Sleep Medicine Reviews xxx (2011) 1e9

Contents lists available at ScienceDirect

Sleep Medicine Reviews journal homepage: www.elsevier.com/locate/smrv

THEORETICAL REVIEW

Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management Drew Dawson a, Janine Chapman*, Matthew J.W. Thomas b Centre for Sleep Research, Level 7, Playford Building, University of South Australia, City East Campus, Frome Road, Adelaide, SA 5000, Australia

a r t i c l e i n f o

s u m m a r y

Article history: Received 30 December 2010 Received in revised form 22 May 2011 Accepted 23 May 2011 Available online xxx

In this review we introduce the idea of a novel group of strategies for further reducing fatigue-related risk in the workplace. In contrast to the risk-reduction achieved by reducing the likelihood an individual will be working while fatigued (e.g., by restricting hours of work), fatigue-proofing strategies are adaptive and protective risk-reduction behaviours that improve the resilience of a system of work. That is, they increase the likelihood that a fatigue-related error will be detected and not translate into accident or injury, thus reducing vulnerability to fatigue-related error. The first part of the review outlines the theoretical underpinnings of this approach and gives a series of ethnographically derived examples of informal fatigue-proofing strategies used in a variety of industries. A preliminary conceptual and methodological framework for the systematic identification, development and evaluation of fatigueproofing strategies is then presented for integration into the wider organisational safety system. The review clearly identifies fatigue-proofing as a potentially valuable strategy to significantly lower fatiguerelated risk independent of changes to working hours. This is of particular relevance to organisations where fatigue is difficult to manage using reductions in working hours due to operational circumstances, or the paradoxical consequences for overall safety associated with reduced working hours. ! 2011 Elsevier Ltd. All rights reserved.

Keywords: Fatigue Workplace Safety Error Management

Introduction In recent decades, fatigue-related accidents and injuries have been increasingly subject to regulatory scrutiny and many developed countries have now identified fatigue as a work place hazard. In an attempt to reduce the risks associated with fatigue, researchers have extensively studied the ways in which fatigue alters the likelihood of human error and, as a consequence, the likelihood of subsequent accidents and injuries. In general, three clear lines of evidence have associated increasing levels of fatigue with decrements in cognitive function1e3; impaired task performance4,5; increases in error and accident rates,6,7 and ultimately, reduced safety.8e10 Despite these advances, however, we do not yet have a strong body of evidence directly linking increased cognitive Abbreviations: Fatigue, for the purposes of this review all references to fatigue imply mental fatigue unless specifically indicated otherwise; FPS, fatigue-proofing strategy; FRMS, fatigue-risk management system; HOS, hours of service; PSWM, prior sleep/wake model; SMS, safety management system; SOP, standard operating procedure; TEM, threat and error management. * Corresponding author. Tel.: þ61 8 8302 2453; fax: þ61 8 8302 6623. E-mail addresses: [email protected] (D. Dawson), janine.chapman@ unisa.edu.au (J. Chapman), [email protected] (M.J.W. Thomas). a Tel.: þ61 8 8302 6624; fax: þ61 8 8302 6623. b Tel.: þ61 8 8302 1966; fax: þ61 8 8302 6623.

impairment to accident and injuries in a simple monotonic manner. If we are to effectively regulate work places in order to reduce fatigue-related risk, it is important that we have a much clearer understanding of the relationship between fatigue, the nature of the subsequent impairment, and the way in which fatigue-induced errors of cognition are translated into unsafe behaviours and to accidents and injuries. Without such an understanding it will be difficult to develop systems that reduce fatigue-related risk effectively without imposing significant and perhaps unnecessary operational costs. Traditionally, the regulatory framework for managing fatiguerelated risk was compliance with prescribed hours-of-service limits to shift and break durations. Recent years have seen a shift away from prescriptive rule sets towards approaches that incorporate fatigue management within the general context of the safety management system (SMS).11e13 The SMS methodology for fatiguerisk management can be represented using Reason’s hazard control framework, in which an accident or incident is the end result of a longer causal chain of events rather than the immediate error of the human at the controls.14,15 According to Reason’s model, safety is compromised when a hazard is able to penetrate successive layers of defence, permitting a trajectory of accident opportunity. Adverse events are seen to arise from a combination of active failures, which relate to the unsafe acts committed by people who are in direct

1087-0792/$ e see front matter ! 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.smrv.2011.05.004

Please cite this article in press as: Dawson D, et al., Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management, Sleep Medicine Reviews (2011), doi:10.1016/j.smrv.2011.05.004

2

D. Dawson et al. / Sleep Medicine Reviews xxx (2011) 1e9

contact with the system, and latent conditions, relating to the more distal factors within a system that have the potential to increase safety risk. Enhancing safety therefore involves the identification and management of both active failures and latent conditions within a system of work, and gaining a clear understanding of how the two interact to determine outcomes. Helmreich’s16,17 model of threat and error management (TEM) builds on this concept by providing a more detailed focus on the specific individual and team countermeasures to specific hazards. In this instance, the term ‘error’ is used to describe the active failures at task level, and the term ‘threat’ maps directly onto the latent condition component by describing the factors that an individual or team must identify and manage in order to maintain safety. Importantly, the TEM conceptualises hazard management as a process of joint responsibility between organisations and individuals, and places significance on the ubiquitous nature of error in socio-technical systems. This, in turn, emphasises the need to build better mechanisms to manage the inevitable occurrence of error during normal everyday operations. Taken together, these models represent a holistic approach to error management, which involves limiting the incidence of errors and also creating systems that are better able to tolerate or contain error. In relation to fatigue-risk management, this represents a shift from passive systems of prescriptive and compliance toward a more

dynamic combination of risk-prevention; early identification of potential risk, and effective strategies to mitigate the likelihood of risk translating into a fatigue-related incident. Application of SMS to fatigue-risk management Integrating the principles underlying the Reason14 and Helmreich16,17 models, an SMS-based approach to fatigue was first articulated by Dawson and McCulloch.11 Based on the defences-indepth framework, this model conceptualises a fatigue-risk management system (FRMS) as a series of defensive layers that can be positioned at four points along the potential event trajectory. Each of these points provides an opportunity to identify and prevent fatigue-related accidents at different levels of control, offering a proactive rather than reactive approach to fatigue management (Fig. 1). Level 1 of the trajectory is the degree of sleep opportunity provided by a specific work pattern. Most qualitative rostering guidelines18e20 or more recent quantitative fatigue-modelling tools21e24 ostensibly provide an index of the relative sleep opportunity associated with the pattern of work and a global index of the level of work-related fatigue associated with a schedule or roster. By using Level 1 tools it is possible to ensure that employees are provided with an adequate opportunity to rest and recover and, at

