For a long time, nurses and nursing assistants were among the only women working at night in many countries (Gadbois 1981; Estryn-Béhar and Poinsignon 1989). In addition to the problems already documented among men, these women suffer additional problems related to their family responsibilities. Sleep deprivation has been convincingly demonstrated among these women, and there is concern about the quality of care they are able to dispense in the absence of appropriate rest.
Organization of Schedules and Family Obligations
It appears that personal feelings about social and family life are at least partially responsible for the decision to accept or refuse night work. These feelings, in turn, lead workers to minimize or exaggerate their health problems (Lert, Marne and Gueguen 1993; Ramaciotti et al. 1990). Among non-professional personnel, financial compensation is the main determinant of the acceptance or refusal of night work.
Other work schedules may also pose problems. Morning-shift workers sometimes must rise before 05:00 and so lose some of the sleep that is essential for their recovery. Afternoon shifts finish between 21:00 and 23:00, limiting social and family life. Thus, often only 20% of women working in large university hospitals have work schedules in synchrony with the rest of society (Cristofari et al. 1989).
Complaints related to work schedules are more frequent among health care workers than among other employees (62% versus 39%) and indeed are among the complaints most frequently voiced by nurses (Lahaye et al. 1993).
One study demonstrated the interaction of work satisfaction with social factors, even in the presence of sleep deprivation (Verhaegen et al. 1987). In this study, nurses working only night shifts were more satisfied with their work than nurses working rotating shifts. These differences were attributed to the fact that all the night-shift nurses chose to work at night and organized their family life accordingly, while rotating-shift nurses found even rare night-shift work a disturbance of their personal and family lives. However, Estryn-Béhar et al. (1989b) reported that mothers working only night shifts were more tired and went out less frequently compared with male night-shift nurses.
In the Netherlands, the prevalence of work complaints was higher among nurses working rotating shifts than among those working only day shifts (Van Deursen et al. 1993) (see table 1).
Table 1. Prevalence of work complaints according to shift
|
Rotating shifts (%) |
Day shifts (%) |
|
|
Arduous physical work |
55.5 |
31.3 |
|
Arduous mental work |
80.2 |
61.9 |
|
Work often too tiring |
46.8 |
24.8 |
|
Under-staffing |
74.8 |
43.8 |
|
Insufficient time for breaks |
78.4 |
56.6 |
|
Interference of work with private life |
52.8 |
31.0 |
|
Dissatisfaction with schedules |
36.9 |
2.7 |
|
Frequent lack of sleep |
34.9 |
19.5 |
|
Frequent fatigue on rising |
31.3 |
17.3 |
Source: Van Deursen et al. 1993.
Sleep disturbances
On workdays, night-shift nurses sleep an average of two hours less than other nurses (Escribà Agüir et al. 1992; Estryn-Béhar et al. 1978; Estryn-Béhar et al. 1990; Nyman and Knutsson 1995). According to several studies, their quality of sleep is also poor (Schroër et al. 1993; Lee 1992; Gold et al. 1992; Estryn-Béhar and Fonchain 1986).
In their interview study of 635 Massachusetts nurses, Gold et al. (1992) found that 92.2% of nurses working alternating morning and afternoon shifts were able to maintain a nocturnal “anchor” sleep of four hours at the same schedule throughout the month, compared to only 6.3% of night-shift nurses and none of the nurses working alternating day and night shifts. The age- and seniority-adjusted odds ratio for “poor sleep” was 1.8 for night-shift nurses and 2.8 for rotating-shift nurses with night work, compared to morning- and afternoon-shift nurses. The odds ratio for taking sleep medication was 2.0 for night- and rotating-shift nurses, compared to morning- and afternoon-shift nurses.
Affective Problems and Fatigue
The prevalence of stress-related symptoms and reports of having stopped enjoying their work was higher among Finnish nurses working rotating shifts than among other nurses (Kandolin 1993). Estryn-Béhar et al. (1990) showed that night-shift nurses’ scores on the General Health Questionnaire used to evaluate mental health, compared to day-shift nurses (odds ratio of 1.6) showed poorer general health.
In another study, Estryn-Béhar et al. (1989b), interviewed a representative sample of one-quarter of night-shift employees (1,496 individuals) in 39 Paris-area hospitals. Differences appear according to sex and qualification (“qualified”=head nurses and nurses; “unqualified”=nurses’ aides and orderlies). Excessive fatigue was reported by 40% of qualified women, 37% of unqualified women, 29% of qualified men and 20% of unqualified men. Fatigue on rising was reported by 42% of qualified women, 35% of unqualified women, 28% of qualified men and 24% of unqualified men. Frequent irritability was reported by one-third of night-shift workers and by a significantly greater proportion of women. Women with no children were twice as likely to report excessive fatigue, fatigue on rising and frequent irritability than were comparable men. The increase compared to single men with no children was even more marked for women with one or two children, and greater still (a four-fold increase) for women with at least three children.
