In my introductory blog post, I referenced several quotes from W. Edwards Deming, best known for his work in Japan’s successful approach to quality in manufacturing. His 14 points that make up a “System of Profound Knowledge” is still extremely relevant and widely used today. Here is a link that summarizes his points: http://www.managementwisdom.com/wedde14po.html. In that post, I was advocating adopting his approach to the process of safety as well.
That being said, I want to specifically reference Deming’s third point in this blog post. It can be summarized as this: “Cease dependence on inspection to achieve quality. Eliminate the need for massive inspection by building quality into the product in the first place.”
In our recent white paper titled “Predictive Analytics in Workplace Safety: Four ‘Safety Truths’ that Reduce Workplace Injuries,” we state, among other things, Safety Truth # 1 that suggests the following: “More inspections predict a safer worksite.”
A participant in the EHSQ Elite LinkedIn group posed this comment, “These two concepts seem to be implying completely opposite approaches – ceasing dependence on inspections to achieve quality (safety?) vs. more inspections to predict a safer worksite.”
His question is fair and here is my response:
I believe both approaches are very similar and not opposed at all. Like many uncertainties, the terminology is the key here. Typically an inspection in the quality world is a critical evaluation, usually of a finished product, at the end of the production process. Once inspected, the finished product either passes or fails the inspection. Doing more inspections won’t change how any defects came to be or really address the process itself. What is advocated is making the system better so that the outputs are better. For example, if a finished product is failing due to a component, then addressing the component quality is what is needed, not doing more inspections. However, doing inspections will help determine where to focus attention.
In our system, as detailed in our white paper, an inspection is a collection of observations that occur BEFORE defects (in the form of safety incidents) occur. These observations are either safe or unsafe based on the conditions and/or behaviors observed in the inspection. The behaviors and conditions being observed are usually detailed in the safety plan (desired state) and can vary. Examples include such things as “Fall protection at six feet”, “Proper ladder use”, or “Seat belt worn in motor vehicles”. Observations measure how well the safety process (actual state) is working. These leading indicators allow adjustment to the process before negative outcomes or defects (e.g. safety incidents) occur. In other words, collecting worksite observations and acting on the data to drive improvement is “building quality into the system.” These inputs plus process will improve the outcomes – in this case – incidents and injuries.
For example, if you identify a critical task or behavior that is out of tolerance (many at-risk observations and/or high-severity findings) nothing will change unless you make adjustments. An adjustment can range from feedback, to an informal action plan, to a strategic intervention. Once an adjustment is made, only further observations will tell you if what you did actually worked. These constant adjustments help to continually improve your process.
In deriving our safety truths, we compared inspection patterns with lagging indicators (incident data). Inspection count was a big correlating factor. Now, that being said, conducting inspections by itself will not necessarily improve your process, as Deming states. Please bear in mind that this is not the only safety truth we discovered. If you look at the other safety truths, we also identify that you can’t just have an inspector (safety) doing the observations. When this happens, ownership of safety by workers and line managers shifts to the safety department, which is bad. So inspections are great, but getting more out of the safety team is not good by itself. The other safety truths touch on quality of data collected and how findings are acted upon and addressed.
All of this simply ties back to the safety process. If you have no process or if you have a poor process, inspections will only tell you that an issue exists. Acting on the data is necessary in a proactive way in order to be safe.
For more information on our Safety Truths, please visit our Learning Center: http://www.predictivesolutions.com/learning-center/
4 comments
Tony O'Dea says:
May 31, 2012 at 8:18 pm (UTC 0)
cary,
I believe that both you and deming are correct. as deming says, you can’t ‘inspect quality into the process’, but, as you point out, inspections are necessary to reduce the acts and conditions that lead to accidents. To quote a colleague and fellow PS SafetyNet enthusiast- Bill Rose from Hunter Roberts (sorry for the misquote, Bill)- The way to reconcile the two views is by this statement- ‘we need to go from ‘prblem fixing’ to ‘problem solving’. Merely finding and fixing hazards will not improve quality until the underlying causes are identified and corrected.
TJ lyons says:
June 3, 2012 at 10:55 am (UTC 0)
I see the link between Safety and Quality in your thoughts. Deming would focus on every step along the way to unsure quality work was produced. Any weakness in the system CONTRIBUTED to the overall product. In safety we look for any CONTRIBUTOR to an incident in the same way. In the book the Drunkards Walk, a classic look at chance, the writer stated ““The chances of an event depend on the number of ways in which is can occur”. This was welcome to see for it confirms that even the little contributors matter in safety. Remove something from the chain of events and there are no events.
Kopstar says:
January 9, 2013 at 2:12 pm (UTC 0)
I can see the logic in the article above however I think it misses one fundamental point that is intrinsic in Deming’s thinking. The hidden point is that on a behavioural level where inspections or checks of any kind other than on oneself remove individual ownership and responsibility.
Therefore by relying on inspections to pick up actions absolves people (individuals) of the required behavioural change to improve. I believe this is one reason organisations in pursuit of Safety Exemplar (Shell Model) struggle to move from being caculative or reactive into proactive and generative.
Cary Usrey says:
January 9, 2013 at 2:31 pm (UTC 0)
Kopstar,
Thank you for the clafification. It is always tough to get all points across in a brief article, especially when dealing with such a broad topic.
I would love it if you would share more examples of how you feel it is best to become more proactive and generative. What works best, in your estimation?