A Fatigue/Risk Index to assess work schedules

Transcription

1 Somnologie 11: (2007) DOI /s ORIGINAL ARTICLE Simon Folkard Karen A. Robertson Mick B. Spencer A Fatigue/Risk Index to assess work schedules Ein Fatigue/Risk Index, zur Entwicklung von Schichtarbeitsplänen Zusammenfassung Es wird ein Ermüdungs-/Gefährdungs-Index entwickelt, mit dem die Erschöpfung und das Risiko bestimmt werden können, die sich aus einem Schichtarbeitsplan ergeben. Beide Indizien basieren auf Literaturübersichten und berücksichtigen neue Erkenntnisse insbesondere solche, die kumulative Effekte einbeziehen. Der Ermüdungs- und der Received: 20 December 2006 Accepted: 6 March 2007 Published online: 24 May 2007 S. Folkard ( ) Laboratoire d Anthropologie Appliquée (LAA) Ergonomie Comportement et Interactions (EA 4070) Université Paris Descartes 45 rue des Saints Pères Paris, France Tel.: / Fax: / s.folkard@swan.ac.uk S. Folkard Body Rhythms and Shiftwork Centre Swansea University Swansea, UK K. A. Robertson M. B. Spencer Centre for Human Sciences QinetiQ Farnborough, UK Gefährdungs-Index wurden beide aus drei verschiedenen Komponenten entwickelt: (i) ein kumulativer Anteil basierend auf dem Arbeitsmuster bis hin zu jedem beliebigen Schichtplan, (ii) ein Anteil der die Belastung, den Beginn der Arbeit, die Dauer und die Tageszeit während der Schicht berücksichtigt und (iii) eine Komponente, die die Art der Arbeit, die tatsächlich genommenen und geplanten Arbeitspausen berücksichtigt. Die neuen Ermüdungs- und Gefährdungs-Indizes stellen eine einfache Methode zum Erfassen und Vergleichen von Schichtarbeitsplänen zur Verfügung. Sie basieren auf speziellen, mathematischen Modellen für Trends der Ermüdung und für eine Risikoabschätzung. Die Indizes können helfen, konzipierte Schichtpläne auszuwerten und sie erlauben es, Parameter so lange zu ändern, bis die Ermüdung und das Risiko sich in akzeptablen Grenzen befinden. Schlüsselwörter Unfälle Verletzungen Sicherheit Wachsamkeit Schläfrigkeit Schichtarbeit mathematische Modelle Summary This paper describes our development of Fatigue and Risk Indices that allow users to assess the likely fatigue and risk associated with a given work schedule. Both indices are based on literature reviews and take account of recent developments in the field, especially those related to cumulative effects. The Fatigue and the Risk Indices were both constructed from three separate components, namely: (i) a cumulative component based on the pattern of work leading up to any given shift, (ii) a duty timing component concerned with the effect of start time, shift length and the time of day throughout the shift, and (iii) a job type/breaks component which relates to the activity being undertaken and the provision of breaks during the shift. The new Fatigue and Risk Indices provide users with a simple method for assessing and comparing work schedules based on the sophisticated mathematical modelling of trends in fatigue and risk measures. They can inform users decisions about the desirability of particular work schedules, and allow the user to vary the various parameters until fatigue and risk are judged to be within acceptable limits. Key words accidents injuries safety alertness tiredness shift work mathematical models

2 178 S. Folkard et al. Introduction In the mid 1990s, the UK s Health and Safety Executive (HSE) pioneered the development of a method for assessing the risk arising from fatigue associated with the work patterns of safety critical workers [9]. It should be noted at the outset that a very broad concept of fatigue was used in that it included sleepiness due to shortened sleeps, etc., as well as the mental or physical fatigue that might accrue over a continuous period of work and might potentially be ameliorated by rest breaks. The original methodology involved the calculation of a Fatigue Index (FI) and it was intended that the index could be used to provide an assessment of changes in work patterns and to determine whether any particular aspect of the work pattern was likely to result in increased levels of fatigue. The first version of the Fatigue Index consisted of a simple set of rules for evaluating the contribution of various factors associated with the development of fatigue, namely: (i) the length of the shift, (ii) the interval between shifts, (iii) the number of rest days, (iv) the quality of the rest breaks, (v) the variability of the shifts, and (vi) the time of day. Each of the six factors was scored independently and the resultant values were summed to provide an overall index of fatigue. However, the weightings given to the various factors were relatively arbitrary and based on guesstimates made by various experts in the field. It was