Designing better shift systems

0003-6870(95)ooo44-5

Applied Ergonomic

Vol 27. No. 1, pp. 3944, 1996 Elsevier Science Ltd Printed in Great Britain 0003-6870196 $10.00 + 0.00

Designing better shift systems P. Knauth Abt Arbeitswissenschaft HP, Universitiit Karlsruhe, Hertzstrasse 16, D-76184 Karlsruhe, Germany

The results of some intervention studies on the effects of the change from weekly rotating to quicker rotating shift systems are presented. Consequently, the following recommendations for the design of shift systems according to physiological, psychological and social criteria are discussed:

(1) Nightwork should be reduced as much as possible. If this is not possible, quickly rotating (2) (3)

(4)

(5)

shift systems are preferable to slowly rotating ones. Permanent nightwork does not seem to be recommendable for the majority of shiftworkers. Extended wsorkdays (9-12 hours) should only be contemplated when the nature of work and the work1oa.d are suitable for extended working hours, and the shift system is designed to minimize the accumulation of fatigue and toxic exposure is limited. An early sta.rt for the morning shift should be avoided. Flexible working time arrangements can be achieved in all shift systems. The highest flexibility is possible in the so-called ‘time autonomous groups’. Quick changeovers (e.g. from night shift to afternoon shift on the same day) should be avoided. The number of consecutive working days should be limited to five-seven. Every shift system should include some free weekends with at least two consecutive days off. The forward rotation (phase delay, clockwise rotation: morning/evening/night shift) would seem to be most preferred.

Keywords:

shift system,

shiftwork

the same questionnaires twice, in parallel with the experimental groups. In Table 2 the effects of the change to a new shift system on sleep problems, social problems and subjective health are summarized. ‘Improvement’ in this table means that only in the experimental and not in the control group was a reduction of problems observed at the second questionnaire study compared to the first one. The category ‘no significant change’ covers different possibilities; firstly neither in the experimental nor in the control groups was there a significant change with regard to the problem or secondly the problems were reduced or increased in a parallel manner in both groups, i.e. this change could not be attributed to the new shift system. Although from the data on problems in Table I there is no clear picture after the trial period, the majority in all groups voted in favour of the new shift system. (Some shiftworkers did not participate in the first questionnaire study, but as they had also experienced the old and new shift system they participated in the final vote.) In field studies it is never possible to control all conditions. In some of the groups, parallel to the change from a weekly to a quicker rotation, the weekly working time was reduced or a change from a discontinuous to a continuous shift system was achieved. In some groups the new shift system was introduced with an inadequate number of staff and in one group the workers were not allowed to participate in the

The title of this paper does not promise a single optimum shift system, since this just does not exist. All shift systems have drawbacks and advantages. However, we are able to state which ones are more favourable or are better for shiftworkers, and which ones are less favourable. The EC Council Directive 93/104/EC only lays down minimum safety and health requirements for the organization of working time. However, much more can be done for the health and safety of shiftworkers. A very effective way of reducing problems of shiftworkers is to design shift systems according to ergonomic recommendations. Some examples of changes from traditional to such ‘better’ shift systems are presented below. Examples of traditional

and new shift systems

Altogether shift systems were changed in 26 groups of shiftworkers. In Table 2 .those groups with both the same old and new shift system (within one sector) have been combined to form 10 aggregated groups, two from the steel industry, five from chemicals, and one each from food processing, oil and utilities. In all studies questionnaires were filled out under the old shift system and after a trial period of six-12 months under the new shift system. Furthermore, control groups which did not change their shift system filled out

39

Designing better shift systems: P. Knauth

40 Table 1 Effects of the change from weekly rotating to quicker rotating shift systems (‘ t Branch

1st and 2nd questionnaire study n

Sleep problems

Steel industry”

22 28 80 28 52 43 39 25 108 16

= =

Chemical

industryb

Food processing” industry Oil refineryd Waterworks, gasworks, power

stationZ

“Unpublished data (Knauth and Hornberger) bHornberger and Knauth, 1993; Hornberger, ‘Knauth and Kiesswetter, 1987 dKnauth and Schcnfelder, 1990; Schiinfelder, ‘Unpublished data (Knauth and Schmidt)

