About 10 years ago, while I was the Technical Editor for ReliablePlant and Machinery Lubrication magazines, I conducted a survey of our readers. My question was: “what goes wrong in your plant?” I utilized the failure cause categories outlined in the DOE STD 1004 – the standard for conducting root cause analysis in the nuclear power industry (note: DOE STD a great standard that’s applicable to any industry and is available to download as a pdf file at no charge).
I wasn’t surprised by the results – about 80% of all failures can be attributable to human factors. Topping the list is lack of or ineffective procedures. Personnel/Human error came in a close second. Lack of training and supervision combined for 23%. Design errors rounded out the human errors component with just under nice percent. Equipment and material problems (non-human causes) came in at just under 20%. About 70% of the reported human failures are organizational artifacts, where the organization failed to provide properly designed equipment, procedures and training about its proper operation and maintenance and supervision to ensure that work gets done correctly. Let’s delve into the human factors of failure in more detail.
Get Your Cheese In Line!
In his seminal book Human Error (1990), George Reason elegantly analogizes human error to the holes found in swiss cheese. Most organizations have a series of barriers that are intended to prevent people from making mistakes. However, these barriers, which Reason likens to slices of swiss cheese, aren’t perfect. Like swiss cheese, the barriers have holes. These holes could be due to organizational influences, unsafe or inadequate supervision, various precondition for unsafe (or unreliable) acts and the unsafe or unreliable acts themselves. If the holes in the slices of cheese align just right, an accident is the result (Figure 2). Our job in managing the human factors of failure is three part. First, where appropriate and risk-based, we need more slices of cheese – or more potential barriers. Secondly, we want to minimize the number of holes in each slice of cheese. Lastly, we want to minimize the size of each remaining hole.
For most organizations, the barriers to failure start with a policy. A policy that is signed by the ranking executive, often the CEO, and displayed at the organization’s places of business, represents the organization’s public commitment to an objective. When I enter a plant or company headquarters, I typically see framed policies for safety, environment performance, and quality/customer satisfaction. I rarely see a policy for equipment asset management. I find this perplexing. Equipment asset dependent companies rely upon those physical asset to achieve profits, increased shareholder value, safety performance, environmental sustainability performance, quality and customer satisfaction, and any other cooperate dashboard goals. Why wouldn’t the organization have a policy specifying how assets will be designed and acquired, operated, maintained, and, eventually disposed or reused? Publicly committing to asset management in the form of a policy is a great step toward preventing human error. Education and training are the next major slice of swiss cheese on which to focus. Regrettably, most organizations educate and train their people to either comply with regulatory requirements and/or to conform to social norms. Evidence of this fact lies in the fact that when the firm is suffering an economic down cycle, training is typically the first thing to get cut from the budget.
This is a pity, Estebe-Lloret, et al. (2018) found a positive and significant relationship between investing in employee education and training and organization performance. This stands to reason because people are a major slice of cheese in Reason’s Swiss Cheese Model, and training them well can significantly reduce the number and size of the holes. Technology can provide and engineered slice of cheese to reduce the risk of human error. In some instances, the engineered solution comes in the form of automation, which can reduce or, in some instances, eliminate human error, assuming the automation technology is properly engineered, stalled and configured. In other instances, the technology helps the human by error-proofing a system. This is called Poka-yoke in the Toyota Production System (TPS). Error-proofing improves the man-machine interface. It can include sensory cues to guide the human worker. For example, a gauge where the acceptable operating error is shaded green and the unacceptable area is shaded red is a very simple technology that helps to prevent human error. Other engineered interventions are much more elaborate. Procedures and checklist are another really important slice of cheese. Recall that lack of or ineffective procedures is this number one cause of failure according to my 2010 survey.
I generally find procedures in the plant to be significantly inadequate. Too often, I encounter vague procedures that frequently lack necessary fit, tolerance, quantity and quality details. For example, I often see inspection PMs that say “check pressure” or “check temperature” then they should say “verify that temperature/pressure is above/below/between X and/or Y.” It’s unrealistic to expect workers to remember all the details necessary to assure the reliable operation and precision maintenance of plant equipment. Another major slice of cheese in the human factors management model is supervision – active supervision. In many plants, supervisors do just about everything but supervise. Supervision is about setting priorities, allocating resources, coaching, mentoring, performing work quality checks. It’s not about running for parts, hunting for tools or performing the actual work itself. And, supervision occurs on the plant floor, not on email. Often, supervision is the last slice of cheese to prevent human error induced mistakes that compromise asset performance, safety, environmental performance and quality.
Bringing It All Together
"The fact is that 75-80% of the failures in your plant are due to human factors – mostly organizational artifacts. If you wish to improve the reliability of your equipment assets and, thus, your profit, safety, environmental and quality performance, you must tackle human performance. That means that you must take a long hard look at your policies, education and training, technological/engineered solutions, procedures, checklists and supervision to determine if they are adequate to prevent human error and, if not, add slices of cheese, reduce the number of holes in each slice and the size of each hole that you can’t completely eliminate. A good plan is to conduct day in the life of (DILO) studies to determine what causes confusion and a lack of clarity for operators, maintainers and supervisors – you front line team then backtrack to the policies, systems, processes and procedures that represent the root causes. The mistakes occur on the front line, but in 70% of the cases, they’re caused by some latent precursor weakness. Get your data in the DILOs then fix the problems upstream."