Looking Beyond Injury Rates to Prevent Fatalities

Recordable injury rates tell only part of the safety story. Organizations that identify precursor conditions, verify critical controls, and focus on high-energy hazards are better positioned to prevent serious injuries and fatalities.

• By Shawn M. Galloway
• Sep 14, 2026

Picture a facility at the peak of its safety history. Five years without a recordable injury, a corporate delegation arriving in days, a banner already ordered. Then a learning team sits down with a maintenance crew to review one of the most ordinary jobs on site: breaking into a transfer line to change out a filter. Forty minutes of conversation surface what years of clean audits missed.

The low-point drain used to verify the line is empty has been partially plugged for years, and the gauge that would have confirmed it was removed during a piping modification years earlier. So the crews developed a workaround technique: crack the flange slowly, standing to the side, listening for the first hiss for signs of pressure. It has always worked. The product in that line, released under pressure onto a person, would kill or maim before anyone could possibly react. But the paperwork says the exposure does not exist.

That facility is composite, drawn from patterns I have encountered across hundreds of assessments and from the investigation files I have reviewed as an expert witness after fatalities. The details change; the pattern does not. In nearly every file, there is a visible precursor, a filed warning, or a workaround everyone on the crew knew. The organizations were rarely negligent in any cartoon sense. They were sincere, program-rich, well-audited, and blind in a specific, recurring, correctable way.

This blind spot shows up in the national numbers. Over the past two decades, recordable injury rates in the United States fell by more than half, a genuine achievement. Over the same period, the fatality rate barely moved. (See Figure 1: The Two Curves That Should Scare You.) And every facility leader should hold the two curves side by side. In 2024, 5,070 American workers died from workplace injuries, a rate of 3.3 per 100,000 full-time equivalent workers; that rate has oscillated between roughly 3.3 and 3.7 for most of two decades, and the total count per year has hovered near five thousand since the mid-2010s (1). The injury curve dove. The fatality curve drifted. If the same forces drove both numbers, the two curves would have fallen together. They did not, and that divergence is the evidence.

Figure 1: The Two Curves That Should Scare You

Research explains why. A multi-company study of serious injury and fatality exposure found that only a fraction of recordable injuries, roughly one in five, carried genuine fatal or life-altering potential (2). The rest belong to a different population of events entirely. This dismantles the comfortable pyramid logic many of us were raised on, the premise that shaving the bottom of the injury triangle automatically thins the top, which Fred Manuele's research dismantled years ago (3). Frequency and severity are different problems; they have different causes, they demand different controls, and they are discovered by different sensors.

The sensor that detects fatal exposure is the precursor: a situation in which high energy is present and the control standing between that energy and a human being is absent, compromised, or reduced to paperwork. A plugged drain, a missing gauge, and a crew reading a flange hiss are precursors. They had never produced an injury. Just one afternoon of ordinary bad luck could result in a funeral. Facilities with real prevention capacity define their precursors, hunt them during normal work, trend them, and review them at the leadership table with more attention than the injury rate receives. Facilities without it discover their precursors during an investigation report.

Different sensors lead to different controls. The events that kill share a physical signature: gravity from height, objects in motion, stored mechanical or hydraulic force, electricity, pressure, temperature, toxic chemistry. Energy-based safety research led by Matthew Hallowell and the Construction Safety Research Alliance has shown that the controls worth trusting against high energy share three properties (three-property standard): they target the energy source directly, they remain effective without depending on a human doing the right thing in the moment, and their presence is verified in the field before work proceeds (4). Hold your own critical controls up to that standard and prepare for discomfort. A rule is a hopeful mandate. A warning in a procedure is a hopeful suggestion. Training is a hope with a certificate attached. These have their place in the hierarchy, but when the energy is high enough to kill, hope is not a control.

What does this mean for the professional responsible for a facility? (See Figure 2: Four moves matter most).

First, inventory the handful of tasks on your site where fatal energy is present; the list is shorter than the procedure manual suggests, and that focus is the point.

Second, hold each critical control on that list to the three-property standard, then walk the field to confirm the control physically exists and functions, because verification steps quietly become difficult to impossible to perform as equipment changes over the years.

Third, make hunting precursors in routine work a standing rhythm, using learning-team conversations that ask crews how the work actually goes rather than how the procedure says it should; a close call with fatal physics deserves the same rigor as an injury, because the difference was luck.

Fourth, put precursor findings, verification results, and control health on the leadership agenda ahead of the recordable rate, so a site can be celebrated for its injury record and challenged for its precursor backlog without contradiction.

Figure 2: Four Moves That Matter Most

One warning as you begin: Success protects precursors. Every day in a long streak makes the workaround seem more proven, makes the uncomfortable question sound more disloyal, and makes the approaching celebration a reason to wait. Mature facilities treat a quiet stretch as a hypothesis to test and sends learning teams into its most routine work, precisely because drift does its best work when it is quiet. Immature facilities order a banner.

Prevention capacity is one of seven organizational capacities I examine in my forthcoming book, The 7 Capacities for Safety Excellence, which asks why safety programs stopped preventing fatalities and what executives must build instead. But no facility needs to wait for a book to take its own blood pressure. If your evidence of prevention looks the same whether or not a fatality-capable exposure was part of your routine work, you are measuring the program but confusing it with the capability. When did your leadership team last review, a precursor finding from a job that has never caused harm? Can your crews name the last workaround they surfaced voluntarily, without an incident forcing it into view? And if a learning team handed you the plugged drain nine days before your ceremony, what would your facility, as it operates today, do in response?

This article originally appeared in the September 2026 issue of Occupational Health & Safety.