How to Run a Control-Health Review in 8 Checks Before the Board Meeting

A board can receive a clean safety dashboard and still miss a drifting control, because injuries appear late while barrier decay starts earlier. A control-health review exists to close that gap by asking whether the controls behind the numbers are still fresh, owned, and proven in the field.

Across more than 250 cultural transformation projects in 30+ countries, Andreza Araujo has seen leaders trust rate stability for too long. In Safety Culture: From Theory to Practice and The Illusion of Compliance, the recurring warning is simple: a tidy report does not prove a live control. When that warning is ignored, the board may praise a calm chart while the worksite keeps carrying the same exposure.

ISO 45001:2018 expects organizations to manage change, and James Reason's latent failure lens explains why that matters, because the event that reaches the board room usually started much earlier, inside supervision, maintenance, design, staffing, or change management. The board does not need more noise. It needs a tighter way to see whether the safety system is still telling the truth.

Key Takeaways

• A control-health review separates outcome rates from control proof, which keeps the board from mistaking calm numbers for safe work.
• The best review looks at one serious-risk control family, not the entire safety program at once.
• Field proof matters more than paper proof, because the workplace is where the control either works or fails.
• Action aging and named ownership are stronger board signals than a large stack of closed items.
• The review becomes useful only when it repeats on a fixed rhythm and drives a real escalation rule.

What you need before starting

• One control family linked to serious harm.
• One recent incident or near miss that touches that control.
• One field verification point that the board can ask about.
• One owner who can change the system, not just report on it.
• One slot in the next board pack.

If you want the metric logic behind the review, start with Control Health Metrics: 5 Blind Spots Boards Miss. That article maps the blind spots. This one turns them into a repeatable board routine.

Step 1: Choose one control family and one serious-harm path

Pick a single risk path, not the entire safety program. Focus on energy isolation, traffic separation, fall protection, confined space, or contractor interface. The point is to stop the board from reviewing a hundred fragments at once, because a spread-out dashboard usually hides the one control that has already started to drift.

This is also where the practical audience matters. A board chair, CEO, finance leader, or EHS director does not need a prettier summary. They need one clear control family that can be followed from the field to the board pack without losing its meaning.

Step 2: Separate outcome, control, and freshness

Write three columns: what happened, what should prevent it, and when that control was last proven. Outcome measures tell you that something occurred. Control measures tell you whether the barrier exists. Freshness tells you whether the barrier was checked recently enough to trust it. A board that mixes those three things can congratulate itself while the control ages quietly in the field.

That distinction is one reason this review is more useful than a generic KPI update. Outcome tells the story of the past, control tells the story of the current barrier, and freshness tells the board whether the barrier still deserves confidence.

Step 3: Demand field proof, not paper proof

Ask for one piece of evidence from the workplace, not from the binder. If the control is a guard, ask for the inspection point. If it is isolation, ask for the field lockout point. If it is traffic control, ask for the route separation that workers can actually see. Paper can confirm intent, but only the field confirms function.

Across more than 250 projects, Andreza Araujo has seen how quickly a board can be comforted by tidy documentation. The trap is not bad faith. It is distance. Once the board relies only on the report, it stops seeing how the control behaves under routine pressure.

Step 4: Test the oldest high-risk action

Choose one overdue action and ask why it is still open. An open action is not automatically a failure, but an open action on a serious control should trigger a question about ownership, priority, and escalation. Andreza Araujo's leadership work shows that many organizations stay polite about delay until delay becomes part of the culture. The board should not need a fatality to notice that routine.

This step is where James Reason's model becomes practical. Latent conditions do not arrive as a dramatic surprise. They stay in the system, age inside normal work, and wait for the same weak point to be reused again and again.

Step 5: Name one owner with real authority

Every control family needs one owner who can move something in the system. If the owner cannot change maintenance priority, staffing, contractor access, or capital timing, then the review has a presenter, not an owner. The board should hear the name, the authority, and the next decision date in the same sentence.

This is where safety decision rights matter. Without them, the board gets status updates, while the real decision stays trapped somewhere below the level that can fix the exposure.

Step 6: Compare the board slide with the field reality

The comparison below keeps the board honest. It shows how easily a neat slide can point in one direction while the field points in another.

That table should create a little discomfort. If the slide looks better than the field, the board is not seeing the control. It is seeing a summary of hope.

Step 7: Write the escalation rule before the meeting ends

The board should know what happens if the control is stale, if verification fails, or if action aging passes the limit. Without an escalation rule, the review becomes a conversation that ends when the agenda ends. With a rule, the review changes work in the next cycle, which is the only test that matters.

A useful rule is simple. If proof fails twice, the item returns with a named owner, a deadline, and a clear stop condition. That keeps the review from becoming an endless loop of reassurance.

Step 8: Put the same control back on next month's agenda

A control-health review only works if it repeats on a fixed rhythm. Put the same control family on the next board pack, ask for the same field proof, and compare the response. If the control is still healthy, the board will see consistency. If it is not, the review will catch drift before the drift becomes an event.

This is the practical difference between oversight and management. Oversight asks whether the control remains trustworthy. Management changes the conditions that make it trustworthy. Both are needed, and both need a schedule.

Recommendation

Start with one serious-risk control family, one owner, and one field proof point, then repeat the review at the same board cadence. If the routine changes nothing in the next meeting, it was only reporting. If it changes what leaders ask, who owns the fix, and how fast the work moves, then the board is finally reviewing control health instead of outcomes alone. If you want the broader governance lens, keep Headline Podcast in the rotation and pair this review with the decision-rights work already in your board pack.

FAQ

What is a control-health review?

It is a board-level routine that checks whether critical controls are still fresh, owned, and proven in the field. It is different from a normal KPI review because it asks about the barrier, not only the result.

Which control family should the board start with?

Start with the control family that can create serious harm if it fails, such as isolation, traffic separation, fall protection, confined space, or contractor interface. The review works best when it follows one path all the way from the workplace to the board pack.

Is this just a different name for a dashboard review?

No. A dashboard review can stay at the level of trend, while a control-health review asks whether the barrier was tested, who owns it, and how recently it was proven in the field.

What is the most common trap?

The most common trap is to treat a low injury rate as evidence that controls are healthy. Low rates can be real, but they do not prove that the control behind them is still strong enough for serious risk.

How often should the review repeat?

It should repeat on a fixed board rhythm, usually monthly or quarterly depending on the risk path and the speed of change. The key is consistency, because a one-time review does not expose drift.