Fatigue on rising was reported by 58% of night-shift hospital workers and 42% of day-shift workers in a Swedish study using a stratified sample of 310 hospital workers (Nyman and Knutsson 1995). Intense fatigue at work was reported by 15% of day-shift workers and 30% of night-shift workers. Almost one-quarter of night-shift workers reported falling asleep at work. Memory problems were reported by 20% of night-shift workers and 9% of day-shift workers.
In Japan, the health and safety association publishes the results of medical examinations of all the country’s salaried employees. This report includes the results of 600,000 employees in the health and hygiene sector. Nurses generally work rotating shifts. Complaints concerning fatigue are highest in night-shift nurses, followed in order by evening- and morning-shift nurses (Makino 1995). Symptoms reported by night-shift nurses include sleepiness, sadness and difficulty concentrating, with numerous complaints about accumulated fatigue and disturbed social life (Akinori and Hiroshi 1985).
Sleep and Affective Disorders among Physicians
The effect of work content and duration on young physicians’ private lives, and the attendant risk of depression, has been noted. Valko and Clayton (1975) found that 30% of young residents suffered a bout of depression lasting an average of five months during their first year of residency. Of the 53 residents studied, four had suicidal thoughts and three made concrete suicide plans. Similar rates of depression have been reported by Reuben (1985) and Clark et al. (1984).
In a questionnaire study, Friedman, Kornfeld and Bigger (1971) showed that interns suffering from sleep deprivation reported more sadness, selfishness and modification of their social life than did more-rested interns. During interviews following the tests, interns suffering from sleep deprivation reported symptoms such as difficulty reasoning, depression, irritability, depersonalization, inappropriate reactions and short-term memory deficits.
In a one-year longitudinal study, Ford and Wentz (1984) evaluated 27 interns four times during their internship. During this period, four interns suffered at least one major bout of depression meeting standard criteria and 11 others reported clinical depression. Anger, fatigue and mood swings increased throughout the year and were inversely correlated with the amount of sleep the preceding week.
A literature review has identified six studies in which interns having spent one sleepless night exhibited deteriorations of mood, motivation and reasoning ability and increased fatigue and anxiety (Samkoff and Jacques 1991).
Devienne et al. (1995) interviewed a stratified sample of 220 general practitioners in the Paris area. Of these, 70 were on call at night. Most of the on-call physicians reported having had their sleep disturbed while on call and finding it particularly difficult to get back to sleep after having been awakened (men: 65%; women: 88%). Waking up in the middle of the night for reasons unrelated to service calls was reported by 22% of men and 44% of women. Having or almost having a car accident due to sleepiness related to being on call was reported by 15% of men and 19% of women. This risk was greater among physicians who were on call more than four times per month (30%) than in those on call three or four times per month (22%) or one to three times per month (10%). The day after being on call, 69% of women and 46% of men reported having difficulty concentrating and feeling less effective, while 37% of men and 31% of women reported experiencing mood swings. Accumulated sleep deficits were not recovered the day following on-call work.
Family and Social Life
A survey of 848 night-shift nurses found that over the previous month one-quarter had not gone out and had entertained no guests, and half had participated in such activities only once (Gadbois 1981). One-third reported refusing an invitation because of fatigue, and two-thirds reported going out only once, with this proportion rising to 80% among mothers.
Kurumatani et al. (1994) reviewed the time sheets of 239 Japanese nurses working rotating shifts over a total of 1,016 days and found that nurses with young children slept less and spent less time on leisure activities than did nurses without young children.
Estryn-Béhar et al. (1989b) observed that women were significantly less likely than men to spend at least one hour per week participating in team or individual sports (48% of qualified women, 29% of unqualified women, 65% of qualified men and 61% of unqualified men). Women were also less likely to frequently (at least four times per month) attend shows (13% of qualified women, 6% of unqualified women, 20% of qualified men and 13% of unqualified men). On the other hand, similar proportions of women and men practised home-based activities such as watching television and reading. Multivariate analysis showed that men with no children were twice as likely to spend at least one hour per week on athletic activities than were comparable women. This gap increases with the number of children. Child care, and not gender, influences reading habits. A significant proportion of the subjects in this study were single parents. This was very rare among qualified men (1%), less rare among unqualified men (4.5%), common in qualified women (9%) and extremely frequent in unqualified women (24.5%).