Fig. 1. Fatigue-risk trajectory. There are multiple layers that precede a fatigue-related incident, for which there are identifiable hazards and controls. An effective fatigue-risk management system should attempt to manage each layer of risk.11 HOS: hours of service; PSWM: prior sleep/wake model; SMS: safety management system

Please cite this article in press as: Dawson D, et al., Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management, Sleep Medicine Reviews (2011), doi:10.1016/j.smrv.2011.05.004

D. Dawson et al. / Sleep Medicine Reviews xxx (2011) 1e9

least in theory, ensure that they are at an appropriate level of risk due to work-related fatigue. Level 2 is the actual amount of sleep obtained. While the provision of an adequate sleep opportunity is necessary, it is not always sufficient to ensure that an employee is adequately rested. Level 2 controls focus on identifying occasions where an adequate sleep opportunity has not produced sufficient sleep to ensure an employee is fit-for-duty. In general, this is achieved by setting minimum sleep and maximum wake durations using simple mental heuristics or more complex software-based models. For a more comprehensive discussion of these approaches, see Thomas and Ferguson25 and Dawson and McCulloch.11 Level 3 relates to the behavioural symptoms of fatigue. Sometimes an adequate sleep opportunity and a nominally adequate amount of actual sleep can still be associated with increased levels of fatigue. This might be due to factors such as sleep disorders,26,27 individual differences in the recuperative value of the sleep requirement,28,29 or idiopathic situational reasons. In these circumstances, techniques for identifying at-risk individuals can use the presence and/or frequency of fatigue-related behaviours (e.g., yawning, struggling to stay awake, degraded task performance) to indicate an elevated level of fatigue-related risk. Many of the fatigue-detection technologies to emerge in recent years30e32 along with traditional self-report scales33e35 provide examples of Level 3 tools designed to identify individuals at risk based on behavioural indicators of fatigue (see also review by Balkin et al.).36 Level 4 is concerned with the assessment and control of fatiguerelated error. As previously discussed, fatigued individuals are more likely to make errors that result in an increased level of fatiguerelated risk. At the task level, it is possible to identify the indicators of impaired cognitive performance associated with elevated fatigue. Thus, it is possible to introduce methods that decrease the likelihood a fatigued individual operating in the workplace will make an error that leads to an accident or injury. An effective Level 4 control should therefore have two key objectives: 1) to reliably recognise the indicators of fatigue-related impaired cognitive performance within the workplace, and 2) to develop and implement formal procedures that serve to compensate for the impaired performance, thus reducing the potential for accident opportunity. Despite the four levels of risk identification and control, all safety systems have the potential to fail at some point in time. If properly monitored, however, fatigue-related incidents can provide the individual and organisation with a potential learning opportunity and insight into how to prevent future events re-occurring. Level 5 provides an incident analysis control mechanism to feedback into system reform. This information is used to improve identification of fatigue-related risk and to continually update Level 1e4 tools in ways that will better prevent future fatigue-related error, incidents and accidents. Summary of strategies for reducing fatigue-related risk Each of the steps in the FRMS provides the opportunity to identify potential incidents, and the presence (or absence) of appropriate control mechanisms in the system. Based on this, there are two possible approaches that can be used to reduce fatiguerelated risk: 1) fatigue-reduction strategies, and 2) fatigueproofing strategies. Fatigue-reduction strategies are the techniques for reducing the likelihood that a fatigued individual is operating in the work place. To reiterate, this is typically achieved through prescription of maximum shift and minimum break duration for individual shifts or work periods, or alternatively, bio-mathematical models for estimating work-related fatigue can be used to ensure an adequate sleep opportunity.37 These controls correspond to Level 1 of the

3

FRMS. Countermeasures based on implementing Level 2 and 3 controls can also be used to reduce the likelihood that a fatigued individual enters or remains in the workplace by identifying and precluding them on the basis of insufficient sleep (Level 2 controls) or the presence of behavioural indicators (Level 3 controls). In contrast, fatigue-proofing strategies (FPS) are techniques for decreasing the likelihood a fatigued individual operating in the workplace will make an error that leads to accident or injury. Despite the implementation of Level 1e3 tools, SMS theory,14,15 along with practical experience, suggests that these controls will never be wholly effective. Whether due to circumstances beyond the control of the organisation, or because the risk of an individual ‘not working’ are considered acceptable, situations will inevitably arise where fatigued individuals are present in the work force. In this case, a different class of controls is required. In the event that an individual is working while fatigued, risk can only be reduced further by reducing the likelihood that a fatigue-related error will cause an accident or injury. This mode of risk management can be categorised as ‘fatigue-proofing’ rather than fatigue-reduction. That is, a process of (re)designing a ‘system of work’ in ways that make it more resilient (error tolerant) to fatigue-related error. In our opinion, the effective management of fatigue-related risk within an organisation requires the systematic use of both fatiguereduction and fatigue-proofing strategies to be adopted and formalised as integral and complementary elements of an FRMS. The development and application of the fatigue-reducing control procedures at Levels 1e3 of the error trajectory have been discussed extensively elsewhere.11,37 However, formal application of Level 4 fatigue-proofing controls within the safety system of the workplace has yet to be addressed. The remainder of this review will present a description and preliminary framework for the development, implementation and evaluation of fatigue-proofing strategies from an error management perspective. Examples of informal fatigue-proofing strategies In our work with industry, we have observed that the use of informal FPSs within organisations is relatively common. They are, however, typically hidden since they represent implicit rather than formal elements of the safety system. This is because formal risk controls for fatigue do not usually encompass the notion of ‘fatigue-proofing’ as a part of the formal SMS, and as a result employees may be culturally blind to their role and function. In many organisations, informal FPSs have evolved as part of the traditional work practices within a work group. To illustrate this principle, some of our observations of simple yet effective informal fatigue-proofing behaviours from a selection of industries are detailed below. Maritime industry Marine pilots may be required to be at sea for several days with unpredictable working times based on tides, weather, and other shipping traffic. In these scenarios, common fatigue-related mistakes include navigation command errors.38,39 We have observed fatigued marine pilots operating on the bridge using an informal FPS to help prevent these errors. Verbal navigation commands are reinforced by a concomitant hand signal in the appropriate direction, accompanied by a pre-emptive request for the helmsman to ‘call back’ the pilot’s command to ensure that what was said was correct; correctly heard, and correctly actioned. This combination of a physical indication of the change-indirection and ‘call-back’ increases the likelihood of ‘trapping’ inadvertent errors. Where a helmsman is provided with and replies with a verbal and physical indication of direction, a fatigue-related error of comprehension is less likely.

