Free
Special Articles  |   July 1997
The Effect of Electronic Record Keeping and Transesophageal Echocardiography on Task Distribution, Workload, and Vigilance During Cardiac Anesthesia 
Author Notes
  • (Weinger) Associate Professor of Anesthesiology, University of California San Diego School of Medicine; Staff Physician, Veterans Administration San Diego Health Care System.
  • (Herndon) Research Assistant, Department of Anesthesia, University of California San Diego School of Medicine, Current position: Resident in Medicine, University of California Los Angeles School of Medicine.
  • (Gaba) Associate Professor of Anesthesia, Stanford University School of Medicine, Staff Physician, Veterans Administration Palo Alto Health Care System.
  • Received from the Department of Anesthesiology, the University of California San Diego School of Medicine, the Veterans Administration Medical Center, San Diego, California, and the Department of Anesthesia, Stanford University, Stanford, California. Submitted for publication November 21, 1996. Accepted for publication February 22, 1997. Supported in part by grants from the Anesthesia Patient Safety Foundation, the University Anesthesia Foundation, the Crummer Foundation, and the University of California, San Diego Academic Senate. Dr. Herndon was supported partially by the Stanford University School of Medicine Medical Student Scholars Program. Presented in part at the Sixty-Seventh Congress of the International Anesthesia Research Society, San Diego, California, March 21, 1993.
  • Address reprint requests to Dr. Weinger: Veterans Affairs Medical Center (125), 3350 La Jolla Village Drive, San Diego, California 92161–5085. Address electronic mail to: mweinger@ucsd.edu.
Article Information
Special Articles
Special Articles   |   July 1997
The Effect of Electronic Record Keeping and Transesophageal Echocardiography on Task Distribution, Workload, and Vigilance During Cardiac Anesthesia 
Anesthesiology 7 1997, Vol.87, 144-155. doi:
Anesthesiology 7 1997, Vol.87, 144-155. doi:
Anesthesia providers spend 10–15% of their total case time keeping the clinical record. [1–3 ] A recent innovation is the use of electronic anesthesia record keeping (EARK). The greatest value of innovations like EARK may be their ability to improve the performance of medical personnel, e.g., by reducing workload, enhancing vigilance, or providing better data. [4 ] If the use of the electronic record allows the anesthesiologist to spend less time on record keeping, more time may be available for direct patient care activities. On the other hand, critics have suggested that anesthetic vigilance could be impaired when the anesthesiologist is removed from the record keeping process. [5,6 ] This assertion has not been borne out by recent studies. [7,8 ]
The administration of anesthesia is a cognitively demanding task involving vigilance, monitoring, and decision-making. Moreover, the anesthesiologist often works in an environment that is crowded, noisy, or poorly designed. [10 ] Because these factors may contribute to adverse events, clinicians and researchers have sought to identify ways to improve anesthesia administration and to improve the design of the anesthesia workspace. Human factors research techniques, such as task analysis and workload assessment, may provide useful objective data on the structure and characteristics of the anesthesiologist's job and the impact of design innovations on task performance.
During the past 20 yr, time-motion studies have investigated the administration of anesthesia in the operating room. [1,2,7,11,12 ] Weinger et al. reviewed the literature and showed that these studies documented changes in practice, particularly with the introduction of new monitoring equipment. [9,10 ] The present study was undertaken to investigate the impact of a new anesthesia technology, EARK, using methodologies developed in previous studies. [3,13 ] The goal was to ascertain how EARK use alters task profile, workload, or vigilance during cardiac anesthesia cases. In addition, this study allowed investigation of the impact of the intraoperative use of transesophageal echocardiography (TEE) on these measures of clinical performance.
Materials and Methods 
After acquiring Institutional Review Board approval and written informed consent, nine CA2 or CA3 anesthesia residents were enrolled in the study. All subjects had a minimum of 5 weeks of previous cardiac anesthesia training at the University of California, San Diego Medical Center (UCSD) and at least 3 weeks of recent experience providing cardiac anesthesia at the San Diego Veterans Administration Medical Center (SDVAMC). All studies were conducted in a single cardiac operating room at the SDVAMC on patients undergoing coronary artery bypass graft (CABG) procedures performed by the same surgical team. Before starting the case, the type of record keeping to be used was randomly selected to be either electronic (ARKIVE[trademark symbol], San Diego, CA [EARK]) or traditional manual charting (MAN). A cross-over experimental design was intended; subjects were to be studied once during each of the record keeping conditions. All subjects had at least 1 month experience with MAN during anesthesia for cardiac surgery at UCSD and had used this EARK system for at least several months on virtually all cases, cardiac and noncardiac, at the SDVAMC.
