A summary of the findings for cardiovascular diseases and other illnesses, injuries, health behaviors, and performance effects are listed below. Findings are discussed for each of the four work schedule categories listed earlier.

FINDINGS ASSOCIATED WITH OVERTIME

Twenty-nine studies and one meta-analysis examined associations between overtime and the outcomes targeted for this report. The studies were conducted in Japan (10), United States (5), Sweden (5), Germany (2), South Korea (2), Canada (1), Hong Kong (1), Netherlands (1), Thailand (1), and United Kingdom (1). The studies used a variety of criteria to group participants based on the number of hours worked. For example, the criterion used to define the group with the lowest number of hours worked ranged widely from 35 to 60 hours per week across studies.

Overtime and Cardiovascular Findings

Table 3 displays the methods and results for the studies examining overtime and cardiovascular findings. Two case-control studies of Japanese workers reported that overtime during the previous month was associated with an increased risk for acute myocardial infarction. Liu et al. [2002] reported that 61 or more hours of work per week and less than 2 days off a month increased the odds by two-fold or more. Sokejima and Kagamimori [1998] observed a U-shaped relationship: as compared with 7 to 9 hours of work per day, higher risk was associated with both shorter hours (less than 7 hours a day) and longer hours (more than 11 hours a day).

Findings for hypertension were inconsistent across four studies. Iwasaki et al. [1998] reported significantly elevated systolic blood pressure in older salesmen (ages 50 – 60) whose combined commute and work hours exceeded 61 hours per week as compared with older salesmen working 57 hours or less. No differences were reported in younger workers (ages 20 – 49 years). Hayashi et al. [1996] observed increased blood pressure in groups of white collar employees working 84 to 96 mean hours of overtime per month as compared with those working 25 to 43 mean hours of overtime. Nakanishi et al. [2001b], however, published the opposite results: white collar workers reporting 10 or more hours of work per day had a lower risk for developing hypertension when compared with workers reporting less than 8 hours of work per day. Lastly, Park et al. [2001a] reported no correlation between blood pressure and work hours in Korean engineers whose work hours during the previous month ranged from an average of 52 hours to 89 hours per week. No participants in this study worked less than 52 hours on average per week.

Overtime and Other Illnesses

Table 4 displays the methods and results for the studies examining overtime and other illnesses. Overtime was associated with poorer perceived general health in three of four studies [Ettner and Grzywacz 2001; Kirkcaldy et al. 2000; Siu and Donald 1995; Worrall and Cooper 1999], increased neck or musculoskeletal discomfort in two studies [Bergqvist et al. 1995; Fredriksson et al. 1999], increased mortality in one study [Nylén et al. 2001], and subfecundity in one study [Tuntiseranee et al. 1998].

Mozurkewich et al. [2000] conducted a metaanalysis of 10 studies published between 1987 and 1997 and reported no association between long work hours and preterm birth. Further analysis of the six higher quality studies suggested a weak relationship between long working hours and preterm birth (Odd Ratio = 1.24 with a 95% Confidence Interval of 1.04 to 1.48). In contrast, Voss et al. [2001] reported that more than 50 hours of overtime during the previous year was associated with less sick time in a Swedish study.

Associations with diabetes mellitus in two Japanese prospective health studies were contradictory. Kawakami et al. [1999] reported that 50 or more hours of overtime per month increased the risk for development of diabetes mellitus as compared with 25 hours or less. In contrast, Nakanishi et al. [2001a] reported that 11 hours or more a day was associated with a reduced risk as compared with less than 8 hours. Both studies collected work hour data at the initial contact and did not examine the influence of working long hours over the course of several years.

In summary, overtime was associated with increased morbidity and mortality in 8 of 12 studies, and one meta-analysis suggested a possible weak relationship between overtime and preterm birth.

Combined relationship of pressure to work overtime and rewards in Dutch postal workers was examined by van der Hulst et al. [2001]. Rewards included salary, job security, and career opportunities. They reported that high pressure to work overtime in combination with low rewards was associated with a 3-fold increase in the odds for somatic complaints as compared with a reference category of low overtime pressure in combination with high rewards. In contrast, high pressure in combination with high rewards did not differ from the reference category. Ninety-five percent of the sample worked less than 50 hours per week.

