Originals
Task-related increases in fatigue predict recovery time after academic stress
2016 Volume 58 Issue 1 Pages 89-95
Details
2016 Volume 58 Issue 1 Pages 89-95
Objective: The aim of this study was to investigate the time course of recovery after an academic exam as a model of high workload and its association with stress-related fatigue. Methods: Thirty-six medical students (17 females, 19 males) filled out diaries during an exam phase, starting 2 days prior to the exam, and a control phase 4 weeks after the exam for 14 days, respectively. Fatigue, distress, quality of sleep, and health complaints were assessed. Recovery time was determined for each individual and variable by comparing the 3-day average with the confidence interval of the control phase. Recovery time was predicted by Cox regression analyses. Results: Recovery times of all variables except health complaints were predicted by stress-related fatigue. Half of the individuals had recovered after 6 days, and 80% of the individuals had recovered after 8 days. Conclusion: The time necessary for recovery from work demands is determined by fatigue as a measure of resource depletion.
(J Occup Health 2016; 58: 89–95)
Phases of intense work including high workload and long works hours, such as is typical for the period prior to an academic exam, tax the individual's resources and thus are associated with increases in fatigue. This has, for example, been observed in experimental studies on mental work1) and in the course of an intense workday2) but also for jobs characterized by high workload in general3). In this context, it is assumed that mental and physical resources are limited and thus prone to depletion under high workload conditions, as has documented, for example, for mental effort1, 4, 5). A central sign of this depletion is subjective fatigue, “the state of weariness following a period of exertion, mental or physical, characterized by a decreased capacity for work and reduced efficiency to respond to stimuli”6). Next to fatigue, high workload has also been found to be associated with increases in subjective health complaints and distress, both in employees and students7, 8). Similar results have also been found for the quality of sleep, where high workload was associated with a deterioration of the quality of sleep in employees as well as in students during an exam period9, 10). It should be noted that in students, these strain reactions are not only a result of high workload but are additionally aggravated by anticipation of an exam and the threat of failure11). All in all, we therefore assume the period of exam preparation including the exam day to be associated with poorer health and well-being.
According to the effort-recovery model, these strain reactions are curable by recovery, i.e., by termination of the demanding activities12, 13). If, however, the recovery process cannot be completed and recovery is therefore insufficient, then not only are the strain reactions perpetuated, but the individual also has to display greater effort to overcome the diminished cognitive capacity associated with mental fatigue, thus additionally taxing his/her resources14, 15). Consequently, insufficient recovery has been found to be associated with, for example, increased accident risk16), prolonged fatigue17), and health complaints18).
For recovery to be successful, two conditions have to be met. First, there has to be sufficient time for recovery to come about. Second, the external and internal conditions have to enable recovery. The latter primarily includes physical termination of the demanding task and a mental detachment from work, i.e., cessation of a mental preoccupation with work19). This is necessary, as a continued mental preoccupation with the topics associated with stress prolongs the stress response and thus disrupts recovery20). Whereas the importance of mental detachment has been extensively studied in the last decade21), less is known about the amount of time necessary for the completion of recovery.
Classical experimental research on rest breaks during cognitively demanding work has shown that there is an optimal duration for rest breaks, depending on the duration of work and the scheduling of the breaks22). Rest breaks that are too short will not be fully restorative and therefore not lead to a full recovery of mental performance. Rest breaks that exceed a certain amount of time, on the other hand, will not add to recovery. Less is known regarding longer recovery periods, i.e., the number of “free”, non-work days necessary for recovery after longer phases of work (e.g., two day work shifts, a work week, etc.). This, however, has both practical and theoretical relevance. The practical relevance is obviously the number of non-work days that should be provided to an employee for complete recovery to occur. The typical 2-day weekend has been shown to contribute to recovery, though results are conflicting regarding whether it enables full recovery23, 24). In the case of more intense phases of work, it is suggested that in some instances, 2–4 days are necessary for recovery23, 25). The theoretical relevance of studying the time necessary for recovery is to ascertain whether the amount of time needed for complete recovery is dependent on the resource depletion associated with work, as suggested by the aforementioned studies on intense work and recovery23, 25). It is reasonable to assume that the more one draws on one's resources, the longer it will take to replenish them.
