Avoiding the Crisis du Jour

by Paul A. Keller, CQE, CQA

It is a pretty safe guess to say that the biggest worry to your operations today may not be the same as yesterday, or last week. Crises come and go, and thankfully so. For generally, we lack the resources to properly confront the crisis of each day. So when the crisis du jour vanishes, to be replaced by the piece de resistance of tomorrow, we can all breathe a sigh of relief and pat ourselves on the back for putting yet another nightmare behind us. Until it comes back to haunt us once again - a leftover du jour, so to speak. Then, of course, its much easier to "solve," since we just remember who got yelled at last time, and figure they need a reminder.

If this sounds too familiar, you have plenty of company. In fact, these engagements are sold out for months in most organizations. What is missing here is the correct approach to problem solving and analysis. For starters, we need to avoid confronting every crisis du jour as if it were indicative of something in and of itself. Many times, it is not.

How can I say this? Because, as a general rule, common cause variation is more prevalent than special cause variation. To deal correctly with the crisis de jour, we must first understand if it is induced by a common cause or a special cause. If related to a special cause, we should pay close attention to the details of its occurrence, since they represent conditions of the process specific to that point in time. If related to a common cause, we should address the system that produces it. As a common cause, the variation evidenced by the crisis in question is an inherent part of the behavior of the process. If we treat a common cause like a special cause, our tampering would tend to increase the amount of variation in the process. (See Deming, Out of the Crisis).

Consider an example. The staff meeting on Monday morning turns into a free for all because the error rate for last week - or infection rate, scrap rate, non-conformance rate (substitute your key process parameter here)- went to hell in a hand basket (or a quicker mode of transport). While it typically sits at 3%, last week saw an unprecedented 4.2%. And people want some answers. NOW! Well, before we start changing our hiring, training, or disciplinary practices, some analysis would be warranted to see if this rate is noteworthy or not.

Charting this key parameter over the course of time on a p chart, we find that its average is indeed 3%. We should expect to have a larger rate than this 50% of the time, and a lower rate than this 50% of the time, with limits that vary depending on the sample size (# of pieces, patients, records audited, etc.), assuming our process is stable. Based on the volume experienced last week of 1000 units, if our process is stable, we should expect it to operate between 1.4% and 4.6% error rate. Any variation within these limits tells us the process has not varied, that it is driven by only the common causes which are inherent to the system. Knowing this, we should not pat ourselves on the back for the non-existent "reduction" to 2% that occurred last month, nor search for the elusive cause of its predictable "crisis-level" of 4.2%.

Using the control chart, we now recognize that the process itself must be improved to reduce the error rate. At this point, we have several options available to us. A Cause and Effect diagram would probably be useful to map out the potential sources of common cause variation. (Consider the 5M and E: methods, materials, manpower, machines, measurement and environment).

However, to properly brainstorm on theses causes of process variation, we should map out the sequential process tasks using a Flowchart. A Flowchart will allow us to appreciate the complexities of the process, which should help in identifying the potential causes of variation at each step. In may be that some paths are not necessary, or do nothing to improve the customer experience, and can be removed. In other cases, paths may lead to undesirable conclusions (such as customer dissatisfaction) and the process would have to be re-designed to prevent this occurrence.

The interdependence of the potential causes identified in the Cause and Effect Diagram can be identified using the Interrelationship Digraph (ID, one of the 7 Management and Planning tools). The ID is particularly useful for identifying root causes which contribute to other sources of variation. A Process Decision Program Chart (PDPC, another 7MP tool) can be used graphically depict various alternatives and countermeasures associated with a process.

Learn more about the Quality Management tools for process excellence in The Handbook for Quality Management (2013, McGraw-Hill) by Paul Keller and Thomas Pyzdek or their online Quality Management Study Guide.