Maintaining an accurate anaesthesia record is one of the more difficult and time-consuming tasks of the anaesthesiologist. This task is often compromised in favour of actual patient care, especially during crisis situations, thus making auditing of critical events less reliable.
Introducing an anaesthesia information management system (AIMS) not only improves the accuracy of the anaesthesia chart but also reduces workload [1–5]. Secondly, AIMS provides a very detailed account of the process of anaesthesia thus allowing reliable auditing [6–8]. It can also provide the physician with smart alarms and guidelines for the treatment of specific patients [9,10]. Last but not least it can generate a detailed billing record, improve operating room (OR) management and reduce anaesthesia-related costs [11,12].
AIMS, however leads to complete transparency of the work of the anaesthesiologist as well as to a change in everyday practice, and may be met with a degree of fear from members of the department, especially those who have less experience with computers. Resistance to the introduction of AIMS can also be expected due to the perceived legal implications of an automated record, and especially of the presence of artefacts. Despite this, a survey conducted among hospitals that keep automatic anaesthesia records has shown that both the anaesthesiologists and the risk managers consider AIMS to be a protective factor against litigation .
The introduction of AIMS, involves a major investment of 20–30% of the price of a modern anaesthesia machine per OR. Furthermore, it requires considerable investment in manpower both for initial implementation and for maintenance. The successful implementation of AIMS requires in addition to co-operation from the anaesthesia department staff, full co-operation with the software supplier, and with the producers of the interfaced medical equipment. The incorporation of the electronic anaesthesia record into the hospital information system requires further co-operation between various departments within the hospital (management, information technology, medical engineering, electricity and maintenance).
In the present article we report our experience in implementing AIMS, focusing on its acceptance by the department staff and the reliability of data collection during the first 6 months of its use.
The Metavision® AIMS (MVOR-X-edition, iMDsoft, Tel-Aviv, Israel) is configured such that a single remote server is connected to all of the medical equipment via a terminal server for each OR (Lantronix ETS16P). Data are transported from each of the medical devices at a frequency of once per minute (single sample) to the main server (Compaq DL380 4 with four PIV 2.8 GHz processors, 2 GB memory and 2 storage disks of 36 GB each + 4 storage disks of 72 GB each). The anaesthesiologist, through a workstation (PIV with 512 MB RAM), mounted on the anaesthesia machine (and connected to the server), enters data on events, drugs and fluids into the anaesthesia record and views the complete anaesthesia record including a preoperative chart and information derived from the hospital information system. An example of anaesthesia chart is depicted in Figure 1.
In-hospital implementation of AIMS followed a schedule devised by the software company. Briefly, a two tier-training program was employed. Initially, clinical application specialists from the software company trained selected staff members to be project manager and four super-users, responsible for all aspects of the system. These, in turn, trained the rest of the staff on a one-to-one basis with repeated sessions (60–90 min each session) as required. The super-users and project manager were also responsible for the detailed customization of the application to local requirements, and for problem solving on a routine basis.
The implementation period lasted 6 weeks, during which the project manager was fully dedicated to the project. The super-users devoted 50% of their time to the project and the other 50% to clinical work.
The Carmel Lady Davis Medical Center is a 450-bed general hospital. The anaesthesia department consists of a staff of 38 physicians, 22 of them faculty and 16 residents. The department provides anaesthesia, critical care and pain management services for the hospital. Anaesthesia is provided by residents, with an attending physician covering two rooms. The system was implemented in the eight main ORs in which around 9000 operations are performed annually (30% unscheduled). The anaesthesia department provides further services in four in-hospital ambulatory ORs and four remote anaesthesia locations, in which AIMS was not implemented due to financial constraints, as well as acute and chronic pain management and an eight-bed general intensive care unit.
The proficiency of the anaesthesia staff in using Windows and Windows-based applications was evaluated using 20 basic tasks (Appendix 1). Anaesthesiologists who failed to perform five of the tasks underwent basic computer training. This training was administered by a hospital computer specialist to groups of 4–5 physicians at a time for 60 min per session.
The satisfaction with AIMS was assessed prior to, 1 week after and 3 months after implementation (go-live day) by a specially designed questionnaire (Appendix 2). Each response was marked on a six-point (0–5) scale: strongly disagree, disagree, partly disagree, partly agree, agree and strongly agree. For the purpose of descriptive statistics results were grouped in three categories: agree with the statement (4–5 on the scale), disagree (0–1) or inconclusive responses (2–3).
