Editor’s Note: Presented here is the second half of an article begun two issues ago reporting on the meeting of the International Committee on Preventable Anesthesia Mortality and Morbidity last June prior to the World Congress.
Simulation in Anesthesia
Dr. Gravenstein described the potential usefulness of using realistic simulators for instruction in the use of crisis prevention and appropriate response to those crises that will still inevitably occur. He reviews the history of simulation in other fields including warfare and aviation. A distinction is made between training devices, including components of systems and simulators, which recreate the entire environment with some degree of realism. In anesthesia, there are numerous computer based systems that ‘simulate ‘ some aspect of anesthesia. While these have great usefulness for teaching various topics, they are not in the category of realistic simulators. Dr. Gravenstein described the system that has been developed in Gainesville. It includes a mannequin, from which invasive and non-invasive blood pressures, breath sounds, heart sounds and carbon dioxide and anesthetic gas concentrations can be recreated. The system is a hybrid in that some variables are recreations of the real thing, e.g. ECG and oxygen saturation. The mannequin includes detail to the level of actual simulated muscle twitching for simulation of neuromuscular blockade monitoring. The entire system is operated by a computer from which various scenarios are enacted, altering the ‘condition’ of the patient and anesthetic situation to elicit responses from the trainee.
Simulators have applications for new residents, allowing to practice routine events and simple critical events. For the experienced clinician, the simulator allows experience in critical events such as anaphylaxis and malignant hyperthermia that would rarely, if ever, be encountered in practice. It is also possible to simulate multiple events and the kinds of distractions that occur in a real operating room situation. Dr. Gravenstein noted the work conducted by Dr. David Gaba at Stanford in recreating and studying the responses to such events. Simulators will also be very useful for training ancillary help, e.g., respiratory therapists and, perhaps, most useful for training anesthetists or assistants in rudimentary skills if extensive educational programs are not available. thus, someone could be trained relatively quickly to handle simple procedures and even manage simple crisis via the simulation environment.
Noting that simulator training is required in aviation and in other disciplines, Dr. Gravenstein feels strongly that this educational and training modality should be introduced into anesthesia. His experience is that it is relatively easy to train faculty as trainers. He estimates that the cost of a simulator in production would be on the order of $150,000 and that this expense would, of necessity, be supported by its use in training clinicians outside the sponsoring department. For comparison, aviation simulators, which cost on the order of 12 million dollars, are operated 22 hours per day. In response to question about the evaluation of the effectiveness of simulation, Dr. Gravenstein described a study conducted in Gainesville. Thirty-two residents were divided into two groups. Each group had the same curriculum with the exception of simulation training being available to one group. The simulator group was found to learn more quickly than the control group, which caught up after several weeks. Although the simulator probably has a role in retraining of anesthetists who may have been out of practice for a while and in testing or credentialing, neither of these possibilities has been examined yet.
An Algorithm for Crisis Management
The AIMS group has developed a strategy for crisis management based on an analysis of the first 1,000 reports they received (see earlier AIMS summary). The objective is to have a strategy that can be recalled easily during a crisis to permit rapid diagnosis and correction of a problem.” The algorithm is based on the acronym COVER/ABCD. For COVER, there are two steps associated with each letter. For each step, there is a subset of considerations: C-circulation (check pulse, bl;good pressure, etc.0; C-color (check color and oxygen saturation); 0oxygen flow; 0 oxygen concentration(via analyzer); V vaporizer; V-ventilation (manually by removing ventilator-check compliance, breathing system, etc.); E-endotracheal tube; E eliminate machine (use self-inflating bag); R redo monitoring equipment (check all instruments for proper function); R review cover. The COVER algorithm is followed by the commonly known AKD Airway, Breathing, Circulation, Drugs. The investigation established this algorithm and a set of rules for assessing how each could be assessed for effectiveness in diagnosis and correction of problems described the AIMS reports. Using and initial set of reports, a first set of rules was established. these were further modified after independent review by each assessor of another group of reports. Of the first 1,000 reported, 72% were evaluated for how the protocol may have worked versus the anesthetist who was actually involved in the event. Of the remaining reports, 16% occurred int PACU, for which the algorithm is not necessarily intended and 12% of reports were not applicable to the algorithm for other reasons, e.g., cross matched blood not available, incorrect premeditation, wrong patient. The protocol was assessed to have performed correctly in diagnosing 99.5% of events with circulation(15%) and ventilation by hand(29%) being the two steps most frequently useful. Nine percent of events fell into the drug category of the ABCD portion of the algorithm. Of the 0.5% of reports for which the algorithm would not have made a diagnosis, the problems included pleural effusion and pheochromocytoma. The algorithm would have been useful for 90% of problems. Of these, the ABCD was effective in 60%. For the 10% that the algorithm would not have been effective, administration of the wrong drug was involved. Since this is only detected after the fact, by definition, correction is not possible.
