Traditionally, the process of investigating incidents, events and negative trends includes a process by which interviews are conducted. It turns out that the interview and note-taking process is one of the most inefficient, time consuming and biased aspects of a root cause analysis or other investigative process.
I became a root cause analysis (RCA) subject matter expert in 1989 and an RCA instructor in 1991. For over 30 years I've used the methodologies for assessments, audits, investigations, apparent cause evaluations and root cause analyses at commercial nuclear plants, the US Nuclear Regulatory Commission, the Department of Energy (DOE) and the Department of Defense (DOD). For two decades I used the traditional RCA approach of pulling together a team of 5 to as many as 20 subject matter experts, and send them out to interview a long list of personnel. However, these traditional investigations were taking weeks and months to complete. Around 2010, I began searching for ways to "lean" out the RCA process and develop more modern, efficient and accurate methods that were more in line with today's appetite for speed and flexibility. Here are some of the largest inefficiencies I discovered, just in the interview process.
One-on-one Interviews did not lend themselves to pursue cause and effect analysis with the person being interviewed. Causal analysis is essential to getting to the deepest-seated causes.
Personnel being interviewed often felt as though they were being "interrogated." We would document their answers in our notebooks and they did not get to see what we were writing. This dynamic often led to interviews that were less than productive because the interviewees did not trust the process, so their responses lacked honesty and candor, and in some cases they became uncooperative.
One-on-one interviews lended themselves to being biased by both the person being interviewed and the RCA team member. Interviewees could make statements that could not be checked or validated until later. They could also attempt to steer the RCA towards a non-productive path. RCA team members also applied their own biases to what they thought they heard; it's human nature.
The order of the personnel being interviewed was not carefully considered. For example, interviewing supervisors and managers before the workers could taint or influence the investigation with the management "party line."
The one-on-one interviews involved a series of questions that were often crafted before having gathered the necessary information. For example, a well-developed timeline, or before having read the applicable requirements. The rush to interview personnel often led to a lack of focused, evidence-based questions, which led us down rabbit holes that were unproductive.
The one-on-one interviews did not include a visual way to chart the methodical coverage of key aspects of the investigation. For example, questions to determine the effectiveness of the "defenses" that should have prevented the event (i.e. the programmatic requirements and physical barriers such as personal protective equipment). Also, questions to evaluate other important areas that we should be investigating; areas that come from our understanding of the "Anatomy of an Event."
Copious pages of interview notes from each RCA team member were then brought back into a conference room. All of those bits of information had to be discussed and integrated. Imagine each bit of information as a small puzzle-piece, and the RCA team had to assemble a giant puzzle from all these pieces. Each RCA team member relayed what they thought they heard, often injecting their own bias that went unchecked because the other RCA team members were not there. They also had to recall information that was provided days or weeks before, with less than perfect notes to refresh their memories. The overall process of integrating interview notes was often arduous and sometimes it even got confrontational.
The analysis of the information was often not well structured. Sometimes we could assemble an Events & Causal Factors Chart. Fishbone charts became unwieldy with more than 10 lines of inquiry and are not very useful in identifying root causes. (I stopped using the Fishbone in 2007). More often than not, the RCA team crafted a narrative that best explained the causes for the event that took place. The results were easy to challenge or second-guess and often difficult to defend.
My search for ways to "lean" out the RCA process and develop more efficient and accurate methods led me to a complete re-invention of how RCAs are conducted, which I call BlueDragon Hyper-Integrated Causal Analysis. One of the major changes is how information needed to be transferred from the subject matter experts in the affected organizations to the RCA team. Here then, is our tip of the week.
Tip of the Week:
Replace the one-on-one interview process with facilitated causal analysis sessions. The exception would be if an individual interview was needed because it may involve a sensitive subject. Here are the basics:
Schedule a series of interviews with a representative sample of the affected organizations. For example, 3 to 6 subject matter experts.
We no longer call those "interviews" because that is a misnomer. They are "facilitated causal analysis" sessions.
Follow a disciplined process for scheduling these sessions: the workers from all affected organizations must go first (one organization at a time). Then supervisors, and lastly the managers.
The purpose of causal analysis sessions will be to answer the lines of inquiry that will be displayed on the BlueDragon chart. We can also have causal analysis sessions using a virtual whiteboard.
The answers from the group are placed on the chart in a transparent manner. This is to generate cooperation and trust.
There are no notes taken during the process. We take the answers and place them directly on the chart for validation.
Answers are validated with the group in real time, to avoid bias from any one individual. And the RCA team is never separated, to avoid bias from team members as everyone hears everything at the same time.
The RCA team's goal is to identify the root causes and contributing factors by guiding the participants through the causal analysis process; to get them to cooperate and provide as many answers to as many of the lines of inquiry and to go as deep as they can with their level of knowledge.
Each group that comes in will get to see and validate the answers that are already on the chart and we continue to build the chart in a transparent manner. With this transparent approach, it is important that workers go first and managers last, as managers may taint the board with the "party line."
We can schedule 4 to 6 sessions per day and engage (obtain answers from) over 100 subject matter experts in a matter of 2 or 3 days. It would take weeks to complete that many one-on-one interviews, and more weeks to process all the notes
Along with other aspects of Hyper-Integrated Causal Analysis (HCA), these changes in how interviews are conducted have dramatically accelerated the transfer of information from the organization's knowledge base to the finished HCA chart.
For examples of HCA in action and the dramatic impact that these changes have made in the time it takes to conduct the RCAs, visit the following pages:
Human Error - a worker received a shock from a 208v, 20a current: https://www.dle-services.com/hca-human-errors
Equipment Malfunction: 7 motor failures over the span of 11 months: https://www.dle-services.com/hca-equipment-failures
Programmatic Issue: recurring deficiencies in the site-wide Procurement Program: https://www.dle-services.com/hca-org-prog-issues
Negative Performance Trend: a proactive review of a trend of dropped objects from heights: https://www.dle-services.com/hca-proactive