Anaesthesia records (cases) during the first 6 months after implementation of AIMS were studied using the ‘Query-Wizard’ tool, which is a part of the software package. Staff compliance was evaluated by registering any handwritten charts as well as by measuring the rate of missing basic demographic data. The database was initially queried for all cases of anaesthesia, and for all records of basic demographic data (ASA grade, intravenous (i.v.) line insertion site, anaesthesiologist's name, type of surgery and start of surgery). Data were exported to Excel (Microsoft Corporation, Redmond, WA) for the calculation of the rate of missing basic demographic data (pivot table). Automatically collected data from all non-cardiac surgeries was evaluated for artefacts marked by the anaesthesiologist during the case (noise and motion artefacts of the monitors) and for extreme values of heart rate (HR) and oxygen saturation whether they were registered as artefacts or not. The database was queried for all automatically recorded data that were registered as artefacts, as well as for HR values of <20 or >190 bpm or oxygen saturation levels of <80% (as such values have to be either clarified or marked as error), and then for HR >170 bpm and oxygen saturation <85%. As data are recorded by the system once per minute (single digital sample from the monitor) and as the standard for documenting physiologic parameters during anaesthesia is every 5 min we investigated extreme values only if they persisted for 5 min or more. We further scrutinized the new system as used for Caesarean sections by comparing the AIMS database to the hospital records of Caesarean sections. We evaluated the rate of missing manually inserted data on delivery and the administration of oxytocin in that stressful environment.
All of the data were collected from the AIMS database using the Query-Wizard tool that is a part of the Metavision® software package. Data were exported to Microsoft Excel for analysis and described as proportions. Staff responses (scale 0–5) between periods were compared using the non-parametric Wilcoxon signed ranks test and reported as medians and frequency distributions. Overall satisfaction with AIMS was computed as the mean of eight questions (Appendix 2) and its reliability was measured using the Alpha-Cronbach coefficient (0.75 and 0.77 at 1 week and 3 months, respectively). Comparison of score between males and females was performed using the U-test and correlation of score with experience and age was analysed using the Spearman's rank correlation coefficient. All statistical analyses were performed using SPSS statistical software (version 11.5) and P-value <0.05 was considered statistically significant.
The anaesthesia staff is aged between 30 and 58 yr (median 41) (Table 1). Eleven of the 38 anaesthesiologists (29%) did not use a computer on a routine basis prior to implementation of AIMS and 13 (34%) required basic computer training, following which all could perform the 20 basic tasks. During the 6-week implementation process six of the staff (16%) required extra training sessions in the use of the MVOR system.
Thirty-one physicians completed all questionnaires and their responses are presented in Table 2. The median response regarding increased workload was 3 (scale 0–5) before go-live day and dropped to 2 a week after implementation (P =0.001). Three months later the response was similar. Both 1 week and 3 months after implementation only 10% were highly concerned about increased workload (score 4–5) as opposed to 41% before. The median response before go-live day regarding decreased attention to patients was 3 (scale 0–5) (Table 2) which dropped to 2 (P =0.003) 1 week after implementation and to 1 after 3 months (P < 0.001). At 3 months only 3% were highly concerned (score 4–5) as opposed to 21% 1 week after go-live day and 46% immediately preceding it. The median response regarding medicolegal concern was 2 both 1 week and 3 months after implementation (P =0.46) (Table 2). Responses show that the staff did not feel a need to extend the period of double charting (median 0 at both time points P =0.31). The median score regarding pre-implementation training was 4 both 1 week and 3 months after implementation (P =0.8), however 30% and 40% respectively did not think they had received adequate training prior to implementation (Table 2). The median overall satisfaction score at 1 week after and at 4 months after implementation was 3.8 and 3.9, respectively (P =0.1). There were no correlation between age, experience or gender and the overall satisfaction with AIMS.
Evaluation of implementation success
During the first 6 months following implantation a total of 4074 patients underwent 4429 anaesthetic procedures for which an electronic anaesthesia chart was recorded. On the go-live day, double charting was used. During the next 6 months a manual paper chart was used in only six additional cases (due to the computer not being connected to electricity in four cases and failure to log-on in two).
HR (as an example of an automatically recorded vital sign) had to be entered manually in 29 cases (0.7%) due to failure of the interface between monitor and system (information cable disconnection, terminal server disconnected from electricity and system maintenance).