The aim of the present study was to investigate the effect of recovery time and its dependency on resource depletion. We chose to study recovery from academic stress, as it is a naturally occurring phase of intense work stress with a precise end point (an exam) and a recovery phase that is similar for all individuals. Also, the exam's preparatory phase resembles typical occupational demands in that work is characterized by (a) long work hours, (b) mental work, and (c) high time pressure. In addition, as stated before, academic stress leads to similar strain reactions as job stress, including health complaints, distress, and sleep problems. We deliberately selected a recovery period during ongoing academic activities, thus resembling something like “everyday life”, rather than a recovery period during holidays, as holidays may booster well-being beyond the level expected by usual recovery from work due to the higher occurrence of pleasurable activities26). Based on the previous discussion in the literature, we assume that (a) full recovery after a period of high workload ending with an exam should take at least 3 days and (b) that the recovery time is positively related to resource depletion at the time of the exam, which is operationalized as an increase of fatigue.
Medical students were approached by a doctoral student asking them to participate in a study on recovery from exam stress. Of the 50 diaries handed out, 38 were returned, and two were incomplete; this left us with 36 individuals (17 men, 19 women) participating in the study, which constitutes a response rate of 72%. Their mean age was 27.9 years (SD=5.7, age range 20–42). Only one student was taking medication; this student was taking medication against blood clotting. Regarding living arrangements, 28% of the students lived alone, 25% lived with their parents, and 47% lived with a partner. The majority of the participants (69%) were physically active at least once a week, and 14% of them were smokers.
VariablesFatigue and distress were measured with a widely used standardized German questionnaire assessing various aspects of mood experienced in the current situation27). Individuals scored adjectives associated with specific mood on a 6-point Likert-type scale, with the response categories “barely”, “a little”, “somewhat”, “quite”, “predominantly”, and “completely.” In response to the statement “This is how I felt today.” Fatigue was assessed with 8 adjectives (tired, in need of recovery, rested, exhausted, recovered, spent, drowsy, weary), and distress was assessed with 4 adjectives (tense, calm, nervous, relaxed). The internal scale consistencies in the present study were Cronbach's Alpha (CA)=0.83 for fatigue and CA=0.80 for distress. Quality of sleep and health complaints were assessed with a standardized German questionnaire assessing stress and recovery28). Individuals scored items on a 6-point Likert-type scale with the response categories “not”, ”barely”, “somewhat”, “quite”, “predominantly”, and “completely”. Quality of sleep was measured with 4 items (“I fell asleep peacefully”, “my sleep was restful”, “I woke without an external cause”, “I slept restlessly”). Health complaints were assessed with 4 items (“I had health complaints”, “I had a headache”, “I felt physically unwell”, and “I felt physically weary”). The internal scale consistencies in the present study were CA=0.85 for sleep and CA=0.79 for health complaints.
DesignThe study was approved by the ethics committee of the Medical University of Vienna (issue number 1058/2012). Students willing to participate in the study received a booklet containing a section on demographic variables as well as two diary sections to monitor well-being during the exam/recovery period and during the control period, which started 4 weeks after the exam. Participants were instructed to fill out the diary starting 2 days prior to the exam in the evening before dinner time for a time period of 14 days. The two days prior to the exam and the day of the exam were defined as the exam period, the 11 days following the exam were defined as the recovery period. Twenty-six days after the exam, participants were prompted to fill out the diary for 14 days as a control period. The control period was during a time of normal study activity during the semester, with four exceptions (11%), in which the control phase fell within the winter holidays. Depending on the stage of the curricula, exam durations varied between 1 and 5 hours. The median duration of exam preparation was 12 weeks, varying from 4 to 20 weeks. Students studied for their exam approximately 8 hours per day, including weekends, in addition to other academic tasks. Failure to pass the exam required a repetition of the exam, typically within several months; in extreme cases, a student may have to repeat a whole academic year. Thus, an academic exam is a strong incentive for intense preparatory work.
StatisticsIndividual recovery times were determined equally for each of the four outcome variables as follows: First, the means of 3 consecutive days of the exam and recovery periods were calculated to determine the average level of the respective variables for every day of the exam and recovery periods excluding the first and last day. This was done to ensure that the average level of the variable and not a single day potentially reflecting an outlier determined the recovery time. Second, as a benchmark for recovery, the 99% confidence interval (CI) of the 14 days of the control period was calculated. Full recovery was assumed to have occurred when the (3-day average) level of the respective variable was identical or smaller (larger) than the upper (lower) boundaries of the 99% CI. This was determined for each individual. In other words, full recovery was assumed to have occurred when the respective value of the recovery period was equal to the average level of the control period. This procedure resulted in four new variables specifying the days to recovery (recovery time) for each outcome variable.
To determine an individual's resource depletion associated with the exam, i.e., the increase of fatigue in the exam period, the average level of the respective variable was calculated for the two days preceding the exam and the exam day; from this number, the average level of the control period was subtracted, resulting in a strain amplitude associated with the exam period.
