How Dr. Thomas Krause Thinks About Incident Investigation Before Compliance Creates Confidence

Episode 11 of the Headline Podcast, published on December 3, 2025, brought Dr. Thomas Krause into a conversation about leadership quality, felt safety value, and the limits of conventional incident investigation.

The central argument is that compliance can close a report without explaining the conditions that made safe work difficult, so an investigation should trace decisions, trust, and system design before it assigns responsibility to the person closest to the event.

Why a compliant investigation can still be weak

A compliant investigation is still weak when it proves that a process exists but does not explain how earlier decisions shaped the conditions in which people worked.

Organizations often treat the completed form as evidence that learning has occurred. The form may contain a timeline, a violated procedure, a corrective action, and a manager's approval, yet those fields can describe the event without testing the system that produced it.

Dr. Thomas Krause challenged that habit on the episode because an investigation that stops at the final action can make the event look smaller than it was. The question is not whether a rule was available. The question is why the rule was difficult to follow, what competing demand shaped the choice, and which decision made that condition normal.

This is why Headline's earlier work on evidence before incident conclusions matters here. Evidence should slow the rush toward a neat explanation, especially when the first explanation places the full burden on one worker.

What the episode says about leadership quality

Leadership quality predicts investigation value because leaders decide which questions can be asked, which evidence is protected, and whether uncomfortable findings can change the work.

Krause said on Headline Podcast, “We were surprised to learn the strongest predictor of success was the quality of leadership given to the initiative.” That observation changes the role of the investigation sponsor. The sponsor is not only approving a report. The sponsor is shaping whether the review can reach decisions made 1 year or 5 years before the event.

A serious sponsor protects time for interviews, asks what the system made difficult, and refuses to accept a finding that cannot explain the gap between the written method and the work as performed. That work requires more than authority. It requires the willingness to hear evidence that may implicate planning, staffing, procurement, maintenance, or production priorities.

For a safety director, the practical test is visible in the first 30 days after an event. Did the investigation team gain access to the relevant decision records? Did the sponsor ask an operational leader to explain the trade-off? Did the team preserve contradictory evidence instead of smoothing the story?

Why blaming the final action misses the system

The final action is only one layer of causation, so a useful investigation also examines the decisions, interfaces, resources, and assumptions that narrowed the worker's options.

Krause put the problem plainly on the show: “Incident analysis always looks like it was the employee's fault for not following the procedure, but look deeper and you see following it was made very difficult by system factors set by decisions made a year, or five years, ago.”

The statement does not remove accountability. It makes accountability more complete. A worker may still have made a poor choice, but the organization must also account for the conditions that made that choice more likely, less visible, or harder to challenge.

James Reason's work on latent failures supports this wider lens because harmful events often emerge when several weak conditions align. A report that names only the last action leaves the earlier layers intact. It also teaches the workforce that honesty creates personal exposure while system owners remain outside the review.

What trust changes in the evidence

Trust changes the evidence available to an investigation because people disclose more accurate information when they believe leaders will respond to facts rather than punish the speaker.

On the episode, Krause noted that an organization could have 60% of people say they trust their supervisor and still sit in the 90th percentile, leaving 40% of people outside that trust relationship. The point is not that one survey number defines culture. The point is that a favorable average can conceal a material minority whose information never reaches the decision path.

Investigators should therefore compare the official account with the experience of people who worked around the event. Ask which warning was known, who could raise it, what happened after earlier concerns, and which part of the work plan was treated as fixed even when conditions changed.

The comparison should protect identities where necessary, but confidentiality should not become a reason to avoid action. A trustworthy process tells people what was heard, what will change, who owns the change, and when the organization will verify it.

Before and after the Krause lens

The Krause lens moves incident investigation from report completion toward decision traceability, which makes the difference between a closed file and a changed exposure easier to see.

This comparison also clarifies why a polished report can be dangerous. It may create confidence without changing the conditions that produced the event. The organization then repeats the same investigation pattern and calls the repetition consistency.

How to ask better questions after an event

Better questions connect the event to a decision path, so investigators can identify where information, authority, resources, or control reliability broke down.

Start with the work as it was planned, then describe how it was actually performed. Ask what changed, who knew, and what authority existed to pause or redesign the task. The sequence should include the worker, supervisor, planner, maintenance owner, contractor interface, and leader who accepted the residual risk.

Use 3 question groups. First, ask what the team expected to happen. Second, ask which condition made that expectation unreliable. Third, ask which decision could have detected or corrected the drift earlier. This creates a causal map without treating the worker's last action as the whole explanation.

The same logic applies to worker voice in safety. Participation is not proven by attendance. It is proven when information from the work reaches a decision and the decision changes the exposure.

What safety leaders should measure next

Safety leaders should measure whether investigations change decisions and controls, not only whether reports are completed within a target time.

A useful review can track 4 evidence fields. Record the decision that shaped the exposure, the control that failed or became difficult, the owner who can change the condition, and the verification that will show whether the change held.

Review those fields at day 30 and day 90. At day 30, confirm that the chosen action reached the work and that workers can describe the new condition. At day 90, test whether the control remains usable under normal pressure, changed staffing, and non-routine work.

Headline's discussion of visible felt leadership adds a useful leadership test. Leaders do not make safety credible by repeating the value. They make it credible when the response to evidence is visible, timely, and connected to the work.

Recommendation

Use every serious incident review to produce a decision trace, a system owner, and a field verification date before the report is considered complete.

For the next investigation, appoint a sponsor who is senior enough to remove barriers but close enough to understand the operation. Give the team 1 clear question beyond compliance, such as why the control became difficult to use or why an earlier warning did not change the plan.

Require the final report to identify the last action and at least 3 earlier conditions that shaped it. Then name the owner, the change, and the evidence that will be reviewed after 30 and 90 days. If the team cannot name those elements, the organization has documented the event but has not yet learned from it.

Dr. Thomas Krause's message is useful because it turns leadership from a declaration into an investigation condition. A leader's value for safety becomes credible when people can tell the truth, the system can absorb the truth, and the work changes because of it.

Listen to the full conversation with Dr. Thomas Krause on Headline Podcast and use its central question in your next review: which earlier decision made safe work harder than the report admits?