1 = t = t = =

L

t

L dayoff morningshift E = eveningshift q

q

N = nightshift

Figure la Old discontinuous shift system in the steel industry (additional days off are not shown)

I

I

=

day

M

E

E

E

off

M = morning shift E = evening shift N = nightshift

Figure lb New rapidly rotating, the steel industry

= = ? = = = =

‘ J ’ worsening)

Subjective health

Vote in favour of new shift system after trial period

= =

100% 95% 96% 54% 90% 91% 59% 84% 86% 87%

I = = = = = =

(n=25) (n=39)

(n=31) (n=27)

1992

T

M

Social problems

‘=’ no significant change;

1994

design of the new shift system or to choose between alternatives. This was one of the reasons why only 54% of this group voted in favour of the new shift system (for further details see corresponding publications).

M

’ improvement;

discontinuous

shift system in

Within the framework of this paper only a few shift systems can be presented in detail. The first group in Tuble 2 changed from a three-shift system (Figure la) to a rapidly rotating shift system (Figure Zb), all shiftworkers voting in favour of this new system. In the old shift system no social contact was possible in the second and third weeks (Figure la), whereas the new shift system provided two or more free evenings between Monday and Friday in three of four weeks. This is also true for the new shift system of the second group in Table I. This group changed from the traditional continuous shift system with four teams and seven consecutive night, evening and morning shifts (Figure 2~) to a new system with nine subteams and quickly rotating shifts (Figure 2b). The advantages of this new system are forward rotation (explanation below), short periods of night, morning and evening shifts, three days off after the night shift period, at least one free evening every week between Monday and Friday, and a more regular pattern of the whole shift system. However, the fact that the shiftworkers always have to work six days in a row is disadvantageous. The next example of a new shift system has shorter working periods, but is very irregular. The first group in the chemical industry (group 3) worked for many years under the old shift system shown in Figure 3~. The 24 hours on Sunday are covered by two 1Zhour shifts instead of the usual three S-hour shifts. Under this arrangement the shiftworkers ‘gain’ an additional free Sunday in the second week. Although at first glance seven days off in every fourth week seems to be an advantage, in reality this was not the case. The shiftworkers were very tired at the end of the seven consecutive morning shifts and even more so at the end of the 20 working days. There were so exhausted that they needed three-four days to recover. On the last free Sunday they already felt that burden of the coming 20 working days. Therefore, they only used a small part of the free week for leisure activities. The new shift system was designed by the shiftworkers themselves,

Designing better shift systems: P. Knauth

41 The speed of rotation

Regarding the speed of rotation, it is possible to classify roughly all shift systems into three categories - permanent shift systems (e.g. permanent night shift) - slowly rotating shift systems (e.g. weekly rotating shift system, Figures la, 2a, 3a) - rapidly rotating shift systems (e.g. Figures lb, 26, 3b). day off !!! 1 morningshift E q evening shift N = night shift

Figure 2a Old continuous

shift system in the steel industry

Th

( Fr / Sa / Su (

In Table 2 a simplified summary of arguments for or against certain types of shift systems is presented (more detailed discussions are published by Wilkinson, 1992; Folkard, 1992; Knauth, 1993). In general, the endogenous circadian body clock of most people cannot adjust to nightwork within one week (Knauth et al, 1978; 1981; Dahlgren, 1981; Knauth, 1993). Adjustment is only improved where there is total commitment to the tasks and work, where the shiftworker is an extreme evening type or where bright light is used (e.g. Folkard et al, 1978; Hildebrandt et al, 1987; Czeisler et al, 1990). On the other hand, the least disturbance of circadian physiological functions is observed in rapidly rotating shift systems. j

q morningshift E = evening shift N = nightshift n = dayoff

Week

11MO1 Tu We Th Fr Sa Su I I I I I I /

/I

I

’

I

2 3

M

Figure 2b New continuous

I= Ml M2 Nl N2 E

s,hift system in the steel industry

discussions on our ergonomics recommendations for the design of the shift system (Figure 3b). The apparent irregularity and the fact that it has no ‘forward’ rotation of shifts (explanation below) are disadvantages. However, the following aspects are positive: after .intensive

Figure

continuous

Weak/iMO

shift system

in the chemical

Tu

We

Th

Fr

Sa

Su

m = dayoff Ml M2 Nl N2 E

for the design of

Based on the above experiences and on the results of studies by other authors, the effects of key characteristics of shift systems are now discussed.