To determine the trajectory of recovery and association between resource depletion and recovery time, Cox regression analyses were performed with IBM SPSS Statistics (Version 20). The dependent variable was the number of days to recovery for the four outcome variables. The cases for which recovery was observed during the observation period versus not observed were denoted in the status variable. As the independent variable, the amplitude of fatigue was used. A total of 4 Cox Regression analyses were performed. The number of days to recovery for each outcome variable was obtained from the survival plot.
The average individual levels of fatigue and distress during the exam period were increased in all individuals (100%) in relation to the control period, and those of quality of sleep and health complaints were increased in almost all individuals (92%). Specifically, on a group level, the statistical differences between the exam and control periods and their effect sizes (Cohen's d) were t=18.3, p<0.001, and d=1.1 for fatigue; t=13.8, p<0.001, and d=1.1 for distress; t=9.2, p<0.001, and d=0.7 for quality of sleep; and t=8.4, p<0.001, and d=0.8 for health complaints. Group averages for the total exam/recovery and control period are presented exemplarily for fatigue in Fig. 1. The extent of the increase of fatigue as an independent variable varied between 31 and 147% of the associated average level of fatigue during the control period, the lower and upper quartiles being 58 and 101%. Depending on the variable, 83 to 97% achieved full recovery in the observed 11-day recovery period following the exam (Table 1).
Group means and standard errors of the mean for each day of the exam and the control period for fatigue; day 3 is the exam day.
| Fatigue | Distress | Quality of sleep | Health complaints | |
|---|---|---|---|---|
| Increase in fatigue | ||||
| Hazard ratio | 1.14** | 1.20** | 1.12* | 1.05 |
| 95% CI | 1.04–1.25 | 1.10–1.32 | 1.02–1.22 | 0.98–1.14 |
| N (model) | 30 (83%) | 35 (97%) | 32 (89%) | 32 (89%) |
| N (failure to respond) | 0 | 0 | 3 (8%) | 3 (8%) |
| N (failure to recover) | 6 (17%) | 1 (3%) | 1 (3%) | 1 (3%) |
| Total model chi-square | 9.2** | 15.9*** | 6.4* | 1.9 |
| Days to recovery | ||||
| 20% | 5 | 3 | 2 | 2 |
| 50% | 6 | 4 | 4 | 3 |
| 80% | 8 | 6 | 6 | 4 |
The results of the Cox Regression analyses are displayed in Table 1. The increase of fatigue in the exam period compared with the control period predicted the time to recovery for fatigue, distress, and quality of sleep, but not for health complaints. Greater stress-related fatigue was associated with longer recovery time as indicated by hazard ratios >1.
The time course of recovery is illustrated in Fig. 2 using inverse survival plots. The cumulative proportion of individuals showing recovery is plotted against the days to recovery. According to these plots, it took 6 days until at least 50% of the individuals had fully recovered from fatigue; for distress and quality of sleep, it took 4 days, and for health complaints, it took 3 days.
Time course of recovery for the four outcome variables based on inverse survival plots. Y-axis, probability for recovery; x-axis, days after the exam (=recovery period).
The present study investigated the time course of recovery after a major academic exam in medical students as a model for recovery after high workload and intense stress. The exam period was associated with a marked increase in fatigue, distress, and health complaints as well as a decrease in quality of sleep. This is in line with previous studies on the effect of intense work periods2) as well as on the effect academic exams8, 29). All variables started to improve on the day following the exam day. Most, but not all individuals, achieved full recovery in the 11-day observation period.
The average recovery times were 6 days for fatigue, 4 days for distress and quality of sleep, and 3 days for health complaints, thus supporting hypothesis 1. However, it took 8 days for the majority of the individuals (80%) to recover in all variables. Thus, the recovery times observed in the present study markedly exceeded previously reported recovery durations, which were 2 days in a study on recovery from 8- to 10-hour day and night shift work in nurses25), 3–4 days in construction workers with seven consecutive 12-hour day shifts23) and less than 3 days in students after an exam period30). However, these studies had some limitations, as, for example, a maximum of only 3 rest days were considered in the shift-work study, thus limiting the possible time span for recovery25), inferential statistics were not used in the construction worker study, thus limiting the strength of the results23) and measures were assessed only every 3 days in the exam period study, thus limiting its accuracy30). Besides these limitations, the exam stress in the present study may have been more severe, thus causing these discrepancies. In any event, the present study suggests that a period of high workload due to exam stress taxes resources to such an extent that the majority of young, healthy individuals need up to 8 days to fully recover their health and well-being.
