The Impact of Sleep on Operator Performance

As an industry we understand that sleep has an impact on operator performance and capability. For a long time, data from this area came from looking at studies conducted in other sectors that are tangentially related, and to the best of my knowledge we have never seen any research on this topic specifically in the hydrocarbons sector. Fortunately the team at Shiraz University of Medical Sciences has gathered hard evidence based on research with an Iranian Petrochemical facility.

At Yokogawa we have been in the process of researching the limits of short-term memory capacity for production operators, specifically researching what we can do as an industry to ensure that the design of our systems are conducted in a manner that works within the limitations of human short-term memory. It was this research by Yokogawa that led to the discovery of the research at Shiraz University, and we hope to share more of what we have learnt by publishing more articles in the near future.

The Evidence is Clear

Short-Term Memory Test

What stood out to me from the research at Shiraz University was an understanding of the levels of short term memory deterioration over the course of a shift: we can see a clear deterioration between the 'before and after' data in both day-shift and night-shift operators. However, what is compelling is that we can see a much larger disparity in the 'before and after' data in night-shift employees as compared to day-shift employees.

For those who have attended the Advanced Decision Support workshops conducted by Yokogawa, you will have noticed our strong focus on understanding omission and commission errors. Particularly how each type of error contributes to some of the incidents that have been occurring.

Error of Ommission: a mistake that consists of not doing something you should have done, or not including something such as an amount or fact that should be included
— Cambridge Dictionary
Error of Commission: a mistake that consists of doing something wrong, such as including a wrong amount, or including an amount in the wrong place
— Cambridge Dictionary

Commission Errors

Thanks to this research, we now have reliable statistics to understand the linkage between omission errors and sleepiness, specifically the significant increase of omission errors over the course of a shift.

In this data, we can clearly see an increase of commission errors over the course of the shift as the operator is getting more tired, with day-shift operators being slightly more accurate in their work execution. But generally speaking both deteriorating at nearly the same rate.

The data starts to get very interesting when we get to omission errors, notice the significant jump in before and after results for night shift operators?

Omission Errors

It is clear that our operators know how to do their tasks based on the results from the number of commission errors; but as they are getting more sleepy and their short term memory suffers we can see a jump in the number of steps or tasks that they forget.

But the researchers took this one step further by trying to understand the sleepiness trend over the course of multiple days in a shift cycle, fortunately from a research perspective the petrochemical facility in question had the following shift structure: 7 nights, 7 days and 7 days off. This enables us to see the curve over a long period of time as compared to a facility with a 2/2/4 structure.

Shift Work

y-axis: Sleepiness | x-axis: Day

As you can see in the data for the day-shift operators, their performance is consistently good, potentially even improving as their body adjusts. But when we look at the night shift operator, they are able to “power through” the first few nights, but they soon hit a brick wall that is followed by large increases in sleepiness.

To quote the researchers of this study: “Increased sleepiness during the night-shift is related to the natural circadian rhythmicity of sleepiness and indicates lack of adaption of the body to night work.”

So we are left in a situation where “all three cognitive functions (memory, sustained attention, and reaction time) were impaired and showed a significant decrease at the end of both day and night shifts”. This is further compounded on night-shift with the sleepiness levels contributing to an increase in potential errors.

So what can we do about this?

The challenge is that there is not a single change that will solve the issue, and this cannot be solved in a “cookie cutter” manner. In some cases, dare I say, that these changes are not commercially viable for the industry to accept...

When Yokogawa is involved in this space (Advanced Decision Support) it is critical to take a holistic approach, but at the same time always keep the operator at the core. One of the core techniques is called “Work Domain Analysis” which is essentially a method for understanding all of the elements that the operator needs to conduct and at which timing; this is the first step in the “Cognitive Work Analysis” framework. It is based on this that we can then start to understand some of the mental loading that the operator will be under.

In addition to this a detailed ergonomic study of the operations teams work environment, as you can see in some of the data above it is abundantly clear that we need to provide the optimal working environment for our operators in a manner that enables them to reduce the risk of errors, and this is particularly important over the critical night shift period. Ranging from control room layout, through to lighting and sound.

How many companies are also considering how they support their operators in a truly holistic approach, from the perspective of work life as it relates to the operator performance? Should a company coach on overall health and well being? Should a company provide healthy and organic food for employees at the canteen instead of soda machines and food that is just average quality? Should we offer incentives to those who sleep better, like Aetna and other leading companies?

To quote the CEO of Aetna, who has quadrupled the companies value in just 6 years while incorporating wellness programs and sleep incentives, "It's all about the business fundamentals," and sleep is as fundamental as it gets. In collaboration with Duke University, Aetna has measured 69 minutes of improved productivity per employee per month after investing in health and mindfulness, which he feels is directly connected to the company's success over the last few years (watch the interview here). How would 69 extra minutes of productivity translate to the cognitive performance of an operator? The improvement potential could be very significant.

Additional Thoughts:

  • What would those omission and commission errors look like on a graph if we conducted the test multiple times throughout the shift?
  • What would the sleepiness curve look like if we improved the control room with regards to lighting (although there is research on this topic already), adaptive workload management, or giving operators a 30 minute nap mid-shift?
  • Would moving the staff to a 4/4/4 shift roster impact the long term results?

Questions like these are becoming more relevant each and every day with increased competition worldwide, and there is no doubt that the most forward looking companies in our industry will be addressing them sooner than later. What do you think?

Thank you to Shiraz University of Medical Sciences for their valuable research. Please inquire about Advanced Decision Support and our upcoming workshops to learn more about improving performance by looking to the operators first.