Please cite this article in press as: Dawson D, et al., Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management, Sleep Medicine Reviews (2011), doi:10.1016/j.smrv.2011.05.004

4

D. Dawson et al. / Sleep Medicine Reviews xxx (2011) 1e9

Aviation industry Long-haul airline pilots will typically undertake flights of 14e16 h duration. Despite best intentions, there are occasions when pilots are not able to sleep in-flight and may be required to land the plane fatigued.40 Based on the existing literature,1,41 we know that if a non-typical event occurs during descent it may produce time-pressured decision-making, which can lead to poor cockpit performance and significantly increase the risk of an error. We have observed an informal FPS where pilots self-identify themselves as ‘fatigued’ to their co-pilot when not adequately rested. They then commence preparation for descent several minutes earlier than is usually required. This strategy helps to ensure that all critical calculations and decisions are undertaken earlier in the descent, preventing time-pressure and thus enhancing safety. Health care Doctors performing on-call duty are particularly susceptible to the detrimental effects of sleep inertia, or the impaired cognition, grogginess and disorientation that commonly occur on awakening from any sleep period.42 Cognitive performance immediately on waking is associated with impairments in short-term memory, counting skills, speed of processing, and number fact and lexical retrieval.43 An informal fatigue-proofing technique used in this instance is for the caller to give the instructions for the on-call appointment, and then follow this up by asking the doctor if the call has woken them. If this is affirmed, the instructions will be repeated again until the caller is satisfied that they have been correctly heard, understood and processed. Power industry In the power industry, ‘line teams’ are often required to attend emergency power outages at all hours in relatively unprotected scenarios. In our discussions with these groups, a frequently raised topic was the way in which they informally apply higher-thanusual levels of jocularity and teasing in the early hours of the morning. This strategy allowed supervisors to identify fatigued coworkers based on their irritability in responding and refusal to engage in teasing behaviour. They were then typically assigned to tasks at a lower risk of a fatigue-related error than those who exhibited more robust team behaviour. In this context, it would appear that employees use increased jocularity and teasing as a sensitive mood ‘probe’ for identifying fatigued co-workers in order to assess their ‘level of impairment’ and to reduce the likelihood they are engaging in higher-risk activities. Common themes within informal fatigue-proofing strategies The FPSs that develop informally within organisations are individually tailored and targeted to both the situational and cultural context of the workplace. Gaining a full understanding the contextual operation of each strategy is important due to the unique characteristics of the critical task being performed, the associated indicators of fatigue, and the specific impact and consequences of fatigue-related error associated with the task. However, our examples suggest that across industries, there are two common themes underlying the formation of FPSs. The first theme relates to the pre-signalling of elevated risk, where workers informally communicate their elevated level of fatigue to others to increase the level of risk mitigation. Examples from maritime and aviation demonstrate that the selfidentification and communication of fatigue can: a) alert others

to the need for confirmation on critical decisions and procedures, and b) allow additional task preparation time to prevent time pressure. These strategies therefore appear particularly useful in conveying to others the need for extra vigilance, reducing the chance of the fatigued individuals’ error leading to an adverse outcome. The second common theme seen in informal FPSs is the high level of scrutiny for potential error; that is watching for and identifying the indicators of elevated levels of fatigue-related risk in others. Observations from the power industry show that workers are able to preliminarily assess the deterioration in mood and social interaction that is often attributable to fatigue-related impairment and act accordingly. In the situation of on-call doctors, the caller is able to identify the potential cognitive impairment and risks associated with sleep inertia and put preventive strategies in place. Thus, strategies based on high levels of error scrutiny may be of particular benefit where fatigued individuals are unaware or unable to assess their own levels of impaired alertness. However, it is important to note that the pre-signalling of risk and error scrutiny are not necessarily exclusive component of an FPS in practice. One potential difficulty with the self-assessment of fatigue is that while people can estimate their level of fatigue or alertness with some degree of reliability, there is little evidence to support the notion that individuals can use this subjective assessment to make sound judgements about the concomitant level of risk.11 We can see from the aviation example that when pilots presignal their level of fatigue, this in turn alerts the co-pilot that extra vigilance is needed. Similarly, when marine pilots self-identify as fatigued and request additional verbal and hand signalling checks, the helmsman is then alerted to the increased likelihood of error and error scrutiny will be enhanced. The pre-signalling of risk and a high level of error scrutiny should therefore work synergistically towards the development of FPSs in workplaces where possible. Fatigue-proofing as an integral part of the SMS In each of the examples, we can see the informal development of an implicit standard operating procedure (SOP) that has the potential to identify and control fatigue-related risk using the principle of ‘fatigue-proofing’. The introduction of FPSs can increase the organisational awareness of fatigue and the likelihood that cost-effective risk controls can be put in place effectively. This has considerable potential to enhance safety when operational circumstances may preclude traditional fatigue risk reduction strategies derived from Level 1, 2 & 3 in the FRMS. However, in many organisations, the creation and persistence of FPS are often haphazard in nature and intent. They are often observed and passed on through long-standing workplace customs and undocumented mentoring systems, thus operating as a ‘guerrilla’ element of an organisational SMS. Due to this, there is a lack of clear criteria for determining what is, or is not, the appropriate or effective strategy in a given situation for a specific group of employees. Furthermore, informal FPSs may operate without management endorsement. In this instance, the informal status of a fatigue-proofing SOP may carry considerable risk due to: 1) delayed, partial and/or inaccurate transmission across the organisation 2) failure to quickly identify and eliminate ineffective or dysfunctional strategies It is therefore critical that where informal FPSs can be identified they are clearly articulated, evaluated for efficacy and where appropriate, standardised and documented so as to form part of the formal SMS for the organisation.

