Continue reading »]]> Relationship Based Safety: Moving Beyond Culture and Behavior was published in the ASSE Safety Professional December 2012 issue. This six part series will break down the main concepts and explain the concept of Relationship Based Safety (RBS), a new safety management perspective based on findings from the scientific community on the application of Complexity Theory to organizations.  We will cover:

1. Introduction into Relationship Based Safety: What is it? Where did it come from? How is it different than anything else we’ve seen?
2. Complexity Management, Relationship Psychology, and Relational Coordination: How do these theories affect the way we think about safety management?
3. How do these concepts affect our understanding of how to improve communication and collaboration to achieve safety excellence?
4. Practical Drift: the human tendency to drift away from standard procedures without documentation. How is it spotted before an accident? How is it managed?
5.  Putting RBS into place: Practical applications and implementation.
6. The implications for safety leadership: More reliance on collaboration and interdependence requires competence in people skills.

Introduction to Relationship Based Safety (RBS)

Relationship Based Safety (RBS) adds value to the safety management field by providing a more complete understanding of the influences that affect risk assessment, communication quality, and collaboration. Collaboration includes a work group’s ability to recover from failure, and to trust each other.

We define RBS as a safety management approach that recognizes that the quality of relationships in the work place affect safety performance because they determine the openness of communication, level of task coordination, and mutual support during the work process. RBS becomes even more important as uncertainty and time constraints increase because high functioning relationships reduce misunderstandings, allow for divergent perspectives, and encourage full commitment to personal accountability.

During my 20 years of experience and research in the safety culture arena, I frequently came across employees who told me the root cause of a tragedy was lack of trust and open communication. “Management just didn’t listen when we tried to tell them what was wrong.” This seemed a little too touchy feely so I continued the search for more “substantial” root causes. In the meantime, my management clients frequently complained that employees simply didn’t make safety a priority and wouldn’t follow procedure. Basically, each group was pointing at the other as the source of the problem.

As I continued to search for answers I came across complexity management theory, relationship psychology, and relational coordination. Each seemed to hold a piece of the puzzle and led to my starting to use the term “Relationship Based Safety (RBS).”

RBS does not exclude the importance of technical and management systems, or the contributions of behavioral and culture-based approaches. Both culture and behavioral safety recognize the importance of relationships in shaping human behavior—but neither go quite deep enough. We are going to show that the quality of relationships in your organization may be the principal influence in your safety performance beginning with the importance of developing and maintaining an ongoing conversation between senior management, supervision and employees.

Top Down and Sideways Conversation

You will note we are using the word “conversation” as opposed to “communication.” This is purposeful as emails, videos, and texting do not have much impact when it comes to accident prevention. The following safety study provides substantial proof on the benefits of investing in creating a structure to promote relationship building through conversation.

Dominic Cooper (2010), a respected researcher, author, and behavioral psychologist, did a study in construction safety with 14 subcontractors employing 47,000 Third-Party Nationals from India, Indonesia, Malaysia, Nepal, Philippines, Sri Lanka, Turkey and the A19 UAE involved in the construction of 2 X LNG Super Trains. The results guide us quite nicely into the benefits of relationship-based safety.

In the study safety performance showed steep improvement over three years using a safety observation process, but the reasons for its success is the real story. Table 1 shows the reductions in the TRIR rate.

Table 1. Total Recordable Incident Rates (TRIR) by year

Table 2 shows that managerial Safety Leadership impacted employees’ safety behavior, by some 85.6 percent.

The Corrective Action Rate (which employees attributed to management commitment) impacted safety behavior by around 2l.5%. Observer Visible Ongoing Support (VOS) records account for a further 32.4% improvement. (VOS was the documentation submitted by observers regarding the amount of contact they had with each of the management levels). Adding Front-Line Managers Safety Leadership face-to-face conversations into the equation accounted for an additional l9.5% improvement. Middle and senior managers safety leadership conversations accounted for a further 6.7% and 5.5% respectively. 

Table 2. Impact of Management System Activities on Safe Behavior

“The Corrective Action Rate is also a measure of managerial Safety Leadership, as it is they who control the resources for these to be attended to. In practical terms, this means attending to any Corrective Actions reported and reinforcing the perception of demonstrable Safety Leadership with the workforce is very important to improve safety performance.” (Cooper: A22)

This supports other important studies like Zohar’s (2000) and O’Dea (2003) findings that increasing the frequency of management –subordinate safety interactions positively influences safety performance. The study results also show that in construction it is front-line management (supervisors closest to the workers) that has the most influence.

How does RBS add to Behavior Based Safety (BBS)?

As Dr. Cooper’s study shows behavior based safety observation programs can be very successful, but the observation and data collection represent a small percentage of the reported causes. As many companies know BBS observation programs can fail. Large amounts of effort can be spent on programs that produce poor results, including “pencil whipping”, which can lead to a lack of trust and subsequently a lack of action on the observation data by management.

The programs are successful when they are used as a vehicle to promote dialogue followed by corrective action. In Cooper’s study, the interaction that proved most impactful was between supervisors and employees. RBS provides the perspective that it is possible to get the positive interaction independent of a formal behavioral observation process by instituting the expectation that supervisors and managers get out on the floor and make talking to employees their first priority

RBS includes regular audits, walkthroughs, and observations especially for hazard identification and positive social interaction. An observation process can provide a successful structure as long as we expand the concept beyond behavior observation to promoting the development of mutual respect, shared goals and shared knowledge.  We will be discussing the role of these three elements in future blogs. Research has proven them to be essential to safety excellence.

Simply stated RBS differs from BBS by replacing behavior as the core or foundation of loss prevention with the concept of interdependence. What is interdependence? Originally introduced by Stephen Covey into organizations, he defined it as a relationship of mutual dependency. You are reliant on and responsible for each other. We will be talking more about this concept in our next blog.

Practical Applications of RBS

Practical applications from this can be applied to many industries:

• Implement daily conversations between supervision and employees to collect input on hazards, changes in procedures, and concerns.
• Conversation and tracking of corrective actions – Severe hazards need immediate correction, others must be fixed or action plan for addressing the hazard communicated within 30 days.
• For each level of management (Senior, Middle & Front-line) identify roles and responsibilities for communication. The Cooper study showed that senior management communication with employees had less impact than front line supervision, thus frequency of communication is role dependent.
• Select an instrument, such as a perception survey, and dedicate personnel to track effectiveness of conversation on levels of trust and communication.
• Maintain a consistency and focus of safety priorities on an ongoing basis. Do not rely on one-time planning session. Depending on the complexity of the task, several check-ins may be required.
• Select one area of focus to achieve excellence such as housekeeping or elimination of skin contaminations. Watch the success in one area spread to other areas.

Next: Complexity Theory: A deeper understanding of why conversation is a primary driver in safety performance.

Biography

Rosa Carrillo, President of Carrillo & Associates, is a thought leader in transformational leadership for environment, safety and health. She brings 20 years of industry experience with all levels of the organization. Her results and many publications create instant credibility with leadership and the workforce. She is fluent in English and Spanish and is at ease working across many cultures. For more information, please visit her website: http://carrilloconsultants.com

References

Cooper, D. (2010) Safety leadership: application in construction site, Geiornale Italiano de Medicina del Lavoro ed Ergonomia. Vol 32, N. 1: A18-A23

Covey, Stephen (1989). Seven habits of highly effective people.

O’Dea A, Flin R. The role of managerial support in determining workplace safety outcomes. Contract Research Report 044/2003. Sudbury: HSE Books, 2003.

Zohar D. A group-level model of safety climate: Testing the effect of group climate on micro-accidents in manufacturing jobs. Journal of Applied Psychology 2000; 85: 487-596.

Continue reading »]]> The term Lean Production was popularized by Womack et al. to name a Japanese production system that had been developed by the automaker Toyota Motor Company several years after the Second World War. The Toyota Production System (TPS) was inspired by Ford’s mass production system, but deeply deviated from it to suit the socioeconomic reality in Japan after the Second World War (Pestana, 2011). It is about delivering the best value possible to the customer with the minimum resources possible.

Lean strives to achieve this goal through elimination / reduction of waste. The seven types of waste defined by Lean are (Aveta, 2012):

1. Overproduction – This type of waste refers to procedures that should be finished because the requirements have been fulfilled. Instead, the procedures were continued despite the accomplishment of the required goals. This type does not only include product amounts made in excess, but also things produced too early, as well as excess transportation costs.  It is important to know the exact demand of your product or service through research to reduce overproduction waste.
2. Unnecessary Transportation – means having too many transports for a certain material or work in progress transportation.
3. Waiting/Queuing – This type of waste could be described as the inactivity period generated by a processing machine, by a worker who stopped working, or by a function that needs time to be finished.
4. Extra Processing – This type of waste refers to additional procedures that are made after the product is completed because of fabrication defects or bad storage and handling.
5. Motion- Motion waste is the pointless movement of various employees, raw materials, or machines from one place to another.
6. Inventory – Some of the inventory may be unhelpful in the current production order or it can provide only indirect help, which is usually not enough. The inventory range can be from raw state to work-in-progress and finished products.
7. Defects – Products which do not respect the standards of quality imposed by the client are considered defective.

Lean has many principles and tools. One of its key tools is “Value Stream Mapping / VSM”. It employs a flow diagram documenting in high detail every step of a process. Many lean practitioners see value stream mapping as the fundamental tool to identify waste, reduce process cycle times, and implement process improvement. Some organizations treat the value stream map as the hallmark of their lean efforts (ASQ, 2009).