Siu and Donald [1995] also reported a relationship with overtime pay. Men from Hong Kong who received no payment for overtime reported more health complaints when compared with men who received payment.

Mizoue et al. [2001] examined the relationship of overtime and sick building syndrome among Japanese municipal employees working in an environment with few workplace smoking restrictions. Thirty hours of overtime or more during the previous month was associated with a 2.6-fold increased risk for symptoms of general malaise and irritation of the mucous membranes and skin.

Fredriksson et al. [1999] examined the combined influence of domestic workload and overtime in workers from a broad range of occupations in Sweden. Additional domestic workload increased the cumulative incidence or prevalence risk for disorders of the neck in men and women who were working overtime.

Overtime and Injuries

Table 5 displays the methods and results for the studies examining overtime and injuries. Two studies reported that overtime was associated with higher on-the-job injury rates in construction workers or health care workers [Lowery et al. 1998; Simpson and Severson 2000]. Åkerstedt et al. [2002], however, reported no relationship between more than 50 hours of work per week and work-related fatalities in a 20-year prospective Swedish study.

Overtime and Health Behaviors

Table 6 displays the methods and results for the studies examining overtime and health behaviors. Studies by Nakamura et al. [1998] and Shields [1999] reported that overtime was associated with increased odds for unhealthy weight gain in men. Shields also reported that changing from a 40-hour workweek to longer working hours raised the odds for smoking in both men and women. In contrast, Park et al. [2001b] found no difference in smoking across three groups of engineers whose work hours ranged from a minimum of 52 hours per week to a maximum of 89 hours.

Differences in alcohol consumption also varied among studies. Shields [1999] reported a Ushaped relationship: women who either reduced or increased their average hours worked per week during the previous two years increased their odds for higher alcohol consumption. Trinkoff and Storr [1998] reported higher alcohol consumption in nurses was associated with working more overtime shifts per month. Park et al. [2001b] reported no differences in alcohol use among engineers whose work hours ranged from minimum of 52 hours per week to a maximum of 89 hours.

Mizoue et al. [2001] found a significant decrease in the percentage of workers who participated in regular sports activity as overtime hours increased. Studies by Shields [1999] and Kageyama et al. [1998], however, reported no significant relationship between long working hours and leisure-time physical activity.

Overtime and Performance

Table 6 displays the methods and results for the studies examining overtime and performance. Proctor et al. [1996] investigated 248 United Auto Workers working day and evening shifts. The researchers reported poorer performance on tests of cognitive function (e.g., Trail-making Test, Wisconsin Card Sort Task, Symbol Digit Substitution Task, Visual Reproduction, Pattern Memory, Vocabulary Task) and executive function (the ability to prioritize and plan tasks) for those individuals who worked overtime as compared with those who did not. Kirkcaldy et al. [1997] reported that as work hours increased in health care workers, automobile crashes and on the job "accidents" increased.

FINDINGS ASSOCIATED WITH EXTENDED WORK SHIFTS

Twelve field studies and three laboratory studies examined associations between extended work shifts and the outcomes targeted for this report. The field studies were conducted in United States (4), Australia (2), Sweden (2), United Kingdom (2), France (1), and Germany (1). The studies compared a variety of extended shift schedules: 12-hour day shifts and 12-hour night shifts; 8- hour rotations and 12-hour rotations; 8-hour and 10-hour rotations. Injuries and performance across the 1st hour through the 12th hour of long shifts were also examined. Table 7 displays the methods and results for studies that examined extended work shifts.

Extended Work Shifts and Illnesses

Lipscomb et al. [2002] reported that working 12 or more hours per shift was associated with increased risk for back disorders in nurses when compared with an 8-hour shift. Prunier- Poulmaire et al. [1998] reported that a 12-hour fast rotation (shift change more than once a week) was associated with increased leg pain, and visual complaints, as compared with day shift. In addition, the 8-hour 3-shift rotation showed increased risk for more leg pain, as well as more cardiovascular and gastrointestinal complaints, when compared with day shift. In contrast, Johnson and Sharit [2001] reported that a 12-hour fast rotation was associated with better perceived general health and fewer gastrointestinal complaints when compared with a fast 8-hour 3-shift rotation.