A similar time span for improvement was found for employees on vacation, the health and well-being of whom peaked on the eighth day31). However, improvements during vacation do not only reflect a respite effect (i.e., being relieved of work stress), but also reflect the results of positive leisure activities improving mood and well-being on their own accord26). In contrast, the post-exam time in the present study occurred during normal study time, and thus in all probability, it did not provide additional incentives to promote well-being. In this sense, the present study reflects “everyday” recovery brought about by relief from high workload and stress.
The increase in fatigue induced by the exam period was used as estimate of resource depletion, a greater increase in fatigue indicating a greater depletion of resources. This is in line with the fact that (mental) fatigue is associated with a variety of cognitive and performance impairments in regard to, for example, complex cognitive activities32), accuracy1), reaction time33) and subjective task engagement1). Though the notion of resource depletion has been criticized as not taking motivational factors into account and changes in incentives can in fact reduce fatigue1), recovery is seen as the main pathway to improve fatigue and related impairments, irrespective of it origin34).
Fatigue associated with the exam period predicted recovery time in three of the four outcome variables: fatigue, distress, and quality of sleep. A greater increase in fatigue was associated with longer recovery times, thus supporting hypothesis 2, that is, that a greater resource depletion necessitates a longer time for resource replenishment. A relationship between resource depletion and recovery time is also suggested by animal studies, which found stress intensity to be positively associated with the recovery time of biological stress markers35). The failure to find an association with the recovery of health complaints may be due to the relatively smaller effect of exam stress on this variable as well as the shorter recovery time, limiting variance in reactivity as well as recovery time and, in consequence, reducing the chance of finding a significant association between resource depletion and recovery.
There are several factors that may explain the time-dependency of recovery: mental detachment, sleep quantity, and sleep quality. Mental detachment reduces the strain associated with a continued preoccupation with previously experienced stress19). It has been found that stress affects mental detachment negatively, thus presumably increasing the time required for successful detachment to occur. Sleep quality has also been found to be negatively affected by stress, making sleep less recuperating and thus delaying recovery36, 37). One the other hand, high workload may also lead to shorter sleep durations, resulting in a sleep debt, the restoration of which may require several nights38, 39).
Some tentative implications of the current findings for occupational health can be stated. Firstly, the positive association between resource depletion and time for recovery implies that the time allocated for recovery (the number of non-work days) should depend on the extent of work demands and/or work hours in the work episode preceding the recovery phase. In future studies, this association could possibly be quantified more precisely. Secondly, the interindividual differences in resource depletion found in this study, resulting from a similar, though not identical, episodes of (academic) work stress, suggest that similar tasks may exhaust individuals to a different extent, thus making it necessary to allocate recovery times depending on the individuals need for recovery. Indeed, individual differences in resource depletion and recovery were observed in one of the first studies on work and recovery ever published40). Thirdly, the time necessary for recovery after an intense period of work may exceed the number of days usually available for recovery, typically being two days. In these instances, longer recovery periods should be considered. At this point, however, the exact relationship between work demands, individual resource depletion and recovery times cannot be quantified and awaits future studies.
The following limitations need to be addressed. High workload due to exam preparation and anticipation of exam stress coincided in the present study, presumably resulting in greater stress than high workload alone. Thus, the underlying stress in the present study may be higher than the average stress experienced in usual job situations. This may also explain the unusually long recovery times. The cause for increases in fatigue during the exam phase was not determined. Both the extent of workload as well as individual characteristics may account for these increases. For almost all study participants, the recovery and control periods fell in usual study time within the academic semester. However, we did not assess what individuals actually did during these periods. It may well be that individuals engaged in more recovery activities in the post-exam phase than in other times of the semester, thus promoting recovery. We did not record students that where approached but not willing to participate in the present study, thereby limiting the possibilities to document a potential selection bias. However, the interindividual differences in age, strain, and recovery duration were large and generally sufficient to find associations. On the other hand, we cannot rule out that individuals consenting to participate were interested in the study because of individual difficulties with stress and/or recovery, potentially contributing to longer recovery times. Finally, the present study used a paper-and-pencil format, thus supplying no definite information on response times, and we cannot rule out the possibility that, for example, some participants filled out missed days retrospectively, thereby subjecting responses to recall bias.
To conclude, recovery from work stress is a time-dependent process, the required time for recovery being determined by the amount of fatigue experienced during work stress. Thus greater resource depletion necessitates longer recovery times. In the present study, recovery time for fatigue varied between 5 and 8 days, indicating that severe work stress may indeed require recovery periods clearly exceeding the length of a weekend.
Acknowledgments: We appreciate the help of Michael Kundi in analyzing the data.