The main drive behind the use of VSM is to enhance value adding by increasing productivity through waste elimination or reduction. Such enhancement of value adding will result in enhanced quality (Hines & Rich, 1997). If improved quality is a byproduct of VSM, could safety benefit as well? Is it possible that VSM could result in safety enhancement (i.e. Lean Safety)?

I think the answer is “Yes”. VSM can be utilized to enhance safety but to do that we need to find where the location of safety is in relation to the ultimate goal of VSM, which is the elimination / reduction of waste. Hines & Rich (1997) state that in relation to value, there are three types of operation:

1. Non-value adding (NVA): pure waste and involves unnecessary actions which should be eliminated completely
2. Necessary but non-value adding (NNVA): may be wasteful but are necessary under the current operating procedures
3. Value-adding (VA): involve the conversion or processing of raw materials or semi-finished products or service requests through the use of manual labor.

If I want to place safety under one of the above three categories, it will be (NNVA)! Yes, that is true and logical despite all the facts we know about the ethical and economical role of safety. However, when we refer to value in line with the above categorization, safety is NNVA. This conclusion may look negative but it is not because safety will be given a “Necessary to Operation” status, so let me call it “Necessary to Operation / NTO” rather than NNVA. The following diagram illustrates the three types of operations within process context.

There are two important issues need to be carefully considered when putting safety as NTO:

1. The positioning of safety as NTO is a management decision / commitment
2. To increase the effectiveness of putting safety as NTO, we need to truly integrate NTO into the stream (workflow) of the value in question

As an example, if I want to improve an automated stream of a specific value through the elimination of NVA to improve productivity, improve value, quality and enhance safety (all in one kick), I need to design that stream workflow so that the NTO is given the status of being a validation factor. What this means is that when the stream reaches a point where NTO is required, the stream can’t progress unless such NTO is met. In other words “NTO” will be embedded though true integration into the value stream as a validation factor. Such true integration is more than just “addition” or a “bolt-on”. If I can use the analogy of my Toyota car, I would say that: 

• The car seat belt as a sub-system is added but not integrated into the car system because I can start my car engine and drive away without putting on my safety belt
• The car gear sub-system is truly integrated into the car system because starting the car engine will not be allowed to start unless the gear is on “P” position

True integration of NTO (safety) into a stream is illustrated by the following diagram:

The position of Safety NTO within an existing workflow can be achieved through a comprehensive study and re-engineering of the workflow in question. Effective participation of all those who are involved in the re-engineering process is essential for the success of the embedment of safety NTO into workflow.

Moreover, technology can facilitate such embedment. An example of the utilization of technology to integrate a safety aspect into a workflow is the integration of permit-to-work (PTW) into maintenance service request where workflow is re-engineered in a way that makes PTW a validating factor. If it is not satisfied, the system will not create a work execution order. The following table describes the key steps of such workflow: 

To conclude, I would like to emphasize that Lean and its tools can give safety a significant control status not only by considering it as NTO but also by integrating it into value stream map in a manner that makes it a necessity.

It is obvious that the focus of this approach is the workflow (procedure), which means that other safety enhancement tools like training / awareness, monitoring / measurement and continuous improvement are still very important.    

References:

Ana Catarina Viriato Maia Ferreira Pestana (2011). Application of Lean Concepts to Office to Related Activities in Construction. Faculty of San Diego State University

American Society for Quality (2009). Profitable Applications of Value Stream Mapping Tutorial [Online]. Accessed on 22 February 2013 at: http://asq.org/learn-about-quality/lean/overview/value-stream-mapping.html

Aveta Business Institute (2012). Explaining the Seven Types of Lean Waste [Online]. Accessed on 9 March 2013 at: http://www.sixsigmaonline.org/six-sigma-training-certification-information/articles/explaining-the-seven-types-of-lean-waste.html

Michael Fisher, (1999),”Process improvement by poka-yoke”, Work Study, Vol. 48 Iss: 7 pp. 264 – 266

Peter Hines and Nick Rich (1997). Lean Enterprise Research Centre, Cardiff Business School, Cardiff, UK. International Journal of Operations & Production Management, Vol. 17 No. 1, 1997, pp. 46-64

Peter Hines, Nick Rich, John Bicheno, David Brunt, David Taylor, Chris Butterworth, James Sullivan, (1998),”Value Stream Management”, The International Journal of Logistics Management, Vol. 9 Iss: 1 pp. 25 – 42

Richard L. Daft and Arie Y. Lewin, (1993). Where are the Theories for the New Organizational Forms?  Organization Science, Vol. 4 No. 4

Wikipedia (2013). Dialectical Materialism [Online]. Accessed on 18 February 2013 at:  http://en.wikipedia.org/wiki/Dialectical_materialism

Biography

Amjad is a qualified QHSE Professional (MIIRSM, Grad-Iosh, Honor Degree in Electrical Engineering, NEBOSH National Diploma, PG Certificate in TQM & Organizational Excellence (M.Sc. in progress)), with around 14 years of HSE/Q experience within the properties development, facilities management and utilities sectors. He currently serves as the QHSE Manager in Dubai for Farnek L.L.C. (Total Facilities Management) – a provider of comprehensive facilities management services (http://www.farnek.com/home.php).

No matter how much traditional safety “stuff” you do, it doesn’t necessarily mean you have a functional safety system.  Without such a system, however, safety results are largely dependent on luck and good intentions.  This is not a strategy for safety success.

Webster tells us that a system is “a …

Continue reading »]]> What Is a Safety System?

No matter how much traditional safety “stuff” you do, it doesn’t necessarily mean you have a functional safety system.  Without such a system, however, safety results are largely dependent on luck and good intentions.  This is not a strategy for safety success.

Webster tells us that a system is “a regularly interacting or interdependent group of items forming a unified whole.”   Like any effective system, safety systems must include goals and objectives with plans (P), for how every level of the organization will contribute to those goals and objectives by doing (D) specific activities. To provide for accountability and to ensure those planned activities actually happen and that they are effective, they must also be measured and evaluated – commonly known as check (C) in Deming parlance. These three steps are vital for any organization that hopes for continuous improvement by acting (A) on what they learn in their ongoing system evaluations.  Systems are not static, however, but are constantly transforming themselves as they learn and continually improve.  Most commonly such systems are known as PDCA systems.

 PDCA Safety System

Since most organizations want to stay in business, they have long recognized the importance of nurturing a PDCA system for their business imperatives.  Inexplicably, however, many of these same organizations fail to manage safety in the same manner.

Check: The Missing Link

Time and again the author has witnessed organizations start on a safety management system only to bog down in the details without completing the process.  Most commonly they lack an effective “check” step.  The check step goes by many different names but is absolutely essential for any functioning system.  Dr. W. Edwards Deming, often credited as the father of the plan, do, check, act cycle, actually preferred “study” to check but eventually yielded to the wishes of his Japanese customers.  Six sigma programs use terms like “measure, analyze and improve.”  The ANSI standard for health and safety management systems (ANSI/AIHA Z10-2012) calls for an “evaluation and corrective action” step and the DOE uses the term “feedback and improvement” for the same process.  Whatever you call it, a check step is vital for establishing accountability, as well as for acquiring the feedback necessary to fix and continuously improve the entire system.

Unfortunately some organizations appear reluctant to self assess safety in a meaningful manner.  As a result even companies that establish excellent safety goals and objectives, complete with comprehensive (and often expensive) implementation schemes, fail in their efforts.  They fail because they lack a process to measure and analyze their progress.  Safety assessment is left to the safety staff, if it is done at all.  Managers may know their accident rates but have little understanding of how well (or poorly) they are doing in regard to their organizational safety goals and objectives.  Therefore they don’t know what to fix, improve, do away with or celebrate – leaving only hope that needed corrective actions and process improvements are implemented.  I’ve seen this “don’t ask don’t tell” safety approach so often I’ve actually given it a name – the Plan, Do, Hope, Pray (PDHP) process.

Why the Missing Link?

There are, no doubt, many reasons for the PDHP approach seen in so many organizations, but safety personnel are often part of the problem.  Commonly the safety staff hands off a set of traditional lagging indicators to a passive management.  These metrics are then accepted by management as all the evidence needed to support the staff’s conclusion (“just tell us if it’s safe”).   It sadly surprising how many managers/supervisors don’t know what is really going on in the field and are content to read reports, issue memos, give speeches, and then hope their operations are performing safely.

There Is a Better Way

In the author’s many years of organizational assessment, the best (i.e., safest) organizations employed a variety of leading and lagging indicators but often went beyond the numbers to include management discussions (not safety staff sermons) on what managers are seeing in the field and specifically what they are doing to ensure improvement based on those observations.  Managers (again, not the safety staff) were also expected to discuss any accidents, incidents, near misses and negative trends as well as the status of associated corrective actions.  These discussions stressed concrete actions to improve the safety system and help ensure continuous improvement.

The Importance of Observations

You can only go so far with performance indicators.  Even the best crafted leading and lagging indicators are no substitute for actually getting out, observing work and interacting with the workers.  Every organization must have a real time way to inform itself of how well, and safely, its operations, especially high consequence operations, are conducted. In high reliability organizations such as commercial nuclear power this is known as operational awareness. An effective safety system check step is not possible without it and you can’t get it from behind a desk.  Unfortunately many organizations don’t seem to understand the importance of operational awareness, not only to safety, but to their very existence.

