Many research projects were also under way with this transplantation program, and it became apparent that many data points were missed during the early morning hours. However, our group was still able to produce many peer-reviewed publications and gain some recognition. At one international meeting, I was asked to give a presentation on “how to do 2 AM research.” This presentation led to a review of the effects of fatigue, one of the major causes of dysfunctional behavior by physicians in the operating room environment and the cause of many medical errors. Recently the Institute of Medicine extrapolated the incidence of adverse events in hospitalized patients from 2 large studies and concluded that at least 44,000—and maybe as many as 98,000— Americans die each year as a result of medical errors (1).

Despite an extensive volume of research data in the area of sleep deprivation and performance, information relating directly to physicians is inconsistent. However, when these data are combined with those obtained from aviation and industry in general, valid conclusions can be made.

The human being is a complicated physiological machine that is prone to err. The incidence of human error is increased by fatigue, sleep deprivation, and stress (2). As performance decreases, errors of omission occur with increasing frequency. This may be demonstrated by minor errors, or “slips.” These “slips” may be exhibited as recording errors or recent memory loss, and they rarely lead to a major event. However, as performance further deteriorates errors of commission occur: these are “mistakes” where the planning process itself is flawed. These mistakes may lead to technical or judgmental errors, such as selecting a wrong or inappropriate technique that results in an adverse outcome (for example, administering succinyl choline to a paraplegic). These mistakes, of course, may have devastating results.

Fatigue is the inability or unwillingness to continue effective performance and is caused by excessive workload, stress, sleep loss, and circadian disruption (3). Fatigue and sleep deprivation are different entities. Fatigue is more responsible for performance changes than are circadian rhythm disruptions, and the degree of fatigue can be affected by environmental conditions (4). Cognitive function deteriorates more than physical performance, and fatigued individuals demonstrate impaired learning and thought processes, memory defaults, and interpersonal dysfunction.

Sleep deprivation may be seen if <5 hours of sleep occur in a 24-hour period. It can cause measurable deficits in cognitive function; however, motivation can compensate for decreased performance by increasing an individual's effort and arousal (3). The overall effects of sleep deprivation are decreased efficiency, instability, recent memory deficit, difficulty in thinking, depersonalization, and inappropriate humor. Continuous working adversely affects cognitive function and mood to the detriment of the person.

Mental performance increases between 8 AM and 2 PM and then gradually declines, reaching a low point between 2 AM and 5 AM. This performance is limited by the body's circadian rhythm, which governs body temperature, hormonal processes, and general performance. Body temperature is at its lowest between 2 AM and 5 AM and is associated with sleep, decreased performance, and worsened mood. Alertness closely follows the body temperature rhythm (5).

There is tremendous individual variation in the physiological responses to sleep loss. In a controlled trial on Boeing 747 aircrews, one group on long-haul flights took 40-minute naps and the other group took no rest. In the no-rest group, decrements in reaction time and psychomotor vigilance were noted, as well as twice as many micro-events (brief sleep events, or “head nods”) compared with the rest group. It was concluded that brief naps act as an acute safety valve (3). In anesthesiologists, it has been demonstrated that 2 to 3 hours of sleep in a 24-hour period is better than no sleep at all. However, after a person is awake for 24 hours, a 30-minute to 2-hour nap is followed by a period of sleep inertia. This is a period of impaired vigilance. A nap of at least 4 hours is required to avoid this phenomenon (6).

Toward the end of a prolonged work period, the phenomenon of “end spurt” may be seen. This occurs when a task is considered 90% complete and causes increased vigilance, but “let down” will occur if the procedure is prolonged beyond the expected finish time (7). Industrial psychologists recommend that 4 hours be a maximum duration of work without a break.

Sleep loss affects new-skill performance more than automated skills. Under conditions of sleep deprivation, cognitively exciting tasks appear to be least affected. Performance on cognitive tasks is also enhanced when vigorous exercise is taken during rest periods, but this effect diminishes as sleep deprivation increases.

A stressful environment impairs vigilance and adversely affects complex monitoring tasks. Complex performance—as measured by the Multiple Task Performance Battery Test, which includes monitoring of warning lights and meters, mental arithmetic, problem solving, target identification, and tracking— declines with sleep deprivation. Short bursts of exercise lead to better performance scores (8). The number of targets monitored has an effect on performance; decrements occur when 16 targets are monitored compared with 8 (9).

Vigilance requires a state of maximal physiologic and psychological readiness to react. It requires a cognitive skill level that can rapidly and reliably assess a large volume of information. Vigilance, alertness, complex memory, decision making, attention, selection of information, and conscious effort all are vulnerable to compromise with sleep deprivation.

Complex monitoring tasks, stress, and fatigue are all affected by environmental conditions. As task complexity increases in anesthesia management, record keeping deteriorates, and carelessness and lapses in vigilance occur. The performance of experienced airline pilots is less affected by increasing workload than that of new graduates. Presumably, this would also apply to anesthesiologists.

Noise may cause distraction during critical periods. It negatively affects information processing and short-term memory, interferes with effective verbal communication, and masks taskrelated cues (10). However, it may also act as an “activator,” and appropriate background music may prevent performance decrements over time (11). Ambient temperature also may affect performance: overheated, dry rooms cause performance deterioration, and extremely cold temperatures cause distraction and reduce manual dexterity. A temperature range of 17°C to 18°C (62.6°F to 64.4°F) with a moderate humidity (50%) is recommended for best work (12).

Physical condition and personal habits can have a significant influence on vigilance and monitoring performance. Preventable human errors are a major contributor to poor outcomes. If one is feeling fatigued then it is likely that one is experiencing a vigilance decrement (13).

Small doses of caffeine can have a positive effect on performance but may increase tremor and anxiety. Alcohol in small amounts (20 to 35 mg/dL) significantly impairs performance, as does the effects of a hangover.

In a landmark case in New York, the grand jury found that long working hours of residents had contributed to the death of patient Libby Zion. The state legislature reacted by prohibiting physicians from working >24 consecutive hours without an 8- hour rest period and from working >80 total hours per week (14). The European community has placed restrictions on the number of hours a physician can work.

In an effort to provide policymakers and the community with an easily grasped index of the relative impairment associated with fatigue, Dawson and Reid (15) expressed fatigue-related impairment as a blood-alcohol equivalent. Forty subjects took part in 2 counterbalanced studies comparing the effect of fatigue with that of alcohol. They found that after 24 hours of sustained wakefulness, cognitive psychomotor performance decreased to a level equivalent to that at a blood alcohol concentration of 0.10%.

The following guidelines can be made from the data available. If possible, avoid any surgery between 2 AM and 5 AM. If this is not possible, automate as much of the task as possible and design a clear protocol so that mistakes are reduced and errors are limited to those of omission only. After 24 hours of continuous work, it would seem prudent to have a mandatory rest break of at least 5 hours. The medical community should address this issue before the legal community decides the guidelines for us.

Teach us to live that we may dread So few hours spent in bed. Let us sleep and we may save Our patients from an early grave. —Asher (16)