3a Old

morning shift (6 hours) morning shift (12 hours) night shift (6 hours) night shift (12 hours) evening shift

industry

short periods of morning shifts, short periods of night shifts, short periods of evening shifts, in every week at least two free evenings between Monday and Friday, two longer free weekends, after each period of night shifts, at least three days off, and a maximum of five consecutive working days. Ergonomics recommendations shift systems

day off = = = = =

Figure

industry

= morning shift (6 hours) = morning shift (12 hours) q night shift (6 hours) = night shift (12 hours) = evening shift

3b New

continuous

shift system

in the chemical

Designing better shift systems: P. Knauth

42 Table 2 What is the best ‘compromise’ Permanent nightwork

Shift system

Disruption circadian rhythms

of

Accumulation sleep deficits

Weekly rotating 0

0 yes

of

yes

0 yes

and negative effects (such as increased fatigue, reduced performance or reduced safety). The available evidence suggests that ‘there can be no general answer to the question as to the maximum desirable duration of a shift’ (Folkard, 1992, p. 1461). Summarizing the recommendations in the literature, we can formulate recommendation no. 2 (Wallace, 1989; Tsaneva et al, 1990; Ong and Kogi, 1990):

shift system Rapid rotating 0 least disruption

0

Weeks without free evenings

yes

yes

Performance during night shift

0 somewhat better

reduced

0

0 no

yes 0

Recommendation 0

Extended workdays contemplated, if:

no 0

0

no 1

I

Nightwork should be reduced as much as possible. If this is not possible, quickly rotating shift systems are preferable to slowly rotating ones. Permanent nightwork does not seem to be recommended for the majority of shiftworkers.

(9-12

hours)

should

only be

the nature of work and the workload are suitable for extended working hours the shift system is designed to minimize the accumulation of fatigue there are adequate arrangements for the cover of absentees overtime is not added toxic exposure is limited it is likely that a complete recovery after work and a high acceptance of the working time arrangements are possible.

reduced

A lack of adjustment is also reflected in the typical shorter duration of sleep during the day after night shifts. If there are several night shifts in a row there is likely to be a bigger cumulative sleep deficit towards the end of a span of night shifts (Foret and Benoit, 1978; Kieswetter et al, 1985; Folkard et al, 1990; Escriba et al, 1992). However, after a single night shift (or two night shifts) the sleep deficit can be compensated at once. After a week of morning shifts a cumulative sleep deficit may also occur. Most shiftworkers value leisure time during evening hours higher than at all other times of the day, because the evening is best for social contacts in our society. However, a week of evening shifts and a week of night shifts hinders social contacts. In rapidly rotating shift systems more regular social contacts can be made. Only performance is likely to be somewhat better on permanent night systems, provided that the individual’s body clock adjusts, for example after 10 or more night shifts in a row. This solution can only be accepted in situations where safety is paramount (Folkard, 1992), but is unacceptable in normal industry, because it would isolate the shiftworker from the rest of the society. In conclusion, the best ‘compromise’ shift system seems to be the rapidly rotating shift system. We conclude: Recommendation

no. 2

Timing of shifts An early start for the morning

shift (Knauth,

1993)

(1) reduces sleep before the morning shift (because the majority of shiftworkers go to bed at the usual time, due to social pressures or to the ‘forbidden zone’ for sleep onset (Lavie, 1986); furthermore, some shiftworkers have a disturbed night sleep before the first morning shift, because they worry as to whether they will wake up in time, (2) increases fatigue during the morning shift, the risk of errors and accidents in the (3) increases morning shift. Other aspects like timetables of public transport, job content, time for onset of sleep after night shifts, meals together with the family and wishes of the shiftworkers have to be taken into consideration when fixing the starting and finishing times of morning, evening and night shifts. As there is no optimal solution, it would seem desirable to consider flexible working time arrangements, which have in fact already been introduced even in three-shift systems. Experiences with so-called ‘time autonomous working groups’ in the Netherlands and in Germany have shown that this type of working time arrangement after an initial phase of learning how to handle conflicts within the group - has always been beneficial to the workers and the company (e.g. De Haan, 1990). In conclusion we can formulate recommendation no. 3:

Duration of shifts

Recommendation

While traditionally most shift systems have utilized 8 hour shifts, some unions and workforces favour longer shifts, since they reduce the number of shifts worked per week, which means an increase in the number of rest days (Tepas, 1985). There may be both positive effects (such as reduced hours spent in commuting or bigger blocks of days off)

-

no. 3

An early start for the morning shift should be avoided. Flexible working time arrangements can be achieved in all shift systems. The highest flexibility is possible in socalled ‘time autonomous groups’.

