Job Safety

November 4, 2005

Nuclear Fatigue and the NRC

Think of the nuclear power industry and you think of controlled risk, of operators in huge control rooms monitoring dials and computer screens, all day every day, until something happens. And nothing much happens – a perfect formula for the conditions that promote fatigue and inattention.

The Nuclear Regulatory Commission “NRC” is deeply concerned with these issues though. They have done some of the best studies available on fitness for work testing. They are acutely aware that maintaining operator vigilance is very important.

In the following article, extracted from their web archives, particularly note the attention paid to the under representation of fatigue as a causal factor in reported incidents. This is a common problem in industry.

Extract from the NRC report: ASSESSMENT OF THE NRC’S “POLICY ON FACTORS CAUSING FATIGUE OF OPERATING PERSONNEL AT NUCLEAR REACTORS” (See NRC – Reading Room)

“Incidence of Fatigue-Related Events

Only a limited number of events at nuclear power plants have been directly attributed to fatigue. This may be in part be the result of the levels of defense-in-depth at nuclear power plants which redesigned to reduce the potential for personnel errors to have consequential effects on plant safety. However, the staff is not able to state with certainty the actual number of events that result from fatigue, and any estimates should be interpreted with caution. In fact, plant incident reports typically do not contain much of the critical information needed to determine the contribution of human error. As noted in an EPRI report concerning control room operator alertness, “it is often necessary to rely on anecdotal evidence when presenting the case for the critical importance of operator alertness in the safe and efficient operation of a nuclear power plant” (EPRI, 1990).

One reason that the staff believes that the number of events attributed to fatigue may be underrepresented is that the research literature and operational data suggest that the conditions of shift work in nuclear power operations are such that one would reasonably expect personnel to be at risk of fatigue-induced impairment.

This research includes the following examples:

• Studies show that personnel who work more than 12 hours a day are at increased risk of personnel error (Folkard, 1997; Dawson and Reid, 1997; Rosa, 1991). The NEI data concerning the use of deviations from the policy statement indicate that thousands of 24 person-hours are worked by personnel when they are at increased risk of impairment (see Section 6).

• Several studies show that nuclear power plant personnel exhibit circadian variations in alertness, and there are variations in the incidence of nuclear power plant personnel errors and events that coincide with these circadian variations in alertness (Bobko, 1998; Cox and Cox, 1996; Maloney, 1992).

• Studies show that personnel who are fatigued have impaired ability to maintain their attention (Harrison and Horne, 2000; Williamson, 2000; Bobko et al., 1998; Dawson and Reid, 1997; Dinges, 1995; Dinges, 1992; Rosa, 1991). The staff reviewed the Human Factors Information System (HFIS) data for 1997 through 1999 and found more than 5,000 instances of less-than-adequate independent verification, self-checking, and awareness or attention.

These data were compared with HFIS data on findings related to the use of overtime. This analysis revealed that nuclear plants with repeated findings concerning use of overtime have a 50-percent higher incidence of HFIS causal factors related to fatigue. Another reason that the staff believes that the number of events attributed to fatigue is underrepresented is that event investigation methodologies may not adequately address fatigue as a root cause, as indicated by the following factors:

• Depth of assessment – Most incidents at nuclear power plants are not subjected to an in-depth analysis that would identify fatigue as the underlying cause. Licensee event reporting requirements (10 CFR Parts 50.72 and 50.73) have not included causes of human performance problems at a level that would necessarily identify fatigue.

• Root cause assessment tools – There are no accepted criteria or structured approaches for evaluating the role of fatigue in accidents (Rosekind et al., 1997). As a result, when events are subjected to root cause assessment, fatigue may still not be identified. McCallum and Raby (1995) assessed investigation procedures employed by the NRC, the Federal Aviation Administration (FAA), and several international transport authorities. They found that the existing procedures do not adequately address the factors underlying fatigue as a causal element in cases in which the initial screening suggests fatigue as a factor.

• Lack of objective proof – When conducting a root cause analysis of events that involve personnel error, it is difficult to conclude that fatigue is a cause because there is little objective proof, absent the person sleeping, that the individual was impaired by fatigue. Even when nuclear plant personnel have been found with their eyes closed, they have asserted that they were not asleep, and investigators have concluded that the individual was “inattentive” (e.g., Peach Bottom, 1989).

• Ease of substantiating event causal factors – Fatigue degrades an individual’s abilities but does not necessarily cause the event. For example, the alert individual recognizes an error in a procedure, whereas the fatigued individual does not and implements an incorrect procedure. As a result, an investigator would focus on objective contributing factors (e.g., the procedure error) or describe the behavior (e.g., cognitive error) without citing a contributor, such as fatigue, that is difficult to substantiate.

• Accuracy of post-event observations – When individuals are debriefed following an incident, they may appear alert because of the stimulation of responding to, or potential consequences of, the event. Impairment from fatigue would not be readily apparent in such circumstances.

• Accuracy of self-assessment – Although self-assessment of fatigue can often indicate the level of fatigue, research suggests that other factors may influence such self-assessments (Wylie et al., 1996; Dinges, 1995 ). In addition, studies have shown that individuals may believe that they are relatively more alert than indicated by physiological indices (Wylie et al.,1996; Dinges, 1995; Rosekind and Schwartz, 1988).

• Veracity of self-assessment – For various reasons, individuals may be reluctant to
acknowledge that they were fatigued at the time of an event involving personnel error, including the implication that they were not fit for duty (Horne and Reyner, 1995).

One outcome of these challenges to identifying fatigue as a causal factor is that the investigation identifies the observable effects or consequences of fatigue, rather than fatigue itself. Other agencies and investigative bodies have come to similar conclusions concerning the attribution of fatigue to events. A letter from Jim Hall, Chairman of the NTSB, to DOT Secretary Rodney E. Slater, dated June 1, 1999, included the following statement.

Fatigue has remained a significant factor in transportation accidents since the Safety
Board’s 1989 recommendations were issued. Although generally accepted as a factor in transportation accidents, the exact number of accidents due to fatigue is difficult to determine and likely to be underestimated. The difficulty in determining the incidence of fatigue-related accidents is due, at least in part, to the difficulty in identifying fatigue as a causal or contributing factor in accidents. There is no comparable chemical test for identifying the presence of fatigue as there is for identifying the presence of drugs or alcohol; hence, it is often difficult to conclude unequivocally that fatigue was a causal or contributing factor in an accident. . . .

Although the data are not available to statistically determine the incidence of fatigue, the transportation industry has recognized that fatigue is a major factor in transportation accidents. Similarly, the DOT has concluded that fatigue statistics that are founded solely on accident reports underestimate the true extent of the problem (DOT, 65 FR 25545). In addition, the staff has learned that the Air Force Safety Center is revising the documentation to be used by accident investigation teams since they now believe that fatigue is underreported as a factor (Palmer et al., 1996). Also, a U.S. Coast Guard study suggests that direct measurement of fatigue may underestimate its true extent (Maritime Safety Committee, 1997). After the Coast
Guard revised its procedures for investigating events, they found that the contribution of fatigue was 20 times greater than previous estimates.

Summary
• Few events at U.S. nuclear power plants have been attributed to fatigue.
• The number of events attributed to fatigue should be interpreted with caution and cannot be reported with certainty.
• Many factors challenge the ability of event investigators to identify fatigue as a causal factor.”

The full text can be downloaded from the NRC site.
The “Reading Room” link below is well worth a visit. Try searching on a relevant topic and you will be amazed at the amount of material that is available.

NRC – Reading Room