Please cite this article in press as: Dawson D, et al., Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management, Sleep Medicine Reviews (2011), doi:10.1016/j.smrv.2011.05.004

D. Dawson et al. / Sleep Medicine Reviews xxx (2011) 1e9

Methodological framework for integrating FPS The majority of fatigue research to date has focussed on quantitative methodologies and empirical approaches to establish the effects of fatigue on basic elements of task behaviour (e.g., response times, decision making, etc). As our examples show, however, informal FPSs may be complex and subtle behaviours that are deeply embedded within the culture of the organisation. Employees, therefore, may not initially recognise their behaviours as ‘fatigue-proofing strategies’ per se, but simply the ‘way we do things around here’. FPSs may be implemented unconsciously with little understanding of why they have evolved or as to their purpose. As such, the fatigue-proofing concept may be difficult to describe quantitatively without significant consultation, discussion and critical reflection. The identification and classification of FPSs will therefore require careful application of a mixed-methods approach that is sensitive to the context from which they are drawn.44 To integrate fatigue-proofing as a formal element of the organisational SMS, we outline four main phases. The first two phases relate to the ‘harvesting’ of existing informal FPSs currently operating at both the local work group level and in other parts of the organisation or other relevant organisations. The third phase outlines the methodology for identifying common areas of fatiguerelated risk and developing further FPSs to address them. The final phase relates to evaluation and assessment of FPSs in practice, to inform their wider dissemination into organisational policy and procedure. Given the complex and exploratory nature of fatigueproofing, we suggest the use of both qualitative and quantitative techniques, and an interactive, action research-based paradigm to facilitate a co-operative process of knowledge-gathering, problem diagnosis and implementation within the work setting and wider SMS.45 Harvesting fatigue-proofing strategies Phase 1: identifying FPS at local level The first step in this process would be to initiate discussion around the ideas underlying the principles of fatigue-related error, preferably as a component of staff fatigue-awareness training. A brief assessment of current staff knowledge and information needs may be useful in order to tailor the training sessions to the specific requirements of the workplace. Because effective FPSs rely on the ability to recognise the indicators of fatigue-related impairment in both the self and others, an important educational focus should be the ways in which workplace-specific skills and functions are commonly affected by fatigue. For example, sessions may address behavioural indicators such as a decline in social interaction, attitude and mood as demonstrated in our observations from the power industry, as well as inconsistent performance and degradation in attention, reasoning and accuracy.46 In conjunction with the training sessions, discussions at work-place level are required to develop a clear working definition of the nature and operation of an FPS. Again, these discussions may be introduced as part of the normal safety discourse and utilise methods such as tool-box meetings47 to provide a mechanism for communication, consultation and planning in which all employees can contribute. When a comprehensive understanding of the FPS concept is established, field-based qualitative methods such as focus groups and semi-structured interviews can be applied to elicit knowledge of the current strategies familiar to, and applied within the work group.48 As a reasonable level of domain knowledge can be assumed, it would be useful to focus the themes for discussion on specific areas of relevance, for example the experience and

5

perception of fatigue-related impairment in the self and others; how fatigue-related impairments are handled on a personal and group level, and if the strategies in place seem effective. In addition to the qualitative interviews, more focused ‘case study’ group discussions relating to actual fatigue-related incidents in the workplace may be a valuable tool for generating further data. The act of visualising or ‘brainstorming’ specific events during case study analysis may cue additional information that would otherwise have been lost or forgotten.49 As informal FPSs are often implemented unconsciously, however, the qualitative knowledge-eliciting techniques should be complimented and enhanced by a series of parallel work place observations. An ethnographical approach in this instance is useful to generate understanding of the workplace culture through the employee’s lived experience, which will help to contextualise the examples communicated through interviews and discussion groups.50 However, it should be noted that in some situations, indepth observation may be problematic or impossible to implement, for example, in the case of observing a pilot’s performance in a single-seater aircraft. The degree of flexibility afforded to these methods is therefore dependent on the nature of the workplace setting. If an ethnographical focus is unfeasible, it may be beneficial to employ supplementary knowledge-eliciting techniques, such as open-ended questionnaires or opinion surveys. The overriding aim of the initial harvesting phase is to generate the richest possible knowledge base that is both representative of the work group in question and responsive to the specific operational needs of the organisation. A note on dysfunctional or unsafe strategies Taken together, the data collected in Phase 1 will identify the informal FPSs used by individuals or workgroups. The next stage would be to collate them and to prioritise the implementation of those that are likely, at least in theory, to produce the most significant level of risk reduction for the work group. However, in categorising the FPSs, it may become apparent that some of the informal strategies currently in place are dysfunctional, unsafe or potentially dangerous. An example of this comes from a study of South American long-distance truck drivers who were asked via survey to identify behaviours that they regularly employed to reduce fatigue-related risk.51 While some of these were demonstrably effective (e.g., additional self-managed breaks), others such as ‘winding down the window’ or ‘playing loud music’ were not, and could also potentially act as a further distraction. More seriously, truck drivers have also reported consciously increasing their driving speed when tired, with the rationale that the ‘danger’ element of speeding would make them more alert. Rather than representing a potential FPS, this is clearly a critical safety hazard. In cases such as this, the data collection in Phase 1 can help to alert the organisation to the hazard and preventative regulation can be implemented accordingly. Informal FPSs for which the benefit is more ambiguous but have no associated counter-intuitive or unsafe connotations can be subject to further evaluation later in the process. Phase 2: identifying FPS outside the work group Having identified the FPSs currently in place at the local level, a second phase of harvesting should involve identifying the informal FPSs used by similar groups of employees outside the work group. Techniques analogous to those employed in Phase 1 are also applicable here. However, given the informal way in which these strategies evolve and disseminate, there is likely to be considerable local variation in their use and benchmarking may