Designing better shift systems: P. Knauth Distribution of leisure tim,e within the shift system According to the Council Directive 93/104/EC, there should be a minimum daily rest period of 11 consecutive hours per 24-hour period (Article 3). However, there are shift systems in use with only 8 hours between the end of one shift and the start of the following shift. This causes a reduction in sleep between the shifts, because from these 8 hours the time for travelling, eating, washing and other activities has to be subtracted (Knauth and Rutenfranz, 1972; Knauth et al, 1983; Totterdell and Folkard, 1990). Therefore, increased fatigue has to be expected in the second shift, as Saito and Kogi (1978) have shown. A second aspect regarding the distribution of leisure time refers to the number of consecutive working days. To avoid an accumulation of fatigue, it is better to have shorter working periods each followed by days off. However, the workload, the organization of breaks and exposure to unfavourable environmental conditions vary according to the different places of work. Although it is difficult to define a limit for the maximum number of consecutive working days, Koller et al (1991) recommend limiting the number of consecutive working days to between five and seven. In European culture, the utility of leisure time on Saturdays and Sundays is higher than that from Monday to Friday afternoon for the majority of workers (e.g. Wedderburn, 1981; Hornberger and Knauth, 1993). Some social events, like traditional Friday and Saturday ‘pub nights’ or church-going on Sunday, may lead to a reduction of sleep and even to an increased accident risk in the following morning shift (Monk and Wagner, 1989; Folkard and Totterdell, 1991). Therefore, every shift system should provide enough time for leisure activities and recovery at weekends, some free weekends with at least two consecutive days off (not curtailed by the sleep after the last night shift). Recommendation -

-

no. 4

Quick changeovers (e.g. from night shift to afternoon shift on the same day) should be avoided. The number of consecutive working days should be limited t,o five to seven. Every shift system should include some free weekends with at least two consecutive days off.

Direction of rotation Shift systems, which first move from morning shift to evening shift and then to night shift, have a forward rotation (phase delay, clockwise rotation), whereas the counterclockwise rotation (night to evening to morning shifts) is called backward rotation or phase advance. The forward rotation prolongs the day of change from one shift to a different type of shift and therefore corresponds better to the endogenous circadian rhythm, which has a period of more than 24 hours (Wever, 1979). The majority of shiftworkers who have had experience with both kinds of shift systems have voted in favour of the forward rotation (Landen et al, 1981;

43 Czeisler et al, 1982; Knauth and Kiesswetter, 1987; Epstein et al, 1991). In a cross-sectional study, pure advancing shift systems resulted in higher levels of fatigue and more sleep disturbances between shifts, compared with pure delaying systems or mixed systems (Barton and Folkard, 1993).

Forward

rotation

of shifts

(phase

delay

]I

Conclusions Wilkinson (1992) recommends permanent night shifts, because ‘degrees of adaptation can normally be achieved which are considerably better than those found on weekly rotating regimes’ (p. 1440), which may in turn improve sleep and performance. However, Folkard (1992) and Knauth (1993) point to the remaining cumulative sleep deficit on the last night shift of many consecutive night shifts. Furthermore, Knauth (1993) and Folkard (1992) argue that permanent night shifts for most people are not acceptable, because they hinder social contacts, whereas quickly rotating shifts make more regular social contacts possible. However, Folkard recommends permanent night shifts in situations, where safety is paramount. All three authors agree the traditional weekly change of shifts to be the worst solution. There is a clear need for controlled longitudinal studies (also including field studies with bright light). In the discussion of Table I it has already been stated that in field studies, in general, several parameters are changed in a parallel manner; a rigid experimental control is rarely possible. In addition to the change from weekly to quicker rotating shift systems, the following factors might have positive effects on the outcome of questionnaire studies: an adequate manning of the teams, a favourable organization of additional shifts (e.g. fixed for one year in advance) and the reduction of weekly working time. It could also be that, after a trial period, shiftworkers vote in favour of a new shift system because they do not want to change again. However, in a Dutch oil refinery (Jansen et al, 1983) and in a German paper mill (unpublished) the shiftworkers undertook two successive trials with two new shift systems. In the final vote, in both cases, they chose the first and not the second alternative. The judgement of the workers may be negatively affected if there is no participation of the workers, too few persons per team, a flexible organization of additional shifts which only takes the requirements of the company into consideration, change from a discontinuous to a continuous shift system, or special unfavourable sequences within a shift system not allowing for enough sleep. Despite all the methodological difficulties, there is a need for controlled longitudinal field studies, comparing permanent shift systems and quickly rotating ones. References Barton,

