

Fictional exercise: Evidence confirms confusing route identification contributed to an incorrect selection.

Reminder

It depends on attention while ambiguity remains.

Mechanism

The reviewed change should reduce the identification problem itself.

Test your decision

What is a stronger action basis?

  1. Make an unreviewed valve modification during the briefing.
  2. Review the identification and selection arrangement, then verify its effect in the task.
  3. Issue only a generic “be careful” email.
Compare the explanation

Review the identification and selection arrangement, then verify its effect in the task.

An action should address the supported mechanism through the proper review.

Open this interactive lesson →

LESSON 7

Close the loop in the workplace

Distinguish circulation from applied learning.

Safety team reviewing photographed evidence and an isolated work area
Existing SHAREhub photographic illustration. The written fictional case supplies the condition and decision evidence.

Close the loop in the workplace

Sending a report is dissemination, not proof of learning. Identify who needs the finding, which task or document changes and how the change will be used. A person receiving a long PDF may not recognise the one implication for their role.

Translate findings into the relevant operating, maintenance, competence or change arrangements. Check implementation and effectiveness, then return the result to people who raised the issue. Closure should show how the original weakness was addressed, with unresolved parts retained rather than hidden by the email date.

What does the incident learning loop look like?

The learning loop should connect each event to control improvement, verification and changes to the wider management system.

A useful learning-loop model is: event -> report -> triage -> investigate -> identify weak controls -> action -> verify -> communicate -> update risk profile, training and audit. The important point is that the loop does not end when the report is issued. It ends only when the organisation has changed something meaningful and checked that the change works.Investigating accidents and incidents, HSG245

In a mature system, the loop also feeds sideways and upward. Sideways means similar teams, sites and contractors receive the learning before the same weakness appears elsewhere. Upward means senior leaders see themes, repeated weak controls, overdue actions and resource gaps. Downward means workers see that reporting creates change, which improves trust and future reporting.

Loop stageWhat should happenCommon failure
EventIncident, near miss, unsafe condition, audit finding or weak signal occurs.Only injury events are treated as learning opportunities.Investigating accidents and incidents, HSG245
ReportWorkers can report quickly without fear or complex barriers.Reports are delayed, filtered or discouraged.Investigating accidents and incidents, HSG245
TriagePotential consequence, recurrence and barrier impact determine priority.Actual harm alone determines investigation level.Investigating accidents and incidents, HSG245
InvestigateFacts, timeline, task, human factors and barrier condition are examined.Investigation starts with blame or a preferred answer.Investigating accidents and incidents, HSG245
Identify weak controlsFind which controls were absent, weak, bypassed, misunderstood or unverified.Focus remains on what the person did wrong.Investigating accidents and incidents, HSG245
ActionActions strengthen design, process, competence, supervision or assurance.Actions are generic reminders or repeated training.Investigating accidents and incidents, HSG245
VerifySomeone checks the action is implemented and effective in real work.Actions are closed when a document is uploaded.Investigating accidents and incidents, HSG245
CommunicateLearning reaches workers, supervisors, contractors and similar risk owners.Lessons are emailed but not translated into task-level understanding.Investigating accidents and incidents, HSG245
Update systemsRisk profile, training, audits, procedures and standards are updated.The event report sits outside the management system.Investigating accidents and incidents, HSG245

Applied case

Fictional exercise: A report is emailed widely; the affected handover procedure remains unchanged.

Circulated

People were sent information.

Applied

The relevant task arrangement changed and was checked.

Test your decision

What is still missing?

  1. Translation into the affected work arrangement and verification of use.
  2. More recipients on the same email.
  3. A claim that receipt proves competence.
Compare the explanation

Translation into the affected work arrangement and verification of use.

Learning requires changed and verified practice where relevant.

Open this interactive lesson →

LESSON 8

Transfer the mechanism, not the headline

Evaluate whether another organisation’s event is relevant.

Safety team reviewing photographed evidence and an isolated work area. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Avoid / redesign: Eliminate shared reversing by redesigning delivery; use one-way routes and segregated pedestrian paths. Control the task: Eliminate the wet-cleaning/pedestrian overlap by rescheduling or using a dry method where suitable. Verify before release: Stop reversing or pedestrian access if barrier, visibility or agreed traffic route fails; reopen after physical control and briefing are verified.

Transfer the mechanism, not the headline

External learning starts by comparing mechanisms, not matching company names or products. A different industry may share a dependency, handover problem or loss-of-containment pathway. Conversely, similar equipment may operate under different conditions.

Record the relevant similarities, differences and questions requiring review. Do not copy another site’s limit, modification or emergency distance into local instructions. Use external learning to test local assumptions and decide whether competent assessment or a controlled change is needed. The story becomes useful when its applicability is examined.