At least one value of the automatically recorded HR was marked as erroneous in 553 of cases (12.5%) (Fig. 2). In 334 cases (7.5%) only one or two erroneous values were marked. In 97 cases (2%) five or more erroneous values were marked. The quality of the automatic data was further evaluated by assessing all cases in which very extreme data were recorded (HR <20 or >190 bpm and oxygen saturation < 80% in patients who did not undergo cardiac surgery) (Fig. 2). A total of 1522 extreme data points were found in 774 different cases (17.5%). Of these, in 67 cases (1.5%) more than four such data values were found. In 40 of these cases these measurements were marked as artefacts. Of the 27 cases in which extreme physiologic data were recorded, hypoxemia during either induction or following extubation was nearly always the cause. When using less stringent criteria for extreme values (HR > 170 and SP O2 <85%) an additional 26 cases were found in which extreme data were recorded for 5 or more minutes. Most of these patients had elevated HR (children and patients with supra-ventricular tachydysrhythmias).
Quality of manually entered data
Manually entered data that were evaluated included the type of surgery, the name(s) of the anaesthesiologist(s), the ASA grade and i.v. line insertion site (Table 3). The event of start of surgery, which is also manually recorded, was marked in 4175 (94%) of the cases.
According to the hospital register, 281 Caesarean sections were performed during the 6-month study period. In 259 of them (92%) the type of surgery was manually entered by the anaesthesiologist (12 of 22 omissions occurred during the first month after implementation). In 263 cases the event of delivery was manually entered (94%). Of the 19 cases in which delivery was not recorded, 13 occurred in the first 6 weeks after implementation. The administration of oxytocin was recorded in all cases of Caesarean section, irrespective of whether the type of surgery or the event of delivery had been recorded.
In this study, it is shown that even in a department in which 34% of the staff had minimal computer experience, and in which at least half do not work daily with AIMS, the introduction of a commercial AIMS could be accomplished effectively during a short time period.
Moreover, our findings indicate that the acceptance rate of such a system is high and that the users, who initially were wary of the system, have fully accepted it. In a previous report, Coleman and colleagues describe the implementation process of AIMS (Arkive®) in Duke University Medical Center . Unlike us, the Duke team had previous experience with automated anaesthesia records, and the same Arkive® was in use for cardiac anaesthesia for the 4 yr prior to implementation. Despite this difference the success of our implementation is comparable to that reported by the Duke University team, who also assessed the staff attitudes 2 yr later and reported satisfaction with AIMS. Quinizio and colleagues have recently reported the acceptance of the NarkoData® AIMS system in the University Hospital of Giessen . They found that after 5 yr of use most users were satisfied with the system but that those who perceived themselves to be well trained were more likely to be satisfied. While we found that 40% of our staff did not feel that they were fully trained, this did not seem to influence overall satisfaction.
In the past, AIMS had to depend on the computer and information technology capability of each hospital. Benson and colleagues have described in detail the process of modifying the NarkoData® in order to adapt it to the requirements of their department. They reported that a period of over 2 yr elapsed from the initial contract with the software company to actual implementation . In comparison, the process in our department from contract to full implementation took only 6 months. While implementation of the present system still required information technology expertise, maintenance and development rely almost entirely on the built-in tools of the system, requiring only slight modification.
Achieving an accurate and detailed anaesthesia record is one of the main objectives of implementing AIMS. During the first 6 months of use we found that omission of manually inserted data was low. The administration of oxytocin was recorded for patients who underwent Caesarean sections (although we cannot be sure that the accurate dose was recorded). Even parameters not routinely recorded prior to implementation, such as i.v. line insertion site and the event of delivery in Caesarean sections were omitted only infrequently (in 12% and 6% of cases, respectively). The tradition of accuracy in manual records differs from place to place. It has been shown that physicians smooth the manual anaesthesia chart [1,2,17]. Although we did not analyse our handwritten anaesthesia charts, our impression is that manual insertion of data improved significantly after AIMS introduction. We found that automatically recorded data were very reliable and the prevalence both of artefacts and of extreme values of HR and oxygen saturation was relatively minor.
AIMS can be used for the development of computerized clinical decision support systems and for data analysis. Benson and colleagues describe the process required to alter the database of the NarkoData® in order to allow for statistical analysis of the data . The AIMS implemented in our hospital (Metavision®) can support both functions. Data retrieval for the current study was performed using the Query-Wizard tool, which queries the SQL database. The Event-Manager tool is being used for the development of computerized clinical decision support systems.
In conclusion, despite the fact that our department of anaesthesia was not previously computerized and that a third of the staff was barely computer literate, the present system was successfully implemented in a short time period to the satisfaction of the anaesthesia staff. Both manually and automatically recorded data are accurate, and the data can be easily analysed.