There are three types of safety observation in common usage:

1. Traditional OSHA-type inspections that focus mainly on conditions
2. Behavioral-based observations that are typically performed by employee peers and focus on a set of pre-determined “critical behaviors.”
3. Observations of work and work processes – commonly known as management walkarounds but sometimes performed by safety staff, safety committee members and others.  (See Safety Management by Walking Around Part 1, Part 2, and Part 3 in earlier SafetyCary articles).

The three types of field observations are not mutually exclusive and all three can add safety value as well as provide useful input to the safety review implied in the safety system check step.  The remainder of this article, however, will address #3, observation of work and work processes.  The author has found this type of observation the most valuable both in identifying root cause safety issues and in gaining assurance that process safety and safety critical operations are appropriately addressed.

Why Observe Work?

Managers have an obligation to understand how safely the work they are responsible for is performed. Lack of such an understanding has been a recurrent and principle factor in recent tragedies from the Columbia and Challenger disasters to the Deepwater Horizon.  A genuine understanding of safety performance cannot be delegated or achieved solely from accident statistics.  Nor can this understanding be gained through traditional compliance inspections that focus on conditions or some predetermined subset of behaviors.  A deeper understanding of the work, and the systems and processes that support it (or not), is needed and should be the principle goal of work-focused walkarounds.  Conditions and behaviors (see observation types 1&2 above) are not ignored but the emphasis is on observing work.  The goal is to understand how the work is actually performed and then to partner with the employees to gain additional understanding of the work, its hazards and the adequacy of the controls – including the use, misuse or nonuse of procedures, training, equipment, environmental factors etc.  Well structured walkaround programs emphasize the importance of partnering and actively listening to those performing the work.  Worker engagement is prompted with questions such as, “What is the worst thing that could happen on this job?”  This cooperative and “fault free” approach helps ensure employee cooperation in finding safer and better ways to perform the work.  Effective walkarounds are thus performed with employees – not to them (a tip of the hat to Alan Quilley for this turn of phrase).

How important is operational awareness?  Consider the Deepwater Horizon where 11 employees were killed and tens of thousands of barrels of oil spilled into the Gulf of Mexico for 87 days. 

Just seven hours prior to the Deepwater Horizon explosion, top managers from both BP and Transocean were on the drilling platform walking around. They were focused, however, on conditions such as fall protection devices, housekeeping and trip hazards, and a specific employee behavior, glove use.  Missing from their walkaround was any serious attempt to find out how the very critical well capping work was progressing, thus tragically missing an opportunity to deal with the many safety issues that were affecting the capping effort.  Subsequent investigation by the Chemical Safety Board found BP overly focused on common personal injuries and behavior – to the detriment of process safety as well as more serious safety issues.  This finding was very similar to an earlier finding at BP’s Texas City refinery where an explosion killed 15 in 2005.

Summary

There are no silver bullets in safety, but a systematic PDCA approach is essential to success.  And you don’t have a working safety system without a vigorous check step.  Observations, especially observations of work, put reality into your check step data.  Without this reality-based safety review you can hope your work and critical processes are safe, but you’ll never really know – at least not until it’s too late.

* Image courtesy of Deepwaterhorizon News, Video and Gossip – Gawker

Biography

Mr. Loud’s (jjl7280@aol.com) over 40 years of safety experience includes 15 years with the Tennessee Valley Authority (TVA) where he served as the supervisor of Safety and Loss Control for a large commercial nuclear facility and later as manager of the corporate nuclear safety oversight body for all three of TVA’s nuclear sites.  At Los Alamos National Laboratory he headed the independent assessment organization responsible for safety, health, environmental protection, and security oversight of all Laboratory operations.  Mr. Loud is a regular presenter at national and international safety conferences.  He is the author of numerous papers and articles.  Mr. Loud is a Certified Safety Professional (CSP) and a retired Certified Hazardous Materials Manager (CHMM).  He holds a BBA from the University of Memphis, an MS in Environmental Science from the University of Oklahoma and an MPH in Occupational Health and Safety from the University of Tennessee.

Continue reading »]]> In the first article of this series, “Reflections on Safety Barriers”, we discussed some general principles of safety barriers and proposed a wide operational definition. The second article, “Assessment of Safety Barriers”, described one possible way of doing a hazard and safety barrier assessment, including a practical example. In the third and, for now, final part of this series, we’d like to discuss some possibilities of how to strengthen the management of safety barriers in an organization.

Knowing and understanding barriers

Safety management is often approached from a somewhat negative perspective. One of the essential elements of good safety culture and good safety management is knowing your challenges, usually based on risk assessments, incident reports and the like. Nothing wrong with that, but it’s also important to know your strengths, and the barriers you have put in place may just be that.

Norwegian railway safety regulations stress that “Barriers shall be identified, and it shall be known in the organization what barriers have been established and what function these have” (Sikkerhetsstyringsforskriften § 2-3). While this requirement may be rather idealistic and hard to fully comply with to the letter, we do agree with the bottom line. Understanding of barriers in all their variations; be it Man, Technical or Organizational (MTO) is an essential factor. After all, how can you manage something you don’t know about? Awareness of some of the misunderstandings discussed in the first article may be an important factor.

Incident investigation

Probably the most basic feedback tool on barriers is incident investigation. Having models like the Swiss Cheese Model in the back of our mind, one may, after all, reason that if an incident has happened, then one or more barriers appear to have failed.

Jernbaneverket has chosen PRISMA[1] as its taxonomy for root causes. This taxonomy is structured after the MTO-principle and if the investigation is of a good enough quality, it’s often easy to relate identified underlying or root causes to barriers that have failed in one way or another. Let’s look at a hypothetical, yet life-like, example.

Rail traffic safety depends a great deal on the signaling system. Missing a ‘red light’ can mean a critical mistake and bring a train in the path of another train, annihilating the carefully planned separation in time and space. This is a reason that much of the rail infrastructure is protected against this type of event, attempting to stop a train as quickly as possible after passing the signal. A SPAD (Signal Passed At Danger) is probably the most important kind of precursor incident to a collision between trains[2] and will generally be carefully investigated.

The diagram below shows a causal tree of a fictional case where a driver misses the signal because of two direct causes: he was distracted and the signal was hard to see. Working down towards root causes (very simplified) we find that the driver was texting while driving (a conscious violation of a known rule) and that the signal was hard to see because of three underlying factors: there was vegetation that covered part of the signal due to neglected maintenance, the signal was placed in a curve, and at the moment of the incident the driver was blinded by a low standing sun. The blue boxes below the causal tree indicate the PRISMA categories. Relating these to barriers we see weaknesses in Design, Maintenance and the performance of the job. We even can consider altering the standard design of the signal (e.g. by attaching a larger back screen) such that sunlight isn’t so big a factor anymore.

And of course we don’t need to limit ourselves to information from one incident. By looking at trends in your safety statistics you may find out that some locations are more vulnerable than others, or that a particular (root) cause occurs frequently indicating a weak spot, e.g. in the management system.

Find the weak spots in your barriers actively

Systematic and proactive safety management means that we’re not waiting for incidents to happen before we take action. One way of knowing the weak spots in barriers is obviously through analysis and assessment as discussed in the second part of our article. But there are many other methods.

Leading indicators such as audit, inspection and observation programs, maintenance records, competence management and training and many others provide useful information. Matching the HSEQ-database and other databases to barrier assessment/identification may be a goldmine of information.

One may even consider dedicated monitoring of certain critical barriers. An example is presented below. Starting in 2012, Jernbaneverket monitors a special routine where trains are allowed to pass a signal that is temporarily out of order. This means that technical barriers are temporarily disabled and that safety mainly depends on good communication routines. By monitoring these cases it’s easy to identify weak links in the infrastructure where failures are more frequent[3], and start actions to improve the situation. As an example, one could name the second and third highest locations. These two locations are related and the diagram clearly shows that most events occurred in the same period, January 2012. A bottleneck in supply of spare parts and shortcomings in the routine to handle the damage were revealed. Both were remedied.

Gathering this data for the first time also made it possible to correlate the data to available information in the maintenance database and information about the (less frequent) occurrence of SPADs. There turned out to be a significant correlation for a number of locations, so this new monitoring enhances the possibility for risk management.

Recovery factors

Traditional incident investigation tends to look at what went wrong. A rather unknown, and little used, ‘enhancement’ of investigations is to look at what went well. This can in many cases give you information about barriers that worked, in some cases you may even discover barriers that you weren’t aware of before and which you can further improve on and can be strengthened.

The PRISMA taxonomy additionally provides a number of so-called recovery factors. A recovery factor is something that explains why an incident didn’t develop its full potential; why an incident didn’t lead to injury or greater damage. Recovery factors are the things that influence the course of an event positively and save the day, so to say. Recovery factors may be mechanisms put in place willingly to detect failure and counteract in a timely manner, but quite often recovery factors are not planned beforehand.

Looking at recovery factors means that one doesn’t just look at what has happened, but one takes into account what potential there was in an incident and why it didn’t get that far. Emergency response routines are a typical example. Evaluating those in relation to incident investigations often gives valuable information about possible improvement.

The PRISMA taxonomy (adopted for Rail) contains the following recovery factors:

If we revisit our SPAD incident we can elaborate on the causal tree. The SPAD incident could have resulted in a collision between two trains with a number of possible outcomes (people dead or injured, material damage to rolling stock and infrastructure, delays) not specified in the diagram. This, however, didn’t happen because the automatic train control system (ATC) detected the passing of the signal at danger in time and stopped the train within the safe distance. This clearly was a recovery factor that was planned beforehand.