Each case studied involved a similar set of patient procedures and a similar level of anesthesia attending supervision. Peripheral intravenous and arterial catheters were inserted before entry into the operating room. A pulmonary artery catheter was inserted via a sheath introducer into the right internal jugular vein before induction of anesthesia. All patients received a standardized high-dose opiate-based anesthetic. An Ohmeda Modulus II anesthesia machine, Space Labs PC II monitor with built-in ST segment analysis and trend plots, Datex capnograph/agent analyzer, and a Lifescan processed electroencephalograph monitor completed the anesthesia monitoring array (Figure 1). A TEE probe (Hewlett Packard, Waltham, MA) was inserted immediately after endotracheal intubation. During anesthetic induction, an attending anesthesiologist was always present. After insertion of the TEE and after an initial TEE examination (using a Hewlett-Packard Sonos 500), the resident generally functioned autonomously (i.e., only occasional attending anesthesiologist presence until termination of cardiopulmonary bypass).
Figure 1. The layout of the operating room was standardized for all cases. During EARK cases, the subjects used an ARKIVE[trademark symbol] computer terminal to enter annotative data. During MAN cases, the EARK terminal was disabled and removed from the anesthesia resident's line of sight. The observer used the computer to record the anesthesia resident subject's tasks, workload, and response latency to illumination of the alarm light.
Figure 1. The layout of the operating room was standardized for all cases. During EARK cases, the subjects used an ARKIVE[trademark symbol] computer terminal to enter annotative data. During MAN cases, the EARK terminal was disabled and removed from the anesthesia resident's line of sight. The observer used the computer to record the anesthesia resident subject's tasks, workload, and response latency to illumination of the alarm light.
Figure 1. The layout of the operating room was standardized for all cases. During EARK cases, the subjects used an ARKIVE[trademark symbol] computer terminal to enter annotative data. During MAN cases, the EARK terminal was disabled and removed from the anesthesia resident's line of sight. The observer used the computer to record the anesthesia resident subject's tasks, workload, and response latency to illumination of the alarm light.
×
The physical layout of the anesthesia workspace remained unchanged throughout the study (see Figure 1). During the EARK cases, the subjects used an ARKIVE[trademark symbol] computer terminal containing preconfigured templates specific to cardiac anesthesia drugs, fluids, and procedures to facilitate entry of annotative data. Alphanumeric data were entered on a touch-screen display. Most vital signs were continuously recorded and displayed on the EARK terminal. During MAN cases, the EARK terminal was disabled and removed from the subject's line of sight.
Before starting the case, subjects were explicitly instructed that patient care was paramount and that they could terminate the study at any time. An Apple Macintosh (Cupertino, CA) computer, for use by the observer, was placed in the operating room next to the TEE monitor and adjacent to the medication cart. The observer's position, facing the anesthesia machine and the head of the operating room table (see Figure 1), allowed a clear view of the subject without interfering with patient care activities. The collection of data began on successful placement and calibration of the pulmonary artery catheter and was concluded on initiation of cardiopulmonary bypass. Task characteristics, workload profile, and vigilance were recorded using previously described techniques, [3 ] which will be described briefly.
Time-Motion Analysis. A single observer (O.W.H.) resolved the activities of each subject into 32 task categories (Table 1) using custom software. Each task occurrence was recorded by clicking with a mouse on the appropriate button on the computer display. The software then automatically logged the time and task initiated. If two tasks occurred simultaneously, the observer recorded the dominant task first and then toggled between the two tasks based on the frequency each dominated the subject's time. Teaching of the subject by the attending anesthesiologist was coded as “attending conversation.” Data collection was suspended when the subject was out of the operating room on a break. “Recording” was defined as either writing on the hand-charted record or entering data on the EARK. If a subject looked at either the hand-charted record or EARK display without entering new data, the task was recorded as “observe monitors.”
Table 1. Task-specific Workload Factor Scores* 
Image not available
Table 1. Task-specific Workload Factor Scores* 
×
Vigilance Probe. The latency of response (reaction time) to an “alarm light” was used as a measure of vigilance. [3,14,15 ] A small (1 cm) bright circular red light mounted on the side of an 8 x 12 cm black box was placed on top of the electrocardiograph monitor. Subjects were instructed to respond with either a verbal or manual indication as soon as they detected illumination of the light, which occurred at 8- to 12-min random intervals. On detection of the illuminated light, it was promptly extinguished, and vigilance latency (in seconds) was recorded.
Subjective Workload Assessment. At random intervals, between 10 and 15 min, the computer prompted the observer first to score the subject's workload and then to query the subject as to his or her workload. Workload was assessed using a Borg Workload Scale (a visual analog type scale going from 6 = no exertion to 20 = maximum exertion). [3,13,16 ]
Data Analysis 
Because of the failure to attain a completely balanced design (see Results), all cases in which EARK was used were analyzed as a single group, and similarly, a second group consisted of all of the MAN cases, independent of anesthesia provider. The data from each case were segmented into “preintubation” and “postintubation” periods. The preintubation period began on completion of pulmonary artery catheterization and ended with successful endotracheal intubation. The postintubation period started immediately thereafter and ended with initiation of cardiopulmonary bypass. Statistical significance was set at the P < 0.05 level, and, where appropriate, data are presented as mean +/- SEM.
