A root cause analysis (RCA) changes something only when it turns evidence about an incident into owned corrective actions—and checks whether those actions work. A report that stops at “someone made a mistake” may name the trigger without explaining why the system allowed it. The practical goal is to identify supported causes, address them, and measure whether the risk has fallen.
Why many root cause analyses fail to change anything
An RCA is not a search for one person to blame or one tidy explanation. An incident may have several contributing causes, and a useful analysis explains how they combined. OSHA warns against stopping at carelessness or a missed procedure without investigating why those conditions existed. Its workplace guidance frames the point plainly: “Addressing underlying or root causes is necessary to truly understand why an incident occurred, to develop truly effective corrective actions, and to minimize or eliminate serious consequences from similar future incidents.” OSHA’s incident investigation guidance and its 2016 root cause analysis fact sheet concern workplace safety; requirements discussed in the fact sheet apply to covered Process Safety Management and EPA Risk Management Program contexts, not every incident.
As an Amazon Associate I earn from qualifying purchases.
Healthcare guidance reaches a similar prevention-focused conclusion in a different operational setting. AHRQ says the “sole objective” of investigating an adverse event or near miss is preventing future adverse events, and cautions that blame can discourage staff from reporting hazards and near misses. Its CANDOR investigation guide is healthcare-specific, not a universal compliance rule.
What a useful RCA report contains
Use a structure that lets a reader follow the path from event to evidence, from evidence to causes, and from each cause to a measurable response. Keep established facts separate from assumptions, and preserve uncertainties or conflicting accounts instead of smoothing them into a confident narrative.
- Scope and event statement. State what happened, when and where, the actual or potential consequence, and why the review was opened. Define the event under review so readers know what the analysis does—and does not—cover.
- Evidence and chronology. Identify the records, interviews, observations, equipment or process information used. Build a timeline of the sequence and relevant context, including handoffs and changing conditions. AHRQ recommends gathering information iteratively, interviewing people involved, observing the context, and reconstructing events chronologically.
- Causal analysis. For significant points in the sequence, ask what happened, why it happened, and what allowed it to happen. Use an analysis tool to organize inquiry, then test candidate causes against evidence.
- Findings and contributing factors. Name each supported factor in plain language, point to the evidence, and explain its relationship to the event. Consider management systems, equipment, process design, training, workload, and the environment when relevant. Do not label a person’s action as the endpoint of the analysis.
- Corrective action plan. Map each identified cause to one or more actions. For every action, specify the change, accountable owner, due date, necessary resources or dependencies, and evidence that it was completed. CMS guidance for root cause analysis with performance improvement projects says to develop at least one corrective action for each root cause.
- Effectiveness check and learning. State what will be measured, the baseline if available, the review period, who will check the results, and what happens if the measure does not improve. AHRQ notes that a solution cannot be evaluated as effective without a measurement strategy; observations, audits, or other data collection can provide evidence.
How to analyze causes without settling for the first answer
Tools help teams ask better questions, but none proves a cause by itself. AHRQ’s quality-improvement resources and CMS guidance describe several approaches; choose according to the event, available evidence, and question being investigated.
Five Whys
Follow a causal chain from an observed failure toward the conditions that enabled it. Do not stop because you have asked “why” five times: the number is a prompt, not a rule. An event with several interacting causes may need several lines of inquiry. Test whether the proposed cause is supported and whether addressing it would plausibly prevent recurrence.
Rank #2
Fishbone diagram
Use a cause-and-effect diagram to sort candidate factors into categories and reduce the risk of focusing prematurely on one person or trigger. Treat the diagram as an organizer for hypotheses, not evidence that every listed factor contributed.
Timeline and structured event review
When sequence, handoffs, or context matter, start with a chronology grounded in records, interviews, and observation. AHRQ’s CANDOR approach places information gathering and chronological reconstruction before consensus and solution discussions, helping teams distinguish what is known from what is inferred.
Fall-out analysis
AHRQ describes fall-out analysis as a resource for practices examining performance and quality-measure outcomes. It can fit a question about missed steps or variation in a care process, but it is not a universal method for every incident investigation.
CMS’s guidance offers useful causal tests: would the event have occurred without the proposed cause, and might the problem recur if that cause were addressed? These questions do not replace evidence; they help expose explanations that are merely descriptive or too narrow. CMS guidance for performing RCA with PIPs discusses these tests and tools including Five Whys and fishbone diagramming. AHRQ also lists them in its root cause analysis tools.
Turn each supported cause into an action someone can verify
An action plan should make the link between finding and response visible. If an analysis identifies a cause but the plan has no action aimed at it, the report is unfinished. If an action does not trace back to a supported cause, explain why it belongs in the plan or reconsider it.
Prefer changes that make the safer process easier or make the hazard harder to reach, where practical. CMS identifies engineering controls, process simplification, and standardization of equipment or processes as stronger types of corrective action. The right option depends on the setting, cause, feasibility, and applicable requirements; the label alone does not guarantee effectiveness.
Retraining, discipline, and writing a new policy should not be treated as automatic fixes. AHRQ’s healthcare guide cautions that recommendations such as these do not consistently produce sustained improvement. If a procedure was not followed, investigate whether it was current and workable, whether tools or workload pressures affected compliance, and why the gap was not detected earlier. A short-term correction may address an immediate factor while a broader system change addresses an underlying cause.
For each action, make accountability concrete: name an owner, due date, dependencies, and completion evidence. AHRQ emphasizes accountable owners and measurement strategies. CMS’s recommendation of at least one corrective action per root cause is a practical completeness check, not a guarantee that the selected actions will work.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Measure whether the change reduced risk
Completion is not the same as effectiveness. Installing a safeguard, revising a workflow, or delivering training shows that an action occurred; it does not establish that the conditions behind the event have changed.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitches- Choose a measure tied to the cause. Use a leading measure that checks whether the intended process or control is in place, an outcome measure that tracks the relevant failure or harm, or both.
- Set a baseline and review period. Record the starting point when available and define when and how often results will be reviewed. Do not invent a baseline if reliable data are unavailable.
- Assign the check. Name who will collect or review data, using observations, audits, records, or other suitable evidence.
- Decide what happens next. Specify how the team will respond if the measure fails to improve, worsens, or reveals an unintended effect. That may mean revisiting the causal analysis or revising the action.
Without these details, a report can document activity but cannot credibly show whether the risk declined. AHRQ’s CANDOR guide describes measurement as necessary to evaluate a solution and recommends using observations, audits, or other data collection.
A final quality check before sharing the report
- Can a reader distinguish observed facts, reported accounts, and assumptions?
- Does the chronology show how the event unfolded, including relevant handoffs and conditions?
- Does each causal finding cite evidence and explain its connection to the event?
- Did the team investigate why a procedure was not followed rather than treating noncompliance as the final answer?
- Does every root cause have at least one mapped corrective action, with an owner and due date?
- Does the plan specify how effectiveness will be measured and what happens if results disappoint?
- Is the report focused on prevention and learning rather than blame?
For deeper healthcare-specific reading, the 2020 book Root Cause Analysis in Health Care: A Joint Commission Guide to Analysis and Corrective Action of Sentinel and Adverse Events is listed by AHRQ PSNet with ISBN 9781635851618: AHRQ PSNet’s bibliographic listing.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