Please cite this article in press as: Dawson D, et al., Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management, Sleep Medicine Reviews (2011), doi:10.1016/j.smrv.2011.05.004

6

D. Dawson et al. / Sleep Medicine Reviews xxx (2011) 1e9

provide a useful additional approach to risk reduction.52 In addition to discussions and interviews with frontline staff, it would also be valuable at this stage to include higher level expert, professional and stakeholder input. Specifically, semi-structured focus groups may facilitate useful discussion around professional boundaries and norms in relation to fatigue-risk management, and gain insight into the organisational factors and incentives that may impact the integration of standardised FPSs into the SMS. Developing new fatigue-proofing strategies Phase 3: identifying common fatigue-related error and developing further FPSs While the first two phases identified the informal FPSs currently in use, it is also possible that there are potential but as yet undefined FPSs that might be usefully implemented. The third phase should involve identifying common fatigue-related errors and an active process for developing FPSs to address the errors that pose the greatest risk. Data on areas of common fatigue-related error can be initially identified by consulting the record of incidents and near-misses within the organisation, and expanded using the brainstorming and group session techniques described in Phase 1 and 2. When areas of need have been ranked and prioritised, each common area of fatigue-related error can be the topic of more indepth qualitative investigation to determine ways to formulate optimum strategies. Congruent with the common themes highlighted in our earlier examples of informal FPSs, important themes for discussion may include the nature of the specific error and ways to explicitly pre-signal the level of elevated risk or increase levels of error scrutiny. These discussions should purposefully target staff at different levels throughout the organisation in order to gain a full range of engagement and experience with error management. The information gathered from these sessions will provide the basis for novel FPSs to be generated. Evaluation As there are likely to be considerable variations in the type and effectiveness of potential strategies, stringent evaluation is essential prior to FPSs being standardised and incorporated into the formal SOPs of an organisation. Activity at Phase 1-3 serves to identify the most promising FPSs to implement at the local or organisational level. Phase 4 will then assess the utility and effectiveness of the initial stages to ensure that subsequent procedure has clear empirical underpinnings and a considerable degree of evidence-based support. Outcome and process evaluations described below are recommended to provide justification for the continuing use or deletion of FPSs at local levels and externally. Measuring outcomes We offer two suggestions for the experimental outcome evaluation of FPSs. First are simulator studies, designed to assess the usefulness of FPSs in a created environment representing the functions of real workplace equipment under operational conditions. Simulators are typically designed for use in complex systems, and are common in the aviation, marine and surface transportation industries as tools that explore interactive behaviours in a fail-safe environment.53 Simulator experiments can provide the levels of consistency and control required to compare and contrast the effectiveness of FPSs, particularly when the nature of study involves a manifest breakdown in performance. An additional benefit is the ease with which rarely observed, out-of-course and safety critical scenarios can be represented. This approach may therefore have

particular utility in industries where observing employees using qualitative field-based methods is unfeasible, and will also provide a useful means of generating quantifiable evidence of the efficacy of FPSs prior to testing them in a more applied setting. The second suggestion for testing FPSs is a workplace pilot trial. Ideally and where practicable, this would involve a longitudinal cluster randomised design, where workgroups or sites are allocated to experimental (FPSs in place alongside standard procedure) or control (standard procedure only) groups. Due to the complex nature of the organisational SMS, it is likely that multiple outcome measures are required to effectively assess the impact of FPSs in the workplace. The primary outcome measure, such as the change in the number of fatigue-related errors and incidents, should ideally be standardised by exposure, for example the number of workerhours exposed to the hazard or workplace. However, depending on the error frequency rate and the size of the work group being tested, other outcome measures such as near-misses may be particularly useful in determining the effectiveness of FPSs.54 Secondary outcome measures of interest may include changes in knowledge and attitudes towards fatigue-related safety, and changes in reporting of relevant data, as these would provide information useful to the conceptual development of the trial. Understanding processes Process evaluations, which explore the ways in which the intervention is implemented, can provide valuable insight into why a strategy may fail, or the causal mechanisms of how a successful strategy works. A process evaluation nested inside a workplace trial may be particularly useful for identifying contextual factors associated with variations in outcome.55 The observational techniques described in Phase 1 would be useful to detect variations in the delivery and application of FPSs, or to determine the cause of any unexpected consequence of the strategy in action. Implementation focus groups can also be conducted to discuss barriers and facilitators to the pragmatic use of FPSs that may not have been apparent during the initial knowledge-gathering phases. Although the additional fieldwork should not be viewed as a substitute for quantitative evaluation, the implementation analysis phase can provide valuable insight into how and why particular strategies may be unfeasible for particular contexts or work situations, or how promising FPSs can be refined and optimised prior to their standardisation. These new insights can then feedback into the action research cycle to streamline further FPS development and inform the design of further experimental and applied workplace research. Conclusions on the mode of approach As a general rule, the size and nature of the work group and/or organisation will determine the mix of qualitative and quantitative approaches required. Larger and more diverse workplaces will initially generate more complex fatigue-related error profiles. In these instances, careful qualitative investigation is needed to supplement the quantitative techniques to ensure the strategies are contextually representative, appropriate and effective. As lack of effect may reflect implementation failure rather than genuine ineffectiveness of the safety control, it is particularly important to undertake sufficient process evaluation alongside experimental work to limit variation in implementation and outcome. The mixed methods approach outlined here offers a robust framework for: 1) identifying the scope of informal FPSs already in operation at local level and ironing out dysfunctional or unsafe strategies; 2) harvesting existing FPSs outside the work group to allow comparison of best practice across organisations; 3) prioritising areas likely to present the greatest level of fatigue-related risk and developing

Please cite this article in press as: Dawson D, et al., Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management, Sleep Medicine Reviews (2011), doi:10.1016/j.smrv.2011.05.004