In 2004 Dutch railway infrastructure manager commissioned an analysis of 89 near-miss incidents with track workers (near collisions) reported in 2003. The aim of the analysis was to see if some common root causes could be found. One finding was that it was possible to clearly identify recovery factors in 45 of the cases. The two most important recovery factors were related to successful unplanned interventions by employees:

• Track workers noticed a nearing train and saved themselves or their colleague by pulling out of the tracks in time.
• Train drivers noticed the work crew and signaled by ‘honking’ which prompted the track workers to clear the track.

Identifying these mechanisms gave the possibility to strengthen them, e.g. as a part of training. Interestingly, the ‘honking’ action wasn’t described beforehand, neither in formal training, nor regulations; this was something that people ‘just did’.

In our experience many safety professionals do focus very much on things that went wrong. While there’s a lot to learn from things that went well, as illustrated by the examples above, looking at recovery factors (barriers that worked) is an element of incident investigation that is still rather unknown, even though it is facilitated by other methods besides PRISMA, and should be used more widely.

—+—

This concludes the final part of this three-part article on barriers. We realize that space limitations caused us to leave many important and interesting things unsaid. We hope to have given a few suggestions for reflection and possible improvement in organizations. Feedback and discussion is warmly welcomed by the authors.

——————————————————————————————————————-

Biography

Carsten Busch has studied Mechanical Engineering and after that Safety. He also spent some time at Law School. He has over 20 years of HSEQ experience from various railway and oil & gas related companies in The Netherlands, United Kingdom and Norway. These days he works as Section Head Safety and Quality for Jernbaneverket’s infrastructure division.

Beate Karlsen has studied Occupational Health and Safety at Haugesund and at the Stavanger University. She has been in various OHS functions in Jernbaneverket and works currently as Senior Advisor Safety in Jernbaneverket’s infrastructure division.

Continue reading »]]> Wallflowers are under appreciated.  I’m not talking about the rock band fronted by Jakob Dylan, son of 60’s rock legend Bob Dylan.  No, I’m talking about the unassuming, shy, serious, soft-spoken, thoughtful folks that would rather read a book than attend a party.  You know, the sort of introspective-types that tend to spend their energy quietly going about their days inconspicuously.  These docile listeners are commonly labeled “introverts.”  It’s a shame that society mistakenly overlooks these quiet ones’ contributions and overly acknowledges and rewards the aggressive, outspoken, hard-charging social go-getters.  Think about it:  the social spot-light shines brightest on the accomplishments of successful attention seekers like Donald Trump, Shaquille O’Neal and Dr. Martin Luther King, Jr. – all unashamed extraverts.  At the same time, society doesn’t often immediately recall the quiet, yet significant, accomplishments of Warren Buffet, John Stockton and Rosa Parks – all accomplished introverts.

Not to put too fine a point on it but think about the three aforementioned introverts for a moment.  The business world has benefited greatly from Warren Buffet, the NBA is tremendously grateful for John Stockton’s legendary career and the contribution of Rosa Parks to the Civil Rights movement is immeasurable.  These “wallflowers,” “introverts,” “quiet ones” or whatever you may call them, are outstanding in their own unassuming way. Their contributions to the world speak just as meaningfully as their noisier, more recognizable counterparts.  Great accomplishment is not just relegated to attention seeking extraverts.  Meaningful contribution to life’s conversation echoes as clearly from the wallflowers of the world as it does from the soap box addicts that constantly campaign for your attention.  So what does all this have to do with safety?  Glad you asked!

The message of workplace safety is boldly declared loudly and clearly from company safety representatives each day.  Businesses hire these subject-matter extroverts to represent the company’s interests in keeping people safe and the good ones do just that – loudly and clearly!  These toastmasters of risk that enthusiastically perform their craft well will observe workplace hazards, grab the nearest megaphone and shout a warning from the roof tops for all to hear.  This is actually a good thing; however, many organizations carelessly delegate the important duty of identifying workplace hazards to only company safety representatives and sadly ignore the softer-spoken, discounted warning from the more numerous employees at large.  These shyer hazard-identification voices, if overlooked, are a huge missed potential for reducing risk.  That won’t do!

So true it is that safety professionals are typically the ones most ideally prepared and equipped to identify workplace hazard.  I won’t challenge you on that point; I am one of those safety professionals!  It is also true that when an organization assigns the task of hazard identification to others besides the safety representative, workplace injuries reduce in a fantastic way.  The safety gurus and data heads at Predictive Solutions have proven this point to be true and call this phenomenon “Safety Truth #2.”  The top of the graph shown below happens to represent actual data collected by Predictive Solutions illustrating organizations that assign only the safety professionals to seek out hazard in the workplace.  See how many unwanted incidents those businesses experience?  Now take a look at the bottom of the graph, as diversity is increased.  When hazard identification is assigned to a diverse group that extends beyond just the safety specialist, those unwanted incidences begin to shrink.  You can’t ignore those results.  Troubling workplace incidents reduce when a diverse group of employees join together to find and remove hazard from the workplace.  Now that’s a truth I like!

Some savvy practitioners of loss control would classify the results of achieving Safety Truth #2 as effective employee involvement in the safety program.  I completely agree.  To those same safety practitioners that tirelessly preach this truth and message of safety day-in and day-out, garnering management commitment and soliciting employee involvement to safety, I say preach on!  Let us unite our voices to strengthen our message and gather more and more confederates to our cause.  And, while we’re uniting and hollering, as we safety extroverts like to do, let us not forget to include the wallflowers in our efforts to weed hazard out of the workplace.  Their quieter, softer-spoken but legitimate help will go a long way and toward keeping organizations safe and hazard free.

You can’t deny the business world would be poorer without investing icon Warren Buffett, the NBA would be seriously lacking without the assists-and-steals leader John Stockton, and I don’t even want to think of where we’d be right now without the quiet-yet-bold actions of Rosa Parks.  These high-achieving introverts leave a powerful impression on alert admires of their individual realms.  In the realm of safety let us also empower the safety introverts and listen to their hazard identification message loud and clear.  As we pay attention to these safety wallflowers, the truth of hazard reduction will be realized and greater workplace safety will be achieved.

Live Safety!

Biography

Rob Loose is a Safety Professional with Manufacturing, HR and Health Care management experience. Rob has worked for MAU Workforce Solutions since 2006 supporting both HR and operations functions.  Currently he serves as MAU’s Corporate Safety & Risk Manager.  A 1998 graduate of Brigham Young University in Provo, UT, Rob now calls North Augusta, SC home where he lives with his wife and  two kids.   Rob is a member of the Augusta, GA chapter of ASSE.

Continue reading »]]> Barrier assessment in practice – an example

In the first part if this trilogy, Reflections on Safety Barriers, we argued that focus on safety barriers is one way to deal with some of the shortcomings of traditional risk assessments. We further talked about general principles and posed a wide operational definition and discussed some misunderstandings. In this second part we’d like to take a look at a practical example and discuss a possible method.

Background

In June 2011 we had a fire at Hallingskeid[1] in an artificial snow protection tunnel up in the mountains leading to the loss of two train sets and causing severe damage to railway infrastructure (not to mention the delays). Following this incident we had to evaluate preventive actions we had considered in the past and for the future. What better way than barrier assessment to see what was in place, what could possibly put in place, and how good these barriers intended to control the fire hazards actually are?

No method was immediately available, but we had already been thinking about a barrier-based approach to assessments and been toying around with some ideas, so the project gave a positive boost to try out some ideas. The approach chosen was very much inspired by two wider known tools: fault tree analysis (FTA) and the BowTie XP tool.

We experimented briefly with a traditional FTA where we included barriers in the various branches. FTA is a very powerful tool, but we experienced difficulties to phrase events/failures and barriers in such a way that they made sense and weren’t expressed in ‘forced’ statements. One possibility is building all branches with a combination of hazards and barrier failures combined by an AND gate, but this makes for an awkward tree and neither did we like to see all our barriers phrased in negative statements.

Our tool of choice would have been BowTie XP because of its ease of use. One disadvantage of this piece of software, however, is the non-linearity of the branches in diagrams which makes them more difficult to understand. They are not really intuitive and require training and continued experience. An additional problem was that we had neither the budget for acquisition of the software, nor the time for the process with procurement and ICT that alas is mandatory in a public company.

Method for hazard and barrier assessment

Hazard and barrier assessment (HBA) was considered a good systematic tool to identify hazards and barriers and assess possible weaknesses in the system. The HBA we ‘developed’ had to be qualitative, easy to use, explain and understand (thus be visual) and be cheap (i.e. done within the possibilities of available generic software). We settled for kind of a hybrid of the FTA and BowTie methodologies which we refined as we went on.

The process for this HBA is as follows (numbers corresponded to the diagram below):

1. Choose the unwanted event you want to assess, often some kind of incident/accident. Be specific about the situation, object or job in order to clarify your scope and to keep the assessment within workable limits (so choose “fire in the workshop” instead of “fire”).
2. Identify possible consequences of the unwanted event.
3. Identify possible barriers that prevent or mitigate these possible consequences of the unwanted event. These can be specific for one or more consequences, or common for all consequences.
4. Identify possible initiating events or direct causes of the unwanted event. To do this it’s practical to go through all hazards/possible sources of energy as well as through the various sub-systems.
5. Identify possible barriers that can stop these initiating events from growing into the unwanted event.
6. Identify possible barriers that prevent the initiating event from occurring.
7. Identify possible (underlying) causes that may cause the failure of the (in steps 3, 5 and 6) identified barriers. This may be several causes per barrier.
8. Identify possible barriers against these causes for barrier failure.
9. Repeat the previous steps as often as necessary/practical.