Time-Motion Analysis. The actual and percent of total case time spent on each task category were calculated. Average duration of occurrences of each task (called “dwell time”) also were determined. Because subjects usually divided their attention among multiple tasks (e.g., conversing or observing), multiple episodes of a single task (for example, observing the TEE) were interspersed with other tasks (e.g., observing the monitors). The reported dwell times are the average across groups (e.g., EARK vs. MAN) of the average dwell time within each case of uninterrupted episodes of each specific task.
For each dependent variable, comparisons between the primary treatment condition (record keeping method) were accomplished using a two-way mixed analysis of variance (ANOVA)[17 ] with one between subjects factor (EARK vs. MAN) and one within subjects factor (tasks) for the preintubation and postintubation phases. For comparison of the dwell times associated with tasks (whether within or between groups), only tasks which occurred in all 10 cases in both groups were included in the analysis. Newman-Keuls post hoc tests [17 ] were used to ascertain significant differences within and between groups. This approach also allowed a statistical comparison between tasks within the different segments of the cases, independent of record keeping condition.
Vigilance Latencies. Because of a non-normal distribution of the vigilance latency data, Mann-Whitney tests [18 ] were used to ascertain statistically significant (P < 0.05) differences between groups (MAN vs. EARK) and across time periods (pre- vs. postintubation). To examine the impact of performance of specific tasks on vigilance latency, the task initiated closest to the time of light illumination was used in the calculations of “task-related” response time.
Subjective Workload Assessment Analysis. The subjective workload data of the two groups were similarly divided into preintubation and postintubation segments. Between and within group differences were assessed using two-way ANOVA followed by Newman-Keuls tests. [17 ] The correlation between observer rating and subject self-assessment of workload was determined using linear regression analysis.
Link Analysis. A link score between two task categories was defined as the number of incidents in which two tasks occurred in succession. For example, the link score between the categories “observe monitors” and “recording” was calculated by adding the number of incidents when “observe monitors” was followed immediately by “recording” with the incidents when “recording” was followed by “observe monitors.” Dwell times were not considered in this analysis. A link percentage was calculated by dividing each link score by the total number of links occurring in that case. Only link pairs that occurred in all 10 cases in both groups were included.
Workload Density Analysis. Workload density was calculated during specific events. A time window was created for each event (self-reported workload, detection of the vigilance light, and when specific tasks were initiated), beginning 1 min before the event and ending 1 min after the event. Workload density was calculated by first multiplying the amount of time spent on each task performed in the window by that task's workload factor score (to be discussed). [13 ] Then, the sum of these values were divided by the length of the window (2 min) and multiplied by 100.
Factor scores were calculated from a separate study in which questionnaires were sent to 49 UCSD anesthesia providers. [13 ] The 39 respondents (response rate of 80%) scored each specific task (e.g., “observe monitors” or “laryngoscopy”) in terms of its perceived difficulty (1 = easy/low, 2 = medium/modest, 3 = hard/high) on three different indices: mental workload, physical workload, and psychologic stress. A factor analysis was used to generate a single factor score (mathematically combining the three workload indices), which reflected the perceived overall workload associated with each clinical task (Table 1).
Linear regression analysis was used to ascertain relationships between workload density and subjective workload or vigilance latencies. Average workload density values were calculated during the tasks of “intubation,”“recording,”‘observe/adjust TEE,’ and “observe monitors,” and a comparison was made within and between groups using ANOVA.
Results 
Ten EARK cases and 10 MAN cases of similar duration (195 +/- 16 and 163 +/- 12 min; P > 0.05) were undertaken by nine anesthesia residents. The intended crossover design was unsuccessful because of logistical and clinical scheduling constraints. Two subjects each were studied one, two, three, and four times, respectively. In one case, a subject was inadvertently allocated to MAN when EARK was indicated by randomization. There were no significant unexpected or critical events during any of the cases.
Time-Motion Analysis 
The two record keeping groups revealed a similar distribution of tasks performed before intubation (Table 2). The percent time spent record keeping during induction was minimal in both groups. In 16 of the 20 cases studied, no manual record keeping occurred before intubation. There were virtually no differences between MAN and EARK groups in the percent of time spent on induction tasks. When all cases were included (MAN and EARK), the predominant induction tasks were bag and mask ventilation (24.8 +/- 2.0%), observation of the monitors (18.6 +/- 1.0%), and administration of intravenous medications (9.0 +/- 0.8%). Dwell times for the tasks of laryngoscopy and intubation were significantly longer than the dwell times for all other tasks analyzed (Table 3).