D. Dawson et al. / Sleep Medicine Reviews xxx (2011) 1e9

novel FPSs, and 4) evaluating the practical operation of formal FPSs, with an empirical and contextual implementation focus. Perhaps most importantly, the combination of quantitative and qualitative evaluation considerably strengthens the internal validity of the process by allowing cross-validation of findings, generating strong evidence on which to base future system reform. What this approach assumes about the organisational SMS Much of the previous discussion has assumed certain pre-existing elements of the general SMS that will incorporate this process. While a comprehensive discussion of this issue is beyond the scope of this document, it is important to identify at least some of the more salient factors. This will, at minimum, discourage organisations from over extending their efforts in a way that proves futile in the short term, and allow them to set reasonable expectations to develop at a rate that remains within the overall envelope of the SMS. Ability to address hazards using a risk-based framework The principles of risk management as defined by ISO 3100056 are probably the most critical piece of safety infrastructure on which fatigue-proofing is predicated. In brief, this requires that an organisation is able to identify and rank hazards based on the likelihood and consequence of (fatigue-related) error. By doing so, an organisation is able to determine the relative priority of different errors, hazards, and the relative efficacy of intervention in terms of reduced risk. Without such a framework it is difficult to determine the opportunity cost of addressing different errors or hazards systematically across work groups or organisations. Ability to identify ‘at risk’ individuals Another key assumption in the FPS approach is the capacity to reliably recognise the indicators of fatigue-related impairment. This applies to self-identification in order to pre-signal an elevated level of risk, and identifying cues in others through means of increased error scrutiny. In doing so, the appropriate FPS can be employed, shifting the individual from a ‘normal’ to ‘fatigued’ mode of operations (for a further discussion of this see Dawson et al.37). Tools at Levels 1, 2 and 3 of the FRMS can be used to identify fatigue-related risk at both the group level (Level 1: sleep opportunity) and individual level (Levels 2 & 3: actual prior sleep-wake behaviour and behavioural indicators, respectively), thus providing clear guidelines on when a shift to fatigued modes of operation is necessary. An alternative approach would be for an organisation to make the ‘fatigued’ mode the default operational mode. In this respect, all individuals would be assumed to be tired and fatigue-proofing SOPs would be used irrespective of objective indicators of fatigue. It could be that if used in conjunction with the original FPS development framework, this approach may act as a further ‘defence in depth’ layer in the system, while introducing a degree of procedural flexibility if required. Ability to formalise and enforce the use of FPSs An effective fatigue-proofing approach requires that the organisation have an advanced safety culture, and a high degree of staff and management flexibility and trust. Organisations with a high reliance on prescription and compliance may struggle with an SMSbased approach to fatigue-proofing. For example, effective risk management depends crucially on establishing an open reporting culture.57 Without this principle it is unlikely that an individual will self-identify as fatigued or voluntarily offer information regarding a fatigue-related error, which could produce a failure to implement

7

FPSs correctly and may potentially pose a greater risk to fatiguerelated safety. It should be noted that anonymous reporting of errors may offer a practical solution to identifying common fatigue-related errors in organisations that lack sufficient maturity in their approach to safety. However, promotion of a trusting organisational culture that works openly and collectively to strengthen defences and recover error is an essential requirement of an effective FPS approach, in addition to a highly resilient SMS.

Summary and conclusions Fatigue-related risk in the workplace is a serious safety hazard. Drawing upon the principles of Risk Management, SMS and TEM theory, the FRMS states that fatigue-related risk can be managed using a ‘defences-in-depth’ framework. Moving away from traditional hours-of-service restrictions, the FRMS highlights two distinct approaches to reducing fatigue-related risk: fatiguereduction strategies, relating to Levels 1e3 on the errortrajectory, and fatigue-proofing strategies, relating to Level 4. The present review offers a comprehensive conceptual and methodological framework for integrating FPSs into the organisational SMS, thus creating an additional layer of defence aimed at minimising the likelihood that a fatigue-related error will lead to an adverse safety outcome. A main strength of this approach lies in the ecological validity of using effective FPSs identified and evaluated by staff currently engaged in carrying out a specific task. The framework we propose for FPS development is theory-based, consultative and respects the expertise and experience of staff. It acknowledges the considerable behavioural plasticity of most complex task behaviour and respects the capacity of employees to adapt their behaviour and work safely when fatigued. Where it is possible to identify fatigue-related errors and build effective fatigue-proofing SOPs that are highly error-tolerant, it is also possible to significantly reduce the likelihood of an error translating into an accident or incident without unnecessary effects on operational cost and/or flexibility. This is very attractive in industries where reductions in working hours can impact more negatively on organisational or employee income than the benefits associated with reduced fatigue-related risk. Most importantly, it also provides a means of resolving the impasse over reducing working hours as the sole means of reducing fatigue-related risk. In many occupations, employees are acutely aware of the complex behavioural ecology that surrounds ‘working while fatigued’ and the potentially paradoxical effect associated with restricting hours of work. For example, a reduction in working hours for junior doctors may well lower levels of fatigue-related risk, but can also concurrently increase risk due to reduced availability of health care providers or increased concurrent patient load.58 Thus, FPSs can provide employees with a well-targeted approach to risk reduction without the paradoxical effects associated with restricted working hours. As the FPS framework is a novel concept, however, we caution readers that recommendations at this stage should be considered provisional and, in line with the cycle of action research, subject to ongoing refinement on the basis of post-implementation evaluation prior to wider dissemination. One limitation of this approach is its potential to be exploited disingenuously by organisations that lack the safety culture and developmental maturity to identify, evaluate and implement effective FPSs within their organisations. If fatigueproofing is seen as the only strategy that needs be used to mitigate fatigue-related risk then the organisation has not effectively grasped the principles of risk and safety management. Where an organisation has successfully implemented the ‘defences-in-depth’ model,

Please cite this article in press as: Dawson D, et al., Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management, Sleep Medicine Reviews (2011), doi:10.1016/j.smrv.2011.05.004

8

D. Dawson et al. / Sleep Medicine Reviews xxx (2011) 1e9

we envisage no major problems with the integration of FPSs as one of the many hazards to be managed under the organisational SMS. In conclusion, the preceding review offers a preliminary framework and guidelines for the future development, evaluation and implementation of FPSs, a novel approach aimed at strengthening the defensive layers within an organisational FRMS. This approach will provide specific lessons about the challenges likely to be faced in sustaining and refining FPSs on an ongoing basis. In combination with the insights gained from data gathered from outside the local work group and organisation, it also has the potential to provide more valuable and generic lessons about the role of fatigue-risk management within the SMS as a whole.

