How to Build an Incident Investigation Charter Before Interviews Begin

What should an incident investigation team agree before it interviews the first witness? The answer is not a longer form. It is a short charter that defines the event, the decisions under review, the evidence standard, and the people who can change the controls.

An investigation charter prevents the first confident explanation from becoming the final explanation. It also keeps a serious inquiry practical, because the team can test each finding against the same scope instead of collecting disconnected statements. The guide below is designed for a supervisor, EHS manager, or appointed incident lead who needs a usable charter before the first interview.

What you need before starting

Prepare the initial notification, the known time window, the work activity, the immediate controls already taken, and the names of people who can preserve records. Do not write a cause statement yet. At this stage, the team should separate what is known from what is suspected.

Open a working document with four headings: confirmed facts, open questions, evidence to protect, and decisions that may have influenced exposure. This simple structure follows the discipline described in Safety Culture Diagnosis: Learn how to do your own by Andreza Araújo, where a useful diagnosis depends on observable evidence rather than a polished declaration. It also connects with the first-hour controls for preserving incident evidence, which should begin before the investigation meeting.

Step 1: State the event without explaining it

Write one sentence that describes what happened, where it happened, when it happened, and what exposure or harm occurred. Keep the sentence factual. For example, a maintenance worker was exposed to stored energy during a valve intervention at the north process unit during the morning shift.

Then record what the sentence does not claim. It does not say that the worker failed to follow a rule, that the permit was inadequate, or that training was missing. Verification is simple. Ask whether two people who were not present could read the sentence and identify the same event. A common error is to insert a conclusion into the event description, which makes later evidence serve the opening assumption.

Step 2: Define the investigation boundary

Set the start and end points of the inquiry. The boundary may begin with the task design, a shift handover, a contractor change, or a maintenance decision, and it may end with emergency response, medical care, or the restoration of the control. The correct boundary follows the decisions that shaped exposure, not only the seconds in which contact occurred.

Write two boundary statements. The first says what the team will examine. The second says what it will not examine unless new evidence requires expansion. Verification comes from the question, “Could a decision outside this boundary have made the event more likely?” If the answer is yes, widen the boundary. A narrow boundary that excludes planning, supervision, or change management can produce an accurate description of the last action while missing the conditions that made it likely.

Step 3: Name the decisions under review

List the decisions that kept the task moving. These may include accepting a schedule, approving a work method, assigning a competent person, releasing equipment, changing a sequence, or continuing after a control became unavailable. The purpose is not to create a list of guilty people. It is to locate the choices that controlled exposure.

Give each decision a possible owner, the information available at the time, and the control that the decision was expected to protect. A decision owner is not always the person who performed the last task. As Andreza Araújo argues across more than 25 years of multinational EHS leadership, accountability becomes credible only when it reaches the person who can change the operating condition. Verify the list by asking which owner could approve a different barrier, resource, sequence, or stop decision.

Step 4: Set one evidence standard

Agree what will count as evidence before the team begins collecting it. Include direct observation, equipment condition, permit and isolation records, training or competence records, shift logs, planning messages, photographs, system data, and consistent witness accounts. Not every source will answer every question, so the charter should state which sources are preferred for each type of claim.

Use a claim table with three columns: claim, supporting evidence, and evidence still missing. This prevents a plausible story from being treated as a verified finding. The standard is working when a reviewer can distinguish an observed fact, a reported account, an inference, and an unresolved question. A common error is to use the phrase “the team believes” as if it were evidence. Belief can guide a search, but it cannot close one.

Step 5: Choose the interview sequence

Plan interviews according to evidence risk, not hierarchy. Speak first with people whose memory may be affected by the worksite changing, whose records may disappear, or whose actions are central to the event. Protect the interview environment by explaining the purpose, the expected confidentiality boundaries, and the difference between a fact question and a blame question.

Use open prompts before testing a specific possibility. Ask the person to describe the task, the expected condition, the change they noticed, and the decision that followed. Compare the sequence with the dialogue structure used for a behavioral observation conversation, but do not turn an incident interview into a routine coaching conversation. Verification requires a planned order and a written reason for any later change. A common error is interviewing the most senior person first, then allowing that account to frame every later statement.

Step 6: Assign roles and decision rights

Name the incident lead, evidence coordinator, interview lead, technical specialist, operations representative, and final decision owner. One person may hold more than one role in a small operation, but the charter should still show who performs each responsibility. The final decision owner must have authority to approve changes that affect production, maintenance, staffing, or capital.

Write the limits of each role. The incident lead coordinates the inquiry, while the evidence coordinator protects records and the technical specialist tests equipment or process questions. The operations representative supplies context without controlling the conclusion. Verification comes from a short responsibility check in which every participant can say what they own and what they cannot decide. A common error is assigning the EHS manager responsibility for controls that only operations or engineering can authorize.

Step 7: Test the charter against serious-risk exposure

Before interviews begin, ask whether the event could have produced a serious injury or fatality, even if the actual outcome was less severe. This question changes the quality threshold. A minor injury can reveal a failed isolation, a missing engineered safeguard, or a decision path that would produce a worse outcome under slightly different conditions.

Record the credible high-consequence exposure and the barrier that should have prevented it. Do not inflate the event. Use the same disciplined reasoning applied in the five-question test for whether evidence changed the control. Verification requires a named barrier, an owner, and a piece of evidence showing whether the barrier existed and worked. A common error is to close the inquiry because the injury was treatable, even though the failed control could not reliably prevent a fatal outcome.

Step 8: Set the review cadence and closure test

Choose when the charter will be reviewed, who receives interim findings, and what must be true before the investigation closes. A short review after the first evidence pass can expose scope gaps while records and memories are still available. A second review should test whether the proposed actions change the decision or barrier that allowed the exposure.

Define closure with observable conditions. The team should be able to show that the event statement is factual, the boundary is adequate, key decisions have owners, evidence supports each finding, and actions have resources, due dates, and verification methods. Use the four evidence breaks that allow a known hazard to return as a final challenge to the action plan. A common error is closing when the report is approved rather than when the changed control has been checked in operation.

Final checklist for the incident lead

Before the first interview, confirm that the charter answers these questions.

• What exactly happened, without a cause statement?
• What time window and decisions are inside the investigation boundary?
• Which evidence is protected, and what remains unverified?
• Which decisions and barriers have named owners?
• What is the interview order, and why is it appropriate?
• Could the event have produced a serious injury or fatality?
• Who can approve changes, and how will the team verify them?
• What conditions must be met before closure?

An incident investigation charter is useful because it makes the inquiry harder to steer with confidence alone. It gives the team a common definition of fact, a visible route from evidence to decision, and a closure test that reaches the field. That is the difference between documenting an event and changing the conditions that could repeat it.

For practitioners who want a broader structure for turning safety analysis into operational leadership, Andreza Araújo's Make The Difference: Be a Leader in Health & Safety is available through the Andreza Araújo store.