Monday, September 17, 2012

Probability vs. Performance: the Heinrich accident triangle revisited

H. W. Heinrich changed the world of safety fundamentals forever with his pioneering work in the 1930’s. One of his concepts that continues to make me think is his accident triangle (pyramid), a concept that we all are familiar with.

So many near misses lead to a lesser number of first-aid injuries and thence onward through the logic to recordables and ending in the inevitability of a fatality. This inevitability of disaster has always bothered me. If I cross the railroad tracks too many times I will die, or drive to work, or something else like that. I am not a fatalist, and thus my pondering is on what exists that will enable the industrial work place to overcome this fatalistic teaching. My work with companies and individuals that have done both well and poorly in safety always leads to individual behaviors as a foundational key, after workplace conditions, training and safety standards are addressed. In many cases, the “behavior foundation” appears to be made of sand. It is not a firm foundation on which to build a zero-injury culture. So how does one attack this foundation of the triangle? Years of thought and effort in this area have led to a whole different level of sub-foundation as shown below. 

If our new safety triangle (pyramid) is built on the “stone foundation” of excellent fundamentals that modify behaviors and actions, can we limit (reduce) the base of improper activities that lead to 90+ percent of the injuries in Heinrich’s pyramid? Each time I have rebuilt the foundation, the results have been similarly excellent. The following sub-foundation fundamentals significantly reduced injuries.
·         Visible upper management leadership in safety
·         Noticeable involvement of middle management
·         Focused supervisory performance
·         Active participation of hourly employees
·         Training that both teaches and reinforces this type of foundational excellence 

The above principles tie in very well with the quality function’s 6-sigma initiative for significant improvement; DMAIC (define, measure, act, improve, control). Only in our safety case it is:
·   Define the correct behaviors that eliminate unsafe acts and injuries
·   Train all personnel in these behaviors
·   Measure that they are indeed doing these correct behaviors
·   Reward their accomplishments of these correct behaviors

The key then is not to focus on compliance, or reward “acceptable injury levels/goals,” or on any other single tactic approach. The key is to concentrate on the foundational fundamentals that eliminate the activities/behaviors that in turn move us out of the chance probability of the Heinrich inevitability triangle. We must error proof the fundamental safety processes and deliver safety accountabilities at each level of the organization. We then engage in a daily practice of the upstream activities that deliver downstream performance.

The Doc


  1. I think you are seriously confused. The accident triangle was established through empirical evidence and subsequently confirmed by Bird, the HSE (UK), and others. It is reproduceable indicating there is some sort of scientific basis. Alas, these are all lagging indicators but are very easy to quantify and measure.

    Your "foundation" consists of leading indicators and you offer no manner in which to measure and quantify these indicators nor the empiracle evidence to support your conclusion. A method to measure leading indicators consistently, accurately and in a quantifiable manner has been driving safety professionals to distraction for many years now. If you were to offer a method that would provide a means to measure leading indicators I think you would be on to something.

    Your foundation identifies some basic components of a good safety management system but is incomplete. One would be better off to follow OHSAS 18001.

    Talking about "behaviours" while ignoring the well established and proven hierarchy of controls is the downfall of behaviour based safety. (See BP for a "glowing" example of this.) Engineering and administrative controls are an essential component of safety that cannot be replaced by BBS which all too often companies try to do because they are told that all past accident causation theory models are wrong.

    BBS should be nothing more than management doing its job to make sure that employees have the correct tools and work environment to do the work, that they have the correct knowledge and skill set to do the work, and that they do what they are supposed to do. I have been doing safety management for over 30 years and have concluded the biggest enemy to safety is lazy management, which is a behaviour but never discussed as such. Blaming the worker for "bad behaviour" is behaviour BS.

  2. To: D. Johnson Re your comment posted to the recent blog article Probability vs. Performance: the Heinrich accident triangle revisited