Table 2. Percent Time on Task: Preintubation Period 
Image not available
Table 2. Percent Time on Task: Preintubation Period 
×
Table 3. Dwell Time* on Tasks in the Preintubation Period 
Image not available
Table 3. Dwell Time* on Tasks in the Preintubation Period 
×
After intubation, compared with the MAN group, the EARK group spent less time record keeping, adjusting and observing the TEE, and conversing with the attending anesthesiologist but slightly more time observing the monitors (Table 4). The two groups spent similar amount of time on the combined categories of direct patient care, indirect patient care, and observing and monitoring tasks. When all cases were included, the greatest amount of time was spent after intubation observing the monitors (24.7 +/- 1.5%), record keeping (11.5 +/- 0.6%), adjusting the intravenous tubes (8.1 +/- 0.8%), and adjusting or observing the TEE (7.7 +/- 1.1%). Both groups showed similar mean dwell times on nearly all task categories, including recording (Table 5).
Table 4. Percent Time on Task: Postintubation Period 
Image not available
Table 4. Percent Time on Task: Postintubation Period 
×
Table 5. Dwell Time* on Most Common Tasks in the Postintubation Period 
Image not available
Table 5. Dwell Time* on Most Common Tasks in the Postintubation Period 
×
Electronic anesthesia record keeping and MAN subjects spent a similar amount of time inserting the TEE probe into the patient's esophagus and performing the initial examination. The residents spent on average 7.4 +/- 0.7 min of the immediate postintubation period performing this initial TEE examination, accounting for two thirds of the time spent observing or adjusting the TEE during the study.
Subjective Workload Assessment Analysis 
Similar workload scores were reported in the two record keeping groups. Overall, self-reported workload was significantly higher before intubation than after intubation (15.5 +/- 0.5 vs. 11 +/- 0.1; P < 0.001). The observer's workload scores showed the same pattern and correlated closely with subject self-reported workload (R = 0.85; P < 0.001).
Vigilance Latency 
Subjects had longer vigilance latencies before intubation (mean, 57 s; median, 26 s; range, 1–300 s) compared with after intubation (mean, 31 s; median, 15 s; range, 1–240 s; P < 0.001). The two record keeping groups were not significantly different from each other either before (mean/median, 41/26 s vs. 77/22 s, EARK vs. MAN, respectively) or after intubation (30/15 vs. 32/15 s). Vigilance latencies occurring during the initiation of the four most common tasks after intubation were analyzed (Figure 2). Although the latencies during record keeping were not statistically different between the groups, latencies were significantly longer in both groups when subjects were observing or adjusting the TEE compared with record keeping, observing monitors, or adjusting intravenous lines. In contrast, subjects showed faster response latencies if they were observing the monitoring array in which the alarm light was placed.
Figure 2. The latency of response (response time in seconds) to an “alarm light” was used as a measure of vigilance (see text for details). The figure depicts the range of data for each of four task categories for EARK ([square, open]) and MAN (*symbol*). The box contains 50% of all of the data, whereas the dark line depicts the median value in that group. The minimum and maximum values in each group is shown by the upper and lower bars. The response latency during record keeping was not statistically different between EARK and MAN. In both record keeping groups, subjects had significantly slower responses when observing or adjusting the TEE when compared with record keeping, observing monitors, or adjusting intravenous lines (*P < 0.05). Subjects in both groups had faster response latencies when observing the monitoring array ((dagger)P < 0.05 compared with all three other tasks).
Figure 2. The latency of response (response time in seconds) to an “alarm light” was used as a measure of vigilance (see text for details). The figure depicts the range of data for each of four task categories for EARK ([square, open]) and MAN (*symbol*). The box contains 50% of all of the data, whereas the dark line depicts the median value in that group. The minimum and maximum values in each group is shown by the upper and lower bars. The response latency during record keeping was not statistically different between EARK and MAN. In both record keeping groups, subjects had significantly slower responses when observing or adjusting the TEE when compared with record keeping, observing monitors, or adjusting intravenous lines (*P < 0.05). Subjects in both groups had faster response latencies when observing the monitoring array ((dagger)P < 0.05 compared with all three other tasks).
Figure 2. The latency of response (response time in seconds) to an “alarm light” was used as a measure of vigilance (see text for details). The figure depicts the range of data for each of four task categories for EARK ([square, open]) and MAN (*symbol*). The box contains 50% of all of the data, whereas the dark line depicts the median value in that group. The minimum and maximum values in each group is shown by the upper and lower bars. The response latency during record keeping was not statistically different between EARK and MAN. In both record keeping groups, subjects had significantly slower responses when observing or adjusting the TEE when compared with record keeping, observing monitors, or adjusting intravenous lines (*P < 0.05). Subjects in both groups had faster response latencies when observing the monitoring array ((dagger)P < 0.05 compared with all three other tasks).
×
Link Analysis 
The most common task pair combination in both subject groups was “observe monitors-recording”(P < 0.02 compared with all other links;Table 6). The “observe monitors-recording” link was significantly more common during MAN cases than during EARK cases (P < 0.001). Notably, in all subjects, pairing of TEE probe adjustment and TEE monitor observation (5.8 +/- 0.8%) was more common than was pairing of observation of the monitoring array with either TEE adjustment (1.7 +/- 0.3%; P < 0.05) or TEE observation (1.9 +/- 0.2%; P < 0.05).
