Analysing an accident WITHOUT seeking a culprit is essential to genuinely UNDERSTAND what happened and PREVENT recurrence — seeking a culprit blocks learning and misses the real causes. Why NOT seek a culprit: (1) Seeking a culprit BLOCKS the truth. If the analysis aims to designate someone responsible (and therefore to punish), people protect themselves, hide things, minimise, stay silent. Nothing reliable is learned. Conversely, "no-blame" analysis (just culture) frees speech: people explain what really happened, which makes understanding possible. (2) The apparent "human" cause often hides DEEPER causes. Saying "the accident was caused by a worker's error/inattention" is easy but MISLEADING: why was that error possible? Often, organisational or technical causes (pace, lack of training, defective equipment, unsuitable procedure, missing protection) made the error possible or likely. Stopping at the culprit misses those causes — and the accident will happen again with someone else. (3) The goal is to PREVENT, not to punish. Punishing an individual does not fix the system's causes; understanding and fixing the causes prevents recurrence. (4) That does not mean impunity for serious deliberate misconduct, but the analysis POSTURE must be understanding, not accusation. How to trace back to ROOT CAUSES: (1) Gather FACTS (not opinions): what exactly happened? where, when, how, under what conditions? Collect facts quickly (scene condition, witness accounts, material evidence) before they are lost, in a climate of trust. (2) Do not stop at the APPARENT cause. The first visible cause ("he slipped", "the part caught his hand") is only the starting point. You must dig deeper. (3) Use a causal analysis method (Level 3-5): the "5 WHYS" (asking "why?" successively to climb the chain: he slipped → why? wet floor → why? leak not repaired → why? report not handled → why? …); the CAUSE TREE (reconstructing the logical chain of facts that led to the accident, identifying every factor). These methods structure the climb to deep causes. (4) Look for MULTIPLE causes of several natures. An accident rarely has a single cause: it often results from a COMBINATION of factors — TECHNICAL (equipment, protection, materials), ORGANISATIONAL (procedures, training, pace, staffing, maintenance), HUMAN (in their context: fatigue, lack of knowledge, conditions). Identify every contributing factor. (5) Climb back to causes you can ACT on (organisational/technical root causes), because that is where lasting prevention happens. (6) Lead to corrective ACTIONS: each identified root cause calls for an action (repair, protect, train, revise the procedure or organisation) — prioritised, scheduled, followed up, and built into the single document. Analysis is only worthwhile if it leads to actions. Good practices: (1) Analyse QUICKLY (fresh facts) and WITH SEVERAL PEOPLE (crossed views, including those concerned). (2) A climate of TRUST (no blame) for the truth. (3) Also analyse NEAR MISSES (same causes, no harm — learning at lower cost). (4) Seek SYSTEM causes, not a scapegoat. (5) Turn findings into ACTIONS and check their effectiveness. In short: analyse an accident WITHOUT seeking a culprit because seeking a culprit blocks the truth (people protect themselves), hides deep causes (organisational/technical ones that made the error possible) and aims to punish instead of preventing — while the goal is to PREVENT recurrence. To trace back to ROOT CAUSES: gather FACTS in a climate of trust, do not stop at the apparent cause, use a causal method (5 whys, cause tree) to dig, look for MULTIPLE causes of several natures (technical, organisational, human in context), climb back to causes you can act on, and lead to followed-up corrective ACTIONS. Also analyse near misses. "No-blame" analysis, which seeks to understand the system rather than a culprit, is the only kind that truly learns and effectively prevents recurrence — it is a pillar of safety culture and continuous improvement.

Near misses (incidents that COULD have caused harm but did not) are extremely precious: they reveal the SAME failures as an accident, but WITHOUT the harm — analysing them lets you prevent accidents BEFORE they happen. It is a major, often underused prevention lever. What a near miss is: an event where harm was NARROWLY avoided — a caught fall, an object that fell without injuring anyone, an error detected in time, a dangerous situation spotted before the accident. The danger was there, the chain had started, but chance or a reaction avoided the harm this time. Why they are SO IMPORTANT: (1) Same CAUSES, no harm. A near miss results from the same failures as a real accident (missing protection, unsuitable procedure, slippery floor, defective equipment…) — only the consequence differed (by luck). Analysing a near miss therefore reveals a real source of danger before it causes harm. (2) Prevent BEFORE the accident. That is the major benefit: treating the causes revealed by a near miss stops the corresponding accident from happening. You learn and fix WITHOUT anyone being hurt. That is proactive prevention, at the best moment (before harm). (3) They are FAR MORE NUMEROUS than accidents. For one serious accident there are generally many minor accidents and a great many near misses (the "pyramid" idea: the base of near misses is far wider than the peak of serious accidents). That mass of events is a mine of risk information — far richer than accidents alone (rare). Ignoring near misses means depriving yourself of most signals. (4) They flag risks WITHOUT the human cost. Learning from a near miss costs infinitely less than learning from an accident (no injured person, no suffering, no stoppage). It is the cheapest and most ethical way to improve safety. (5) They sustain VIGILANCE and safety culture (dangerous situations are not trivialised). Why they must be REPORTED (and the challenge): (1) Near misses do not "show" like accidents (no harm, no mandatory reporting): they only surface if people REPORT them. Yet without a suitable culture, they stay silent ("nothing happened", fear of blame, trivialisation). (2) You must therefore actively ENCOURAGE reporting: (a) a climate of TRUST and NO-BLAME (thank the reporter, never punish — otherwise nobody reports any more); (b) make reporting easy (a simple, quick channel); (c) SHOW that reports are used (handled, fixed, fed back) — otherwise people stop reporting; (d) value reporting as a positive prevention act, not as denunciation. (3) A culture where reporting a near miss is seen as useful and safe brings up a huge amount of preventive information. How to use them: (1) REPORT (make it simple and safe). (2) ANALYSE like an accident (root causes, no culprit) — a near miss deserves the same analysis. (3) ACT on causes (corrective actions, single-document update). (4) SHARE lessons learned (lessons benefit everyone). (5) Track trends (recurring near misses flag a risk to treat as a priority). In short: near misses matter because they reveal the SAME failures as an accident WITHOUT the harm — treating them lets you PREVENT the corresponding accident before it happens (proactive prevention, with no injured person). They are also FAR more numerous than accidents (a huge, largely underused signal base) and let you learn at lower cost (human and financial). But they only surface if REPORTED: you need a culture of TRUST and NO-BLAME (thank reporters, do not punish them), easy reporting, and proof that reports are handled and useful. Analyse them like accidents (root causes, no culprit), act on causes and share lessons. Using near misses is one of the most powerful and economical prevention levers — learning and fixing BEFORE harm occurs.