Fatigue at the Workplace: Measurement and Temporal Development

Abstract

Fatigue at the workplace has been described as a multidimensional construct, affecting the overall state of the whole organism, which may be a consequence of prolonged work and various psychological, socioeconomic, and environmental factors. In the short term, fatigue may lead to discomfort, diminished motor control, reduced proprioception, increased force variability, and reduced strength capability, resulting in reduced performance, lowered productivity, deficits in work quality, and increased incidence of accidents and human errors. Fatigue may also lead to longer-term adverse health outcomes such as chronic fatigue syndrome, myalgia, and burnout syndromes, and may be a precursor to WMSDs – work-related musculoskeletal disorders. If indeed fatigue is a precursor to WMSDs and other long-term health outcomes, it may then be a relevant biomarker for cumulative exposure to repetitive and/or sustained work, and thus a useful risk indicator and/or a design and evaluation tool. However, little is known of the temporal pattern of fatigue development and its relationships with disorder risks and work performance. The objective of this thesis was to identify and evaluate a battery of fatigue measures for both laboratory and field-based research, and provide insight into fatigue development in work-relevant task conditions.

Six studies were designed to address these objectives. In the first study, measures and analysis methods that detect fatigue-related changes were identified by a group of expert fatigue researchers. The second was an exploratory study focused on the responsiveness of a select number of measures during a workday and multiple workdays in realistic physically demanding residential plumbing work. In the third study, a selected number of conventional and novel measures were evaluated for their reliability and sensitivity in a controlled laboratory setting. This study also addressed the responsiveness of measures during a test battery or during the fatiguing activity (i.e., continuously), and the time between cessation of activity and test battery in which measures remain responsive. The fourth study reported on whether circadian effects were detectable by selected measures, providing insight towards the daylong reliability of these measures. In the fifth study, measures were evaluated in four fatiguing conditions, representing changes in type of contraction, intensity, and body segment. Furthermore, the pattern of fatigue development and the temporal responsiveness of measures were described. Finally, measures were assessed over an 8-hour light precision micropipetting task to investigate temporal responsiveness of measures and fatigue development. Errors were quantified and the effects of scheduled work breaks were reported.

In study 1, fifty-seven measures were identified based on outcomes and/or effects of fatigue in the workplace. Based on the perceived validity, reliability, and practicality in laboratory and field investigations, four measures were recommended for both settings: maximum voluntary contractions, questionnaires and fatigue scales, Borg’s rating of perceived exercise or discomfort, and visual analog scales. On the other hand, twenty-five measures were not recommended for field studies, including methods traditionally recognized as “gold standard” in measuring cellular and metabolic changes.

In study 2, fatigue was documented in realistic physically demanding work while employing a set of measures to provide a comprehensive picture of fatigue development. Not all measures revealed increasing fatigue over the workday or over the workweek, which may be a result of measures reflecting different fatigue processes. Thus, the study reinforced the need of a complementary set of measures, reflecting multiple domains, to measure and interpret the temporal development of fatigue. Two measures, rating of perceived discomfort and grip strength, indicated significant differences within a work day, notably an increase at the beginning and end of the shift (perceived discomfort) and a decrease between mid-shift and end of shift (grip strength). It was speculated that within-day trends were consistent with central fatigue mechanisms. Over multiple workdays, both central and peripheral components displayed a significant day effect. Fatigue accumulation over the workweek was observed with grip strength, physiological resting tremor, and postural tremor measures, particularly between day 1 (Tuesday) and day 4 (Friday).

