Seven reasons why LTIFR does not impress

LTIFRThe public display of a LTIFR (Lost Time Incident Frequency Rate) as close to zero as possible have done very little to impress the critics to this out-dated lagging indicator. And in the African context, the impressive Million Incident Free Person Hours are nothing more than the store front of a retailer in a large shopping mall.

The “DIFR” was originally introduced to the Industrial Safety fraternity by Frank Bird in his book “Management Guide to Loss Control” which was a result of the work and research of Heinrich and his accident sequence theory. Many research papers have indicated the irrelevance of LITFR in the accident prevention domain, and some critics have developed “new and improved” versions of this lagging indicator.

As a perceived measure of safety “performance”, the general standard is to calculate the number of lost time incidents based on hours worked, multiplied by a constant 1 million (or 200 000 by some). The problem lies in the validity of the constant.

In Bird’s original formula, the constant of 1 million was based on the assumption that one person spends 2080 hours at work in one year. In order to facilitate easy calculations, this was rounded to 2000. It further assumed that the average size of the research population was 500 employees in a company, thus deriving at 500 x 2000 = 1 million.

  1. LTIFR is not Labour based

We used to joke about some managers who had no interest in what you do during the day, as long as you arrive at 8 AM and leave after 5 PM. The LTIFR supports this mind-set. In the South African labour context, the average hours an employee spends “at work” are 1 760 hours per year. This creates a constant of 880 000 for an employer with 500 employees. An LTIFR of 1 is actually, 1.136

  1. LTIFR is not risk based

As a measure of a system designed to manage occupational risks, the LTIFR ignores the risk entirely as it pulls a blanket over all activities and regard them as a single “system”. To explain the difference is like standing next to the highway for 8 hours vs running back and forth across the highway for 8 hours. The factory worker and receptionist are equal in “injury” but not in risk exposure. If both were to lose one shift, the LTIFR would be 2, regardless of the fact that the risk is lower in the office than on the shopfloor.

  1. LTIFR is not accurate in reporting period.

Once again the constant assumes an injury reporting period of 12 months. In recent years, adaptations have allowed for an increase in reporting period to accommodate less than 500 employees, but it still fails in accuracy at both poles.

Company A has 500 employees. Applying the “Constant Assumption” it is possible for them to work 1 million hours in 12 months.

Company B has 3400 employees and Company C has 45 employees.

Assuming all three companies have 5 LTI’s, their respective LTIFR’s would be:

  • A = 5.68
  • B = 0.84
  • C = 63.13

The reporting period is thus extended to a pre-determined set of rules based on staff numbers. In the example above, only Company C would be affected and their reporting period would be extended to 72 months. This will change their LTIFR to 10.52

The truth is that an employer with 45 employees require at least 13 years to reach the constant of 1 million hours.

  1. LTIFR is not a measure of performance – it’s a bench marking tool

According to Heinrich’s model, around 88% of accidents are caused by unsafe acts and 10 percent by unsafe conditions. In his original research, he omitted 2% as “Acts of God” due to inconclusive causes.

Assuming this research remains valid, the LTIFR fails to measure success if the bulk of accidents are “uncontrollable and unpredictable”. This gave rise to approaches such as Zero Harm, Safety Differently, BBS and all the other three letter acronyms conjured up by people exploiting the need for recognition in an abstract system.

The real value in LTIFR is the year end social function when that “Pie in the Sky” number is eventually reached, by hook or by crook. It also allows for a competitive analysis of rival operators and their “track record” which ads to the Unique Selling Point of a business.

  1. LTIFR is not severity based

Copping (1993:1) provides a telling example of the inadequacies of LTI based statistics in this respect.

“After a run of nearly two years accident free a company employee slipped on a step and was unlucky enough to fracture a small bone in his foot. He was unable to work for several weeks and an LTI was recorded with a subsequent loss of safety awards to staff.

At about the same time a container was dropped during an off-shore lifting operation. This latter incident had tremendous potential for injury but as luck would have it no-one was hurt. There is no doubt that the lifting incident was much more serious.”

It is almost as if one can say, “If you die, please do it in this shift; we don’t want to lose another shift.”

Whether a minor injury or a fatality, the apples are the pears and we count each as a single event. We do not add a weighting or penalty to a fatality or a catastrophic environmental event.

And we record lost time as the working time lost by the injured person. We ignore the time spend to find and train a temporary replacement, the investigating time, the time to catch up on back log as a result of the lost management time etc etc. But our system works!

It reminds me of the introductory narrative of the first Zeitgheist project, where people are made to believe that there is a man in the sky who sees everything, and watches us all 24/7/365 and gave us 10 rules to live by, which if we do not comply will result in us being cast into a pit filled with fire! But he loves us!  

