Close-up of a worn yellow crane hook and lifting chain against a steel structure

Look beyond the injury rate

Total recordable injury frequency counts a minor injury and a fatality the same way. Looking at actual and potential serious injuries and fatalities gives leaders a clearer view of how well critical controls work. By Mark Holmes.
Mark Holmes
·
November 9, 2026

Mark Holmes is Health & Safety Lead at Cambianz. With senior experience in the resources sector, Mark has led multi-stakeholder programmes across some of the world's most demanding regulatory environments. Recognised as one of Health and Safety at Work's ‘40 Under 40’, Mark’s technical credibility and expertise turns complex health, safety and ESG requirements into clear, workable outcomes.

Across my experiences in the global mining sector, both at the International Council on Mining and Metals (ICMM) and working in the wider resources industry, a recurring question has shaped how I think about safety: what does improving performance actually look like? Which measurements matter when we are trying to achieve better outcomes?

An obvious answer is fewer injuries. Every injury prevented matters. But when our ambition is to improve the prevention of fatalities and life changing harm, the overall injury rate gives us only part of the picture.

Having worked with industry safety data, assurance and performance improvement, I have become increasingly concerned about the weight we place on total recordable injury frequency (TRIF) as evidence that a business is becoming safer. My concern is that the measure does not distinguish the events that caused, or could have caused, serious harm from those that did not.

The same number can tell very different stories

TRIF counts recordable cases relative to hours worked. Each case contributes equally to the calculation, whether it involves a relatively minor injury, permanent disability or a fatality. Two operations with the same number of cases and hours worked can therefore have identical rates, despite the potential for profoundly different consequences for those involved.

TRIF also tells us little about the circumstances behind those cases. It does not distinguish an injury with limited potential for greater harm from one where someone narrowly escaped death. And when nobody is injured, even a potentially fatal event may make no difference to TRIF at all.

For example, consider a case of a heavy component falling beside someone. The difference between a near miss and a fatality might be a small change in where that person was standing. It is the failure that allowed the component to fall that deserves attention in either case.

This lack of differentiation by TRIF is a fundamental issue and should be of high concern across the industry. An injury rate measures the frequency of recorded outcomes. It does not measure how effectively we are controlling the potential for serious harm.

TRIF has the potential to mislead organisations about the true picture of their operational safety. A falling rate can reflect worthwhile improvements, while leaving significant fatal risks unresolved. We need to understand which risks have actually been reduced before drawing broader conclusions.

Understanding what could have happened

Utilising cross-company safety data actually reinforces the importance of looking beneath the headline number. Aggregation helps us see broad patterns, but it can also conceal the differences that matter most, and understanding actual and potential serious injuries and fatalities (SIF) is essential.

Actual SIFs tell us where devastating harm has occurred. Potential SIFs tell us where that harm was a credible possibility, even though the outcome was a minor injury or no injury at all. Reviewing both Actual and Potential SIFs gives us a richer basis for learning. We can examine the hazard, the person’s exposure and the protections that were present or missing, rather than allowing the eventual injury to determine how much attention an event receives.

Discipline is critical in that assessment. The useful question is what could reasonably have happened in the situation that actually existed. Leaders need to ask what prevented this event from being worse. Was it a reliable control doing its job, or were they simply fortunate through chance?

The purpose of better data is better action

My experience of safety data and assurance is that the value of information lies in the decisions it helps us make.

A separate count of potential SIFs is useful to commence the process, but it is more important to understand what those events have in common. Are people being exposed to moving equipment? Is hazardous energy being left uncontrolled? Are safeguards difficult to use, poorly maintained or routinely bypassed? Do similar weaknesses exist elsewhere in the business? These are all questions that an overall injury rate cannot answer.

These types of questions also take us beyond the immediate actions of the person involved. We need to understand more about the context of the incident such as the specific equipment, work arrangements, competing demands and decisions that made the exposure possible. The people executing that work are essential to that conversation.

Where critical control management becomes practical

At ICMM, a critical control management system identifies the protections that matter most in preventing serious harm and gives them explicit management attention. The procedural logic is straightforward:

  • Understand what could kill or seriously harm someone
  • Identify the controls essential to preventing that event or limiting its consequences
  • Define what those controls must achieve, assign responsibility and check that they work in practice.

The challenge is making that happen consistently. For example, a signed permit does not by itself establish that hazardous energy has been isolated; having a traffic management plan does not by itself demonstrate that people and moving equipment are effectively separated.

We need evidence from the work itself and a clear response when the required protection is missing or ineffective. Verification should test the performance of the control, rather than simply confirm that documentation exists.

Learning from actual and potential SIFs helps focus that effort. Proactive risk assessment extends it to serious hazards that have not yet resulted in an incident.

A better conversation about safety performance

A place remains for measuring injury frequency rates as they can help us understand recorded injury trends. But we should be sceptical of using improving TRIF as reassurance that fatal risks are well controlled.

Wide-ranging conversations are required by leaders. What could cause serious harm in this business? What are incidents and everyday work telling us about our weaknesses? And how do we know our critical controls are effective?

These are the kind of questions that connect measurement to the decisions that protect people. Ultimately, leaders who point to their injury rate improving must also explain why and how the risk of serious harm is reducing. That is the conversation we need to have.