In study 3, test-retest reliability ranged between “poor agreement” and “almost perfect agreement”. In terms of sensitivity, action tremor, MMG RMS amplitude, postural tremor, and rating of perceived fatigue were highly responsive. Perceived fatigue remained elevated, relative to baseline, until 11 minutes post-exercise. Postural and physiological tremor persisted from baseline until the third minute of recovery. Action tremor, however, quickly recovered within the first minute of recovery. This current study found that for most of the measures, there were no statistical differences between test battery and continuous measurement, but a few measures were approaching statistical significance. Action tremor and mechanomyography collected during a test contraction, and perceived fatigue assessed by a visual analog scale, were found to be most reliable, most responsive, comparable to continuous measures, and sensitive after the fatiguing activity, and should be considered with other measures of interest, as part of a test battery.

In study 4, only two measures revealed a statistically significant time-of-day effect: mechanomyography of a flexor forearm muscle and action tremor at 30% MVC. These two measures exhibited rhythmicity based on cosinor analysis. Therefore a degree of caution might be required when interpreting daylong fatigue with these two measures, whereas the other measures may not be susceptible to, or detect, significant diurnal effects. Although the remaining measures did not reveal statistically significant time effects, most measures were characterized with similar patterns to those found in previous literature.

In study 5, there was no one universal measure that was common, in terms of responsiveness, in all exercise conditions. Although no single measure was found to be most responsive in all conditions, there were measures responsive in most exercise conditions as either a continuous or test battery measure. This was the case with action tremor. A maximum voluntary contraction, which is dependent on processes in both central and peripheral domains, was similarly responsive. Rating of perceived fatigue, which has been cited as a centrally mediated indicator, was also found to increase with exercise progression in hand conditions. Therefore fatigue measures, reflecting changes to both central and peripheral processes, may be useful in measuring tasks and exercises of varying parameters. In this study, we support earlier investigations on the pattern of fatigue development in isometric and time-varying (e.g., intermittent isometric, concentric) contractions. The temporal responsiveness of central and peripheral measures, on the other hand, may be a better reflection of the intensity of the task. The shoulder intermittent condition was not consistent with the expected pattern for an intermittent isometric contraction. However, the study protocol may have inadvertently generated lower muscle activity, and therefore the extent of fatigue may have been minimal. There remains a need to understand complex combinations of task-dependent factors in both fatigue development and temporal responsiveness.

In study 6, nine measurement parameters revealed significant increases in fatigue over the work period. Traditional field measures (i.e., MVC and EMG) did not lead to extraordinary time effects. Error rates followed similar trends to the 9 significant measurements: an increase from baseline towards mid-morning, a slight decrease prior to the lunch break, a nadir after lunch, and increasing fatigue effects over the course of the afternoon. Error rates, however, might not be a sole consequence of fatigue – cognitive and physical; but might also reflect changes in arousal level. Over the pipetting task, there was interplay between peripheral and central fatigue mechanisms in three body segments: thumb, hand, and shoulder. Fatigue developed at a “local” level (i.e., at the three body segments) and was consistent with expected patterns observed in study 5, particularly if thumb and shoulder actions were considered concentric actions and the grip force was a sustained isometric contraction.

Overall, the collective assessments suggested that rating of perceived fatigue and action tremor, on average, were highly repeatable and responsive in multiple task conditions. Postural tremor or steadiness and maximum voluntary contractions were moderately reliable and responsive. Different forms of tremor may be responsive to different task conditions. Postural tremor amplitude was found to increase over the course of an 8-hour workday in a light precision work task, and over multiple days in physically demanding work. Action tremor, on the other hand, appeared to be responsive at higher work intensities performed at a shorter duration of time. Possibly, action tremor may be more indicative of changes in the periphery, whereas postural tremor reflects changes more central in nature. Consequently, these measures should be considered for inclusion into a test battery for field use.

For the ergonomist or health and safety practitioner, this body of work provides some insight into the utility of a test battery of fatigue measures to complement current task analysis techniques. For workplace researchers, this dissertation provides insight into the temporal development of fatigue in various task conditions and the reliability and responsiveness of select measures in both short and longer-term work-studies. This research might subsequently elicit future investigations in the relationship between work exposure, fatigue development, and performance and longer-term health outcomes.