It has been assessed that if the algorithm had been applied, it would have been more effective than the operator in 23% of the cases reviewed. In 705 of those cases, the algorithm and anesthetist were equally effective. In only less than one percent did the operator do better than the algorithm. (5% of cases were failed intubations, which were not assessed.)
Dr. Runciman believes that anesthetists should be trained to invoke this very simple concept instinctively. By teaching its use during routine scanning, anesthetists would become familiar with it and be prepared to implement it during a crisis. Since they have not validated this approach in any controlled trial as yet, this hypothesis remains to be tested. This issue provoked discussion about the need for intra-observer and inter-observer reliability in studies such as this. Dr. Cheney noted the usefulness of anthropology and health services training of the project director of the ASA closed claims study.
Accident Investigation as a Tool for Studying Anesthetic Mishaps
This topic was presented via a play, ‘An Anesthetic Catastrophe”, performed in three acts by Dr. J.M. Davies (the Investigator, Dr. A.K. Bacon), (Dr. Friend), (Dr. William Runciman, the Professor) and narrated by Dr. J.B. Cooper. The play portrayed an anesthetic catastrophe occurring in a small community hospital, involving a senior anesthetist (the professor) visiting for the day. The hypothetical patient was described as being healthy, undergoing a routine procedure, but suffered a cardiac arrest from which she could not be resuscitated. Following the event, Dr. Friend intervened to lend support to the professor and to guide the process of accident follow-up. The operating room was closed to assure nothing was disturbed, appropriate hospital authorities and the patient’s family physician and relatives were contacted, arrangements were made for another anesthetist to take over the professor’s cases for the remainder of the day and the accident was investigated thoroughly. Concepts were reviewed for general principles to be used in “breaking the news’ and assisting the family. Dr. Friend helped the professor to debrief what happened. Later, he debriefed others on the team to elicit potentially useful facts. The investigator worked with hospital engineering to inspect the piped gas supplies and anesthetic equipment and drugs. Drugs were sequestered for later analysis. It was explained that the process is not to assess blame, but rather to establish what factors may have contributed to the event. The term ’cause” implies “blame so it is not invoked in this discussion. The Investigator created a chronologic list of all the events in search of possible contributing factors. The different factors were categorized as being .active’ or ‘latent’ failures, the latter being ‘resident pathogens’ built into the anesthesia care system. Finally, recommendations were formulated for preventing future occurrences. During the course of this process, communication was maintained with the patient’s family and with the Professor to provide emotional support.
The discussion following the play brought out many issues surrounding the concept of accident investigation. It was generally agreed that such thorough investigations rarely are conducted. Trained accident investigators are not generally available and protocols for conducting such investigations and assuring the kind of appropriate follow-up demonstrated in this play are not widely known. Several references were mentioned, including the report by Dr. Bacon of ‘Major Anesthetic Mishaps Handling the Aftermath’.(14) A set of “Guidelines for Action following an Adverse Anesthesia Event” has been created (contact Dr. Cooper for copies) and is expected to be published soon. (15) Also, the recent Safe Medical Devices Act in the United States has led to dissemination of a protocol for device investigations formulated by ECRI. (16)
Despite the infrequency with which events are thoroughly investigated, several examples were given during the course of the discussions. (17) It was suggested that, given how rarely accidents occur, such an investigative protocol be implemented even for potentially serious events. Dr Runciman feels that critical incident reporting via forms is generally effective for this since most events have fairly simple contributing factors. Dr. Pat McKay emphasized the need to counsel staff, particularly the professor, following this event. Dr. Bacon offered that contacting the spouse of the affected anesthetist could also be useful.
On the subject of event reporting, Dr. Runciman suggested that competition be elicited over reporting the ‘best’ incident. Dr. Cheney reported some success in their institution at active incident reporting, which he attributes to a system using a single nurse who collects information on all events.
Dr. Gravenstein suggested that event reporting can create problems. When failure to extubate following a procedure was one of the events documented, residents apparently interpreted such events to have a negative reflection on their performance. This probably was the reason for a decrease in non-extubation, but a concomitant increase in reintubations in the PACU. Dr. Pierce voiced his strong support for the distribution and implementation of accident investigation protocols as illustrated in the play.
13. APSF Newsletter (Australia). Suggested protocol for management of anesthetic emergencies. A core algorithm “COVER ABCD, 1990,9:1-2.
14. Bacon AK: Major anesthetic mishaps. Handling the aftermath. Current Anaesthesia and Critical Care 1990; 1:253-257.
15. Cooper, JB, Cullen DJ, Eichhom JH, Philip JH, Holzman RS: Administrative Guidelines for Response to an Adverse Anesthesia Event. J Clin Anesth (in press).
16. ECRI: Investigating device related incidents. In Medical Device Reporting Under the Safe Medical Devices Act. ECRI. p. 23-30,1991.
17. Cooper JB, DeCesare R, D’Ambra M: An engineering critical incident Direct current bum from a neuromuscular simulator. Anesthesiology 1990; 73:168-172.