Smith et al. [1998] compared 12-hour day-night rotations with flexible start times and 12-hour rotations with rigid start times, but found no differences in cardiovascular, gastrointestinal, or pain symptoms.

Extended Work Shifts and Injuries

Hänecke et al. [1998] analyzed 1.2 million injury reports from two national databases in Germany and reported a higher risk for injury after the 8th or 9th hour at work for all shifts. The report indicated more pronounced risk for evening and night shifts as compared to days. Macias et al. [1996] examined hospital incident reports for a 30-month period at one university hospital in the United States and reported that needlestick and biological fluid exposure rates increased during the last 2 hours of 12-hour shifts, whereas no increase occurred during the last 2 hours of 8- hour shifts. Johnson and Sharit [2001], however, reported that production workers who changed from an 8-hour to a 12-hour shift did not show an increase in recordable injuries or lost-time incidents after the change.

Extended Work Shifts and Health Behaviors

Trinkoff and Storr [1998] reported increased odds for higher alcohol use in nurses who worked longer rotating or night shifts and increased odds for smoking in nurses who worked extended night shifts. No relationship was reported between working hours and drug abuse.

Extended Work Shifts and Performance

Two laboratory studies reported deterioration in performance with extended shifts. Rosa et al. [1998] compared a 2-week 12-hour day/night rotation and a 2-week 8-hour day-night rotation using a simulated manual assembly task at three repetition rates and three torque loads. They reported that upper extremity fatigue increased more quickly with increasing time on shift and occurred more quickly during night shifts. The highest fatigue levels were found during 12-hour night shifts. Macdonald and Bendak [2000] compared a 12-hour shift to a more standard workday (7.2 hours) in a laboratory study and reported that the longer workday was associated with deterioration in grammatical reasoning and alertness.

In contrast, four field studies reported no differences in their performance measures during extended shifts. Schroeder et al. [1998] reported that air traffic control personnel working four 10-hour shifts did not significantly differ from personnel working five 8-hour shifts on tests of grammatical reasoning, reaction time, and digit addition although performance of both groups declined across the workweek. Similarly, Smith et al. [1995] reported no significant declines in alertness or cognitive performance between 8-hour and 12-hour shifts in nuclear power plant shift workers. Axelsson et al. [1998] reported no significant difference in simple reaction time and vigilance task measures between 8- and 12-hour shifts in Swedish power plant workers. Also, Lowden et al. [1998] reported no consistent differences on simple performance measures, such as reaction time, in shift workers who switched from an 8-hour to a 12-hour schedule.

The way other features of work schedules, work tasks, and work environment influenced the relationship of shift length and performance were examined by three studies. Smith et al. [1998] reported improvements in alertness in 12-hour day/night rotations with flexible start times when compared with 12-hour rotations with rigid start times. The field study by Macdonald and Bendak [2000] reported that a combination of high workload and 12-hour shifts was associated more consistently with increased errors on grammatical reasoning, greater deterioration in hand steadiness and alertness, and more discomfort when compared to high workload during 8-hour shifts.

Brake and Bates [2001] examined extended shifts in combination with heat stress in Australian underground miners. Cardiovascular fatigue assessed by continuous heart rate monitoring showed no difference between 6-hour shifts and self-paced extended shifts (10-hour to 12.5-hour shifts). Additional heart rate monitoring while riding cycle ergometers showed increases during the first half of long shifts, but decreases during the second half of the shift. Based on these results, the authors suggested that the miners had reduced their effort and were pacing themselves in the latter part of the extended shifts.

One study examined the influence of age. Reid and Dawson [2001] conducted a laboratory study of simulated 12-hour shifts and neurobehavioral performance in younger and older participants. Older laboratory subjects were less able than younger subjects to maintain performance across 12-hour shifts.

FINDINGS ASSOCIATED WITH EXTENDED WORK SHIFTS, COMBINED WITH MORE THAN 40-HOURS WORK PER WEEK

Six field studies examined extended work shifts which also had more than 40-hours of work per week. The studies in this section clearly reported that participants worked on average more than 40 hours per week over the course of several weeks. The studies were conducted in the United States (2), Australia (1), Brazil (1), Canada (1), and United Kingdom (1). A variety of 8-hour and 12-hour schedules were compared across the studies. Table 8 shows the methods and results for studies that examined extended work shifts when combined with more than 40 hours per week.

