A Simple But Effective Method of Causal Analysis
A client recently contacted me to ask for a straightforward yet effective method of post-incident root cause analysis. Like many companies, their process had devolved into a blame-game in which the "root causes" in various incidents all came back to employee behavior or "not paying attention". Consequently, the corrective actions "retraining" and "presenting lessons learned" were not preventing similar incidents from happening. No surprise there... On the other hand, the company is small and they didn't want an over-complicated process that would require alot of up-front training to deploy. This process is based on the "Hazard-Barrier-Target Analysis" method, plus 2 of the famous "5-Whys".
A Hazard-Barrier-Target Analysis (Hazard-BTA) process begins with describing the incident, examines how damage occurs, and assesses the adequacy of installed barriers or other safeguards that should either prevent or mitigate an incident. For a damage-causing incident to occur, there must be a target (what is affected), a hazard and less-than-adequate barriers. By identifying possible targets, barriers can be analyzed. A target is something of value; it may be a person, part, equipment, system, procedure or process. A barrier separates the target from the hazard and may be physical (such as personal protective equipment, machine guards or safety valves), administrative (procedures or directives) or personal (employee training and supervision) (Wilson, et al 133). To determine causation, a simple process of determining why the barrier failed or was inadequate will allow us to drill down past the obvious direct causes (like employees doing the wrong thing) to the systemic (root) causes of failure such as improper procedures, poor employee training, lack of resources provided by the company, etc. The goal of this process is to determine actionable corrective actions which will reinforce or replace the deficient barriers and actually prevent a similar occurrence from happening in the future.
EXAMPLE:
Two chemical plant employees were opening a process line at a flange joint in order to unclog the line. As the line was opened, the clog shifted, causing the product inside the line to splash the workers. As a result, each suffered multiple second and
third-degree burns. In this case, a hazard-BTA (Table below) proved to be an efficient assessment tool that provided valuable information in a relatively short period of time.
PROCEDURE:
Fill out description of incident first
What Is Affected: Details of event will dictate what is affected. Other areas: Worker Near-Miss, Equipment damage, Process failure, Facility damage, etc.
Barriers: Something in place to prevent failure. I.e. Physical guard, safety feature, interlock; or work procedure, JHA, technical manual, etc; OR Worker certification, training, supervision, PPE, etc.
Barrier Failure: Which of the known barriers either failed to prevent the accident or was not in place and led to the accident.
Analysis: After determining which barriers failed to perform as intended, ask “Why” twice, to find direct/root cause(s) of the incident, which can be corrected in the future.
Root Causes: (From Example) 1. Management Failure: Management did not provide accurate work procedure; and lack of supervision allowed workers to not vent and drain the pipe before breaking the flange. 2. Engineering Failure: Spill containment under flanges and valves not adequate to prevent process liquid from flowing onto the ground.
CORRECTIVE ACTIONS:
This process results in determining concrete corrective actions that are not punitive in nature. Here the hierarchy of controls should be utilized to prioritize which actions should be completed to yield the most effective deterrent to future failures.
Thus, for this example appropriate corrective actions would be:
Engineering Control: Install drip pans under flanges and valves likely to be serviced.
Administrative Control: Update the line breaking procedure to include venting and draining of pressurized lines.
Administrative Control: Address the management issue of too few supervisors and require supervisors to be present during high-hazard work.
This method is useful for a variety of incidents, but will only be effective if the initial post accident investigation is conducted thoroughly and with all of the people involved in the incident communicating pre-incident conditions without fear of reprisal or discipline. The typical management stance of “Blame the worker” will yield ineffective corrective actions like more training, more work rules and multiple layers of oversight. I advocate the "Learning meeting" approach described in Dr. Todd Conklin's books. Instead of building a timeline of cause and effect data, start from the beginning of the affected employee's day, ask what the work conditions were like, had they done the job before (i.e. what went right all of the times before it went wrong during this incident?) and surface all of the contributing factors that resulted in the incident. Utilize technical resources, co-workers and supervisors to bring their perspectives. And guide the process to yield the most information, the deepest learning and true understanding of how the task is actually done.
Although the majority of accidents are initiated by an action by an employee, not investigating further and determining why the employee thought taking such action would yield a positive result rather than the negative one that actually happened will shed light on how the work is actually done at the company rather than how management or the Safety depratment thinks it is being done. Studies show that this deviation (usually unseen by management) is where most incidents spring from.