Table 6. Percent of Total Task Links: Postintubation 
Image not available
Table 6. Percent of Total Task Links: Postintubation 
×
Workload Density Analysis 
There was a correlation between subjective workload and workload density (R = 0.49; P < 0.0001). In contrast, vigilance latency did not correlate with workload density (R =-0.01). There was an inverse correlation between elapsed case time and the different workload measures (subjective workload, R =-0.53; workload density, R =-0.39; both P <0.0001). Mean workload density during the 2-min window bounding the record keeping task in EARK and MAN cases was the same (53 +/- 1 in both cases). For both groups, higher workload density was associated with observing/adjusting the TEE (62.0 +/- 0.8) than for observing the anesthesia monitoring array (56.9 +/- 0.3) or record keeping. Workload density at the time of intubation was the highest observed (107.0 +/- 2.4; P < 0.001 compared with the other tasks examined).
Discussion 
The present study used established human factors techniques [3 ] to investigate the effect of electronic record keeping and of TEE on indirect measures of clinical performance during CABG procedures. This study provides a detailed enumeration of anesthesia residents' tasks and workload while administering anesthesia for a complex surgical procedure. The results suggest that the ARKIVE[trademark symbol] electronic record keeping system only modestly reduced the time spent record keeping after intubation relative to manual charting, but it did not affect workload or vigilance. In contrast, manipulation or observation of the TEE appeared to increase workload and adversely affect vigilance.
Electronic versus Manual Record Keeping 
All subjects were well versed in the use of the EARK device, were familiar with the cardiac operating room, and performed nearly identical anesthetics. A high anesthesia workload surgical procedure, CABG, was chosen to enhance detection of any differences between the record keeping conditions. Unfortunately, because of scheduling constraints, the intended within-subjects balanced randomized cross-over design was not attained, although a within-subjects analysis of 12 cases (6 subjects) that met the criteria for the intended experimental design was performed, and the results were not appreciably different from those reported herein for all 20 cases.
Little record keeping occurred before intubation; in only 4 of 20 cases was any manual record keeping performed. In the MAN condition, this required the provider to later reconstruct the events and data associated with induction, whereas in the EARK group, all of the vital signs data had been automatically recorded. We did not assess the “quality” of the anesthesia records produced in this study, although previous research suggests that electronically recorded vital signs are more accurate. [19,20 ]
Although the EARK group spent approximately 20% less time record keeping after intubation than did their MAN counterparts, the average time spent on any single episode of record keeping (dwell time) was not different between the two groups. However, less data needed to be recorded during EARK use because the vital signs data were already being logged electronically. In the face of unchanged workload or vigilance, the clinical significance of the reduced time spent record keeping after intubation in the EARK group remains unclear. Both groups spent similar amounts of time after intubation performing direct and indirect patient care activities (see Table 4).
The EARK group spent slightly more time observing their monitors (13 min total actual case time), although the dwell times on this task did not differ between groups. The present data do not permit assessment of cause-effect relationships. Similarly, the clinical significance of the finding that the EARK group spent one third less time adjusting and observing the TEE remains to be elucidated; this relatively minor difference was not statistically significant when expressed in terms of total time after intubation.
In an earlier task analysis study of record keeping, Allard et al. failed to show a statistically significant difference in the time spent on electronic versus manual record keeping. [7 ] However, there are differences between the study of Allard et al. and the present study. Allard et al. studied anesthetics for a variety of less intense surgical procedures (e.g., otolarnygologic, gynecologic, and general surgical procedures). Additionally, Allard et al. included data collected during the extubation and emergence period. Finally, Allard's group classified record keeping as interaction, whether tactile or visual, with the anesthetic record. In contrast, in our study, generating the record was distinguished from observing the already generated record to obtain clinical information.
There are limitations to the task analysis method used in this study. First, the presence of an observer could affect the subjects' performance (i.e., a Hawthorne effect). However, as was seen in a previous study, [3 ] there was virtually no interaction between the subjects and the observer. Additionally, the presence of multiple other personnel in the room and the busy nature of cardiac anesthesia likely mitigated any potential for this type of bias. Second, this real-time methodology can be influenced by observer bias, particularly because this study could not be blinded. However, given the original hypothesis that EARK would beneficially affect task patterns and workload, any bias introduced by the observer would be expected to have been favorable to the EARK condition. A third issue is the use of a “toggling between tasks” technique to document concurrent tasks. This technique may, under some circumstances, produce misleading dwell times and may also be prone to observer bias. These methodologic issues could be addressed by comparing the results of task analysis of actual cases with subsequent analyses of videotapes of the same cases by the same and different observers. It should be noted that such off-line videoanalysis [21 ] has its own set of limitations, including loss of fidelity, inflexibility, and inability to perform real-time workload assessment or secondary task probing.