Extended Work Shifts Combined with More than 40-Hours Work per Week and Illnesses

Lipscomb et al. [2002] reported that the combination of 12-hour shifts and 40 or more hours of work a week was associated with elevated risk for neck, shoulder, and back disorders as compared to five 8-hour shifts per week. In contrast, Mitchell and Williamson [2000] reported fewer health complaints during a 12-hour day/night fast forward rotation when compared with an 8-hour 3-shift weekly backward rotation.

Tucker et al. [1998a] examined early and late start times in workers on 12-hour shift rotations and 8-hour 3-shift rotations. Both work schedules changed shifts more than once a week. The 12-hour shift was associated with more cardiovascular and musculoskeletal complaints than the 8-hour shift. Workers on 12-hour shifts, having early changeover time, reported the most cardiovascular and musculoskeletal complaints.

Extended Work Shifts Combined with More than 40-Hours Work per Week and Injuries

Mitchell and Williamson [2000] reported that injury data from Australian electrical power station workers were similar after changing from an 8-hour shift to a 12-hour shift; two injuries occurred during the 8-hour schedule, and one occurred during the 12-hour schedule.

Extended Work Shifts Combined with More than 40-Hours Work per Week and Health Behaviors

Mitchell and Williamson [2000] reported that 47% of workers on an 8-hour 3-shift weekly rotation reported using alcohol as a sleep aid when compared with 17% of workers on a 12-hour fast rotation. The 8-hour shifts also had a higher percentage of workers smoking.

Extended Work Shifts Combined with More than 40-Hours Work per Week and Performance

Four studies reported some deterioration in performance when 12-hour shifts were combined with more than 40 hours work per week. Novak and Auvil-Novak [1996] reported an unexpected outcome from the focus groups of nurses who worked four 12-hour night shifts per week: nearly all nurses reported an automobile crash or near-miss during the previous 12 months while driving home after working a 12-hour night shift. The nurses reported no job performance effects when they maintained consistent sleep and wake times, but changing from night work to day activities was fatiguing and affected performance. In a field study, Fischer et al. [2000] examined the 2nd, 6th, and 10th hours of 12-hour shifts in Brazilian petrochemical plant workers and reported a significant decline in subjective alertness at the 10th hour for both day and night shifts. Similarly, Mitchell and Williamson [2000] reported more vigilance task errors at the end of 12-hour day and night shifts when compared to the beginning of the shifts in Australian power plant workers, while no effect was reported for an 8-hour schedule. On the other hand, significant improvements were observed for simple reaction time and grammatical reasoning tests given at the end of the 12-hour shift when compared to the beginning. Although Duchon et al. [1997] reported no differences between 8- and 12-hour shifts on cognitive and psychomotor performance in Canadian mine workers, the heart rate findings suggest that the 12-hour workers slowed the pace of their work.

FINDINGS ASSOCIATED WITH VERY LONG SHIFTS

Three studies examined the relationship between very long shifts and immune function or performance. The studies were conducted in Ireland, Japan, and New Zealand. Table 9 displays the methods and results for the studies that examined very long work shifts.

Very Long Shifts and Other Illnesses

Nakano et al. [1998] reported better immune function in drivers who were allowed to work overtime as compared with drivers having workhour restrictions. This Japanese study examined taxi drivers working 48-hour or longer shifts in 1992 and again in 1993, before and during the economic depression.

Very Long Shifts and Performance

A study in Ireland by Leonard et al. [1998] reported declines in two tests of alertness and concentration in medical residents who had worked 32-hour on-call shifts. They reported no significant declines in a test of psychomotor performance or a test of memory. A New Zealand survey of anesthesiologists linked long working hours to self-reported clinical errors [Gander et al. 2000]

Index: Main Page Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7 Table 8 Table 9 Abbreviations Acknowledgments Introduction Description of Work Schedules and Samples Health and Safety Findings Summary Concluding Remarks References