recognised that, although the index contained many of the important factors which relate to fatigue, the method of calculation was in many cases difficult to apply and the individual factors did not always reflect current knowledge concerning the development of fatigue. Consequently the HSE commissioned QinetiQ (then known as the Defence Evaluation and Research Agency) to carry out an assessment of the index to identify its strengths and weaknesses, and this led to the development of a revised version [13]. This second version incorporated information from various studies of shift workers. It retained five of the original six factors (time of day, shift duration, rest periods, breaks and cumulative fatigue), the scores from each of which were summed to provide an overall index for the pattern of work. One feature of the index was that, like its predecessor, it was designed for manual calculation. This inevitably restricted the extent to which it was able to represent the full complexity of the issues related to fatigue. This revised index was subsequently converted into a spreadsheet format that has been used widely throughout British Industry to assess and compare patterns of working. However, it was recognised that there were areas where the FI was still deficient and where improvements could be made. In many cases, this was due to the simplicity of the calculations which were designed to be carried out with pencil and paper. The use of a spreadsheet permits more complex calculations which can therefore reflect more accurately the interaction between the various factors influencing fatigue. The aim of this paper is to describe our recent development of a more up to date version of the FI that takes account of the current knowledge and understanding of factors associated with the development of fatigue,again broadly defined, in the shift work environment. Given the increasing complexity of the calculations that are now required, it is no longer appropriate to calculate the scores manually and the decision was made that the new index would only be available in a spreadsheet format. Further, since the release of the previous version of the FI, there has been a substantial increase in information concerning the trends in the risk of occupational injuries and accidents related to shiftwork. The new index, therefore, allows the users to choose whether the output from the index is expressed in terms of fatigue (FI), or in terms of the relative risk associated with a given pattern of work (Risk Index, RI). Information used in the development of the new index The development of the new Fatigue/Risk Index (F/R I) was based on (i) feedback from existing users of the FI (mainly related to the user interface and to the definition of terms), (ii) recent information relating to fatigue and shift work, and (iii) a review of the literature relating features of work schedules to the risk of occupational injuries and accidents. Since the development of the previous version of the FI, further information relating to fatigue and shift work, covering a wide range of issues, had become available. Recent studies of the impact of chronic sleep reduction, over a period of one to two weeks, have provided greater insight into how fatigue accumulates over several days and, to a certain extent, into the pattern of recovery. Other data have become available relating to time of day, and in particular to early starts, and these have established the trend in alertness associated with duties starting at different times of day. In addition, there is more information on the impact of time of day on sleep duration. Finally, further published data relating to breaks have confirmed previous findings that breaks are beneficial, although the optimum timing and duration have yet to be established [17]. A review of the literature on the risk of injuries and accidents relating to shift work was undertaken to determine whether there was a link between accident risk and the patterns and timing of work. This highlighted the shortage of good epidemiological studies investigating accident risk in the workplace that also controlled for the a priori risk. Nevertheless, from the studies that did manage to take account of confounding factors, it was possible to identify consistent trends that could