How should external major accident learning be used?

External events should be tested against local vulnerability instead of being treated as interesting history.

For Diploma and process safety learners, external learning is a professional discipline. Major accidents such as Flixborough, Bhopal, Texas City, Buncefield and Anacortes are not there to memorise dates. They help organisations ask whether similar patterns exist locally: uncontrolled change, weak permit systems, poor start-up discipline, alarm and instrumentation weaknesses, contractor interfaces, fatigue, poor emergency planning or failure to learn from previous warnings.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65

The practical test is not 'have we briefed the lesson?' but 'have we checked whether the same failure route could occur here?' That requires scenario review, barrier mapping, audit sampling, maintenance and proof-test checks, operator interviews and evidence that any gaps are closed.

External learning themeLocal question to askEvidence expected
Management of changeCould a technical or organisational change bypass our review process?MOC sample, change-risk assessment and close-out evidence.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Start-up and abnormal operationsAre start-up hazards understood and supervised?Start-up procedure, competence checks and control-room evidence.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Alarm responseDo operators know what action follows a critical alarm?Cause-and-effect logic, drills and alarm review.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Permit and isolationCould work start with incomplete isolation or unclear authority?Permit audit, isolation certificate and field verification.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Contractor interfaceDo contractors understand site major accident hazards?Induction, bridging document and supervision record.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Emergency responseWould escalation, evacuation and communication work in real time?Exercise report, action tracker and role competence.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65

For named historical events, distinguish the investigation finding from the local lesson. CSB Texas City investigation and CSB Anacortes investigation are the US primary investigation routes; HSG245 (2004) is an investigation method, not an account of these later events. Use the listed Buncefield investigation for Buncefield. For each proposed transfer record the precise finding, local equipment/organisational similarity and evidence required before adopting a measure. US findings are not UK law; remaining named-case source-to-lesson mapping needs technical review.

Applied case

Fictional exercise: A published event concerns a different liquid but a similar shared shutdown dependency.

Similarity

A common dependency can make the mechanism relevant.

Difference

Operating conditions may prevent direct copying of controls.

Test your decision

What is useful?

  1. Copy the other site’s settings.
  2. Dismiss it solely because the liquid differs.
  3. Review whether the same dependency exists locally and how it is controlled.
Compare the explanation

Review whether the same dependency exists locally and how it is controlled.

Transfer the causal question, not an unreviewed technical answer.

Open this interactive lesson →

LESSON 9

Protect knowledge that is not in the manual

Recognise tacit and explicit process knowledge.

Process operators monitoring equipment from a UK industrial control room
Existing SHAREhub photographic illustration. The written fictional case supplies the condition and decision evidence.

Protect knowledge that is not in the manual

Controlled drawings and procedures preserve explicit knowledge. Experienced staff may also know how abnormal conditions appear, where an interface is unreliable or why a past decision was made. That experience should be examined and captured where it matters rather than remaining dependent on one person.

Memory is not automatically authoritative. Compare recollection with current equipment and controlled information. Record the rationale and limits, and route any proposed instruction through its responsible owner. A useful knowledge system preserves context without converting an informal workaround into approved practice.

What types of knowledge should be controlled?

A mature system controls both formal information and practical experience because both can affect safety-critical decisions.

Formal information is easier to see: drawings, P&IDs, specifications, permits, inspection records, procedures, risk assessments, design studies and management of change files. Tacit knowledge is less visible: which valve sticks after a shutdown, which alarm is usually the first sign of instability, which drain route has trapped liquid before, which contractor needs close supervision, or why an old design assumption was accepted.

The aim is not to write down every memory. The aim is to identify knowledge that could affect safety, quality, environmental protection, legal compliance or major accident prevention and then make sure it is controlled in a way that survives absence, retirement, restructuring and change.

Knowledge types and capture methods
Knowledge type Typical examples How to capture and maintain it
Controlled technical information P&IDs, layout drawings, relief calculations, hazardous area drawings, equipment files and operating limits. Use document control, technical authority review, revision control and field verification.
Operational knowledge Shift logs, abnormal operation notes, start-up lessons, alarm response and safe operating envelope experience. Use structured shift handover, operating reviews, procedure updates and simulator or scenario training where relevant.
Maintenance and inspection knowledge Corrosion history, recurring defects, overdue inspections, temporary repairs and spare-part substitutions. Use asset registers, inspection databases, maintenance history reviews and technical escalation rules.
Incident learning Near misses, leaks, alarm floods, permit findings, fires, releases, injuries and emergency exercises. Use investigation systems, action tracking, effectiveness review and learning briefings.
Tacit and role-based knowledge Experienced-worker judgement, local plant cues, unofficial workarounds and practical recovery knowledge. Use interviews, mentoring, overlap during handover, role profiles, job aids and supervisor verification.
External learning Major accident reports, regulator alerts, sector learning, professional guidance and supplier safety information. Use a formal external-learning review route with relevance screening and assigned actions.