J. and Folkard,

S. 1993 ‘Advancing 59-64

systems’ Ergonomics 36 (l-3),

versus delaying

shift

44 Czeisler, C.A., Johnson, M.P. and Duffy, J.F. 1990 ‘Exposure to bright light and darkness to treat physiologic maladaptation to nieht work’ The New Enaland J Medicine 323 (18). 1253-1259 Czeiiler, C.A., Moore-Ede,“M.C. and Coleman, R:M:‘1982 ‘Rotating shift work schedules that disrupt sleep are improved by applying circadian principles’ Science 217 (4558), 46tl-463 Dahlgren, K. 1981 ‘Long-term adjustment of circadian rhythms to a rotating shiftwork schedule’ Stand J Work Environ Health 7, 141-151 De Haan, E.G. 1990 ‘Improving shiftwork schedules in a bus company: towards more autonomy’ in Costa, G., Cesana, G., Kogi, K. and Wedderhurn, A. (eds) Shiftwork: Health, Sleep and Performance Peter Lang, Frankfurt, pp 448454 Epstein, R., Tzischinsky, 0. and Lavie, P. 1991 ‘Sleep-wake cycle in rotating shift workers: Effects of changing from phase advance to phase delay rotation’ 20th International Conference on Chronobiology, Tel Aviv Escrlha, V., Perez-Hoyos, S. and Bolumar, F. 1992 ‘Shiftwork: its impact on the length and quality of sleep among nurses of the Valencian region in Soain’ Int Arch OCCUDEnviron Health 64, 125-129 ’ Folkard, S. 1992 ‘Is there a “best compromise” shift system?’ Ergonomics 35 (12), 1453-1464 Folkard, S., Arendt and Clark, M. 1990 ‘Sleep and mood on a “weekly” rotating (7-7-7) shift system: some preliminary results’ in Costa, G., Cesana, G., Kogl, K. and Wedderhurn, A. (eds) Shiftwork: Health, Sleep and Performance Peter Lang, Frankfurt, pp 484-489 Folkard, S. and Totterdell, P. 1991 ‘Circadian variations in performance. Implications for abnormal work hours’ International Symposium ‘Shiftwork and Job Demands’ Paris, 11-12 July, Abstracts, pp 10-14 Folkard, S., Monk, T.H. and Lohban, M.C. 1978 ‘Short and longterm adjustment of circadian rhythms in “permanent” night nurses’ Ergonomics 21, 785-799 Fore& J. and Benoit, 0. 1978 ‘Shiftwork: the level of adjustment to schedule reversal assessed by a sleep study’ Waking and Sleeping 2, 107-l 12 Hildebrandt, G., Deitmer, P., Muug, R. and Poellmann, L. 1987 ‘Physiological criteria for the optimization of shift work (relations to field studies)’ in Oginski, A., Pokorski, J. and Rutenfranz, J. (eds) Contemporary Advances in Shiftwork Research, Medical Academy, Krakow, pp 121-131 Hornberger, S. and Knauth, P. 1993 ‘Interindividual differences on the subjective valuation of leisure time utility’ Ergonomics 36 (l-3)) 255-264 Jansen, B., van Hirtum, A. and Thierry, H. 1983 ‘The inconvenience of shiftwork (case based on the innovative compensatory functions model)’ European Foundation for the Improvement of Living and Working Conditions, Dublin Knauth, P. 1993 ‘The design of shift systems’ Ergonomics 36 (l-3), 15-28 Knauth, P. and Kiesswetter, E. 1987 ‘A change from weekly to quicker shift rotations: a field study of discontinuous three-shift workers’ Ergonomics 30, 1311-1321 Knauth, P. and Rutenfranz, J. 1972 ‘Untersuchungen iiber die