Workload and Vigilance 
There were no differences between the two study groups using the workload measures studied. The Borg scale has been shown in a previous study to be sensitive to detection of different workload levels between phases of the anesthetic and level of anesthesia provider experience. [3 ] Subjective workload during these cardiac cases was significantly higher than that reported previously for experienced providers doing routine ambulatory general anesthesia cases (preintubation 15.5 +/- 0.5 vs. 11.3 +/- 0.7 and postintubation 11.0 +/- 0.1 vs. 9.6 +/- 0.3, cardiac vs. ambulatory, respectively, both P < 0.001). Although subjective workload assessment has limitations, [22 ] it is easy to use, relatively nonintrusive, and has proven effective in laboratory and real-world settings. [3,22–24 ]
The two groups also did not differ in their vigilance latencies to the alarm light. The use of a single visual cue, placed in the anesthesia monitoring array, has limitations as a measure of vigilance. First, it is not a clinically relevant stimulus, and thus it is possible that, during high workload situations, it was intentionally ignored. During this circumstance, the test measures workload or “spare capacity” rather than measuring vigilance. [25 ] On the other hand, in this study, response to the light required little effort, and response latencies were generally short. In addition, the results of Loeb, [8,15 ] using a more clinically relevant signal detection task, are similar to our own. Tasks associated with visual orientation away from the monitoring array (e.g., observe TEE or laryngoscopy) might have been expected to be associated with longer response latencies. Yet, experienced providers should routinely divide their attention between multiple information sources, even during high workload tasks. Thus, despite these limitations, we believe the alarm light is a reliable and relevant probe of vigilance in this task environment.
Why was EARK apparently of little demonstrable help to the anesthesiologist during complex cases? It is possible that deficiencies of the ARKIVE's human-computer interface* limited any benefit of EARK in either this study or that of Allard et al. [7 ] For example, the ARKIVE's mechanism for entry of annotative text, fluids, drug infusions, and other data has been described as awkward and time-consuming.*
The use of electronic record keeping appears to shift the anesthesiologist's recording tasks from the observe-remember-manually record sequence to a higher cognitive level (observational) supervisory task. [26 ] Some have suggested that with EARK, there is a danger that the anesthesiologist may be taken “out of the loop” or may become less attendant to changes in the physiologic variables. [5,6 ] The present data do not support the assertion that EARK impairs vigilance. It appears that anesthesia residents readily adapted their monitoring strategy to accomodate EARK use without affecting overall clinical performance. More sophisticated analysis techniques will be required to ascertain what data streams are being observed at any given time, how that information affects the anesthesiologists' mental model of the clinical situation, and what actions are initiated as a consequence.
Impact of the Transesophageal Echocardiography 
To our knowledge, no previous study has examined the impact of TEE on anesthesia task patterns or workload. Observation and adjustment of the TEE was the fourth most common task performed during the postintubation prebypass phase. Approximately two thirds of the time devoted to the TEE occurred during the initial insertion and examination. The residents only occasionally looked at the TEE monitor display during the remainder of the study period. This lack of attention to the TEE during most of the prebypass period may be a result of the TEE display not being included in the residents' scan patterns, perhaps because the TEE was nearly 180 [degree sign] opposite the anesthesia monitoring array (see Figure 1). Alternatively, the TEE may not have provided useful information in the postintubation, prebypass phase of these CABG cases.
The use of the TEE may decrease vigilance to changes in other clinical data. Subjects in both groups had significantly longer vigilance latencies when they were observing or adjusting the TEE compared with other common tasks. Response times were approximately one third slower than during record keeping and nearly 10 times slower than during monitor observation. The TEE's position in the anesthetic workspace may have contributed to the longer latencies. A future study may incorporate an additional alarm light placed near the TEE machine to test this hypothesis. A study of the effect of alternative placement of the TEE display (e.g., overlayed on the primary vital signs monitor) would also be informative.
Link analysis lends additional support to the theory that the TEE draws the anesthesia provider's attention away from the anesthesia machine and monitors. When using the TEE, subjects mainly alternated between observing the TEE display and adjusting the TEE probe. They were three times more likely to follow TEE observation by TEE adjustment (or vice versa) than to follow either with observation of other monitors. The use of the TEE was also associated with higher workload densities compared with most other postinduction prebypass tasks. The impact of TEE on subjective workload was not evaluated in this study.
Study Limitations and Future Directions 
There are many clinically important questions about electronic record keeping that remain unanswered. Does the use of EARK reduce the anesthesiologist's awareness of the patient's clinical status? Would an anesthesiologist using EARK be less likely to recognize the onset of a critical event such as malignant hyperthermia and respond appropriately? These questions are difficult to answer and are beyond the scope of the present study. Also, this study's results may only apply to longer, high workload anesthetics. Future studies should be performed to investigate the usefulness of electronic record keeping in shorter anesthetics, particularly those in a busy ambulatory surgery center. It remains to be seen whether study of a different electronic record keeping system with a more “user friendly” interface would demonstrate greater benefits of EARK.