Applied case

Fictional exercise: A retiring operator knows why an unusual handover restriction was introduced; the current file contains only the restriction.

Explicit

Documents state information that can be retrieved.

Tacit

Experience can explain context but still requires examination.

Test your decision

What should be captured?

  1. Only the operator’s retirement date.
  2. The rationale, supporting records and present applicability through the controlled knowledge process.
  3. An informal workaround as a new instruction.
Compare the explanation

The rationale, supporting records and present applicability through the controlled knowledge process.

Preserve decision context while checking it against current evidence.

Open this interactive lesson →

LESSON 10

Keep knowledge aligned with the plant

Identify how change can break corporate memory.

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Avoid / redesign: Avoid unnecessary transfer or route to safe destination. Control the task: Independent high-high trip with controlled design and end-to-end test. Verify before release: Planner records basis; hold transfer if destination is uncertain.

Keep knowledge aligned with the plant

A lesson can become misleading when the plant changes and its context is lost. Link information to the equipment, configuration, task and review basis it concerns. Retain a traceable rationale for superseded decisions rather than allowing an obsolete instruction to remain indistinguishable from the current version.

Changes in people and contractors can also remove knowledge. Handover should transfer critical history and unresolved concerns, not just file locations. Retrieval matters: a record that exists but cannot be found during a decision may offer little practical protection.

Where does corporate memory matter across the lifecycle?

Corporate memory matters at every stage from concept design through to decommissioning because each stage creates knowledge that later stages rely on.

Knowledge created during design is often needed decades later by operators, maintainers, emergency planners and project teams. If the basis of safety is not retained, later decisions may treat safety-critical assumptions as optional preferences. Likewise, information discovered during operation, inspection and incident investigation should feed back into design standards, procedures, maintenance plans and training.

The lifecycle view prevents ownership gaps. Engineers may create design knowledge, operators may discover practical limitations, maintenance teams may see degradation first, contractors may introduce temporary change and leaders may decide whether resources are available to act. Knowledge management joins these functions together.

Lifecycle knowledge controls
Lifecycle stage Knowledge that must be retained Control expectation
Concept and design Design basis, inherent safety options, hazardous inventory, layout and safeguard assumptions. Record assumptions, calculations, rejected options and the reason for selected controls.
Construction and commissioning As-built status, deviations, punch-list risk, pre-start-up checks and training needs. Verify installed plant against design intent before hazardous material or energy is introduced.
Normal operation Safe operating envelope, alarm response, operating trends and operator feedback. Use controlled procedures, shift logs, supervision and trend review.
Abnormal operation Trip recovery, start-up/shutdown behaviour, utility failure and unstable process learning. Define escalation points, mandatory responses and technical support triggers.
Maintenance and modification Isolation history, inspection findings, repair decisions, temporary equipment and MOC records. Use permit, isolation, handback, technical approval and record update checks.
Decommissioning Residual inventory, unknown materials, degraded equipment, buried services and legacy hazards. Preserve old plant knowledge and verify assumptions before intrusive work.

Applied case

Fictional exercise: A historic incident action is referenced by an obsolete equipment tag after modification.

History

Retain the rationale and scope of the original decision.

Current

Establish how it relates to today’s plant and task.

Test your decision

What needs attention?

  1. Its connection to the current configuration and controlled information.
  2. Automatic deletion of every historic record.
  3. Assuming the old tag proves present applicability.
Compare the explanation

Its connection to the current configuration and controlled information.

Knowledge needs traceable context across changes.

Open this interactive lesson →

LESSON 11

Test learning effectiveness

Choose evidence that answers the original concern.

Safety team reviewing photographed evidence and an isolated work area. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Change at source: Eliminate shared reversing by redesigning delivery; use one-way routes and segregated pedestrian paths. Make it observable: Stop reversing or pedestrian access if barrier, visibility or agreed traffic route fails; reopen after physical control and briefing are verified. Keep useful evidence: Reporting data

Test learning effectiveness

An effectiveness check should revisit the weakness, not simply count attendance at the briefing. If a handover omission contributed, examine whether the revised handover communicates the required condition in realistic work. Choose a suitable sample and acknowledge its limits.

No recurrence may be encouraging but can be weak evidence for a rare event. Combine it with relevant control evidence. If the review finds that an action did not work, reopen the improvement question without rewriting history to claim it was never needed. Learning includes correcting an ineffective response.

