7578409733

*• • •"#* • ***• • •* 0 • •

I

Performance

if sleepiness on (hc night shifl is as widespread and as dramatic as has alrcady been indicated, one would expect to sce pronounced effects on performance, and con$equeni!y on output and safely. One of thc clas-sics in ihis arca is ihe study by Djerner et al (41), who showed that errors in mcter readings over a period of 20 years in a gas works had a pronounced peak on the night shift. There was also a secondary peak during the aflernoon (figurę 3). Similarly, Browne (42) demonstrated that telephone operators connected calls at a considerably slower pace al night. Hildebrandt et a! (43) found that locomotive engineers failed to operatc thcir alerling safety device morę often at night than during the day, with a secondary peak around 1500.

Most othcr studies of performance have uscd labora-tory typcs of tests and demonstrated, eg. reduced reac-tion timc or poorer mcntal ariihmetic on the night shift (15). Flight simulaiion studies havc, furthermore, shown that the ability to “fly^M a Simulator at night rnay decrease to a level corrcsponding to that after moderate alcohol consumption (0.05 blood alco-hol) (44). To these rcsulls may bc added those from numerous laboratory studies which have demonstrated that performance on many tasks deteriorates during the night hours (45).

Adjustment across shifts has very seldom been in-vcstigated under practical conditions. Laboratory in-vcstigations, howcvcr, clearly indicatc that adjustment does occur, although it may take up to iwo wecks. Frc-ąucnily. the body temperaturę rhythm adjusts in par-altel.

The impression of the night shift deterioration of performance is mainly based on fairly simple psycho-motor typcs of tasks. There is, howcvcr, some specu-lation that high-!cvel coenitive tasks, because of a high mcmory load. might show a dirtcrently phascd rhythm (45). The latter would not, howcver, apply to the siiua-tion wherc slcepiness has come close to actual slecp. sińce any type of activity would then be interfered with.

Time of day

Figurę 3. Number of meler rcading errors accumulaied over »9 years for ihrcc thrcc-shift workers. flars indieaie The ihree shifis. (Figurę redrawn from rcference 4i|

Another important point is that common sense and availab!e data suggest that the output from a produc-tion process will not be affected by night.work as long thc major determinant of the production (Iow is machines rather than peopłe. Thus, it seems rather un-likely that sleepiness induced by thc nightshift would affect output in all occupations.

A morę important arca of impact may bc safely. If sleepiness is severc enough, interaction with thc cn-vironment will cease. and, if ihis interaction coincidcs with a critical need for action. an accidcm may ensue. Such potential performance lapscs duc to nightwork sleepiness were scen for several of the locomotive engineers discussed carlier (31). The transport arca is. in fact. where most of thc available accident data on night shift slcepiness has been obtained. Thus Harris (46) and Hamelin (47, 48) convincingly demonstrated that single vehiclc accidcms havc, by far, the greatest prob-ability of occurring at night (carly morning). Most of ihesc accidents arc lhoughi to bc duc to sleepiness. Wiih respect to air transport Ribak et al (49) found miliiary flight accidents to bc incrcascd in the carly morning, and Price & Holley (50) argucd that also many civil air transport accidents may be causcd by fatigue duc to work scheduling. Finally, a number of spcaacular nuckar accidents (including those at Cher-nobyl and Harrisburg) have been panly attributed to fatigue-inducing work schedules (51).

As with sleepiness, (hc main reason for night shifl deterioration in performance is circadian rhythmicity and słeep loss (45).

Modifying factors

Several factors influence the adjustment to shift work. One such factor is thc direction of rotation of the shifl schcdulc. Since the frcc-running (spontaneous) period of the human slccp-wake cycle avcrages 25 h and sincc it can be emraincd by cnvironmental timc cucs only wiihin I to 2 h of the free-run period, phasc dclays are easier to accoinplish than phase advanccs (52). For thc rotating shift worker ihis situation implies that schedules that dclay, ie, rotaic eloekwise (morning-afternoon-night) shoutd be preferred (o those that ro-tatccounterclockwise. There has been, however. very fcw practical tests of this theory. Still, Czeisler et al (53) have demonstrated that a change from counter-clockwise to dockwise rotation, together with a change from 7-d to 21-d rotation, improved production and well bcing for three-shifi workers. Orth-Gomćr (54) found ihat a change in the same direction among rapid-ly (1 d) rotating police officers reduced blood pressure and improved well betng.

The kngth of a work shift Ls another parameter that one would expect to influence at least sleepiness and performance. In the laboratory it is usually the casc that performance falls with timc if learning effects arc climinatcd (55). Still. in one study of połicemen, Pca-cock ct al (56) found no effeets of a change from 8 h

V

y

• # •

l >

•er-11\*. :>r

' A

If

i;

•» •

(ninc shifts acroi free-two days-th cver, the distribuuun timc. Two other stu. shift workers (58 j cently Rosa &Cc j experiment and demt indeed, produced hi night shifts. In ; i truck drivers, Ha 4 relation between hot an initial “warro »»■: with an incrcase <t Asmaybeexp : (4 on, 8 off) on shi ness and poor pe^^r't parenlly, rotatinj i (hc individual th Another type of i crews on transmi workhours dispta also the time ref< shifts. As with other have dcmonstrat (61). The disturt also in flight simuid In some occupatic worksite until m < others, physicians i are often arc spent * ten pronounced, and although the pra : are still unknowr Systems may be fout officers in the rr~^m-Among indivi< 1 to slecp disturbi : graphic studies (renc supcrficial slecp ii i studies by Forel ; (66) also indical.w related to generał wc Koller ct al (67) ; carlier among shi Dahlgren (68) foLi.J work on the rhythm shifts. Ncilher di ‘ 10-week period <    »

and day work in a gr found ihal the amot turbances in presc « to thcir previou< < minaull et al (71) former shifl workers disturbances app« i directly related to servc that Angersbac carlkr occurrenc- -thrcc-shift worker

i

70