Volume 8, No. 1 • Spring 1993

“Vigilance” Discussed by ASA Panel

David W. Edsall, M.D.

The topic of ‘vigilance’ was directly addressed for the first time at the 1992 ASA Annual Meeting by a panel of experts on vigilance from the anesthesia and psychology academic communities. Unfortunately, fewer than 50 people attended the informative and interesting presentations a result not unexpected due to the almost total lack of discussion of the topic in anesthesia textbooks or journals. In contrast, vigilance is one aspect of the science of human factors which has its own societies and scientific journals. Thousands of peer review articles and dozens of complete textbooks or textbook chapters address the issue of vigilance. (1) The anesthesia community is becoming increasingly interested in human factors, as evident from recent review articles (2) topics of society meetings, (3) and task forces developed by the Anesthesia Patient Safety Foundation.

The origins of this panel came out of discussions about automation of the anesthesia workstation and whether this would have positive or detrimental effects on vigilance by anesthesia providers. The panel composed four speakers who addressed various aspects of the vigilance issue.

“Sustained Tension”

Joel Warm, Ph.D., Professor, Department of Psychology, University of Cincinnati, gave an introductory tutorial for those not familiar with the scientific concept of vigilance. Vigilance is sustained tension toward the occurrence of a signal to which one is expected to respond. A signal is an event to which one is to respond. For example, repetitive vital signs would be events. A vital sip in an unsafe range would be a signal. The ability to sustain attention toward the signal decreases with time and is called the vigilance decrement. This decrement has completed its deterioration in about 30 minutes time with half the deterioration occurring in 15 minutes.

The decrement in vigilance performance is measured by the detection and response to signals and is represented by the following formula: P = f (M,S,U,B,C) where P = Performance; M = Sensory Modality (i.e., visual vs. auditory signals); S = Signal Salience (i.e., volume of auditory signal); U = Stimulus Uncertainty; (i.e., where or when a signal will appear during the watch); B = Background Event Rate (i.e., frequency of vital signs, background noise levels, etc. “The more one has to look for signals, the less likely one is to detect them;’); C = Signal Complexity (the task can be either too complex or too simple, either of which will result in an increased vigilance decrement).

The classic vigilance paradigm occurs when: (1) the task is prolonged and continuous, (2) the signal is infrequent and a periodic; (3) the signal is easily recognizable; and (4) the observer’s response cannot affect the future signal rate.

Central to the study of vigilance is the Signal Detection Theory where performance is measured in terms of detection of a signal (a) and the response criteria (p) used by the observer in responding to the signal. It appears that most of the decrement in vigilance over time does not come from a decrease in perception (a) but rather from a shift to a more conservative response criteria. “The signal is heard but not listened to.” This lack of response may be due to many issues such as boredom, distraction, education, etc. Signal Detection Theory suggests that the human being is quite poor at detecting signals but can be quite good at decision-making once appropriate training has occurred. Therefore, training can affect vigilance. Automated devices, on the other hand, are excellent at detecting but quite poor at decision-making, especially when dealing with unique or complex signals.

Negative Factors

Vigilance is also impacted by stress, fatigue, high or low work load, emotional depression, noise, extreme temperatures, and many other stress inducing task environments which have been shown to increase the vigilance decrement.

The human short-term memory capacity is quite small five to ten items maximum lasting 15 to 20 seconds. Otherwise, the information needs to be converted to long-term memory and this generally requires mental effort. Automation can help reduce the stress of overloaded short-term memory or excessive short-term to long-term memory consolidation. Trend displays, if appropriately used, can reduce short-term memory workload and increase vigilance.

Dr. Raja Parasuraman, Ph.D., Professor of Psychology at the Catholic University of America, presented more information on the effect of automation in vigilance. Examples from the transportation industry demonstrated where automation has resulted in increased efficiency and decreased work load resulting in better performance. The paradox was that as automation becomes more effective human vigilance suffered because the Rehhood of a signal decreases. Thus, one must decide whether to use an automated system that fails very infrequently resulting in boredom, or to use a manual system where the likelihood of failure is higher but the operator is more vigilant and may be able to correct the condition.

Negative issues resulting from automation are: (1) decreased manual skills for the operator; (2) decreased human-to-human communication if the work situation requires a team approach; and (3) increased complacency as the automated system continues to enhance its performance. As reliability of an automated system increases, complacency increases. The solution may be to incorporate an unreliability factor in a controlled and safe manner. In other words, keep the human being in the loop with a certain frequency of manual and problem solving tasks. These tasks could be real or simulated events.

Some emphasis is being placed on selection of individuals who have a personality profile more suited to vigilance tasks. Clear and consistent differences in vigilance can be demonstrated between various human groups, such as male and female or introvert and extrovert personalities. However, the ability to predict which individual will perform better at the vigilance task has not yet been developed.

Recording devices, especially automatic recording devices, are an essential tool in the analysis of both signal detection and response performance. Without such tools, analysis of failed responses is extremely difficult. Recording devices are especially useful in cataloging a chronology of the events and signals providing information about event rate, signal rate, and response.