  1. LTIFR is not compliance based

Measuring the success or failure of a management system without considering the level of compliance to legal requirements by the managers of the system, is pretty much useless. In the current practice, no weighting or penalty is added where a company only complies with 80% of the legal requirements applicable to their operations.

Yet many companies believe that the law sets minimum requirements and they prefer to honour industry best practice over and above the minimum standards. This in itself should become a penalty to the calculation of the LTIFR. If you do more than what is required, why are 88% of your accidents still caused by human error?

  1. Human Reliability and Performance Shaping Factors

According to Wikipedia, Human reliability (also known as human performance or HU) is related to the field of human factors and ergonomics, and refers to the reliability of humans in fields including manufacturing, medicine and nuclear power. Human performance can be affected by many factors such as age, state of mind, physical health, attitude, emotions, propensity for certain common mistakes, errors and cognitive biases, etc.

All these items fall under the commonly defined “Operator Error” or “employee’s own fault”

This paradigm shoots tunnels in the “Zero Harm” drive everyone is in amazement of. As Homo Sapiens, we are not infallible. And as a result, 88% of all accident causes will remain until planet earth leaves its orbit or we start employing Robots.

Safety is a System problem; not a component problem, or is it?

A system is a set of interacting or interdependent component parts forming a complex or intricate whole. Every system is delineated by its spatial and temporal boundaries, surrounded and influenced by its environment, described by its structure and purpose and expressed in its functioning.

Alternatively, and usually in the context of complex social systems, the term is used to describe the set of rules that govern structure or behaviour, such as a Safety Management System.

When we change the term from “Safety Management System” to “System Safety Management”, we change the way we do things entirely. What needs to be managed? What are we managing?

In our context where SHEQ is now, we are managing a system.

What we need is to start managing the safety of a system.

Once we do that, we can align our IFR methods to measure failures based on actual exposure to a real risk, and not some illusive perceived hazard which we simply tolerate because we think it is “part of the process.”

For a discussion on how System Safety and Safety Engineering can reduce your long term SHEQ spend and increase your productivity, contact Cygma SHEQ’s Safety & Reliability Engineering team.

Related Posts Plugin for WordPress, Blogger...
The following two tabs change content below. is Africa's largest independent SHEQ Magazine, hosting over 2 000 articles and news items since 2007. is owned by the Cygma Group, a global provider of risk management and compliance solutions. is registered as a digital publication with the ISSN.

Latest posts by sheqafrica (see all)


About the Author

sheqafrica is Africa's largest independent SHEQ Magazine, hosting over 2 000 articles and news items since 2007. is owned by the Cygma Group, a global provider of risk management and compliance solutions. is registered as a digital publication with the ISSN.

2 Comments on "Seven reasons why LTIFR does not impress"

  1. Koos Duvenhage | 26 April 2017 at 13:02 |

    OK but what the article are not doings is showings us what it is are the proactive indicators we can thus use for measuring safety performance.

    Management systems and compliance are but a mere element of what is required to manage safety effectively.

    But for da rest, I does agree wif yous. Accident stats are not a leading indicator. This is just reactive. Also, the absence of accidents doesn’t mean you are doing things right.

  2. Hi Koos
    There are two phrases in your comment above that draws attention to the reality of H&S.
    1. Measure Safety Performance
    The HSE in the UK published a Guideline on the topic.
    2. Manage Safety Effectively
    It is a common saying that safety in not rocket science. But can you imagine if NASA had the same attitude? Safety should be rocket science. In the late 50’s NASA had an objective. “to put a man on the moon.” They spared no expense to make sure not one failure would prevent them reaching that objective. The objective was reached. (Yes there was failures too, but it was never used as a measure of success. It was used as a lesson learnt to change the process.
    H&S tends to use LTIFR to “justify” accidents based on their root causes, but seldom measure repetative root causes, i.e. the same thing causing more than one incident.
    Another example is the Air Travel industry. With thousands of people flying accross the globe on an hourly basis, the number on incidents are less than we have in mining in South Africa, or on our roads on an Easter weekend. You do not see the airlines reporting number of incidents per 1 000 000 flight hours claiming to be an achievement. They use an Exponential Moving Average, which ensures that older accidents contributes to the safety index by a lesser degree than newer ones. The farther an accident is back in time, the weaker its impact. Their period of monitoring is 30 years. See the JACDEC website.

Comments are closed.

Facebook IconLinkedInLinkedInLinkedIn
error: You are attempting to breach copyright laws. Please see our Terms And Conditions of Use.