What is Severity Rate?

Understanding the Core Concept

The Severity Rate attempts to answer: When we fail to prevent an incident, how badly does it hurt the worker and the business? You can think of it as the "depth" of the wound, while LTIFR is the "number" of wounds.

However, in Australia, "Severity Rate" is not a monolith. It represents different mathematical concepts depending on your jurisdiction or industry sector. This creates significant confusion when benchmarking across organisations.

The three primary definitions you'll encounter are:

Average Duration Definition

Severity Rate = Total Days Lost ÷ Total Number of Lost Time Injuries

If you have 10 injuries and 100 days lost, your rate is 10. This tells you that, on average, an injured worker is off for two weeks. This definition is common in internal corporate reporting to understand medical severity of injuries.

Exposure-Based Definition (Australian Standard)

Severity Rate = (Total Days Lost × 1,000,000) ÷ Total Hours Worked

This normalises lost time against the size of your workforce, creating a rate per million hours worked. It's the standard traditionally favoured by Australian state regulators and Safe Work Australia.

OSHA Definition (US Influence)

Severity Rate = (Total Days Lost × 200,000) ÷ Total Hours Worked

Many Australian subsidiaries of global firms, particularly in oil and gas, use this US standard. The 200,000 figure represents 100 employees working 40 hours a week for 50 weeks. If you're benchmarking your Australian site against a US counterpart, you must know which multiplier is being used—or your comparison will be off by a factor of five.

The "Lost Day" Construct

To calculate Severity Rate accurately, you must define what counts as a "lost day." This is where your data becomes problematic and inconsistent across organisations.

Calendar Days vs. Scheduled Days

Scheduled Days: You only count days the worker was rostered to work. If they're injured on Friday and return Monday, and you don't work weekends, you might record zero lost days. This approach views severity through a productivity lens—days of production lost.

Calendar Days: You count every day the worker is medically unfit, including weekends. If they're injured Friday and return Monday, this is two days lost. This approach views severity through a human impact lens—days of life impaired.

Most modern Australian safety management systems align with the calendar day approach for internal severity tracking because it better reflects the extent of injury, even if it inflates the "lost productivity" metric slightly.

The 180-Day Cap

What happens if a worker is permanently incapacitated? Do you count days lost for the next 20 years?

Standard practice (derived from OSHA and adopted widely in Australia) is to cap the count at 180 days. Once a worker has been off for six months, you stop counting days for the Severity Rate.

This cap creates a critical statistical distortion. A worker off for 180 days with a complex fracture contributes the exact same weight to your Severity Rate as a worker who is permanently paralysed. The metric fails to distinguish between "severe temporary harm" and "catastrophic permanent harm."

How to Calculate Severity Rate: Practical Example

Let's examine a practical example for a medium-sized Australian civil engineering firm:

• Workforce: 150 Full Time Equivalent (FTE) employees
• Hours Worked (Annual): 150 × 40 hours × 50 weeks = 300,000 hours
• Incidents:
  • Worker A: Cut finger (Medical Treatment, 0 days lost)
  • Worker B: Sprained ankle (LTI, 5 days lost)
  • Worker C: Hernia (LTI, 45 days lost)
• Total Lost Days: 5 + 45 = 50

Australian Standard:
SR = (50 × 1,000,000) ÷ 300,000 = 166.6

OSHA Standard:
SR = (50 × 200,000) ÷ 300,000 = 33.3

Average Time Lost:
ATLR = 50 days ÷ 2 incidents = 25 days per injury

You can see immediately why context is critical. A rate of 166.6 looks catastrophic if you mistake it for an OSHA rate, but it's merely the mathematical outcome of the 1,000,000 multiplier.

The Frequency-Severity Paradox

This is the most important concept to understand. If you rely on Severity Rate to tell you if your site is safe, you are falling victim to a dangerous statistical illusion.

Conventional wisdom told you that if you reduce the frequency of minor injuries (cuts, strains), you will inevitably reduce the severity of major ones (fatalities). This is the Heinrich Pyramid theory.

This is false.

Modern research has proven that high-frequency injuries (slips, trips) and low-frequency catastrophes (explosions, structural collapse) have completely different causal mechanisms.

Slips and trips are caused by individual attention, housekeeping, and minor hazards. Catastrophes are caused by engineering failures, complex system interactions, culture, and production pressure.

You can have a Severity Rate of zero (because no one has slipped or tripped) while your gas monitoring system is failing, your rock bolts are corroding, and your pressure valves are stuck.

Case Study: Grosvenor Mine Explosion (2020)

In May 2020, the Grosvenor underground coal mine in Queensland suffered a catastrophic methane explosion, severely burning five workers.