Practice points 1 Fatigue-related risk is a workplace hazard. In recent years, managing fatigue-related risk has moved away from traditional prescriptive regulation towards approaches that embed fatigue management within a safety management systems framework. 2 Integrating the principles of defences-in-depth and error management theory, an effective fatigue-risk management system should attempt to manage each layer along an error-risk trajectory. The tools for managing risk at each level of the trajectory can be categorised into fatigue-reduction strategies and fatigue-proofing strategies. While a clear framework exists for the operation of fatigue-reduction strategies, fatigue-proofing strategies have received less attention. 3 Informal fatigue-proofing strategies are in operation within organisations but rarely form part of the formal safety management system. This lack of criteria may carry considerable risk due to haphazard transmission across the organisation and failure to identify dysfunctional strategies. 4 To standardise and document effective fatigueproofing strategies within the wider safety system, mixed-methods comprising field-based qualitative and quantitative techniques are required to harvest existing informal strategies, and to develop new strategies in additional areas of common error risk. 5 Stringent assessment based on measurement outcomes and process evaluation is essential to generate evidenced-based support for integration into procedure and policy at both local and external level.

Research agenda 1 Informal and potential FPSs should be classified and carefully evaluated for future development and wider dissemination. 2 As the FPSs implemented in one workplace may have significant potential in another, it is essential that broad general human factors and cognitive psychology principles underlying FPSs across organisations are collated and analysed. 3 Where the cost of implementation is low, the benefit high, and fatigue is one of many causes of error, it may be beneficial to alter the SOP so that standard practice is fatigue-tolerant. An important focus is to identify situations where FPSs may be generalised and implemented universally.

References 1. Harrison Y, Horne JA. The impact of sleep deprivation on decision making: a review. J Exp Psychol Appl 2000;6(3):236e49. 2. Lim J, Dinges D. A meta-analysis of the impact of short-term sleep deprivation on cognitive variables. Psychol Bull 2010;136(3):375e89. 3. Williamson AM, Feyer AM. Moderate sleep deprivation produces impairments in cognitive and motor performance equivalent to legally prescribed levels of alcohol intoxication. Occup Environ Med 2000;57(10):649e55. 4. Gillberg M, Kecklund G, Åkerstedt T. Sleepiness and performance of professional drivers in a truck simulator- comparisons between day and night driving. J Sleep Res 1996;5(1):12e5. 5. Philip P, Taillard J, Klien E, Sagaspe P, Charles A, Davies WL, et al. Effect of fatigue on performance measured by a driving simulator in automobile drivers. J Psychosom Res 2003;55(3):197e200. 6. Dinges DF. An overview of sleepiness and accidents. J Sleep Res 1995;4(2, Suppl.):4e14. 7. Folkard S, Lambardi DA. Modeling the impact of the components of long work hours on injuries and “accidents”. Am J Ind Med 2006;49(11):953e63. 8. Åkerstedt T. Consensus statement: fatigue and accidents in transport operations. J Sleep Res 2000;9:395. 9. Folkard S. Black times: temporal determinants of transport safety. Accid Anal Prev 1997;29(4):417e30. 10. Williamson A, Lombardi DA, Folkard S, Stutts J, Courtney TK, Connor JL. The link between fatigue and safety. Accid Anal Prev 2011;43(2):498e515. *11. Dawson D, McCulloch K. Managing fatigue: it’s about sleep. Sleep Med Rev 2005;9(5):365e80. 12. Gander P, Hartley L, Powell D, Cabon P, Hitchcock E, Mills A, et al. Fatigue risk management: organizational factors at the regulatory and industry/company level. Accid Anal Prev 2011;43(2):573e90. 13. McCulloch K, Fletcher A, Dawson D. Moving toward a non-prescriptive approach to fatigue management in Australian aviation: a field validation. Canberra, Australia: Civil Aviation Safety Authority; 2003. *14. Reason J. Managing the risks of organizational accidents. Aldershot, UK: Ashgate Publishing Ltd; 1997. *15. Reason J. Human error. Cambridge, UK: Cambridge University Press; 1990. *16. Helmreich RL, Merrit AC, Wilhelm JA. The evolution of crew resource management. Int J Aviat Psychol 1999;9(1):19e32. *17. Helmreich RL. On error management: lessons from aviation. BMJ 2000; 320:781e5. 18. Federal Aviation Administration. Flightcrew member duty and rest requirements. Washington, DC: Docket No. FAA-2009-1093; 2009. 19. National Transport Commission. Guidelines for managing heavy vehicle driver fatigue. Melbourne, Vic: NTC; 2006. 20. Australian Council of Trade Unions. Health and safety guidelines for shift work and extended working hours. Melbourne, Vic: ACTU OHS Unit; 2000. 21. Dawson D, Fletcher A. A quantitative model of work-related fatigue: background and definition. Ergonomics 2001;44(2):144e63. 22. Belyavin AJ, Spencer MB. Modelling performance and alertness: the QinetiQ approach. Aviat Space Environ Med 2004;75:A93e103. 23. Folkard S, Åkerstedt T. A three process model of the regulation of alertness and sleepiness. In: Ogilvie R, Broughton R, editors. Sleep, arousal and performance: problems and promises. Boston, MA: Birkhäuser; 1991. p. 11e26. 24. Mallis MM, Mejdal S, Nguyen TT, Dinges DF. Summary of the key features of seven biomathematical models of human fatigue and performance. Aviat Space Environ Med 2004;75:A4e14. 25. Thomas MJW, Ferguson SA. Prior sleep, prior wake, and crew performance during normal flight operations. Aviat Space Environ Med 2010;81(7):665e70. 26. Smolensky MH, Milia LD, Ohayon MM, Philip P. Sleep disorders, medical conditions, and road accident risk. Accid Anal Prev 2011;43(2):533e48. 27. Terán-Santos J, Jimenez-Gomez A, Cordero-Guevara J. The association between sleep apnea and the risk of traffic accidents. N Engl J Med 1999;340:847e51. 28. Di Milia L, Smolensky MH, Costa G, Howarth HD, Ohayon MM, Philip P. Demographic factors, fatigue, and driving accidents: an examination of the published literature. Accid Anal Prev 2011;43(2):516e32. 29. Ferrara M, De Gennaro L. How much sleep do we need? Sleep Med Rev 2000;5:155e79. 30. Dinges DF, Mallis MM, Maislin MA, Powell JW. Evaluation of techniques for ocular measurement as an index of fatigue and the basis for alertness management. Washington, DC: US Dept of Transportation; 1998. Report DOT HS 808762. 31. Johns M, Tucker A, Chapman R. A new method for monitoring the drowsiness of drivers. Proceedings of the International Conference on fatigue management in transport operations: Seattle; 2005. 32. Wierwille WW, Ellsworth LA, Wreggit SS, Fairbanks RJ, Kirn CL. Research on vehicle based driver status/performance monitoring: development, validation, and refinement of algorithms for detection of driver drowsiness. National Highway Traffic Safety Administration; 1994. Final Report: DOT HS 808 247. 33. Åkerstedt T, Gillberg M. Subjective and objective sleepiness in the active individual. Int J Neurosci 1990;52:29e37.