The identified elements are drawn in chronological order, such that a preventive barrier precedes an initiating event (unlike with BowTie XP). Causes for barrier failures are drawn on a ‘lower level’ than the barrier they affect in order to create a clearer view. The result will be something that strongly resembles a BowTie diagram. We decided to color code the various items such that it’s easy to separate events/causes and barriers and also distinguish between the various types of barriers (e.g. M, T and O). We used the Visio package, but other graphic packages should do nicely and a roll of brown paper with colored sticky posts will be perfect as well.

The various elements (failures, causes, events and barriers) identified in the assessment can be described in further detail in a table. These detailed descriptions discuss the factors that affect the quality of the barriers, the probability of initiating events and the party responsible for the barrier, among other things.

The model can be built into infinite detail, of course, because it will be often possible to elaborate with new barriers and new possibilities for failures. Further than a few levels, however, is hardly ever practicable because the diagram will become unreadable, or the additional information is not really worth the effort. As a general stop-rule one shouldn’t go further than elements that the organization can control, or influence. One possibility to handle a higher degree of detail is to include this into the descriptions/assessments of the various items in the table. In several cases we found it more practical to describe possible barrier failures as part of the assessment of the robustness of the barriers in the table, rather than drawing additional branches to the diagram.

One important difference with FTA is that we chose not to use OR and AND gates in the diagram. In principle the diagram shows possibilities without taking into account all dependencies between various branches or elements. This may cause some minor challenges in some cases as illustrated in the example below. These can for example be resolved by combining or splitting elements.

Essential in the process is the involvement of relevant competence. This is one way of minimizing and identifying uncertainties. The same can be said about systematic and structured critical checks to ensure as great as possible completeness (e.g. by going through all possible energy sources/hazards and barriers).

An example

An example is provided below. Since space is limited we’ll do a simple version of a comparable situation from the Hallingskeid fire assessment – applied to your garage at home. The example is heavily simplified and all of the elements could be seriously elaborated.

As the central element we chose the unwanted event of a fire in your garage which is located on a good distance from your home or other buildings. Consequences of this event could be: loss of your garage (7.1) and/or car in it (7.2), and material damage (7.3, for instance damage to car or garage, but not complete loss of either). There are a couple of possible barriers against these consequences: the fire can be detected with a smoke detector (6.1, although presumably only few of us have one in the garage), you can detect it yourself (6.2), and when the fire department (6.3) is called in they can limit the damage depending on how fast they can get to you and if water is available[2]. Interesting, of course is that these barriers are not independent from each other. Detection without follow-up is useless while the best fire department around the corner is of little help if they’re not called by someone[3].

On the left side of the unwanted event we identified four main sources that could lead to a fire in the garage:

1. a failure in the electrical installation (e.g. a shortcut),
2. a failure in your car (e.g. an overheated brake),
3. an ignition during some do-it-yourself activities that you do in your garage, or
4. an external source (e.g. arson or lightning strike).

Now your garage will probably be constructed in such a way that it has some form of natural resistance against catching fire just too easily, by choice of materials. This has been placed as a barrier (5) between the various sources of ignition and the unwanted event, but this barrier can degrade, for example through aging, wear and tear (5.1), or because of the fact that you use the garage as a storage for various inflammable materials (5.2) like paint, gas and paper boxes.

You may discover that it’s pretty hard to have barriers against some hazards, like arson (4.1) in our example, and maybe we should have added ‘financial losses’ as another consequence to the right and ‘insurance’ as a barrier against this. For other hazards the barriers are clearer: if you are skilled in your DIY-activities (3.3) and you use good tools (3.2), the probability of a fire will be lower and if something happens you’re probably in the vicinity and can extinguish the spark before it develops into a real fire (3.1). Also are both your garage (including the electrical installation) and your car designed in such a way that they won’t cause fire spontaneously (1.1 and 2.1). Degrading through age and use (1.1.1 and 2.1.1) will increase the probability of failure, however, which can be countered by regular checks and maintenance (1.1.2 and 2.1.2).

This was just a really quick run through of the example. In real life you want to do a more thorough assessment on how serious the various hazards are and how good your various barriers are and maybe you will want strengthen them, e.g. by storing you paint in a steel cupboard or replacing that old switch you ‘fixed’ with duct tape last year.

Experiences so far

Concluding this article, our experiences with this ‘new’ way of doing hazard and barrier assessment so far are positive. It’s easy to use, both as a support for a structured brainstorm session and for more (qualitative) analytic purposes. The visual diagram makes it very easy to present, explain and understand and the use of various colors, icons and symbols enables elaboration in the diagram while a corresponding table provides even more detailed information about the elements. An additional advantage is that a generic hazard/barrier assessment can quickly be made specific for one location, and updated after changes to the system have been made.

One clear disadvantage is that there is no software that supports easy maintenance of the assessment. The way we had to work with a Visio/Word combination as the software basis was cost-effective at the moment, but may not be that in the long run.

The ease of use, the database structure behind it, the possibility to expand/collapse branches and other user features make for a strong argument to use the commercially available BowTie XP tool[4] – despite the drawback with regard to non-linearity/being not chronological.

Another challenge may be the thing with dependencies. Our aim is to gather further experience in the course of 2013 by applying this kind of hazard/barrier assessment to other types of unwanted events (e.g. occupational safety) and refine the method through discussion with more safety professionals inside and outside our organization.

—+—

Why do so many organizations continue to take a reactive approach to safety? Why is it so hard to maintain safety as a priority?  Why do supervisors and managers who truly care about safety behave in ways that contradict their values? Remarkably, how we measure safety is a primary root …

Continue reading »]]> The Case for Leading Indicators

Why do so many organizations continue to take a reactive approach to safety? Why is it so hard to maintain safety as a priority?  Why do supervisors and managers who truly care about safety behave in ways that contradict their values? Remarkably, how we measure safety is a primary root cause for all of these problems.  Incident rate, lost time rate, severity rate and other lagging indicators are poor measures of safety.  Such measures tell us how many people got hurt and how badly, but they do not tell us how well a company is doing at preventing accidents and incidents.

While lagging metrics are necessary, adding leading metrics enables better management of safety.  Leading metrics should focus on proactive activities on the part of all employees—measures that track what people are doing daily to prevent accidents.

The idea that shifting the focus to leading indicators will improve safety management is not new.  Some organizations have made good strides toward developing leading indicators and reducing their dependence on lagging indicators. But what should those indicators be and how can we create meaningful measures?

The quality of the leading indicators is important.  When it comes to measurement, we often choose the obvious or easy measure.  For example, if we want a measure of safety training we track what percentage of employees attended essential training classes.  While attending safety training is clearly a prerequisite, just attending training doesn’t guarantee that workers truly learned what they need to learn to work safely on the job.  Tracking attendance at safety meetings suffers the same flaw.

Even the behavior-based safety (BBS) measures that organizations choose to focus on are often the least meaningful.  Tracking number of observations completed provides interesting information, but if that is the only measure of your BBS process then you are missing the most critical information.  What has improved as a result of your BBS system?

If we are going to take the time to create leading indicators of safety then we should take the time to make those metrics meaningful and impactful. The following criteria can be helpful in guiding you toward better metrics.

Leading indicators should:

• Allow you to see small improvements in performance
• Measure the positive: what people are doing versus failing to do
• Enable frequent feedback to all stakeholders
• Be credible to performers
• Be predictive
• Increase constructive problem solving around safety
• Make it clear what needs to be done to get better
• Track Impact versus Intention

The last criterion is important.  Always ask yourself: what are we trying to accomplish with this component of our safety system?  Then ask if your metric is assessing whether you accomplish what you set out to do.  Are you measuring the impact or just the good intention?

It Ain’t Easy

Shifting your focus to leading indicators is more than just swapping incident rate for another measure.  It means shifting from one measure to multiple measures.  Most organizations use a composite scorecard or dashboard that captures the most important proactive components of their safety system, including one or two lagging indicators.  Of course, more measures means more time. Identifying and tracking leading indicators takes time.  Incident rate is easier, but easier is not better.  The shift to leading indicators also requires getting comfortable with softer measures.  Many organizations shy away from leading indicators because the metrics are often more subjective and easier to manipulate (pencil whip).  Both are true but imperfect leading indicators are better than lagging indicators alone (and do we really think that incident rate can’t be manipulated?).

The move to leading indicators really is a journey.  It is not a “one and done” event. It is advisable to use good criteria for leading indicators (like the ones above) to create some measures and then start using those measures.  Over time you can improve them to best meet your needs and minimize concerns. There is no such thing as a perfect leading metric, so don’t waste time trying to find or create one.

One Size Does Not Fit All 

Every organization will have its own set of leading indicators that best fit their industry and where they are on their safety journey.  That said here are some categories of leading indicators that are common.

• Hazard identification and remediation.  Measure how effectively hazards get reported and, more importantly, how quickly those hazards get remediated.

• Safety Interactions.  Measure the frequency and more importantly, the quality of management-frontline interactions and peer-to-peer interactions.  These may come from BBS systems or other sources, but the interactions should be feedback interactions around safety and be skewed heavily toward positive interactions (as measured by the frontline).

• Pre-job Safety Meetings.  Create criteria for what makes a really effective pre-job safety briefing (tool box talk, pre-shift meeting, or what-ever “just-before-we-do-the-work” meetings you use) and then use those criteria to measure and improve the quality of these important meetings.

