Last month, I offered the concept that post-accident investigations are a critical step in accident prevention. I stressed that the analysis should be
RANDY K. LOGSDON
Last month, I offered the concept that post-accident investigations are a critical step in accident prevention. I stressed that the analysis should be focused on system errors or flaws that contributed to the accident. Also, the analysis should proceed using an objective methodology that answers the question: ìHow did our accident-prevention system fail?î Two key concepts in this statement are ìobjectiveî and ìmethod.î
The methodology used for a thorough accident analysis should be developed before the fact. Also, personnel directing the analysis must understand the purpose of the methodology and be familiar with its application. The reasons for using a defined methodology are to ensure objectivity and consistency, and to make sure the product is accurate and thorough.
It is easy, especially with some events for which the cause seems to be self-evident, to shortcut the process ó to develop hasty conclusions. Our minds work quickly. In the process, one may develop theories that should be explored. However, it is important, even in these cases, to apply the method to its conclusion.
Over the years, a number of methods sporting interesting acronyms have been developed to direct this methodology. These include: Critical Path Method, Failure Modes and Effects Analysis, Fault Tree Analysis and others. Each has its strengths and weaknesses, applications and misapplications. Investigation of these techniques can lead to your development of a methodology suitable for your organizational needs. There also are commercial systems available that may suit your needs. As you develop your methodology, consider these important elements.
Define the undesired outcome
Describe the damage. Quantify the severity of the real or potential injury, property damage or interruption.
This should include physical evidence, photographs, measurements, drawings, written procedures, records and witness statements.
Based on the evidence, what are the possible scenarios? More than one scenario may be plausible. One interesting aspect of this phase is that finding the true scenario is not as important as discovering the system flaws. Repeatedly ask ìWhy?î
Why did the event occur? Why did those factors that lead to the event occur? Continue to follow the ìwhyî to the point where the organization cannot control the ìwhy.î Consider contributions to the accident that include human aspects and behaviors; the environment; the process and process interactions; tools, equipment and materials; education and training; and the application of rules and policies.
Identify system flaws
Any one of the plausible scenarios could have caused the accident to occur. Identify which accident prevention system failed in each scenario. The most important system flaws are those that affect the most people. Look at each of the accident prevention systems you employ. Consider the possibility that a system failure may actually be the absence of a prevention system. (See the lists in last month's column.)
Correct system deficiencies
Based on the identified system flaws, develop corrective strategies designed to mitigate the system flaws. The priority should be based on how effective the change will likely be and how easily it can be implemented and managed. Generally, engineering controls are considered to be most effective, followed by work practice controls and training. Other less effective controls include warning labels and signage.
Audit the modified systems
Check to ensure that the modifications are applied and that they are effective.
You may have noticed that this discussion has not covered the issue of discipline. The focus is strictly on identifying errors and correcting the systems that permitted those errors. Where deficiency in human behavior is identified, appropriate actions may be required to ensure that repeat failures do not occur.
Ultimately, you are likely to find that a systematic approach to accident analysis will provide insight not only into what actually happened, but what could have happened. This analysis may help you identify flaws in the system that have not yet resulted in a tragedy.