3 A Fatigue/Risk Index to assess work schedules 179 form the basis of a risk index. Trends in the relative risk of accidents were identified for the following factors: (i) shift type (morning, afternoon, and night shift) for 8 hour shifts, (ii) time on duty throughout the course of the night shift, (iii) time of day, (iv) consecutive day shifts, (v) consecutive night shifts and (vi) consecutive hours on duty [3 5]. Some additional information on relative risk was also available for the effect of breaks, prophylactic naps and the direction of rotation. From the results of this review, it became clear that there were considerable differences between fatigue and risk in terms of the nature of the trends shown for certain features of work schedules. Although it may be possible, in the long term, to reconcile many, if not all, of these differences, it was decided that the new version of the Index should include two separate indices, one related to fatigue (the Fatigue Index) and one related to risk (the Risk Index (RI)). The derivation of the Fatigue and Risk Index Both the Fatigue and the Risk Index were constructed from three separate components, namely: (i) A cumulative component. This relates to the way in which individual duty periods or shifts are put together to form a complete schedule. The cumulative component associated with a particular shift depends on the pattern of work leading up to that shift. (ii) A component associated with duty timing, i. e. the effect of start time, shift length and the time of day throughout the shift, and (iii) A job type/breaks component.this relates to the content of the shift, in terms of the activity being undertaken and the provision of breaks during the shift. The Fatigue Index Cumulative fatigue This component is based on the amount of sleep loss that is likely to have accumulated over the preceding schedule of work, and on a derived relationship between sleep loss and increased levels of fatigue. Predicted values for this component were compared with the increase in fatigue over consecutive duty periods from aircrew studies [15]. The loss of sleep associated with the pattern of duty was derived from the QinetiQ studies of aircrew [12, 15, 16] and train drivers [10]. Two main cases were addressed, the impact of early start times and of late finishes/overnight work. In the case of an early start time, our data indicated that start times prior to 13:00 were associated with a reduction in sleep duration, with the extent of the reduction depending on how early the duty started. A trend line in the form of a quadratic function fitted the mean values extremely well and was used to estimate the reduction in sleep associated with a given start time. The estimated loss of sleep following a late finish or overnight duty was somewhat more complicated and relied on data from (i) the reported sleep times of Britannia crews on duties ending at various times from the early evening through to mid-morning [12] and (ii) the duration of sleep as a function of time of sleep onset for German and Japanese workers based on diary records and summarized by Kogi [7]. These data are plotted in Fig. 1. The fitted estimate of sleep duration was based on the mean of the aircrew data and the averaged shiftwork data. This was then expressed as a function of duty end time, using the relationship between sleep onset and duty end time from the aircrew data. The resultant predictions for the duration of sleep varied between eight hours (no sleep loss) for duties ending at 18:00 to three hours (five hours sleep loss) for duties continuing through to 12:00. The only other adjustments to sleep duration based on the pattern of duty were made to accommodate quick returns. This was achieved by constraining the duration of sleep to be no greater than the length of the rest period less commuting time and less a further hour to allow for personal needs. The increase in fatigue related to sleep loss, together with the subsequent recovery of alertness was derived from the results of laboratory studies by Belenky et al. [2] and Van Dongen et al. [20]. Exponential functions were fitted to the trends in the Psychomotor Vigilance Task (PVT) [21] over seven consecutive days with restricted overnight sleep and to the recovery over three days following the period of sleep restriction. The output from the PVT was converted to the output scale used for the Fatigue Index, namely the KSS using information on the time course of changes in the PVT over a 28-hour period without sleep [8]. This was compared with the prediction of sleepiness based on the timing of wakeful- sleep duration (h) Fig. 1 Japanese workers German workers aircrew time of day (h) of bedtime The duration of sleep related to bedtime