• Near Miss Reporting.  So much can be learned from near miss reporting it is worth measuring and refining this important component of safety.

• Action Items from Incident Investigations.  The silver lining of incidents is the opportunity to learn and prevent future incidents. Tracking how quickly and effectively those preventative actions are completed is a good leading indicator.

Impact-orientated leading indicators like these may take more time to create, however they allow for better, more proactive safety management and will ultimately drive better results.

Making Leading Indicators Work

A familiar saying is: “what gets measured gets done”.  As we have just argued, what we measure is important because it sets the stage for what people do.  However, it is not completely true to say what gets measured gets done.  Many of us routinely measure our weight and don’t do anything to improve it.  Similarly many of us measure our speed (with our car speedometer) and don’t drive the speed limit.  In truth, measures don’t change behavior, consequences do. Proactive safety activities must be reinforced if they are to persist.  A positive accountability system that holds people accountable for the leading indicators will ensure safety targets are met.

Keep in mind; the measures are less important than the discussion around what was done to achieve the measures (i.e., the behaviors) and on-going assessment of impact. Frequent, brief conversations about the leading indicators at all levels of the organization are a key to improvement. Frequent conversations establish the safety activities as priorities and allow real-time coaching for improving performance.  The focus on impact ensures the activities are truly driving safety results and are therefore worth the time and effort.  The focus on impact also helps performers see the small improvements they might otherwise not attend to; thereby helping them come into contact with the natural reinforcers of their efforts.

Sound like a lot of work?  It is initially, but like anything else it becomes habitual over time.  No one said safety was easy, but who could argue it’s not worth it?

- - -

Judy Agnew, Ph.D.  Senior Vice President, Safety Solutions

Judy Agnew is a recognized thought leader in the field of behavior-based safety, safety leadership, and performance management, and she is an expert consultant who works with clients to create behavioral interventions that ensure organizations are safe by design. As Senior Vice President of Safety Solutions at Aubrey Daniels International (ADI), Judy partners with clients to create behavior-based interventions that use positive, practical approaches grounded in the science of behavior and engineered to ensure long-term sustainability.

Judy has presented at major safety conferences, including the American Society of Safety Engineers, National Safety Council and Behavioral Safety Now, as well as other key corporate events. She is frequently interviewed for national and trade publications and has been featured in Occupational Health and Safety and Industrial Hygiene News to name a few. She is the author of two highly regarded safety books, Removing Obstacles to Safety (with Gail Snyder). and Safe by Accident? Take the Luck out of Safety: Leadership Practices that Build a Sustainable Safety Culture (with Aubrey Daniels). Judy is also the recipient of 2011 Outstanding Contribution Award from The Organizational Behavior Management Network, which recognizes her significant contributions to the field of behavior analysis.

For more information, please visit Judy’s website: http://aubreydaniels.com/judy-agnew-phd

It seems that everyone wants to claim near misses for their cause. That should be good news – near misses present valuable opportunities to learn and improve. Unfortunately, sensible discussion on near misses can all too often degenerate into a debate about …

Continue reading »]]> What are near misses? The disputed territory between lagging and leading indicators.

It seems that everyone wants to claim near misses for their cause. That should be good news – near misses present valuable opportunities to learn and improve. Unfortunately, sensible discussion on near misses can all too often degenerate into a debate about whether they are lagging or leading indicators of safety.  Turn this around though and maybe that is a clue to debunking some of the fallacious arguments involving lagging and leading indicators. Let’s explore.

Near What?

As a starting point: are we talking about a near MISS or a near HIT?

• It was a miss, but it was a close escape – a near miss, with “near” describing the miss.
• It nearly hit – a near hit, with “near” meaning almost.

The choice is yours. “Near miss” is conventional and well understood, that is why I choose to use it. However, having read Daniel Kahneman’s book Thinking, Fast and Slow it occurs to me that psychologically (considering cognitive ease) “near miss” may sound too good, too comfortable, too safe. “Near hit” is a less common phrase, with “hit”  more likely to trigger a mental reaction than “miss” . If so, it may draw greater attention and reaction to the same event.

Luckily nothing happened – this time

Some say that in a near miss nothing actually happened. They argue that a near miss provides a glimpse into the future – a suggestion of something more serious that might happen on another occasion. The message is that, correctly understood, a near miss is an opportunity to learn. Apply that knowledge to take action to prevent possibly more serious consequences another time. Using this argument, near misses are taken as leading indicators that can be used to help create safety.

But it was an incident

“Near misses describe incidents where no property was damaged and no personal injury sustained, but where, given a slight shift in time or position, damage and/or injury easily could have occurred” (U.S. OSHA definition). The clear message is that, despite no physical harm, something undesirable happened. On this basis a near miss is a lagging indicator.

Is a near miss an unsafe condition?

We can make a distinction between “near miss” and “unsafe condition”. An unsafe condition can exist even when there is no incident – making it a leading indicator. Examples could be corrosion of steel walkways, uninspected pressure vessels, defective brakes, PPE not worn, poor electrical grounding.

Too late?

Classing near misses as a lagging indicator does not necessarily mean too late.  True you cannot go back and prevent that particular incident. But as with all incidents up to and including fatalities, it is still possible, if not an obligation, to investigate to learn from the experience and take remedial action to prevent a recurrence. In a sense the lagging indicator generated by incidents becomes a leading indicator for prevention.

Neither and both

Near misses are, quite simply, indicators. They straddle the descriptions of lagging and leading. They represent something that was unsafe (but you were lucky); they are weak signals that provide evidence in advance of the possibility for injury or damage. What matters is that near misses can be a relatively plentiful and rich source of data for learning and improvement.

We should stop worrying about the terms lagging and leading and use wisely whatever data we can to predict and improve. We miss opportunities for improvement if we get too dogmatic and say that lagging = too late, or leading = too subjective.

PDCA / PDSA

Continual improvement in safety can be achieved by using PDCA, or better PDSA:

PDCA (Plan-Do-Check-Act), also known as the Deming Cycle, is used extensively for continual improvement and has been adopted as the basis for management system standards such as OHSAS 18001.

PDSA (Plan-Do-Study-Act), also known as the Shewhart Cycle, is the version Deming taught (for more details of PDSA see Out of the Crisis and The New Economics). He advocated “Study” as this implies understanding why something improved, or not, whereas “check” suggests more an answer to whether or not all is going to plan.

PDSA can be used for continual improvement:

• to help create what we predict will be a safe work environment  (safety precursors);
• as the basis for an operational definition of safety

Creating safety precursors

Take time to think what is required for safe working. Decide how you will achieve it – consistently. Implement those plans. In a previous article I discussed a practical approach for effective improvement to the control of safety risks. Using PDSA can build on that.

Monitor what you do. How well are you doing on what you say you will do? Reassurance could come for example, through observations from inspections, timely completion of actions and and reporting of near misses . Call these leading indicators if you like. Given the importance of learning from near misses, action taken to encourage people to report near misses is also be an important leading indicator of safety. Note that an increase may represent greater transparency and trust, not a worsening situation. Near miss should be prevented but, whatever the incident was, people should be congratulated for REPORTING a near miss.

Complete a PDSA cycle by taking action to rectify any shortcomings in your methods of achieving safety (see Figure 1). Keep repeating the process. The message is that leading indicators help you predict and improve safety.

Figure 1: PDSA on the work process with examples of leading indicators

How effective are the safety efforts?

Wisely chosen indicators will allow you to improve safety performance. But recognise that leading indicators are a heuristic – they substitute the answer to the difficult question “are we safe?” with the answer to a simpler question “are we doing what we think are the right things?”

Even if your indicators are well chosen and showing excellent performance, you may have missed one crucial failure path that could lead to an accident. This is where near misses, as well as injuries and damage, provide valuable information about weaknesses in the safety system.

So, to judge success in achieving safety in any period of time, it is not sufficient to rely solely on leading indicators. However good you may feel about the effort you have put into creating a safe workplace, the acid test is this: did anyone get injured? The final part of the jigsaw is to obtain data on injuries and occupational illnesses, for example TRIR (Total Recordable Incident Rate), or the rates for LTI (Lost Time Injuries) and DART (Days Away, Restricted, or Transferred).

Figure 2 shows how a TRIR might be shown graphically, with 3-sigma process behaviour limits calculated for a level of performance that appears to have reached a stable level of about 1.0, with annual variation predicted to be between 0.5 and 1.5.

Figure 2: Example of TRIR to report safety performance

Call these lagging indicators if you like, but do not stop there. Crucially, you must study the data to understand what it means. As Nassim Nicholas Taleb says in The Black Swan “You can be very confident about what is wrong, not about what you believe is right”.

The data may be necessary for corporate reporting or other reasons, but comparing results against last year or some industry standard provides little or no knowledge. If you want to learn and improve you must study the details – specifically, your details.

Was it just luck?

If you have had no injuries, or a low rate, why was this? If you have had injuries you need to know the specifics, such as the type and severity of injuries, the type of work, the conditions and location where the injury occurred.

How much do you trust the reporting / recording process? Are thorough incident investigations identifying the cause of incidents (not who to blame)?

Is there good reporting on near misses to give a wider perspective of what is happening? Is that distorted by a disproportionate volume of trivial near miss reports to achieve a target or reward? Does fear restrict reporting on serious near misses?

The important point is that knowledge gained from studying lagging performance indicators must be fed back to improve your leading indicators and ensure that you are controlling actual safety risks. For example, if your incidents involve many soft tissue injuries you may need more focus on ergonomics or material handling. Safety checklists should be amended if necessary and perhaps new leading indicators established to track performance in these areas.