1. Benson M, Junger A, Quinzio L et al
. Influence of the method of data collection on the documentation of blood-pressure readings with an anesthesia information management system (AIMS). Methods Inf Med
2. Reich D, Wood Rj, Mattar R et al
. Arterial blood pressure and heart rate discrepancies between handwritten and computerized anesthesia records. Anesth Analg
3. Allard J, Dzwonczyk R, Yablok D et al
. Effect of automatic record keeping on vigilance and record keeping time. Br J Anaesth
4. Edsall DW, Deshane P, Giles C et al
. Computerized patient anesthesia records: less time and better quality than manually produced anesthesia records. J Clin Anesth
5. Edsall DW. Quality assessment with a computerized anesthesia information management system (AIMS). QRB Qual Rev Bull
6. Benson M, Junger A, Fuchs C et al
. Use of an anesthesia information management system (AIMS) to evaluate the physiologic effects of hypnotic agents used to induce anesthesia. J Clin Monit Comput
7. Vigoda M, Gencorelli F, Lin S et al
. Anesthesia information systems helps identify missed opportunities for perioperative beta blockade. Anesthesiology
8. Benson M, Junger A, Michel A et al
. Comparison of manual and automated documentation of adverse events with an anesthesia information management system (AIMS). Stud Health Technol Inform
9. O'Reilly M, AkkeNeel T, VanRiper S. Using an anesthesia information system to provide physician specific feedback to improve timely administration of prophylactic antibiotics. Anesthesiology
10. Junger A, Hartmann B, Benson M, et al
. The use of an anesthesia information management system for prediction of antiemetic rescue treatment at the postanesthesia care unit. Anesth Analg
11. Lubarsky DA, Sanderson IC, Gilbert WC et al
. Using an anesthesia information management system as a cost containment tool. Description and validation. Anesthesiology
12. Coleman RL, Sanderson IC, Lubarsky DA. Anesthesia information management systems as a cost containment tool. Crna
13. Feldman J. Do anesthesia information systems increase malpractice exposure? Results of a survey. Anesth Analg
14. Coleman RL, Stanley III T, Gilbert WC et al
. The implementation and acceptance of an intra-operative anesthesia information management system. J Clin Monit
15. Quinzio L, Junger A, Gottwald B et al
. User acceptance of an anesthesia information management system. Eur J Anaesthesiol
16. Benson M, Junger A, Quinzio L et al
. Clinical and practical requirements of online software for anesthesia documentation an experience report. Int J Med Inform
17. Cook RI, McDonald JS, Nunziata E. Differences between handwritten and automatic blood pressure records. Anesthesiology
18. Benson M, Junger A, Quinzio L et al
. Data processing at the anesthesia workstation: from data entry to data presentation. Methods Inf Med
Appendix 1: Basic computer and Windows-based skills assessed prior to implementation of AIMS
- Restarting and logging off a computer using the ‘Start’ menu
- Minimizing, maximizing and resizing of a window
- Moving between different open Windows
- Using ‘Windows Explorer’ to find a folder
- Creating a folder within a folder
- Rename a document or a folder
- Moving files between folders
- Copying files between folders
- Deleting files
- Using the ‘Save As’ function
- Finding a file using ‘Search’
- Sending a file to an E-mail recipient
- Uploading and opening a file from a CD
- Opening ‘Word’
- Opening a ‘Word’ document and saving it under a different name
- Printing a ‘Word’ document
- Change the font of a part of a document
- Running ‘Internet Explorer’
- Surfing to a specific Internet site
- Retrieving a deleted document from the Recycle bin
Appendix 2: The questionnaire given to anaesthesia staff
Predictive statements (opinion of the staff after training and before go-live)
- I predict that the workload of the anaesthesiologist will increase following the introduction of AIMS
- I predict that attention given to patient care will decrease following the introduction of AIMS
Evaluation of staff satisfaction (evaluated 1 week and 3 months after implementation)
- Following AIMS implementation workload increased
- Following AIMS implementation attention given to patient care decreased
- AIMS increases stress at the workplace
- I prefer to use manual paper chart rather than AIMS
- AIMS improves anesthesia care
- I prefer to use AIMS rather than manual paper chart
- The paper chart has higher quality of data that the AIMS chart
- The use of AIMS deceases the quality of patient care
Assessment of opinion towards the AIMS chart
- The AIMS chart increases medicolegal risks
- I would rather use double charting for a longer time period
- I received enough training prior to implementation of AIMS