4 180 S. Folkard et al. ness, as used for the derivation of the second component of the index (see below). The fit to the PVT data based on this prediction is shown in Fig. 2.This procedure thus enabled an estimate of cumulative fatigue to be obtained prior to any given shift, and the output to be expressed on the same scale as the other components of the index. When these predictions were compared with the results from aircrew studies [15], it appeared that the cumulative effect was being overestimated. It was therefore readjusted by a scaling factor to correspond more closely with the observed results. Duty timing The three factors that contribute to this component of the index are (i) the start time of the duty, (ii) the time of day throughout the duty period and (iii) the length of the duty period. Estimates of these factors have been derived from several studies of aircrew and train drivers. In the majority of these studies, fatigue assessments were obtained, using the seven-point Samn-Perelli (SP) scale [14] at various points during a sequence of duty periods. response time (s -1 ) mean PVT response time time of day (h) Fig. 2 Fit to PVT data based on transformed KSS score. Higher values indicate quicker responses fit The effect of start time is illustrated in Fig. 3, where the range of the seven-point scale on the y-axis is from one (fully alert, wide awake) to seven (completely exhausted, unable to function effectively). This Fig. illustrates that there are considerable differences between the results of the different studies. It is particularly noticeable, for example, that the increased fatigue associated with start times between 04:00 and 08:00 is less marked for the short-haul aircrew than for either the long-haul crews or the train drivers, and it is unclear how this difference has arisen. The estimation of the effect of start time has been based on the average of the two groups that are in close agreement,i. e.the long-haul crews and the train drivers, From these, a fit was obtained in the form of a circadian rhythm with an amplitude of 0.49 and an acrophase, corresponding to the highest levels of fatigue, of 01:15. This rhythm represents the sum of two components, one of which corresponds to an initial state, presumably related to sleep, whose effect persists throughout the duty period. The second component is the underlying circadian rhythm which varies throughout the duty period. The estimates of this rhythm from the various studies suggest that its acrophase is several hours later than that for the start time, and this is consistent with the results of laboratory studies. For the calculation of this component, an acrophase of 05:15 has been used, and an amplitude, on the SP scale, of Results from the same studies have indicated that the component associated with the duration of a duty period is approximately a linear function of duty length. For train drivers, the rate of the increase in fatigue varies from 0.14 per hour on the SP scale, when no driving is undertaken, to 0.23 per hour for a typical amount of driving. This suggests that the effect of shift length should take into account the nature and intensity of the work being carried out. For this reason, job type, together with breaks, has been included as a separate component, and is discussed below. Finally, a transformation has been applied to convert these results from the seven-point SP scale to the scale of Fig. 3 Fatigue levels at the start of duty. The range of the fatigue scale is from 1 (fully alert, wide awake) to 7 (completely exhausted, unable to function effectively) fatigue (7-point) train drivers short-haul aircrew aircrew on long-haul duty start time (h)

5 A Fatigue/Risk Index to assess work schedules 181 the Fatigue Index, namely the probability (multiplied by 100) of recording a value of seven or more on the Karolinska Sleepiness Scale (KSS) [1]. This transformation was based on a large study of aircrew in which data were collected on both scales simultaneously [11]. Finally, a single value for the whole shift is obtained by integrating over the entire shift. Job type/break While it is clear that the intensity of the work being carried out and the provision of breaks within a shift have a significant influence on fatigue, there is relatively little information on which to base reliable estimates. On the one hand, there are laboratory studies and simulations which have shown that alertness is reduced during continuous periods of activity and recovers after a break. On the other hand, there are data collected in the work place which have shown that subjective levels of sleepiness and fatigue are strongly influenced by the levels of workload. For the construction of this component, we made considerable use of the results of the laboratory simulations carried out by Gillberg et al. [6], in which periods of 100 min of continuous activity, both during the day and overnight, were separated by breaks of 20 min.with respect to field studies, we took account of the increase in fatigue experienced by train drivers [10] and aircrew [12, 15, 16] during duty periods involving