Performance in a Complex Multiple-Task Environment during a Laboratory-Based Simulation of Occasional Night Work
Sauer, Jurgen, Wastell, David G., Hockey, G. Robert J., Earle, Fiona, Human Factors
Disruptions of work-rest cycles may be experienced by human operators in many different work environments. These may occur in various forms. There is the widely researched area of shift work, which involves the completion of work on a regular basis at hours when humans are supposed to sleep. This causes major problems with regard to human performance and well-being, largely caused by variations in the circadian rhythm and an accumulating sleep deficit (Folkard, 1996). Other forms of work organization also increasingly require a less-regular form of night work, such as on-call duties and relief work. This means that after a full working day, people sometimes need to continue working throughout the night (e.g., on-call system engineers or junior hospital doctors). Generally, there is much less research on occasional night work than on shift work.
The present study is concerned with occasional night work (ONW), defined here as nocturnal work on a single night. This form of night work has some similarities with unadapted shift work in that ONW refers to the first night of a shift work schedule, during which resynchronization of the circadian rhythm is strongest. However, there are also significant differences between them. Operators are more likely to conserve cognitive-energetical resources, and hence adopt different strategies, knowing that the first night of shift work is the first of a sequence involving several successive night shifts (Alluisi & Morgan, 1982).
Although the shift work literature has examined very thoroughly the effects of different shift work schedules (e.g., rotation rate, direction of rotation, cycle length), there has been little work on ONW. Nevertheless, research on shift work is of some relevance in the present case, as it has generally shown that performance declines during night work as compared with daytime work (Folkard & Monk, 1979). This was paralleled by a higher accident rate at night and, in particular, by an increase in the seriousness of accidents (Smith, Folkard, & Poole, 1994). It has also been shown that performance patterns during night work vary as a function of the cognitive demands associated with task completion (Tepas, Paley, & Popkin, 1997). For example, there is evidence that cognitive tasks with a high memory load are performed comparatively well at night, whereas performance degradations are observed on low-memory-load tasks (Folkard, Knauth, Monk, & Rutenfranz, 1976). Apart from changes in task performance, night work may also result in adaptive changes of task management strategies (Daniellou, 1986). This refers to how a task is done rather than how well it is done. These adaptive changes may allow the operator to maintain adequate performance levels even during periods of scarcer mental and energetical resources.
A limitation of shift work research and sleep research in general is the use of rather simple tasks. Few studies have taken measures of performance on realistically complex tasks (e.g., Monk & Embrey, 1981). Most studies have employed standard cognitive tasks, such as choice reaction time (e.g., Smith, Totterdell, & Folkard, 1995), simple reaction time (e.g., Wilkinson, Allison, Feeney, & Kaminska, 1989), five-target Sternberg (e.g., Folkard, Totterdell, Minors, & Waterhouse, 1993), and detection-of-repeated-numbers tasks (Ottmann, 1998). The limited work that has been done with more complex tasks reveals that performance is more robust than with simple tasks. For instance, Wilkinson (1964) showed that tasks involving complex decision making were remarkably immune to sleep loss even after 60 hr of continuous work. Decrements on vigilance tasks are also attenuated if the vigilance task is strongly integrated in a complex task environment (Alluisi & Morgan, 1982). Some studies have employed tasks that more closely simulate aspects of a real work environment, such as the simulated assembly line task (SALT; e. …