An operational definition of safety

The requirements for an Operational Definition for Safety are:

• What is the organization’s aim with respect to safety – the ideal?
• What will the organization do to create improvements in safety – the methodology?
• How will you know if you have achieved improved safety – the judgement?

The argument above suggests how leading and lagging indicators fit together in an integrated way as part of an operational definition of safety – a balanced approach, if you will.

The aim of continual improvement is implemented through preventive action, tracked by leading indicators that optimise performance of the safety system. Accidents provide data from which judgements can be made about the effectiveness of the preventive action. The knowledge gained from studying accidents is the basis for action to improve the preventive measures and the leading indicators. The cycle is shown diagrammatically in Figure 3.

Figure 3: PDSA as an operational definition of safety

Preoccupation with Failure

Fortunately, there appears to be widespread agreement about how learning from near misses can help with safety improvement. However, we should not let our enthusiasm about the benefits of near miss reporting and investigations overshadow the point that each near miss represents A FAILURE.

In their book, Managing the Unexpected, Karl Weick and Kathleen Sutcliffe identify a preoccupation with failure as a feature of so-called High Reliability Organizations. They stress that a near miss should be interpreted as a sign that your system’s safeguards are vulnerable. Any lapse is a symptom that something may be wrong with the system: something that could have severe consequences. Their view can be compared to the common cause hypothesis that, for an incident in any one organisation, the same causal path may lead to no injury, a minor injury or a major injury.  We should, as Weick & Sutcliffe suggest, “Interpret a near miss as danger in the guise of safety rather than safety in the guise of danger”.

 Beware of people who talk about rear view mirrors

Some people love analogies. An analogy can be a useful way to explain a new concept, but remember that analogies can be fallacies. Just because there is one similarity does not mean that there are similarities in all respects. Less excusably, analogies are sometimes used to amuse an audience and divert attention from rational assessment.

“Using lagging indicators is like driving a car by looking in the rear view mirror”. Ha, ha!

What is meant of course, but conveniently missed, is that it refers to ONLY looking in the rear view mirror. Driving a car and looking from time to time in the rear view mirror warns you about cars approaching fast and about to overtake you. That improves safety.

Summary:  “Don’t throw the baby out with the bathwater”

I hope this article has shown that we learn something from both the PROCESS of creating safety (via leading indicators) and the safety RESULT (via lagging indicators).

Leading indicators can provide valuable day to day information on how PRECISE we are in doing what we say we should do. However, we should also continually revise our judgement of the effectiveness of our safety efforts by the successes and failures we experience. Lagging indicators can provide essential feedback on how ACCURATE we have been in identifying the necessary safety measures required. You just need to make sure you understand your data and not torture it to make it confess to suit some other motive.

Finally, we should learn from near misses and take action to improve. They are a performance indicator; talk of leading or lagging is a distraction from the fact that each near miss represents a failure…. and an opportunity.

References

Deming, W.E., 1986, Out of the Crisis. Cambridge, MA: MIT

Deming, W.E., 1993, The New Economics. Cambridge, MA: MIT

Kahneman, D., 2011, Thinking, Fast and Slow. London: Penguin Books

Taleb, N.N., 2007, The Black Swan: The Impact of the Highly Improbable. London: Penguin Books

Weick K.E. & Sutcliffe, K.M., 2007, Managing the Unexpected: Resilient Performance in an Age of Uncertainty, 2^nd Ed. San Francisco, CA: Jossey-Bass.

Biography

Nick Gardener spent many years working in the chemical, nuclear and automotive industries. He is now a global risk and HSE consultant working for Risk International Services Inc. He can be contacted at: ngardener@riskinternational.com.

Continue reading »]]> This is the first article in a series of three articles dealing with safety barriers. The first part of this ‘trilogy’ will deal with some general reflections on safety barriers and barrier assessment. The second part will discuss a practical method and example of how this can be done. Thirdly we will discuss some possibilities on how to strengthen the barrier management of an organization.

None of the items discussed pretends to be the ultimate or perfect solution, rather is this meant as an attempt to give an impulse to continuous improvement through reflection, sharing experience and (hopefully) generating some discussion.

General reflections on safety barriers

One traditional way of dealing with safety in a proactive way is through risk assessments. While risk assessment and its many methods (e.g. TRA, SJA, HAZOP, HAZID) are powerful and important tools in managing challenges, there is a growing realization that the application of the well-known approach of risk as a product of consequence and probability may have some serious drawbacks. This applies especially to the more calculative and quantitative varieties and the decision making that often is related to those. We won’t go into detail of this here and now and rather suggest reading some good literature in this area[1].

One problem with assessments that focus too much on the function of consequence and probability is that they may neglect the influence of uncertainties and assumptions on risk. These, however, may be of greater influence than any of the other factors of your assessment. To illustrate this point, Aven uses the simple example of a throw-of-the dice gamble. Here a calculation of expected risk/value will become totally invalid if one abandons the common assumption that it’s a fair throw, that the die aren’t loaded and the game isn’t rigged.

One way to tackle these uncertainties in a better way is focusing on barriers that are (or are not) in place and discussing the strength of these barriers.

What are barriers?

The first thing when discussing and assessing barriers is, of course, an understanding and operational definition. In Norwegian railway safety legislation, a barrier is defined as: “technical, operational, organizational or other planned and implemented measures that are intended to break an identified unwanted chain of events” (Sikkerhetsstyringsforskriften, 1-3). Other standards and legislation contain similar wording; e.g. ISO 17776:2000 (“Guidelines on tools and techniques for hazard identification and risk assessment”) defines a barrier as a “measure which reduces the probability of realizing a hazard’s potential for harm and which reduces its consequence” and explains that “barriers may be physical (materials, protective devices, shields, segregation, etc.) or non-physical (procedures, inspection, training, drills, etc.)”.

When we speak of barriers in our company we do this from a MTO-perspective: Man, Technical and Organization in relation to each other. This fits very well with the definitions quoted above and within this series of articles we will advocate a wide, call it holistic if you will, approach to barriers. Barriers in our operational definition are the things we put or have in place to prevent unwanted events from happening (i.e. control the hazards) and/or reduce negative consequences of these unwanted events to a minimum.

Misunderstanding about barriers

While we don’t claim to ‘own the truth’, we have in our daily practice come across some discussions and misunderstandings about barriers that are wrong, or at least unfortunate.

One of the most important misconceptions is the widespread (especially among regulators) thought “the more the better”. Just not true because of the law of unintended side-effects. More barriers may often mean greater complexity, more possibility for (unexpected) interactions and in effect the creation of new failure modes. This does not mean that one shouldn’t strive for defenses in depth[2], but be aware of the fact that by solving one problem you may very well create new ones.

When talking about barriers it’s important to realize that some barriers are very explicit, like the railing on a bridge. Other barriers, however, are much more implicit like the robust design of your car or your skill in handling the car, forged by formal training and years of experience. This is one reason why some barriers are rather invisible or may not at once strike the beholder’s eye/mind as being a barrier.

There is a misunderstanding (at least among some people we work with on a regular basis) that organizational safety measures aren’t barriers at all. Planning, in fact, is one of the strongest barriers there is. The train timetable makes sure that there are no two trains at the same place at the same moment. Separation in space and time is a very solid safety measure and it’s rather odd not perceiving this as a barrier. Maybe this is because people don’t generally relate the timetable to safety, but rather to punctuality.

Many of the ‘implicit’ barriers aren’t identified in traditional risk assessments because they are ‘hidden’ in the form of assumptions (e.g. competent personnel performing the job) or system design (e.g. most of us take it for granted that they won’t fall through the floor of our offices). You could also consider the competency of the observer. We see many observations for PPE but MUCH fewer in more complicated observations. There is an inverse relation to complexity (i.e.more observations the easier it is to determine).

Another misunderstanding is rooted in the fact that human barriers are seen as weak barriers and therefore cannot be considered as barriers. Humans are often perceived as the weakest link in a system. While humans may have a greater variability than many machines, and are prone to fatigue, distractions and errors of various kinds, it’s also true that humans are flexible, can easily adapt  to situations and quite often are a strong factor in ‘saving the day’. Reason and Hollnagel are but two known names in safety who have recently argued that one shouldn’t regard humans only as the source for failure because humans and their variability are just as much the source of success[3].

How good ar our barrers?

This brings us to the next issue: how good are our barriers anyway? Or to put it otherwise, how big are the holes in our Swiss cheese? And where approximately are these holes located? Knowledge about this is an important factor in dealing with assumptions, uncertainties and unknowns.

The thing with barriers is that at they are hardly ever perfect. Barriers such as the ones listed as examples in ISO 17776 can actually fail from one moment to the next. One can choose to follow a procedure or one can decide to take the shortcut, making the rule-barrier useless.

We talked above about the law of unintended side-effects. Barriers may not only be a barrier against hazards, but also a barrier/hinder for the job one is trying to do. It’s tempting then to take shortcuts or find ways to bypass or avoid the barrier, rendering them useless while we may be depending on them.

This mechanism doesn’t only apply to the ‘softer’ barriers, it also extends to technical/physical barriers that can be effective one moment and then ineffective the next. These barriers can be rendered useless in a whim, for example, when we don’t wear seatbelts or safety goggles or when a safety barrier is bridged.