different levels of workload. The derivation of the formulae for this component is beyond the scope of the current paper. However, the principles that were used for their construction are: (i) That there is a baseline level of fatigue that increases slowly throughout a shift (assuming that no naps are taken). It corresponds to the effect of continuous wakefulness, together with a minimal level of activity that is unaffected by breaks. This aspect is reflected in the Duty Timing component of the Index. (ii) There is an increase in fatigue associated with a period of continuous activity, with the more intense the activity the greater the increase. If there are no breaks of any sort, the increase in fatigue can be represented by a negative exponential function of time. (ii) The effect of a break is to initiate a recovery process. If the break is very short (e. g. 2 min), the effect is to halt,rather than reverse,the accumulation of fatigue. If it is long (e. g. 30 min), fatigue recovers to its baseline level, with 50 % recovery achieved after approximately 15 min. The new version of the Fatigue Index requests information on the nature of the workload, the requirement for continuous attention and on the length and frequency of breaks. This information is used to construct the parameters for the job type/breaks component, using the methodology outlined above. The form of the Fatigue Index The final form of the Fatigue Index is given by: FI = 100 { 1 (1-C) (1-J-T) } where C is the cumulative fatigue component, T is the duty timing component, andj is the job type/breaks component. In this formula, C, J and T correspond to estimated probabilities of KSS scores of seven or more and therefore take values between zero and one. The Risk Index Cumulative effects The estimation of the increase in risk on consecutive shifts has been based on the relative risk data over four successive night and day shifts [4]. The increase is reasonably approximated by a linear trend, representing an increase of over each consecutive day shift and of over each consecutive night shift. To estimate the risk for a shift starting at any time of day, a cosine fit has been applied to these linear trends, resulting in an acrophase at midnight (corresponding to the mid-point of the shift), an average increase of and an amplitude of The increase associated with each successive shift therefore varies between = for shifts centered on midnight and = for shifts centered on midday. When two consecutive shifts occur at two different times of day, the average of the values corresponding to the two different shifts has been taken. There is no information on the time course of the recovery in risk after a sequence of consecutive shifts. Accordingly, the pattern of recovery used in Risk Index has been designed to follow, as closely as possible, that of the Fatigue Index. Nor is there any information on the effect of quick returns on risk, and estimates have again been based on comparisons with the Fatigue Index. A simple formula has been adopted, whereby the increase in risk is linearly related to the amount by which the rest period is less than nine hours. Each additional hour of lost rest is assumed to be associated with a 6 % increase in risk. Duty timing This component is calculated by multiplying the risks associated with two individual factors, namely the time of day and the length of the shift. The risk associated with the time of day was derived from the relative risk on the afternoon (1.1519) and night shifts (1.2788), compared with that on the morning shift (1.0000) [4]. However, these values were based on the average risk over a sequence of shifts, not the risk on the first shift. They have therefore been recalculated, using the sequence effect (see above under Cumulative effects ) to obtain an

6 182 S. Folkard et al. estimate of the risk associated with the first shift. The cosine fit to this risk has an amplitude of and an acrophase, corresponding to the maximum risk, close to midnight, and this fit was used in the calculation of the index [5]. The factor related to shift length was derived from the relative risk on different lengths of shift [4]. The fit that was used for the calculation of the index is shown in Fig. 4, where the risk is expressed relative to the average risk on an 8 h shift, which was set equal to one. As there is very little change in the level of risk during shifts lasting between four and eight hours, the relative risk for these shifts was also set to one. A negative exponential function was used to estimate the lower levels of risk on shifts shorter than four hours, and an exponential function with a positive exponent was used for shifts longer than eight hours. For shifts longer