Prior to the explosion, the mine's conventional safety metrics—LTIFR and Severity Rate—were reportedly acceptable. They were effectively managing "lost time."

However, the Queensland Coal Mining Board of Inquiry found that the mine had experienced repeated methane exceedances (High Potential Incidents) in the months leading up to the disaster.

A methane exceedance that doesn't explode causes zero lost days. Therefore, it contributes zero to the Severity Rate. The Severity Rate was blind to the accumulation of explosive gas risk. If you were looking at the Severity Rate, the mine looked safe. If you were looking at High Potential Incidents, the mine was blinking red.

The Brady Review Finding

Dr. Sean Brady's 2019 review of all fatalities in Queensland mines (2000–2019) concluded bluntly:

"The industry should shift its focus from Lost Time Injuries (LTIFR) as a safety indicator... At best the LTI Frequency Rate is a distraction... at worst it results in early warning signs being missed."

Brady found that the industry has a "fatality cycle" completely uncorrelated with the LTIFR/Severity Rate. You can have your lowest Severity Rate in history the day before your worst disaster.

How the Severity Rate Gets "Gamed"

Because Severity Rate is often tied to KPIs, bonuses, and contract renewals, you will face immense pressure to manipulate the number. Understanding these tactics is essential for data integrity.

The "Light Duties" Loophole

The easiest manipulation method is through "Restricted Work" or "Light Duties."

A worker breaks an arm—this should be six weeks off (42 days lost). Instead, you create a "compliance role" in the office. You pick the worker up, drive them to the office, and have them watch safety videos or file paper with one hand.

Result: Calendar days lost = 0. Severity Rate impact = 0 (or minimal).

The DART rate (Days Away, Restricted, Transferred) was invented by OSHA specifically to close this loophole. DART counts restricted days. However, if you're only reporting "LTI Severity Rate" (as per many Australian contracts), you can legally hide a broken arm by putting the worker in the office.

The "Weekend" Discount

If you use "Scheduled Days" counting, an injury on Friday is statistically "cheaper" than an injury on Monday.

A Monday injury: 5 days lost (Mon-Fri). A Friday injury: 1 day lost (Fri). Same injury, same human impact, but the Friday injury makes your Severity Rate look 80% better.

Unscrupulous managers might argue that a worker injured on Thursday "finished their shift" and "wouldn't have worked the weekend anyway," minimising the count.

Culture of Suppression

When you attach bonuses to low Severity Rates, you incentivise workers to hide pain.

The "Pizza Party" effect: "If we go 100 days without an LTI, everyone gets a bonus." Worker A hurts their back. They know if they report it and take time off, they ruin the bonus for 100 mates. They swallow some ibuprofen and keep working.

The Severity Rate stays low, but the risk (manual handling injury) remains untreated and likely worsens. This creates a "hidden factory" of injury that eventually explodes.

Best Practices for Using Severity Rate Responsibly

If you're forced to use Severity Rate for insurance or tender compliance, you must use it responsibly. Here's how you can rehabilitate the metric.

Transition to "Potential Severity"

Stop measuring the outcome; measure the potential.

Old Way: A 5kg hammer drops 20m and lands on the ground. No one hit. Severity = 0.

New Way: The hammer could have killed someone. Record it as a "Potential Class 1 Consequence" (fatality potential).

Safe Work Australia advocates for this Severity-Based Framework. You should track "Potential Fatalities" (Class 1) separate from "Actual Minor Injuries" (Class 3). This allows you to track risk without waiting for blood.

Adopt the High Potential Incident (HPI) Rate

You should replace or supplement Severity Rate with HPI Frequency Rate:

HPI Rate = (Number of HPIs × 1,000,000) ÷ Total Hours Worked

An increasing HPI rate is actually good news initially—it means your culture is healthy enough to report near misses. A decreasing HPI rate combined with a stable Severity Rate suggests you're fixing problems before they become injuries.

Follow ISO 45004:2024 Principles

The new standard ISO 45004: Guidelines on performance evaluation explicitly advises against relying on single metrics. You need a balanced scorecard that includes:

• Lagging: Severity Rate (for insurance/compliance)
• Leading: Percentage of Critical Control Verifications completed
• Cultural: Results of safety climate surveys ("Do you feel safe stopping work?")

ISO 45004 validates the use of focus groups and interviews as data. "The guys on the floor say the pump is vibrating" is a valid data point, often more predictive than the Severity Rate.

Transparent Definitions

In your annual report, add a "Data Definition" footnote:

"Severity Rate is calculated as (Total Calendar Days Lost ÷ Total Hours Worked) × 1,000,000. Lost days include weekends. Cap of 180 days applies."

This transparency prevents the "200k vs 1M" confusion and builds trust with stakeholders.