What evidence shows organisational learning is working?

Evidence should show lessons are implemented, verified and retained, not merely identified.

Follow one near miss from report to potential-consequence classification, investigation, stronger action and field verification. Then check a comparable asset or site for the same weakness. An action marked complete after a briefing is not evidence that an alarm, barrier or decision route now performs as intended. Investigating accidents and incidents, HSG245

EvidenceWhat it should showWhat to challenge
Reporting dataNear misses, weak signals and incidents are captured across sites and teams.No near misses in high-risk work may indicate under-reporting.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Classification criteriaPotential severity and barrier degradation influence investigation level.Only actual injury drives priority.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Investigation qualityImmediate, underlying and system causes are explored.Actions blame individuals or repeat training.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Action trackerActions have owners, due dates, risk ranking and verification evidence.Actions close by date rather than effectiveness.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Learning briefsLessons reach workers, supervisors, contractors and similar sites.One-way email with no discussion.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Repeat-event reviewRecurring themes and weak signals are trended.Events reviewed in isolation.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65
Management reviewLeaders review serious events, overdue actions and barrier health.Meeting focuses only on injury rates.Investigating accidents and incidents, HSG245 Managing for health and safety, HSG65

Applied case

Fictional exercise: Every employee attended a briefing, but the changed handover has not been examined.

Attendance

It shows participation in an event.

Effectiveness

It examines whether the relevant weakness changed.

Test your decision

What would most strengthen closure?

  1. Attendance alone.
  2. A new cover page for the report.
  3. Evidence that the required condition is now communicated and acted on during the handover.
Compare the explanation

Evidence that the required condition is now communicated and acted on during the handover.

Effectiveness is linked to the mechanism, not distribution statistics.

Open this interactive lesson →

LESSON 12

Assemble a learning handover

Integrate facts, action and retained knowledge.

Process operators monitoring equipment from a UK industrial control room. Photorealistic illustration.
Photorealistic illustrative UK workplace scene. Review checks: Change at source: Avoid unnecessary transfer or route to safe destination. Make it observable: Planner records basis; hold transfer if destination is uncertain. Keep useful evidence: Board risk review samples knowledge risks alongside equipment and procedural controls.

Assemble a learning handover

A useful learning record connects the event, supported causes, relevant controls, agreed changes and evidence of their effect. Keep disputed or unknown points visible. A neat narrative should not take priority over an honest evidence boundary.

State who needs the knowledge and how it will survive changes in people, equipment and documents. Preserve outstanding restrictions until their proper release conditions are established. The final product of this exercise is a fictional learning plan, not a declaration that a real plant is safe or that legal reporting obligations have been satisfied.

What evidence shows that corporate memory is working?

Good evidence shows that knowledge is current, used, understood and converted into safer decisions.

Assurance should trace a credible scenario from hazard knowledge to current control evidence. For example, if a site relies on relief protection, the assessor should be able to see the design basis, the scenario assumptions, inspection and test records, MOC history, maintenance deferrals, alarm response guidance, operator training and leadership review of any impairment.

A strong audit does not ask only whether the site has a document. It samples whether different roles can find the right information, whether the information matches the plant, whether changes have been reflected in training and whether weak signals have been escalated.

Assurance evidence
Evidence area What to sample What good looks like
Process safety information P&IDs, safe limits, design basis, relief assumptions and hazardous properties. Current, controlled, technically reviewed and field-verified.
MOC records Temporary and permanent changes, organisational changes and close-out packs. Hazard review, approval, document updates, training and pre-start-up checks are complete.
Inspection history Corrosion data, defect trends, repairs, overdue actions and temporary repairs. Degradation is understood and risk decisions are visible.
Operating records Shift logs, alarm histories, abnormal condition records and handover notes. Plant state and critical conditions transfer reliably.
Incident learning Investigation reports, action closure and recurrence checks. Learning changes controls and is checked for effectiveness.
Competence evidence Critical-task assessment, role handover and refresher training. People can explain what matters, why it matters and when to escalate.

Applied case

Fictional exercise: The fictional wrong-route event has a verified cause, an implemented label change and an effectiveness review still pending.

Implemented

The agreed physical or procedural change exists.

Learned

Relevant people can use it and its intended effect is examined.

Test your decision

Which status is accurate?

  1. The event is fully investigated only if every hypothesis is called a fact.
  2. Implementation complete; effectiveness remains to be established and communicated.
  3. Everything is closed because the labels changed.
Compare the explanation

Implementation complete; effectiveness remains to be established and communicated.

Preserve separate statuses and the remaining evidence need.

Open this interactive lesson →