than 12 h, this has involved extrapolation beyond the range of the available data. It was assumed that the shifts that contributed to the estimation of incident risk involved average commuting times of approximately 40 min, as this is typical of time spent travelling to or from work. If the user of the index specifies a commuting time either longer or shorter than 40 min, the estimation of the risk associated with duty timing is adjusted appropriately. Job type/breaks Two studies have examined the impact of breaks on occupational injuries [18, 19]. Both studies showed an increase in risk as a function of the time since the last break and the relative risk estimates from the two studies were averaged for the present purposes. An exponential fit to these average values is shown in Fig. 5, where the risk is relative to a value of 1.0 over the first 30 min. This function has been used to estimate the risk compared with an 8h shift relative risk fit length of shift (h) Fig. 4 Relative risk of shifts of different duration. Higher values indicate greater risk relative risk time between breaks (h) Fig. 5 Relative risk of continuous periods between breaks. Higher values indicate greater risk changes in risk associated with continuous periods of work,as specified by the user of the index on the defaults screen. As there is no information on the reduction in risk associated with breaks of different duration, the recovery function used for the Fatigue Index has been adopted. Similarly, there is no information of the impact of workload on risk, and the Fatigue Index has again been used to derive the values for risk. The form of the Risk Index The final form of the Risk Index is given by: RI = C * J * T where C is the cumulative fatigue component, T is the duty timing component, and J is the job type/breaks component. The index has been normalised with respect to a typical two-day, two-night, four-off 12 h shift schedule, where the following conditions apply: (i) the shift changes occur at 07:00 and 19:00, (ii) typical commuting time is 40 min, (ii) the work is moderately demanding, requiring continuous attention some of the time, (iv) a break of 15 min is typically taken every two hours, and (v) the longest period without a break is typically four hours, followed by a break of 30 min. The normalization ensures that, if this schedule is repeated over 21 consecutive cycles, covering a period of 24 weeks, the average value of the index is The average values of C and T (but not J) are also normalized to 1.00 over the same period. Examples relative risk fit In order to demonstrate the dynamics of the Indices we calculated the estimated fatigue and risk values over spans of five successive work periods. These work periods were either Normal days (09:00 17:00), or typi-

7 A Fatigue/Risk Index to assess work schedules 183 cally timed 8-hour shifts (Morning = 06:00 14:00, Afternoon = 14:00 22:00 and Night = 22:00 06:00), or typically timed 12-hour shifts (Day = 08:00 20:00 and Night = 20:00 08:00). In all cases, we assumed (i) a commute time of 40 min (ii) that the work was moderately demanding with little spare capacity and required continuous attention some of the time, and (iii) that 15 min breaks were regularly given after each two hours of continuous work. The resultant estimated fatigue and risk values are shown in Table 1. Note that the output scales for the two indices differ substantially from one another (detailed above). A number of important points emerge from inspection of Table 1. First, irrespective of the timing or duration of the work periods, both fatigue and risk were estimated to increase over successive work periods, reflecting on the cumulative component (C). Secondly, both the estimated mean fatigue and risk varied across the different types of 8- and 12-hour shifts with that on the night shift being highest, reflecting on the duty timing (T) component. Finally, however, it is noteworthy that the 8-hour afternoon shift was estimated to be less fatiguing but more risky than the 8-hour morning shift. This latter point reflects on the fact that the peak in fatigue has been found to occur rather later than that in risk [5]. Discussion We have updated the previous HSE Fatigue Index in the form of a Fatigue and Risk Index to allow organisations to assess their work schedules. The significance of the name change is that, as indicated above, the new index consists of two separate indices, one of which is related to fatigue, and which therefore corresponds more closely with the previous index, whereas the other is related to risk. The new index, a full report describing its development, and a guide to its usage are available from the HSE s website [17]. The HSE has