Workplace incident reporting process flowchart (health and safety)
Workplace incident reporting process flowchart for accidents and near misses: first aid, reporting window, statutory report, investigation, corrective actions.
How it works
Rename the lanes to your real roles
Replace injured person or witness, line manager, health and safety officer, HR and regulator with the roles you actually have. Small organisations often have no dedicated safety officer, in which case name the person who holds the duty rather than deleting the lane — the decisions in that lane still have to be made by someone. Keep the regulator lane even if you rarely use it, because it is what makes the external reporting duty visible on the page.
Write your internal reporting window onto the report step
Decide how quickly an event must reach the line manager and put it on the 'Report the incident to the manager' step. Before the end of the shift is a common and workable rule. Say what happens when the line manager is the subject of the report or is unavailable, and give people a route that does not depend on one person being at their desk.
Fill in your own external reporting rules
Open the 'Serious injury or dangerous occurrence?' decision and replace the generic wording with the categories and deadlines that apply where you operate. In Great Britain, RIDDOR 2013 puts the duty on the responsible person, normally the employer, and separates deaths and specified injuries, which are notified without delay and followed by a report within ten days, from injuries that keep a worker off normal duties for more than seven consecutive days, which are reported within fifteen days. In the United States, OSHA requires a work-related fatality to be reported within eight hours and an in-patient hospitalisation, amputation or loss of an eye within twenty-four hours. Confirm the rules that apply to your sites rather than copying either set.
Set the investigation thresholds
The 'Level of investigation required?' decision only works if the trigger for each branch is written down. Scale it to potential severity, not only to the actual outcome: a near miss that could have killed someone earns the full branch, and a minor first-aid case usually does not. Without written triggers the level gets decided by who happens to be free that week, which is how repeat events get logged three times and investigated none.
Fix where the accident book entry is actually made
The chart shows the accident book and register being updated near closure, which is where the reconciliation happens. The entry itself belongs at the time of the accident. Mark on your version who makes the entry, where the book lives, and how individual entries are kept confidential — accident records contain health data and are read later by people with no need to see the rest of the book.
Give verification and the committee review real dates
Attach a due date and an owner to every corrective action, and a date to the 'Actions complete and effective?' check so the loop back to re-agree actions is triggered by a calendar rather than by the next similar incident. Then set the safety committee cadence and say what it looks at: counts by type, near miss reporting rate, overdue actions and repeat causes. Trend review is the only step in this process that finds the problem nobody reported.
Frequently asked questions
What is the difference between an accident, an incident and a near miss?
Incident is the umbrella term. An accident is an incident that caused harm — an injury, ill health or damage. A near miss caused none but had the realistic potential to, and occupational ill health develops over time rather than in one event. All of them travel the same route in this chart, because an accident and a near miss frequently differ only by where somebody happened to be standing. The point at which they diverge is the 'Level of investigation required?' decision, which is where severity and potential severity are weighed.
How quickly does a workplace incident have to be reported?
There are two clocks and they are not the same. The internal one is yours to set, and before the end of the shift is a common rule. The external one is set by law and varies by jurisdiction. As examples: in Great Britain, RIDDOR 2013 requires deaths and specified injuries to be notified without delay by the quickest practicable means and followed by a report within ten days, while injuries that keep a worker off normal duties for more than seven consecutive days are reported within fifteen days of the accident. In the United States, OSHA requires a fatality to be reported within eight hours and an in-patient hospitalisation, amputation or loss of an eye within twenty-four hours. Check the duties that apply to your own sites and industry.
Do near misses have to be reported to the regulator?
Usually not. Most external reporting duties are triggered by an actual injury, a case of occupational disease, or a specific listed event. In Great Britain those listed events are the RIDDOR dangerous occurrences, which are reportable whether or not anyone was hurt — so a near miss is externally reportable only if it falls into that list. Internally you want every near miss, because they are the cheapest information about your workplace you will ever get. That is why the Not reportable branch in this chart does not stop; it goes to the incident record like everything else.
How deep should the investigation be?
Proportionate to the actual and potential severity, which is why the chart makes it a decision rather than a fixed step. The Full branch gathers evidence and interviews witnesses before any cause is named; the Local branch is a team review run by the line manager; the Log only branch sends a minor event to the register so it still counts towards trends. All routes that investigate end at root cause and contributing factors, because stopping at operator error produces a corrective action that is really just a reminder. If you want the analysis method itself — problem statement, evidence, timeline, hypothesis testing — use the root cause analysis process template.
How is this different from the emergency response or IT incident templates?
Scope. The emergency response flowchart covers the first minutes only: whether it is safe to approach, whether to evacuate or shelter, the roll call, and who gives the all-clear. This chart starts after that and covers reporting, recording, investigating and fixing. The incident management process and the incident response process use the same word for IT service disruption and for security events respectively, and neither involves first aid, an accident book or a health and safety regulator.
Does using this template make us compliant with ISO 45001 or with our reporting regulations?
No. It is a starting structure, not a compliance document, and no diagram can create compliance on its own. What it does is make the obligations visible so you can check them against your own rules. ISO 45001 clause 10.2 covers incident, nonconformity and corrective action, and expects incidents to be investigated, causes determined, action taken and the effectiveness of that action evaluated — which is the reason the verification decision and the loop back to re-agree actions are drawn rather than assumed. Replace the reporting criteria, deadlines and retention periods with the ones that actually apply to you, and have the finished process approved by whoever is accountable for health and safety.