In Escribà Agüir’s (1992) study of Spanish hospital workers, incompatibility of rotating shifts with social and family life was the leading source of dissatisfaction. In addition, night-shift work (either permanent or rotating) disturbed the synchronization of their schedules with those of their spouses.
Lack of free time interferes severely with the private life of interns and residents. Landau et al. (1986) found that 40% of residents reported major conjugal problems. Of these residents, 72% attributed the problems to their work. McCall (1988) noted that residents have little time to spend on their personal relationships; this problem is particularly serious for women nearing the end of their low-risk-pregnancy years.
Irregular Shift Work and Pregnancy
Axelsson, Rylander and Molin (1989) distributed a questionnaire to 807 women employed at the hospital in Mölna, Sweden. The birth weights of children born to non-smoking women working irregular shifts were significantly lower than that of children born to non-smoking women who only worked day shifts. The difference was greatest for infants of at least grade 2 (3,489 g versus 3,793 g). Similar differences were also found for infants of at least grade 2 born to women working afternoon shifts (3,073 g) and shifts alternating every 24 hours (3,481 g).
Vigilance and Quality of Work among Night-Shift Nurses
Englade, Badet and Becque (1994) performed Holter EEGs on two groups of nine nurses. It showed that the group not allowed to sleep had attention deficits characterized by sleepiness, and in some cases even sleep of which they were unaware. An experimental group practised polyphasic sleep in an attempt to recover a little sleep during work hours, while the control group was not allowed any sleep recovery.
These results are similar to those reported by a survey of 760 California nurses (Lee 1992), in which 4.0% of night-shift nurses and 4.3% of nurses working rotating shifts reported suffering frequent attention deficits; no nurses from the other shifts mentioned lack of vigilance as a problem. Occasional attention deficits were reported by 48.9% of night-shift nurses, 39.2% of rotating-shift nurses, 18.5% of day-shift nurses and 17.5% of evening-shift nurses. Struggling to stay awake while dispensing care during the month preceding the survey was reported by 19.3% of night-shift and rotating-shift nurses, compared to 3.8% of day- and evening-shift nurses. Similarly, 44% of nurses reported having had to struggle to stay awake while driving during the preceding month, compared to 19% of day-shift nurses and 25% of evening-shift nurses.
Smith et al. (1979) studied 1,228 nurses in 12 American hospitals. The incidence of occupational accidents was 23.3 for nurses working rotating shifts, 18.0 for night-shift nurses, 16.8 for day-shift nurses and 15.7 for afternoon-shift nurses.
In an attempt to better characterize problems related to attention deficits among night-shift nurses, Blanchard et al. (1992) observed activity and incidents throughout a series of night shifts. Six wards, ranging from intensive care to chronic care, were studied. In each ward, one continuous observation of a nurse was performed on the second night (of night work) and two observations on the third or fourth nights (depending on the wards’ schedule). Incidents were not associated with serious outcomes. On the second night, the number of incidents rose from 8 in the first half of the night to 18 in the second half. On the third or fourth night, the increase was from 13 to 33 in one case and from 11 to 35 in another. The authors emphasized the role of sleep breaks in limiting risks.
Gold et al. (1992) collected information from 635 Massachusetts nurses on the frequency and consequences of attention deficits. Experiencing at least one episode of sleepiness at work per week was reported by 35.5% of rotating-shift nurses with night work, 32.4% of night-shift nurses and 20.7% of morning-shift and afternoon-shift nurses working exceptionally at night. Less than 3% of nurses working the morning and afternoon shifts reported such incidents.
The odds ratio for sleepiness while driving to and from work was 3.9 for rotating-shift nurses with night work and 3.6 for night-shift nurses, compared to morning- and afternoon-shift nurses. The odds ratio for total accidents and errors over the past year (car accidents driving to and from work, errors in medication or work procedures, occupational accidents related to sleepiness) was almost 2.00 for rotating-shift nurses with night work compared to morning- and afternoon-shift nurses.
Effect of Fatigue and Sleepiness on the Performance of Physicians
Several studies have shown that the fatigue and sleeplessness induced by night-shift and on-call work leads to deteriorations of physician performance.
