Root cause analysis (RCA) is a structured investigation of why an event occurred and what changes could prevent a similar event. It connects evidence about a failure or loss to corrective action. A useful investigation explains both the immediate event and the conditions that allowed it to happen. Its value depends on whether the organisation implements effective improvements.
Look beyond the immediate failure
RCA can support accident investigation, reliability improvement, maintenance, quality control and business process improvement. Root Cause Failure Analysis (RCFA) often focuses on equipment failures. RCA can also examine near misses and recurring problems before they cause a major loss.
An investigation can identify several interacting causes. A damaged component may explain how equipment failed, but further investigation may reveal unsuitable design assumptions, ineffective maintenance or weak change control. Avoid forcing these findings into a single chain with one final cause.
HSE guidance distinguishes immediate, underlying and root causes. Its human factors guidance also stresses the need to investigate why human failures occurred. “Operator error” is therefore a starting point for questions, not a complete explanation. [1, 2]
Define the investigation
Make the area safe and provide the emergency response first. Preserve evidence as soon as this can be done safely. Record unavoidable changes made during rescue, isolation or recovery.
Define the event, system boundary and investigation objectives. Consider the potential consequences as well as the actual loss when deciding the depth of investigation. Appoint a team with relevant technical knowledge, investigation skills and operational experience. Include human factors expertise where needed. Give the team enough time, access and independence to challenge existing decisions.
Gather and assess the evidence
Collect photographs, physical items, operating logs, alarm records, maintenance history, procedures and relevant design information. Record where each item came from and how it was handled. Check timestamps before combining data from different systems.
Interview people promptly and use neutral questions. Explore what they knew, what they saw and what conditions affected their decisions at the time. Separate observations from later interpretations. Compare accounts with other evidence and record unresolved differences.
Build a timeline of events and conditions. Identify expected safeguards and establish what each one actually did. Similar failures can suggest useful lines of inquiry, but they do not prove that the same mechanism caused this event.
Test competing explanations
Develop hypotheses that explain the evidence. For each one, ask what evidence would support it and what would contradict it. Use inspection, calculations, tests or further interviews where appropriate. Record gaps and uncertainty. Do not select a cause simply because it fits the first convincing story.
Different tools support different parts of this work. “5 whys” can prompt deeper questions, but five is not a required stopping point. Fishbone or Ishikawa diagrams organise possible causes. Neither method proves causation.
Fault tree analysis can examine combinations of failures. Failure mode and effects analysis can help review component failure modes and their effects. Pareto analysis can identify frequent categories in a dataset, but frequency alone does not establish the cause or safety significance of an individual event. Use these tools to organise and test reasoning against evidence.
Examine physical mechanisms, task conditions, interfaces, supervision and management arrangements as relevant. This is not a fixed sequence that every investigation must follow. Retain important external causes even when the organisation cannot remove them directly; it may still be able to reduce exposure or improve protection.
Select and implement corrective action
Link each recommendation to a supported finding. State the change required, its owner, its due date and the evidence needed to show success. Consider whether the same weakness exists in other equipment, sites or activities.
Prefer actions that remove hazards or strengthen engineering controls where practicable. Training or procedure changes may be necessary, but they need to address the conditions found in the investigation. Assess proposed changes for new hazards and apply the relevant change control process.
Keep temporary precautions in place until permanent actions are effective. Escalate recommendations that need resources or authority beyond the investigation team. An action without an accountable owner is unlikely to produce a reliable result.
Verify effectiveness
Separate completion from effectiveness. A revised procedure can be issued without changing how work is done. Define checks that show whether the action controls the identified problem under real operating conditions. These may include functional tests, observations, audits and review of relevant performance data.
A period without another incident is useful information, but it is weak evidence on its own when events are rare. Review control performance and any unintended effects. Reopen the action if the evidence shows that further improvement is needed.
Case study BP Texas City
On 23 March 2005, explosions and fire at the BP Texas City refinery killed 15 people and injured 180. During startup, a distillation tower overfilled and a flammable release occurred through an atmospheric blowdown system. The resulting vapour cloud ignited. [3]
The US Chemical Safety Board investigated technical conditions alongside human and organisational factors. Its report identified deficiencies across BP, including weaknesses in process safety oversight, resources and learning from warning signs. The findings extended well beyond the immediate startup actions. [3]
This case illustrates why an RCA must examine the conditions that shape work and the safeguards intended to prevent escalation. Restricting an investigation to the final equipment failure or human action can leave wider weaknesses unchanged.
Outputs strengths and limitations
The investigation record should include evidence, the event timeline, hypotheses considered, supported causal findings, uncertainties and recommendations. Retain the action plan and effectiveness checks with the findings.
RCA provides a disciplined way to learn from loss and challenge recurring weaknesses. Its limits include missing evidence, unavailable expertise, restricted scope and insufficient time. Bias can also distort interpretation. State what cannot be established, and distinguish a probable explanation from a confirmed mechanism. The investigation remains incomplete as a learning process until its findings lead to tested improvements.
Hashtags
#RootCauseAnalysis, #RCA, #IncidentInvestigation, #SafetyEngineering, #HumanFactors, #RiskManagement, #TheSafetyArtisan
Sources
[1] Health and Safety Executive. Investigating accidents and incidents HSG245 (2004). Read source
[2] Health and Safety Executive. Human factors in accident investigations. Read source
[3] US Chemical Safety and Hazard Investigation Board. BP Texas City refinery explosion and fire. Report 2005-04-I-TX (March 2007), executive summary. Read source