It’s important to see barriers within the context of the system they are placed in. When observing a system, we have to study it as being the combination of man, machines, procedures and other elements. While it’s possible to see each as a separate item with man as one system and the machine as another, and man not being a part of the machine-system, this view of separate systems is not very useful. A man working with a machine creates a new system that is created from several sub-systems. This gives a much clearer view of systems, and also of barriers.

This then also includes the interactions between the various items. The best of machinery (Technical) is useless without an operator (Man), but if he/she isn’t trained properly (Organizational) the output of this system may not be as desired, except by luck.

Assessing the strength of barriers is something a lot can be said about. To just scratch the surface here, one should consider what kind of barrier we are dealing with: M, T or O. Much depends what the barrier shall guard against and how it works. From there, one can distinguish between static and dynamic or active and passive barriers. There is no general rule which is better. This entirely depends on the barrier’s intended function and should be judged accordingly, but one should be alert about dependencies in barriers – e.g. a warning system by itself may not be enough.

Also one should keep in mind that some barriers are pretty situational. When working with one co-worker we may have an understanding that goes way beyond that what is said and our mutual thought pattern may help us to get a job done in no time, even though our communication is sub-optimal. When working with another “less aligned” colleague, the result may be less satisfying.

Another aspect is the consideration of whose responsibility a certain barrier is. This doesn’t necessarily say something about the ‘goodness’ of a barrier, but it does give an indication of the degree of control that you have over the barrier and this should affect the way you value (and depend on) it.

In some cases you may even be in doubt whether you should regard something as a hazard rather than as a barrier. Say you want to work on road safety of your personnel; you can provide them with the best of training and safe cars which you keep in the best of shape through a good maintenance system. Alas you have little control over potential barriers such as the state of the roads and the competence of other road users. Regarding these as hazards is a choice that depends upon your assessment, and maybe you want to consider additional barriers.

Decision making – what barriers to choose?

Resources tend to be limited, so one really important question that almost always pops up: what shall we spend our money on?

The first requirement in decision making, of course, is recognition of the hazard, the realization that something has to be done and the will to actually set things in motion to create barriers. Intellectually understanding the risk but doing nothing just isn’t enough.

Having come so far, everything of course depends on the specific case; what shall we guard ourselves against. But there are some general rules of thumb in safety that are usually wise to follow:

• Reducing probability tends to be more effective than reducing impact, but one cannot entirely discard the latter, of course. Air bags are smart having, even if we never hope to use them.
• Follow the typical hierarchy in safety measures: removing hazard sources trumps collective measures trumps individual measures trumps personal safety gear. If only because working in a quiet environment is so much comfortable than walking around with ear plugs all day long.
• Building in safety and barriers from the start (design or planning) is cheaper and more effective than adding these afterwards.
• See things in perspective with other objectives in order to avoid conflicts as much as possible. Remember the law of unintended side effects?

Cost/benefit, of course, is an important factor to consider and sometimes one may consider accepting a risk. While this may sometimes be acceptable, may often meet political or moral objections. Also, it’s sometimes the question how well-considered some cost/benefit assessments actually are. Often one sees a penny wise/pound foolish approach where the immediate out-of-pocket costs appear to dominate the decision. After all, it’s very easy to decide implementing the more individually oriented and quick to implement ‘barriers’ (prescribing personal safety gear, making a new rule, enforcing existing rules) instead of preferring the slower, but often more robust technical or organizational barriers that go deeper than the surface, but require more effort, time and follow-up.

Common sense is a rather touchy and messy concept which we want to avoid, yet it has to be said that gut feeling can be a good indicator for a decision that has to be taken. But make sure to critically assess these primary or instinctive reactions towards a decision in a systematic way: what reasons are there for or against. And don’t forget: whatever you will decide upon, it will have effect on other things. Silver bullets in safety are rare, or probably non-existing.

—+—

Much more could and should be said about this theme, but let’s wrap it up for now. Next time we’ll take a look at one way this can be done.

Biography

Carsten Busch has studied Mechanical Engineering and after that Safety. He also spent some time at Law School. He has over 20 years of HSEQ experience from various railway and oil & gas related companies in The Netherlands, United Kingdom and Norway. These days he works as Section Head Safety and Quality for Jernbaneverket’s infrastructure division.

Beate Karlsen has studied Occupational Health and Safety at Haugesund and at the Stavanger University. She has been in various OHS functions in Jernbaneverket and works currently as Senior Advisor Safety in Jernbaneverket’s infrastructure division.

Continue reading »]]> In our last blog article by Dr. Timothy Ludwig, “The Anatomy of Pencil Whipping”, the idea of safety observer quality was introduced. This article, originally published in the VPPPA Leader magazine in 2012, further expands the subject by introducing quantitative measures of observer quality that can be used to both evaluate and improve the data collection process.

Employee involvement is a key component of most safety management systems, such as OSHA’s Voluntary Protection Program. However, many companies struggle to find meaningful ways to get every employee engaged. During my site visits as an OSHA Special Government Employee (SGE) I would often find employee involvement was constrained to small, fringe groups and typically comprised of only a fraction of the entire workforce. Over the years, I became a firm believer that the best and most meaningful way for employees to get involved is through participation in the site or company Hazard Recognition and Control program. Specifically this can be done by making workplace observations related to the health and safety program as it relates to behaviors, conditions, and/or safety activities. Our studies have shown this to be beneficial from an accident prevention perspective. However, once you begin to involve many people in the safety observation program, new issues arise such as observation quality. As such, I want to identify the benefits of adopting this program as well as introduce an effective way to both measure and improve observation performance.

Recently our company has done a study that found the following correlations related to inspections – a collection of workplace observations – and reduced incident rates. We call these our “safety truths”:

1. More inspections and observations result in a safer worksite
2. More inspectors, specifically more inspectors outside the safety function, result in a safer worksite
3. Too many “100% safe” inspections results in an unsafe worksite
4. Persistently high levels of unsafe observations results in an unsafe worksite

Clearly involving more observers increases the chances of identifying at-risk or unsafe behaviors or conditions so that they can be addressed proactively before an incident occurs. Remember that an observation is only a snapshot in time – no one person can be everywhere at all times to capture this. In addition, if only those with a safety function collect the information, then ownership of the safety process shifts to that of the safety department and not with the workers and the front-line supervisors where it needs to reside.

One of the reasons most companies are reluctant to engage workers and managers in the observation process is primarily due to perceived gaps in skill and/or knowledge that could lead to poor results. This is understandable; however, it is patently false. Every worker exposed to risk in the workplace should have a voice to speak the truth about how safe or unsafe things are. That being said, the more observers collecting data, the more ways biases can be introduced. Just like training workers on how to do their jobs or write and follow safety procedures, involvement in the observation process requires feedback. Engaging an observer in feedback allows those in the safety function to transition from the traditional safety “cop” role to that of being a coach. Through their experience, the safety team can express what is safe and at-risk, what is important in the worksite to focus safety observations on, or even share risk assessments so that everyone understands the impact of working safely. By measuring the strengths and weaknesses of the observer, individualized improvement plans can be developed and monitored over time such that everyone continually improves at their own pace.

Here is a great example: You have an observer who is highly engaged in the observation collection process. This person submits twice as many observation cards or inspection forms as the average worker at your worksite. However, this person rarely, if ever, finds anything at risk, despite the fact that the area this observer works averages 2-3 at-risk observations per inspection. Some would roll their eyes at the submitted data. Others would go so far as to label the observer as a “pencil-whipper”. How then can you improve this observer’s observation performance?

Individuals are evaluated in your workplace every day. In fact, I would be willing to bet that each of you reading this has been involved in an annual evaluation process with your manager. What did this this process look like? Typically it consists of these steps:

1. Expectations are communicated. What are your goals?
2. Constructive feedback is provided. Preferably this is done frequently throughout the year and not just at the evaluation!
3. Measure performance and progress against expectations
4. Communicate to the individual what is important to the organization. This traditionally involves a discussion around strengths to sustain and weaknesses to improve

The same can be done for the observation process provided there are expectations established and communicated. Here are some examples on establishing quality criteria around measurable metrics:

1. Participation – count of submitted cards/forms/inspections (average over a given time period)
2. Average count of safe and at-risk observations per inspection
3. Identified severity of at-risk observations
4. Comments included with observations
5. Frequency of 100% safe inspections
6. Focus on pertinent hazards/tasks
7. Closure of unresolved issues

If the only metric you track is the count of how many cards or forms are submitted, then your ability to provide feedback is tremendously diminished. However, as you develop your inspection strategy and plan around meaningful objectives, then you can quickly and easily identify areas of improvement and measure that progress over time. This is due to the fact that as the data set grows the metrics become normalized – averaged out over a large data set. As such, it becomes easier to spot outliers from the “norm” who can then be engaged for improvement.

With the example of the “pencil-whipper” above, many would point to the poor data collected as proof only safety folks should do safety inspections. What is not clear is why this person is doing what they are doing. By soliciting feedback and engaging in a respectful manner you may discover perhaps the person believed that is what was expected of them. Perhaps their manager scolded them for “making them look bad”. Even worse, perhaps workers had submitted at-risk findings that were ignored by management and now have a “why bother” attitude.

As this process is adopted, it changes the conversation of “did you do an inspection” to “let’s talk about your data and how we can improve what we see”. As this process continues, engagement becomes more meaningful, the data is trusted more, and there is less reluctance to act on the findings. As the results improve, more observations are collected which provides greater insight and visibility into the safety of the organization which leads to a reduction in incidents. Notice how the incident rate was never a focus but becomes a benefit to an improved observation process.