Wilkinson, Tyler and Varey (1975) conducted a postal questionnaire survey of 6,500 British hospital physicians. Of the 2,452 who responded, 37% reported suffering a degradation of their effectiveness due to excessively long work hours. In response to open-ended questions, 141 residents reported committing errors due to overwork and lack of sleep. In a study performed in Ontario, Canada, 70% of 1,806 hospital physicians reported often worrying about the effect of the quantity of their work had on its quality (Lewittes and Marshall 1989). More specifically, 6% of the sample—and 10% of interns—reported often worrying about fatigue affecting the quality of care they dispensed.
Given the difficulty in performing real-time evaluations of clinical performance, several studies on the effects of sleep deprivation on physicians have relied upon neuropsychological tests.
In the majority of studies reviewed by Samkoff and Jacques (1991), residents deprived of sleep for one night exhibited little deterioration in their performance of rapid tests of manual dexterity, reaction time and memory. Fourteen of these studies used extensive test batteries. According to five tests, the effect on performance was ambiguous; according to six, a performance deficit was observed; but according to eight other tests, no deficit was observed.
Rubin et al. (1991) tested 63 medical-ward residents before and after an on-call period of 36 hours and a subsequent full day of work, using a battery of self-administered computerized behavioural tests. Physicians tested after being on call exhibited significant performance deficits in tests of visual attention, coding speed and accuracy and short-term memory. The duration of sleep enjoyed by the residents while on call was as follows: two hours at most in 27 subjects, four hours at most in 29 subjects, six hours at most in four subjects and seven hours in three subjects. Lurie et al. (1989) reported similarly brief sleep durations.
Virtually no difference has been observed in the performance of actual or simulated short-duration clinical tasks—including filling out a laboratory requisition (Poulton et al. 1978; Reznick and Folse 1987), simulated suturing (Reznick and Folse 1987), endotracheal intubation (Storer et al. 1989) and venous and arterial catheterization (Storer et al. 1989)—by sleep-deprived and control groups. The only difference observed was a slight lengthening of the time required by sleep-deprived residents to perform arterial catheterization.
On the other hand, several studies have demonstrated significant differences for tasks requiring continuous vigilance or intense concentration. For example, sleep-deprived interns committed twice as many errors when reading 20-minute ECGs as did rested interns (Friedman et al. 1971). Two studies, one relying on 50-minute VDU-based simulations (Beatty, Ahern and Katz 1977), the other on 30-minute video simulations (Denisco, Drummond and Gravenstein 1987), have reported poorer performance by anaesthesiologists deprived of sleep for one night. Another study has reported significantly poorer performance by sleep-deprived residents on a four-hour test exam (Jacques, Lynch and Samkoff 1990). Goldman, McDonough and Rosemond (1972) used closed-circuit filming to study 33 surgical procedures. Surgeons with less than two hours of sleep were reported to perform “worse” than more-rested surgeons. The duration of surgical inefficiency or indecision (i.e., of poorly planned manoeuvres) was over 30% of the total duration of the operation.
Bertram (1988) examined the charts of emergency admissions by second-year residents over a one-month period. For a given diagnosis, less information on medical histories and the results of clinical examinations was gathered as the number of hours worked and patients seen increased.
Smith-Coggins et al. (1994) analysed the EEG, mood, cognitive performance and motor performance of six emergency-ward physicians over two 24-hour periods, one with diurnal work and nocturnal sleep, the other with nocturnal work and diurnal sleep.
Physicians working at night slept significantly less (328.5 versus 496.6 minutes) and performed significantly less well. This poorer motor performance was reflected in the increased time required to perform a simulated intubation (42.2 versus 31.56 seconds) and an increased number of protocol errors.
Their cognitive performance was evaluated at five test periods throughout their shift. For each test, physicians were required to review four charts drawn from a pool of 40, rank them and list the initial procedures, the treatments and the appropriate laboratory tests. Performance deteriorated as the shift progressed for both night-shift and day-shift physicians. Night-shift physicians were less successful at providing correct responses than day-shift physicians.
Physicians working during the day rated themselves as less sleepy, more satisfied and more lucid than did night-shift physicians.
Recommendations in English-speaking countries concerning the work schedules of physicians-in-training have tended to take these results into account and now call for work-weeks of at most 70 hours and the provision of recovery periods following on-call work. In the US, following the death of a patient attributed to errors by an overworked, poorly supervised resident physician which received much media attention, New York State enacted legislation limiting work hours for hospital staff physicians and defining the role of attending physicians in supervising their activities.
