Using the Fishbone Diagram to Find Root Cause
Life science organizations today are faced with increasing regulatory pressure to offer products that are compliant, effective and efficient. While this is certainly the goal for many companies, issues often arise along the way, whether through a non-conformance identification, audit finding, customer complaint or another quality event. This is where root-cause analysis can help.
If we fail to identify the root cause of our problems, the best we can hope for in solving them is to be lucky. The complexity of our regulatory problems and the fast pace of business only add to the challenge of finding root cause. The stakes have never been higher. Customers, regulators, investors and many others expect the right answers fast.
While many people anxiously await the next miracle problem-solving tool, let me suggest an old and reliable friend-the fishbone diagram. This tool, also referred to as the cause-and-effect diagram or the Ishikawa Diagram, was developed by Kaoru Ishikawa, a quality management pioneer, and has been around for more than 40 years. I have found this tool very effective in identifying the root causes of regulatory problems that have a moderate amount of risk and complexity.
The successful application of any tool requires us to know and apply a disciplined process. Be wary of shortcuts. The need for speed can greatly limit success. Getting the wrong answer quickly will be of no value to you. Slow down to ultimately go fast.
Five Steps to the Fishbone Diagram
The fishbone process consists of five distinct steps.
The first step is to create the problem statement. You, and everyone on your team, must understand exactly what the problem is. It is the focus of your investigation. A problem statement is a statement of fact clearly identifying what has a problem (object) and what specifically is wrong (defect). It should be as clear and concise as possible. Beware of vague words.
Note the object (product XYZ) and defect (scratched parts) in this problem statement: Scratched parts for product XYZ increased by 70%.
The second step is to brainstorm possible causes. Be sure that your team follows brainstorming rules (Figure 1). In the example, we ask the team to generate possible responses to the question, "Why are we seeing an increase in scratched parts?"
[media:1746:embed:left]
At the end of this step you should have a good list of possible root causes. They should then be clarified and grouped using some system of categories. The six M grouping (manpower, method, material, machinery, mother nature and management) is very common, but any labels that make sense can be used. For example, PEMME (personnel, equipment, method, material and environment) is popular also.
The third step is to draw the fish and labels and enter data appropriately (Figure2). These data are the summary and categories of possible root causes of the increase in scratched parts. Most likely, the team will continue to present ideas, which can be incorporated until the diagram is completed. Continually asking "Why?" as you transfer the data to the diagram helps you drill down through symptoms to causes. It's also important throughout the process to challenge assumptions, support conclusions, etc. Remember-In God we trust-everybody else, bring data!
[media:1744:embed:left]
The fourth step is to determine the most likely cause. Organizing the data in a fishbone diagram gives you the ability to see linkages, relationships and patterns. At this point, a number of things might happen (Figure 3). During your discussion, the most likely cause may become clear to all. Some causes may drop out for obvious reasons. The evidence gathered during the investigation may point to the cause. If there are several possible root causes with no clear winner, you might ask each team member to select the cause he or she considers most likely and to defend this choice. You could use a multi-voting tool or ask team members to rank their top three causes. There are many ways to conduct this step.
[media:1745:embed:left]
Finally, verify the most likely cause. This last step is simply to prove the root cause is correct. What scientific tests might you conduct? How will you monitor the results? Set up an experiment? Remember, if you can turn the defect on and off, you've identified the root cause.
Congratulations! You found the root cause using the fishbone diagram. However, the job isn't finished. Apart from the upcoming corrective and/or preventive actions (CAPA), you have an investigation report to complete and submit to your company and/or to the proper regulatory authority.
Because you used this tool properly, your investigation went wide and deep. Not only did you prove you found the right cause, you were also able to show and justify what you did not select. The US Food and Drug Administration cares about all these things. Your fishbone diagram (and other documentation) provides the evidence.