    Thank you very much for your in-depth response to this short blog article. I have seldom gotten such a detailed, well thought out commentary on any of the 150+ blog posts to date. I find I am in agreement with much of what you mention. I frequently use Heinrich’s base material because it often leads to controversy and this in turn leads to meaningful engagement on what really makes a difference in safety. Yes, I have seen the material by both Frank Bird and the UK that shows there truly is a relationship between close calls, actions (behaviors) of employees (hourly and salaried) and injuries. In fact countries, industries, companies and work groups all have these kinds of downstream measures we call incident rates. And they are statistically validated.
    As for solutions; I favor a multifaceted approach that goes way beyond single focus initiatives like BBS or regulations. There truly is a complex dynamic that includes solid safety management systems, engagement in appropriate safety activities at all levels in an organization and a relentless pursuit of a zero-incident safety culture across the organization. Not all of this material can be delivered in a short blog article. In the blog article I did allude to some of what I have found to be important to achieving such a zero-incident safety culture:
    • Dr. Dan Petersen’s Six Criteria of Safety Excellence
    o Visible upper management leadership in safety
    o Noticeable involvement of middle management
    o Focused supervisory performance
    o Active participation of hourly employees
    o Flexibility in approach
    o Feedback from the workforce that continues to focus on what needs to improve
    • More research from Dr. Petersen that details 20 statistically proven safety management processes. These are upstream processes that must be error proofed by the hourly and salaried workforce if downstream injury indicators are to improve. Each process solution must include accountabilities (activities/behaviors) that are not treated as the normal fluff of BBS. Rather they are engagement requirements that make a difference to safety performance. Hourly, supervision, middle management and upper management have both responsibilities and accountabilities as a part of this model. I tried to show a few of these as the rock solid subfoundational processes in the last graphic of the blog article: Incident Investigation, Discipline, Supervisor Performance, Communication, Recognition, Accountability.
    • Excellent Safety Management System: the 20 Safety Management Processes are a part of this, as are accountabilities; observations of work; well thought out safety systems that are documented and carefully followed by all in the workforce; safety regulations that are absolutes of a culture of correct; and more that a short article could not deliver
    As the blog article concluded: The key then is not to focus on compliance or to reward “acceptable injury levels/goals,” or to take any other single tactic approach. The key is to concentrate on the foundational fundamentals that eliminate the activities/behaviors that in turn move us out of the chance probability of the Heinrich inevitability triangle. We must error proof the fundamental safety processes and deliver safety accountabilities at each level of the organization. We then engage in a daily practice of the upstream activities that deliver downstream performance.

    D. Johnston you have obviously had a career that includes a passion for excellence in safety. I appreciate and honor your heartfelt comments and commitment. I wonder if there is a way we could actually meet and continue a worthwhile dialogue on achieving safety culture excellence. Thank you for pushing the envelope on this.

    The Doc

  3. I would be interested to know where I can find the 20 safety management processes you speak of in your last response.

  4. The safety management categories measured by Caterpillar's Safety Perception Survey are:

    Attitude Towards Safety
    Awareness Programs
    Employee Training
    Goals of Safety Performance
    Hazard Correction
    Incident Analysis
    Involvement of Employees
    Management Credibility
    New Employees
    Operating Procedures
    Quality of Supervision
    Recognition for Performance
    Safety Climate
    Safety Contacts
    Substance Abuse
    Supervisor Training
    Support for Safety

    For much more detail about the SPS, please visit If you have any follow-up questions, or would like to discuss how the survey is deployed and leveraged, please provide your email address and I will get in touch with you. Thank you for your interest.

  5. Dr. Williamsen,

    I have worked in the mining industry for seven years and I am pursuing my Doctoral Degree from a private, distance college (Walden) and I am working on my problem statement. Please do not judge the quality of the statement; however, any thoughts on the proposed content would be helpful.

    Injury rates in the metal/nonmetal industry have declined since the early 1900s; however, the general problem in the metal/nonmetal industry is that the injury rate has remained steady since 2000 and remains many times greater than the American industry average (Lenne et al, 2012). Hall (1999) reviewed available literature and found no definitive consensus to support whether or not CHS systems have had an impact on the reduction of mine related fatalities and injuries in underground coal operations and no definitive study has been conducted on the affect of CHS systems in the metal/nonmetal industry. The focus of this study is to understand organizational factors influencing safety behaviors in a metal/non-metal mine operation after the implementation of corporate health and safety systems.

    My contact information is,

    Kevin Sleen Jr

  6. I totally agree with your post. While it is true that workers' behaviors could affect the safety process, it should not be the first option in implementing a good safety culture. The pyramid you proposed has all the elements of a good safety culture, though, I would add as a first foundation the application of the hierarchy of controls. I supposed that organisations that reach this level, before unsafe acts, are those ones that pass through the organisational maturity model proposed by Hudson to get into a proactive or generative organisation (Vulnerable, Reactive, Calculative, Proactive and Generative). However, what if the company do not have the resources to reduce the risk at its source by eliminating, substituting or engineering controls, would a contemporary approach to BBS be a solution to this?. Meaning by contemporary approach a model where listening to workers is more important than doing 50 observations per week, for instance.


Note: Only a member of this blog may post a comment.

Connect With Us

Bookmark and Share
/////////////Google analytics tracking script//////////////// /////////////END -- Google analytics tracking script////////////////