Anesthesia Targeted

Matthew Weinger, M.D., Assistant Professor, Department of Anesthesiology, University of California, San Diego, connected the previous presentations more specifically to the field of anesthesiology. Data from aviation and vehicle simulator studies suggest that drug use, including small quantities of alcohol taken within 12 hours of task performance can result in decreased performance. Extreme temperature and fatigue have also been asserted to affect performance in anesthesiology. A description of primary and secondary tasks and their use in the anesthesia environment for studying vigilance and other human factor issues was related to automated record keeping. Several studies now suggest that automated records can decrease the time spent and/or increase the amount of data gathered by the recording task.’ One preliminary study suggests that vigilance seems to be increased during induction when automated records are being used.” Techniques are being developed to study vigilance in the anesthesia environment with the eventual goal of enhancing the performance during high work load (stressful) conditions, as well as low work load (boring) conditions. Another preliminary study seems to indicate that graphical display of vital sign information leads to a. faster signal detection than does the numerical display of the same information.

David Gaba, M.D., Assistant Professor, Department of Anesthesiology, Stanford University, emphasized that vigilance is a necessary but not sufficient condition for appropriate clinical decision-making. It was suggested that an anesthetist may be awake and alert but because of signal complexity, high work load, distractions, or inappropriate signal display, the anesthetist may not address the correct problem.”,” Similarly, a problem may be detected but the response time may be slowed because of difficulty in deciding what to do about the problem. The use of an anesthesia simulator, as in the aviation industry, has been invaluable in identifying some of these issues. Preliminary simulator studies dramatically demonstrated the need for studying anesthesia human factors and vigilance problems in as realistic a setting as possible. Even the most realistic simulators have problems in this regard because the level of expectancy for problems is raised significantly when compared with the real world environment. Beyond vigilance issues of leadership, education, action planning, resource allocation, and communication are other conditions required of the anesthetist in order to provide good patient care.

Both the panel discussion and all four presenters addressed the issue of detection versus response. This issue becomes confused by the third item of the vigilance paradigm in that the signal must be easily detectable. Consider the anesthetist who does not recognize an ST segment change. Is this a vigilance problem? Certainly it could initially be an educational problem or an equipment problem. Many if not most anesthetists state that the recognition of ST segment changes or arrhythmia is not an ‘easily recognizable signal’ as the vigilance paradigm requires. However, for many CCRNs it can be demonstrated that ” task is automatic and, therefore, becomes a vigilance task. Likewise, it might be said that the response to the signs of malignant hyperthermia is not a vigilance issue. Possibly, with appropriate training this response could also become automatic for the anesthetist. Therefore, the study of vigilance includes not only detection but also the response time. In fact, the key component in most vigilance studies is measurement of response time as opposed to the detection of a signal. Training to improve both the detection of signals and increase the response time to a critical event is what this author believes is meant by our profession’s motto: ‘VIGILANCE.’

Dr. Edsall is Chairman, Department of Anesthesiology, Burbank Hospital, Fitchburg, MA, and a pioneer in implementation of automated anesthesia records and information management systems.

References

1. Human Factors. Journal of the Human Factors Society of America, Santa Monica, CA.

2. Weinger, MB, Englund CE: Ergonomics and Human Factors affecting anesthesia vigilance and monitoring performance in the OR environment. Anesthesiolog 1990; 73:995-1002.

3. “Human Performance and Anesthesia Technology.’ The Society for Technology in Anesthesia 3rd Annual Meeting, New Orleans, La. February 17-19,1993.

4. Herndon OW, Weinger MB, Zornow MH, and Gaba DM: The use of automated record keeping requires less time in complicated anesthesia procedures. Anesth Analg 1992; 76:S140.

5. Edsall DW: Documentation time requirement: Computerized information management Systems versus handwritten records. Anesthesiology 1990; 73(3A):A439. (Full paper in press)

6. Kennedy PJ, Feingold A, Wiener, EL, et al: Analysis of tasks and human factors in anesthesia for coronary artery bypass. Anesth Analg 1976; 55:374-377.

7. Zollinger RM, Kreul JF, and Schneider AJL: Manmade versus computer-generated anesthesia records. Surg Res 1977; 22:419-424.

8. Lerou JGC, Dirksen R, vanDaele M., et al: Automated charting and physiological variables in anesthesia: a quantitative comparison of automated versus handwritten anesthesia records. J Clin Monit 1988; 4 (1):37-47.

9. Dirksen R, et al: The clinical use of the Ohmeda automated anesthesia record keeper integrated in the Modulus E anesthesia system. Int j Clin Mon Comp 1987; 4:135-139.

10. Weinger MB, Herndon OW, Paulus MP, Gaba DM, and Dallen LD: Objective analysis of performance workload in the anesthesia cockpit. J Clin Monit 1992; 8:186-187 (Abstract).

11. Gaba DM: Dynamic decision making in anesthesiology realistic simulation for anesthesia crisis management training. Advanced Models of Cognition for Medical Training and Practice, edited by Evans DA and Patel VL. Springer-Verlag, GmbH.

12. Howard SK, Gaba DM, Fish KJ, Yang GS, and Sarnquist FH: Anesthesia crisis resource management training. Aviation, Space, and Environmental Medicine.