This study of cardiac anesthetics also yielded new information about the impact of the use of TEE in the operating room. TEE use in these cardiac cases commanded a significant fraction of the resident's time and attention in the early postintubation period and appeared to increase overall workload. We do not know if these results would be substantially different for private practitioners using TEE during cardiac anesthesia. Can a solo anesthesiologist effectively insert the TEE, conduct an examination, and still maintain appropriate vigilance and situation awareness?[26 ] During what circumstances is the information gained from TEE worth the investment of time and attention by the anesthesiologist? In situations wherein intraoperative TEE cannot be shown to improve anesthesia outcome, would its removal from the operating room decrease workload and perhaps improve vigilance? Again, these important questions are difficult to answer and await future study.
In summary, this study attempted to measure scientifically the impact of two medical technologies on anesthesia task performance. During the conditions of this study, the use of the ARKIVE[trademark symbol] EARK system decreased, albeit modestly, the amount of time spent record keeping during the postintubation prebypass phase of cardiac anesthesia. There was no effect of EARK use on several measures of workload or vigilance. This is also, to our knowledge, the first objective description of the time commitment associated with the intraoperative use of TEE. The suggestion that TEE use may affect adversely vigilance and workload points to the need for further study of this technology's costs versus benefits.
The authors thank Shakha Vora, Sherrie Proctor, and Helen Shen for their technical assistance. The authors also thank the reviewers for the many hours they spent on the manuscript; their extensive comments and suggestions improved substantially the final product of this work.
*Jones W, Garriott M, Meissner T, Swenson M, Forcier H, Gruen D, Weinger MB: Improving the user interface of an automated anesthetic record keeper (abstract). J Clin Monit 1993; 9:230–1.
References 
References 
Kennedy P, Fiengold F, Wiener E, Hosek R: Analysis of tasks and human factors in anesthesia for coronary-artery bypass. Anesth Analg 1976; 55:374-7.
McDonald JS, Dzwonczuk R, Gupta B, Dahl M: A second time-study of the anaesthetist's intraoperative period. Br J Anaesth 1990; 64:582-5.
Weinger MB, Herndon OW, Paulus MP, Gaba D, Zornow MH, Dallen LD: Objective task analysis and workload assessment of anesthesia providers. Anesthesiology 1994; 80:77-92.
Woods DD, Cook RI, Billings CE: The impact of technology on physician cognition and performance. J Clin Monit 1995; 11:5-8.
Noel T: Computerized anesthesia record may be dangerous (letter). Anesthesiology 1986; 64:300.
Noel KR: Controversy in automated record keeping (letter). J Clin Monit 1991; 7:280.
Allard J, Dzwonczyk DY, Block Jr FE, McDonald JS: Effect of automatic record keeping on vigilance and record keeping time. Br J Anaesth 1995; 74:619-26.
Loeb RG: Manual record keeping is not necessary for anesthesia vigilance. J Clin Monit 1995; 11:9-13.
Weinger MB, Englund CE: Ergonomic and human factors affecting anesthetic vigilance and monitoring performance in the operating room environment. Anesthesiology 1990; 73:995-1021.
Weinger MB, Smith NT: Vigilance, alarms, and integrated monitoring systems, Anesthesia Equipment: Principles and Applications. Edited by Ehrenwerth J, Eisenkraft JB. Malvern, PA, Mosby Year Book, 1993, pp 350-84.
Boquet G, Bushman JA, Davenport HT: The anaesthetic machine-A study of function and design. Br J Anaesth 1980; 52:61-7.
McDonald JS, Dzwonczyk RR: A time and motion study of the anaesthetist's intraoperative time. Br J Anaesth 1988; 61:738-42.
Weinger MB, Shen HD, Culp M, Fehrenbacher N, Herndon OW: Real-time workload assessment during anesthesia (Abstract). J Clin Monit 1995; 11:259.
Kankowitz BH, Casper PA: Human workload in aviation, Human Factors in Aviation. Edited by Wiener EL, Nagel DC. San Diego, Academic Press, 1988, pp 157-87.
Loeb RG: Monitor surveillance and vigilance of anesthesia residents. Anesthesiology 1994; 80:527-33.
Borg G: Simple rating methods of perceived exertion, Physical Work and Effort. Edited by Borg G. Oxford, England, Permagon Press, 1977, pp 39-47.
Winer B: Statistical Principles in Experimental Design. New York, McGraw-Hill, 1971.
Howell D: Statistical Methods for Psychology. Boston, Duxbury Press, 1987.
Lerou JGC, Dirksen R, van Daele M, Nijhuis GM, Crul JF: Automated charting of physiological variables in anesthesia: a quantitative comparison of automated versus handwritten anesthesia records. J Clin Monit 1988; 4:37-47.
Cook RI, McDonald JD, Nunziata E: Differences between handwritten and automatic blood pressure records. Anesthesiology 1989; 71:385-90.