Designing

better shift systems:

P. Knauth

Beziehungen zwischen Schichtform und Tagesaufteilung‘ Int Arch Arbeitsmed 30, 173-191 Knauth, P., Emde, E., Rutenfranz, J., Kiesswetter, E. and Smith, P. 1981 ‘Re-entrainment of body temperature in field studies of shiftwork’ Int Arch Occup Environ Health 49, 137-149 Knauth, P., Kiesswetter, E., Ottmann, W., Karvonen, M.J. and Rutenfranz, J. 1983 ‘Time-budget studies of policemen in weekly or swiftly rotating shift systems’ Applied Ergonomics 14, 247-252 Knauth, P., Rutenfranz, J., Herrmann, G. and Pueppl, S.J. 1978 ‘Reentrainment of body temperature in experimental shiftwork studies’ Ergonomics 21, 775-783 Kiesswetter, E., Knauth, P., Schwarzenau, P. and Rutenfranz, J. 1985 ‘Daytime sleep adjustment of shiftworkers’ in Kuella, W.P., Rtither, E. and Schulz, H. (eds) Sleep ‘84 G. Fischer, Stuttgart, pp 273-275 Landen, R.O., VikstrGm, A.O. and dberg, B. 1981 Ordningspoliser i Stockholm. Delrapport III: I Intervention - sociala och psykologiska reaktioner pa forlndrade arbetstider. Rapporter fran Laboratoriet for Klinisk Stressforskning. Karolinska Institutet, Nr. 126, Stockholm Lavie, P. 1986 ‘Ultrashort sleep-waking schedule III. “Gates” and “forbidden zones” for sleep’ Electroenceph. clin. Neurophysiol. 63, 414-425 Monk, T.H. and Wagner, J.A. 1989 ‘Social factors can outweigh biological ones in determining night shift safety’ Human Factors 31, 721-724 Ong, C.N.. and Kogl, K. 1990 ‘Shiftwork in developing countries: current issues and trends’ in Scott, A.J. (ed) Shiftwork, Hanley and Belfus, Philadelphia, pp 417-428 Saito, Y. and Kogl, K. 1978 ‘Psychological conditions of working night and subsequent day shifts with short sleep hours between them’ Ergonomics 21, 871 Tepas, D.I. 1985 ‘Flexitime, compressed workweeks and other alternative work schedules’ in Folkard, S. and Monk, T. (eds) Hours of work. Temporal factors in Work-Scheduling John Wiley, Chichester, pp 147-164 Totterdell, P. and Folkard, S. 1990 ‘The effects of changing from weekly rotating to a rapidly rotating shift schedule‘ in Costa, G., Cesana, G., Kogi, K. and Wedderburn, A. (eds) Shiftwork: Health, Sleep and Performance Peter Lang, Frankfurt, pp 646650 Tsaneva, N., Nicolova, R., Topalova, M. and Danev, S. 1990 ‘Changes in the organism of shift workers operating a day and night 12 hour schedule in carbon disulfide production’ in Coasta, G., Cesana, G., Kogi, K. and Wedderburn, A. (eds) Shiftwork: Health, Sleep and Performance Peter Lang, Frankfurt, pp 324-329 Wallace, M. 1989 ‘The 3 day week: 12 hour shifts’ in Wallace, M. (ed) Managing shiftwork Brain-Behaviour Research Institute, Department of Psychology, La Trobe University, Bundoora, Australia 30833, pp 49-56 Wedderburn, A. 1981 ‘Is there a pattern in the value of time off work?’ in Reinherg, A., Vieux, N. and Andlauer, P. (eds) Night and Shiftwork. Biological and Social Aspects Pergamon Press, Oxford, pp 495-504 Wever, R.A. 1979 The Circadian System of Man Springer, New York Wilkinson, R.T. 1992 ‘How fast should the night shift rotate?’ Ergonomics 35, 1425-1446