promoted its use and the feedback to date has been positive. Unfortunately, however, the spreadsheet used to calculate the index appears to only work on computers configured in English. The outputs from the two indices are on different scales, and both differ from the scale used in the previous version. The Fatigue Index is now expressed in terms of the estimated average probability, multiplied by 100, of a value of seven or more on the KSS, and therefore takes a value between zero and 100. The KSS has been extensively validated,and high scores are known to be associated with a high frequency of microsleeps. The output from the Risk Index represents the estimated relative risk of the occurrence of an incident on a particular shift. As with the Fatigue Index, the risk is averaged over the entire shift. A level of one represents the average risk on a typical two-day, two-night, four-off schedule, involving 12-hour shifts starting at 07:00 and 19:00. A value of two represents a doubling of risk. It should be noted that there are sometimes large differences in the output from the two indices and that a shift with a high value on one index is not always assigned a high value on the other. This is an inevitable consequence of the different information from which Table 1 The estimated fatigue and risk values over spans of five successive normal days and over spans of five successive 8- or 12-hour shifts with various start times FATIGUE INDEX Normal Day 8-hour Shifts 12-hour Shifts Day No. 09:00 17:00 06:00 14:00 14:00 22:00 22:00 06:00 08:00 20:00 20:00 08: Mean RISK INDEX Normal Day 8-hour Shifts 12-hour Shifts Day No. 09:00 17:00 06:00 14:00 14:00 22:00 22:00 06:00 08:00 20:00 20:00 08: Mean

8 184 S. Folkard et al. the two indices have been constructed and, in particular, of the differential effect of time of day.whereas both fatigue and risk are highest on the night shift, the risk of an incident occurring on the afternoon shift is higher than on the morning shift [5]. This contrasts with fatigue which tends to be higher on a morning than on an afternoon shift. The user is therefore able to decide on which index to place the greater weight, although in general it would be wise to ensure that neither index reaches a high value. The new Fatigue and Risk Index, like its predecessor, is intended as a tool to allow organisations, and especially those operating in high hazard situations, to assess their work schedules. It does, however, have some clear limitations. First, it is not intended to take account of individual differences in the phase of their circadian rhythms, or in the build up of cumulative fatigue, etc., rather it is based on averaged results from the various studies. Nevertheless, there are well established individual differences in the phase of circadian rhythms. Further, it is highly likely that similar individual differences occur in other parameters of circadian rhythms, such as amplitude, and indeed in the other components of the Indices such as the susceptibility to cumulative fatigue. Secondly, the Indices are based on rotating work schedules and essentially assume that individuals fail to show circadian adjustment to them. A review of the literature on the adjustment of the circadian rhythm in melatonin to permanent night work indicated that only 7 % of permanent night workers showed good adjustment to the night shift while a further 22 % showed evidence of a level of adjustment which may have been of benefit in coping with consecutive night shifts [17].Thus in the majority of cases (around 70 %), the Fatigue and Risk Indices could be argued to be appropriate assessment tools for permanent night workers. However, for the 30 % of permanent night workers that show at least some adjustment they are likely to be too limiting since these individuals should be able to cope with longer sequences of consecutive nights than those who fail to adjust. Thus the Indices are likely to be conservative and to overestimate the fatigue and risk for the minority of individuals who show circadian adjustment to their permanent schedule. Thirdly, the Fatigue and Risk Indices are designed to enable the user to compare the impact of different patterns of working and provide average values for each duty period. Thus, in their present form the indices are not appropriate for use in post-event analyses (e. g. incident/accident investigation) that require an estimate of fatigue or risk at the moment of the event in question. Rather, they would need to be modified to accept timespecific inputs e. g. the time of the incident itself, the intensity of work, and the pattern of work and breaks during the shift immediately prior to the incident. Finally, the indices differ quite