* The most important references are denoted by an asterisk.

Please cite this article in press as: Dawson D, et al., Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management, Sleep Medicine Reviews (2011), doi:10.1016/j.smrv.2011.05.004

D. Dawson et al. / Sleep Medicine Reviews xxx (2011) 1e9 34. Hoddes E, Dement WC, Zarcone V. The development and use of the Stanford sleepiness scale (SSS). Psychophysiology 1972;9:150. 35. Samn SW, Perelli LP. Estimating aircraft fatigue: technique with application to airline operations. Technical Report No. SAM-TR-82-21. Brooks AFB, TX: USAF School of Medicine; 1982. 36. Balkin TJ, Horrey WJ, Graeber RC, Czeisler CA, Dinges DF. The challenges and opportunities of technological approaches to fatigue management. Accid Anal Prev 2011;43(2):565e72. *37. Dawson D, Noy I, Härmä M, Åkerstedt T, Belenky G. Modelling fatigue and the use of fatigue models in work settings. Accid Anal Prev 2011;43(2):549e64. 38. International Maritime Organization. Fatigue as a contributory factor in maritime accidents. MSC Circ. 565. London: Author; 1995. 39. Allen P, Wadsworth E, Smith A. The prevention and management of seafarers’ fatigue: a review. Int Marit Health 2007;58(1e4):167e77. 40. Roach GD, Darwent D, Sletten TL, Dawson D. Lang haul pilots use in flight napping as a counter measure to fatigue. Appl Ergon 2011;42:214e8. 41. Maule J, Hockey GRJ, Bdzola L. Effects of time pressure on decision making under uncertainty: changes in affective state and information processing strategy. Acta Psychol 2000;104(3):283e301. 42. Leonard C, Fanning N, Attwood J, Buckley M. The effect of fatigue, sleep deprivation and onerous working hours on the physical and mental wellbeing of pre-registration house officers. Ir J Med Sci 1998;176(1):22e5. 43. Wertz AT, Ronda JM, Czeisler CA, Wright KP. Effects of sleep inertia on cognition. J Am Med Assoc 2006;2006(295):163e4. 44. Sinclair MA. Subjective assessment. In: Wilson JR, Corlett EN, editors. Evaluation of human work: a practical ergonomics methodology. London: Taylor Francis; 1990p58e88. 45. Hugentobler MK, Israel BA, Schurman SJ. An action research approach to workplace health: integrating methods. Health Educ Quart 1992;19:55e76.

9

46. Caldwell JA, Caldwell JL, Schmidt RM. Alertness management strategies for operational contexts. Sleep Med Rev 2008;12:257e73. 47. Harper R. Managing industrial safety in South East Texas. J Constr Eng Manage 1998;124(6):452e7. 48. Toseland RW, Rivas RF. An introduction to group work practice. 5th ed. Boston: Pearson/Allyn and Bacon; 2005. 49. Cooke NJ. Varieties of knowledge elicitation techniques. Int J Hum Compt Stud 1994;41:801e49. 50. Lambert H, McKevitt C. Anthropology in health research: from qualitative methods to multidisciplinary. BMJ 2002;325:210e2. 51. Pèrez-Chada D, Videla AJ, O’Flaherty ME, Palermo P, Meoni J, Sarchi MI, et al. Sleep habits and accident risk among truck drivers. A cross-sectional study in Argentina. Sleep 2005;28:1103e8. 52. Camp RC. Benchmarking: the search for industry best practices that lead to superior performance. ASQC: Quality Press; 1989. 53. Gray WD. Simulated task environments: the role of high-fidelity simulations, scaled worlds, synthetic environments, and laboratory tasks in basic and applied cognitive research. Cog Sci Quart 2002;2:205e27. 54. Shannon HS, Robson LS, Guastello SJ. Methodological criteria for evaluating occupational safety interventions research. Saf Sci 1999;31:161e79. 55. Craig P, Dieppe P, Macintyre S, Mitchie S, Nazareth I, Petticrew M. Developing and evaluating complex interventions: the new medical research council guidance. BMJ 2008;337:979e83. *56. Standards Australia and Standards New Zealand. Risk management e principles and guidelines. Technical Committee: AS/NZS 31000; 2009. *57. Reason J. Human error: models and management. BMJ 2000;320: 768e70. 58. Dawson D, Zee P. Work hours and reducing fatigue-related risk: good research vs. good policy. J Am Med Assoc 2005;294:1104e6.

Please cite this article in press as: Dawson D, et al., Fatigue-proofing: A new approach to reducing fatigue-related risk using the principles of error management, Sleep Medicine Reviews (2011), doi:10.1016/j.smrv.2011.05.004