Content of Night Work in Hospitals
Night work has long been undervalued. In France, nurses used to be seen as guardians, a term rooted in a vision of nurses’ work as the mere monitoring of sleeping patients, with no delivery of care. The inaccuracy of this vision became increasingly obvious as the length of hospitalization decreased and patients’ uncertainty about their hospitalization increased. Hospital stays require frequent technical interventions during the night, precisely when the nurse:patient ratio is lowest.
The importance of the amount of time spent by nurses in patients’ rooms is demonstrated by the results of a study based on continuous observation of the ergonomics of nurses’ work in each of three shifts in ten wards (Estryn-Béhar and Bonnet 1992). The time spent in rooms accounted for an average of 27% of the day and night shifts and 30% of the afternoon shift. In four of the ten wards, nurses spent more time in the rooms during the night than during the day. Blood samples were of course taken less frequently during the night, but other technical interventions such as monitoring vital signs and medication, and administering, adjusting and monitoring intravenous drips and transfusions were more frequent during the night in six of seven wards where detailed analysis was performed. The total number of technical and non-technical direct-care interventions was higher during the night in six of seven wards.
Nurses’ work postures varied from shift to shift. The percentage of time spent seated (preparation, writing, consultations, time spent with patients, breaks) was higher at night in seven of ten wards, and exceeded 40% of shift time in six wards. However, the time spent in painful postures (bent over, crouched, arms extended, carrying loads) exceeded 10% of shift time in all wards and 20% of shift time in six wards at night; in five wards the percentage of time spent in painful positions was higher at night. In fact, night-shift nurses also make beds and perform tasks related to hygiene, comfort and voiding, tasks which are all normally performed by nurses’ aides during the day.
Night-shift nurses may be obliged to change location very frequently. Night-shift nurses in all the wards changed location over 100 times per shift; in six wards, the number of changes of location was higher at night. However, because rounds were scheduled at 00:00, 02:00, 04:00 and 06:00, nurses did not travel greater distances, except in juvenile intensive-care wards. Nonetheless, nurses walked over six kilometres in three of the seven wards where podometry was performed.
Conversations with patients were frequent at night, exceeding 30 per shift in all wards; in five wards these conversations were more frequent at night. Conversations with physicians were much rarer and almost always brief.
Leslie et al. (1990) conducted continuous observation of 12 of 16 interns in the medical ward of a 340-bed Edinburgh (Scotland) hospital over 15 consecutive winter days. Each ward cared for approximately 60 patients. In all, 22 day shifts (08:00 to 18:00) and 18 on-call shifts (18:00 to 08:00), equivalent to 472 hours of work, were observed. The nominal duration of the interns’ work week was 83 to 101 hours, depending on whether or not they were on call during the weekends. However, in addition to the official work schedule, each intern also spent an average of 7.3 hours each week on miscellaneous hospital activities. Information on the time spent performing each of 17 activities, on a minute-by-minute basis, was collected by trained observers assigned to each intern.
The longest continuous work period observed was 58 hours (08:00 Saturday to 06:00 Monday) and the longest work period was 60.5 hours. Calculations showed that a one-week sickness leave of one intern would require the other two interns in the ward to increase their workload by 20 hours.
In practice, in wards admitting patients during on-call shifts, interns working consecutive day, on-call and night shifts worked all but 4.6 of the 34 elapsed hours. These 4.6 hours were devoted to meals and rest, but interns remained on call and available during this time. In wards that did not admit new patients during on-call shifts, interns’ workload abated only after midnight.
Due to the on-call schedules in other wards, interns spent approximately 25 minutes outside their home ward each shift. On average, they walked 3 kilometres and spent 85 minutes (32 to 171 minutes) in other wards each night shift.
Time spent filling out requests for examinations and charts, in addition, is often performed outside of their normal work hours. Non-systematic observation of this additional work over several days revealed that it accounts for approximately 40 minutes of additional work at the end of each shift (18:00).
During the day, 51 to 71% of interns’ time was spent on patient-oriented duties, compared to 20 to 50% at night. Another study, conducted in the United States, reported that 15 to 26% of work time was spent on patient-oriented duties (Lurie et al. 1989).
The study concluded that more interns were needed and that interns should no longer be required to attend other wards while on call. Three additional interns were hired. This reduced interns’ work week to an average of 72 hours, with no work, excepting on-call shifts, after 18:00. Interns also obtained a free half-day following an on-call shift and preceding a weekend when they were to be on call. Two secretaries were hired on a trial basis by two wards. Working 10 hours per week, the secretaries were able to fill out 700 to 750 documents per ward. In the opinion of both senior physicians and nurses, this resulted in more efficient rounds, since all the information had been entered correctly.