Mackenzie CF, Martin P, Xiao Y: Video analysis of prolonged uncorrected esophageal intubation. Anesthesiology 1996; 84:1494-503.
Hicks TG, Wierwille WW: Comparison of five mental workload assessment procedures in a moving-base driving simulator. Human Factors 1979; 21:129-43.
Bortolussi M, Kantowitz B, Hart S: Measuring pilot workload in a motion-based trainer. Appl Ergonom 1986; 17:278-83.
Hill SG, Lavecchia HP, Byers JC, Bittner AC, Zaklad AL, Christ RE: Comparison of four subjective workload rating scales. Human Factors 1992; 34:429-39.
Bortolussi M, Hart S, Shively R: Measuring moment-to-moment pilot workload using synchronous presentations of secondary tasks in a motion-based trainer. Aviat Space Environ Med 1989; 60:124-9.
Gaba DM, Howard SK, Small SD: Situation awareness in anesthesiology. Human Factors 1995; 37:20-31.
Figure 1. The layout of the operating room was standardized for all cases. During EARK cases, the subjects used an ARKIVE[trademark symbol] computer terminal to enter annotative data. During MAN cases, the EARK terminal was disabled and removed from the anesthesia resident's line of sight. The observer used the computer to record the anesthesia resident subject's tasks, workload, and response latency to illumination of the alarm light.
Figure 1. The layout of the operating room was standardized for all cases. During EARK cases, the subjects used an ARKIVE[trademark symbol] computer terminal to enter annotative data. During MAN cases, the EARK terminal was disabled and removed from the anesthesia resident's line of sight. The observer used the computer to record the anesthesia resident subject's tasks, workload, and response latency to illumination of the alarm light.
Figure 1. The layout of the operating room was standardized for all cases. During EARK cases, the subjects used an ARKIVE[trademark symbol] computer terminal to enter annotative data. During MAN cases, the EARK terminal was disabled and removed from the anesthesia resident's line of sight. The observer used the computer to record the anesthesia resident subject's tasks, workload, and response latency to illumination of the alarm light.
×
Figure 2. The latency of response (response time in seconds) to an “alarm light” was used as a measure of vigilance (see text for details). The figure depicts the range of data for each of four task categories for EARK ([square, open]) and MAN (*symbol*). The box contains 50% of all of the data, whereas the dark line depicts the median value in that group. The minimum and maximum values in each group is shown by the upper and lower bars. The response latency during record keeping was not statistically different between EARK and MAN. In both record keeping groups, subjects had significantly slower responses when observing or adjusting the TEE when compared with record keeping, observing monitors, or adjusting intravenous lines (*P < 0.05). Subjects in both groups had faster response latencies when observing the monitoring array ((dagger)P < 0.05 compared with all three other tasks).
Figure 2. The latency of response (response time in seconds) to an “alarm light” was used as a measure of vigilance (see text for details). The figure depicts the range of data for each of four task categories for EARK ([square, open]) and MAN (*symbol*). The box contains 50% of all of the data, whereas the dark line depicts the median value in that group. The minimum and maximum values in each group is shown by the upper and lower bars. The response latency during record keeping was not statistically different between EARK and MAN. In both record keeping groups, subjects had significantly slower responses when observing or adjusting the TEE when compared with record keeping, observing monitors, or adjusting intravenous lines (*P < 0.05). Subjects in both groups had faster response latencies when observing the monitoring array ((dagger)P < 0.05 compared with all three other tasks).
Figure 2. The latency of response (response time in seconds) to an “alarm light” was used as a measure of vigilance (see text for details). The figure depicts the range of data for each of four task categories for EARK ([square, open]) and MAN (*symbol*). The box contains 50% of all of the data, whereas the dark line depicts the median value in that group. The minimum and maximum values in each group is shown by the upper and lower bars. The response latency during record keeping was not statistically different between EARK and MAN. In both record keeping groups, subjects had significantly slower responses when observing or adjusting the TEE when compared with record keeping, observing monitors, or adjusting intravenous lines (*P < 0.05). Subjects in both groups had faster response latencies when observing the monitoring array ((dagger)P < 0.05 compared with all three other tasks).
×
Table 1. Task-specific Workload Factor Scores* 
Image not available
Table 1. Task-specific Workload Factor Scores* 
×
Table 2. Percent Time on Task: Preintubation Period 
Image not available
Table 2. Percent Time on Task: Preintubation Period 
×
Table 3. Dwell Time* on Tasks in the Preintubation Period 
Image not available
Table 3. Dwell Time* on Tasks in the Preintubation Period 
×
Table 4. Percent Time on Task: Postintubation Period 
Image not available
Table 4. Percent Time on Task: Postintubation Period 
×
Table 5. Dwell Time* on Most Common Tasks in the Postintubation Period 
Image not available
Table 5. Dwell Time* on Most Common Tasks in the Postintubation Period 
×
Table 6. Percent of Total Task Links: Postintubation 
Image not available
Table 6. Percent of Total Task Links: Postintubation 
×