substantially from one another in terms of the type of occupations on which they have been primarily based. The Fatigue Index is based primarily on fatigue ratings obtained from individuals employed in various types of transport operations while the Risk Index is based primarily on the relative risk of occupational injuries on industrial shift systems. Thus, both indices clearly need to be validated on a wide range of other occupational groups to establish their generalisability. Despite these limitations, the new Fatigue and Risk Indices provides users with a simple method for assessing and comparing work schedules based on the sophisticated mathematical modelling of trends in fatigue and risk measures. They can inform users decisions about the desirability of particular work schedules, and allow the user to vary the parameters until fatigue and risk are judged to be within acceptable limits. Finally, the inclusion of the Risk Index has enhanced the face validity and allows users to make informed decisions based on estimates of the relative risk of injuries and accidents. Acknowledgment The development of this Fatigue/Risk Index was undertaken for the UK s Health and Safety Executive under contract No We are grateful to Debbie Lucas for her help, advice and encouragement. References 1. Åkerstedt T, Gillberg M (1990) Subjective and objective sleepiness in the active individual. Int J Neurosci 52: Belenky G, Wesensten NJ, Thorne DR, Thomas ML, Sing HC, Redmond DP, Russo MB, Balkin TJ (2003) Patterns of performance degradation and restoration during sleep restriction and subsequent recovery: a sleep dose-response study. J Sleep Res 12: Folkard S, Lombardi DA (2004) Towards a Risk Index to assess work schedules. Chronobiol Int 21: Folkard S, Lombardi DA (2006) Modelling the impact of the components of long work hours on injuries and accidents. Amer J Indust Med 49: Folkard S, Lombardi DA, Spencer MB (2006) Estimating the circadian rhythm in the risk of occupational injuries and accidents. Chronobiol Int (in press) 6. Gillberg M, Kecklund G, Goransson B, Åkerstedt T (2003) Operator performance and signs of sleepiness during day and night work in a simulated thermal power plant. Int J Indust Ergon 31: Kogi K (1985) Introduction to the problems of shiftwork. In: Folkard S, Monk TH. Hours of work. Temporal factors in work-scheduling. John Wiley, Chichester, pp Lamond N, Dawson D, Roach GD (2005) Fatigue assessment in the field: validation of a hand-held electronic psychomotor vigilance task. Aviat Space Environ Med 76:

9 A Fatigue/Risk Index to assess work schedules Lucas D, Mackay C, Cowell N, Livingstone A (1997) Fatigue risk assessment for safety critical staff. In: Harris D Engineering psychology and cognitive ergonomics, Vol. 2: Job design and product design. Ashgate Publishing, Co., Aldershot, UK, pp McGuffog A, Spencer M, Turner C, Stone B (2004) Working patterns of train drivers: implications for fatigue and safety. QinetiQ Document Identifier Number QINETIQ/KI/CHS/ CR RSSB Reference Number T Robertson KA (2002) Predicting alertness in future long-range operations: a validation study by ECASS. QinetiQ Report Number QINETIQ/KI/CHS/ CR021119/ Robertson KA, Spencer MB (2003) Aircrew alertness on night operations: an interim report.qinetiq Report Number QINETIQ/KI/CHS/CR021911/ Rogers AS, Spencer MB, Stone BM (1999) Validation and development of a method for assessing the risks arising from mental fatigue. HSE Contract Research Report 254/ Samn SW, Perelli LP (1982) Estimating aircrew fatigue: A technique with implications to airlift operations. Brooks AFB, TX: USAF School of Aerospace Medicine. Technical Report No. SAM- TR Spencer MB, Robertson KA (2000) A diary study of aircrew fatigue in shorthaul multi-sector operations. DERA Report Number DERA/CHS/PPD/ CR Spencer MB, Robertson KA (2002) Aircrew alertness during short-haul operations, including the impact of early starts. QinetiQ Report Number QINETIQ/CHS/PPD/CR010406/ Spencer MB, Robertson KA, Folkard S (2006) The development of a fatigue/ risk index for shift workers. Health and Safety Executive Report no 446. Available at: rrhtm/rr446. htm 18. Tucker P, Folkard S, MacDonald I (2003) Rest breaks and accident risk. Lancet 361: Tucker P, Lombardi DA, Smith L, Folkard S (2006) The impact of rest breaks on temporal trends in injury risk. Chronobiol Int 23: Van Dongen HPA, Maislin G, Mullington JM, Dinges DF (2003) The cumulative cost of additional wakefulness: Dose-response effects on neurobehavioral functions and sleep physiology from chronic sleep restriction and total sleep deprivation. Sleep 26(2): Wilkinson RT, Houghton D (1982) Field test of arousal: A portable reaction timer with data storage. Human Factors 24: