Strong interest in research on issues associated with anesthesia patient safety was again seen in the scientific sessions at the 1990 American Society of Anesthesiolosists’ Annual Meeting. In the four half-day presentations under “Patient Safety, Epidemiology, and Education,”(two oral and two poster), 86 abstracts were offered. Summarized here are only some of the many interesting papers having direct relevance to patient safety. Several other presentations, including some regarding education, also had links to patient safety considerations.

Decreasing Anesthesia Cardiac Arrests

Drs. R. L. Keenan and C. P. Boyan compared the incidence of cardiac arrest during anesthesia in the decades 1969-78 and 1978-88 at the Medical College of Virginia to see if increased safety awareness and enhanced monitoring* have made a difference. The rate of cardiac arrest decreased. Cardiac arrest due in part or totally from anesthesia fell from 2: 10,000 to 1: 10,000 from the first to the second decade studied at this institution. This improvement was due entirely to a decrease in the preventable respiratory category, from an incidence of 0. 84: 10,000 to 15: 1 0,000. Cardiac arrests judged to be unpreventable or preventable nonrespiratory incidents did not change significantly when comparing the two decades. The proposed favorable influence of pulse oximetry was further emphasized by the fact that both of the two preventable respiratory incidents leading to arrest in the latter decade occurred before pulse oximetry was introduced in 1984, but the incidence was decreasing before then as well. The nonrespiratory preventable causes were vaporizer errors or overdoses of inhaled agents in most cases, but included two cases of hyperkalemia and one instance of hypotension due to spinal anesthesia.

Equipment Failure and Misuse

Dr. M. E Vistica and her colleagues from the University of Washington in Seattle examined the role of equipment as a cause of damaging events and adverse outcomes. The data base originated with the A.S.A. Closed Claim Study and consisted of 67 malpractice claims (out of 1, 54 1) in which equipment was implicated as causing a negative outcome. Misuse of equipment was implicated in 57% of these, while 13% were due to equipment failure and 30% were unknown. Catheters, ventilators, and anesthesia machines were the most frequently misused. Breathing system disconnect was the most commonly identified adverse event. “Payments in these closed claims ranged from $430 to $3,500,000 with 70% of all equipment-related claims resulting in payment. following presentation of the paper, there was considerable discussion about “misuse” of the terms “failure” or I ” misuse” Since the end result is statistically the same in terms of payout, the semantics seem to be less important than the result.

Stress in the Attending Anesthesiologist

Dr. J. S. Naulty and colleagues from George Washington Medical Center reported on stress in attending anesthesiologists. Specifically, faculty members were monitored by cardiac monitors while supervising two junior or two senior residents to ascertain. Whether specific events can trigger stress; 2. Whether the true level of the residents influences the perceived level of stress; and 3. If this perceived stress correlates with ST segment changes or dysrhythmias. Attending anesthesiologists less than 50 years of age reported perceived stress more frequently when supervising junior than when supervising senior residents, although heart rate changes were more frequent when supervising senior residents. Attendings over 50 years of age reported a low frequency of perceived stress than younger attendings, regardless of the training level of the residents. The older attendings were able to correlate cardiac changes and perceived stress better than the younger attendings. Perhaps these differences are due to greater experience acclamation, or physiologic alterations associated with maturing in the older attendings.

Failure of Pulse Oximetry

Dr. B. S. A. Gillis and colleagues from the University of Washington examined and reported on the “failure’ rate of pulse oximetry in a PACU location. They accumulated pulse oximetry monitoring data on 2,937 patients. The pulse oximetry was said to have failed when there were two or more 15-minute periods when no values were noted. Using that criterion, the overall failure to display saturation values was approximately 1% with the majority (90%) of these failures at the beginning of the PACU stay. This rate compares favorably to the previously reported failure rate of 0.77% in the operating room. Since these failures of display of pulse oximeters in the PACU were probably at least in part due to emergence delirium and shivering, it was suggested that there may well be a “physiologic” problem with the patient when the device is “failing” to pickup and display pulse oximetry saturation values.

Preventing Transmission of Blood-Borne Infections

Prevention of transmission of AIDS, hepatitis, and other blood-borne infections was also the subject of a study by Dr. E. S. Greene of the Albany Medical Center, who offered arguments for the adoption of national standards of practice for infection control. Use of drugs from multidose vials for more than one patient was cited as a practice that should be abandoned. Among his suggestions for new measures that might reduce transmission of blood-borne disease was the adoption of needleless systems for intravenous administration of drug boluses and infusions.

Implementation of Infection Control Precautions

Dr. A. Rosenberg and colleagues from the Hospital for Joint Diseases Orthopedic Institute and the New York University Medical Center reported a follow-up of a 1989 abstract. They surveyed 653 attendees at a major meeting for infection-related perceptions and practices associated with anesthesia cam Although aware of hepatitis and HIV hazards, there was little improvement in the conduct of their practices. Forty-nine percent of anesthesia personnel still reused vasopressor syringes and 63% still reused common drug syringes. This compared to 47% and 6 1 %, respectively, in 1989. This was done despite 79% of the respondents believing that needles do get contaminated when injecting solutions into I.V. tubing. Fortunately, the investigators report some improvement in the wearing of gloves(47% in 1989 and 79% in 1990), vaccination against hepatitis (53% vs. 6 1 %), and recapping of needles (79% vs. 86%). While there has been some improvement in self-protection on measures, the data indicate that this is not so for patient-to-patient protective maneuvers. It was concluded education must continue.

Transfusion Risks

Dr. Rosenberg and others also retrospectively reviewed records of 233 patients undergoing total hip replacement to determine the most effective method for avoiding the risks induced by homologous blood transfusion in the perioperative period. They found that predonation of two units of autologous blood combined with intraoperative use of the Cell Saver and postoperative reinfusion of shed blood resulted in 96% of the patients avoiding homologous transfusion. Unless the patients had significant cardiovascular disease they were allowed a hematocrit as low as 20% before homologous blood was administered. Postoperative coagulopathy resulting from retransfusion of shed blood was treated with fresh frozen plasma when necessary.

Aspiration Risk

At the Mayo Clinic, Drs. 1. G. Weber, M. A. Warner and M. E. Warner found the incidence of pulmonary aspiration in 105,364 consecutive general anesthetics over three years to be 1:2509 (4: 10,000). Aspiration was rigidly defined as the presence of bilious or particulate material in the tracheobronchial tree. Significant risk factors were emergent cases, age less than two years, and higher ASA physical status classification. Of those 42 patients who did aspirate, one died on the table for other reasons, 26 (62%) developed no symptoms and had no sequelae, and the remaining 15 (36%) were monitored and treated in the intensive care unit for hypoxia and/or pneumonitis. Ten of those required mechanical ventilation. Six patients developed non-bacterial ARDS, two of whom died, and one patient developed Klebsiella pneumonitis. As a direct result of this study, the Mayo Clinic will now discharge from the PACU any patient who has not developed either hypoxemia, wheezing, coughing or radiographic abnormalities within two hours of clinically apparent aspiration.

Gas Machine Failure Detection

In an attempt to evaluate machine checkout habits with and without the use of the FDA checklist, Drs. M. G. March and J.J. Crowley from George Washington University utilized a traveling exhibit of two types of anesthesia machines with a total of eight faults that could be variably introduced. They tested 166 volunteer anesthesiologists on the machine of their choice. The study design called for the anesthesiologist to perform his own checkout procedure on a machine with four previously set faults, then to use the FDA checkout list to repeat his checkout after four different faults mere incorporated into the machine. A short multiple choice quiz was then administered, in order to correlate accuracy of checkout with knowledge about the specific faults. With only one of the faults, the N20-02 ratio override failure, did use of the FDA checklist improve performance significantly. The checklist appeared not to aid the volunteers in finding any other fault. Them was no significant correlation between finding faults and the type or length of practice, or with results on the quiz. The authors suspected that the FDA checklist was not of aid in finding most faults because the volunteers did not use it correctly, and suggest that machine and equipment “failures” might become less frequent as contributors to anesthesia incidents if that checklist is modified to make it mom “user friendly” and if anesthesiologist could be educated to implement correct checkout procedures.

Air Embolism Detection

Dr. J.J. van der Aa and colleagues for the University of Florida evaluated a prototype expert system in diagnosing air embolism as the cause for acute decreases in end-tidal C02 in sheep. Information from monitors of airway pressure, expiratory flow, expired C02, inspired gasses, ventilator settings, and pulse oximetry is fed into the expert system, executed on an IBM-PC/AT, and combined with individual patient data to pinpoint circuit disconnects, leaks, overventilation, or decreased C02 production as causes for abrupt changes in ETCO2. The system correctly identified air embolism without false positives or false negatives in every instance, and did so before alarms on the front-end monitors called attention to the situation. It was also efficient at picking up other causes of acute ETC02 changes, such as mainstem intubation, leaks, and changes in ventilation. The authors concluded that this expert system could be helpful in ruling out other causes of sudden hypocapnia, enhancing the anesthesiologist’s early response to the situation.

Inadequate Pre-Op ECG Evaluation

Dr. R. L. Wollman and colleagues from the Medical College of Virginia compared anesthesiologist’s ECG evaluations to official readings in all patients presenting for anesthesia at another institution over a two week period. They found that 6 1 % of ECGs did not have an official reading at the time the patient came to surgery, and the anesthesiologists’ interpretation of the ECG differed from the cardiologists’ eventual reading of 55.4%. Depressingly, of the ECGs that did have official readings on the chart, the anesthesiologists either failed to note the reading or recorded a different interpretation in 36.7%. In 34.7% of patients presenting without an official reading and 26.3% with a reading, anesthesiologists made errors that were judged “clinically important,” many involving acute ischemia and potentially leading to poor patient outcomes. The authors concluded that cardiologists are obligated to provide timely ECG interpretations on preoperative patients, and that anesthesiologists should be careful to review and document official ECG interpretations.

Evaluating “Mental Workload”

Dr. D. M. Gaba and T. Lee at Stanford University utilized a computerized random presentation of two digit addition problems to measure the mental workload of the anesthesiologist during the course of administering anesthetics in the operating room. Nine residents were tested during the administration of 19 anesthetics, and were instructed to put their patient care tasks first at all times. Delay in answering the problem or skipping them was interpreted as indicating that the resident was busy at other tasks. An independent observer recorded what the residents were doing at the time of presentation of the problems, announced by a beep on the computer. Delay or skipping was most commonly associated with performance of manual tasks, followed in order by attending interaction, conversation, and recording. 13% of the delays were associated with “doing nothing” indicating that mental activity alone is also a defined workload. The authors pointed out that this study reinforces the findings of Cooper, et. al. that teaching in progress in the operating room can be distracting and even a contributing factor in anesthetic incidents and, thus, may impact negatively on patient safety. This methodology could be utilized in comparing workload in different types of surgeries and anesthetic techniques, among different practitioners, and in measuring the effects of stress, fatigue, automation, and alarms on the anesthesiologist’s vigilance .

Allergic Reactions to Anesthetic Drugs

Dr. G. Occelli and colleagues, Hospital Saint-Roch, Nice Cedex, France, demonstrated that skin testing accurately predicted potential allergic reactions in a group of patients who have had a life-threatening anaphylactic reaction to anesthetic medications, both confirming sensitivity to the suspect drug from the past and accurately demonstrating lack of reaction to other medications that were used to subsequently administer uneventful anesthetics. This study suggests that skin testing of patients at risk for allergic reactions can identify potential drug reactions and aid anesthesiologists in their drug choices.

Airway Management

Drs. J. T. Roberts, H. H. Ali and G. D. Shorten from the Massachusetts General Hospital advocate using a bubble inclinometer to aid in evaluation of the airway. This inexpensive device was found to be helpful in determining laryngeal tilt which the authors believe correlates with graded scales of difficulty in using a #3 MacIntosh laryngoscope blade to visualize the larynx.

A prospective study from West Germany (Dr. A. Deller et. al.) suggests that the devastating consequences of airway difficulties can be minimized if anesthesia personnel carefully evaluate mobility of the cervical spine and any anatomic abnormalities of the face, neck and teeth. 8,2 84 patients were evaluated and 2 54 could not be intubated on the first attempt. I 0 1 patients proved difficult on repeated attempts. The most common predictors among these patients were facial anatomic abnormalities, difficulty visualizing the pharyngeal structures with maximal mouth opening, and a small mouth orifice.

Dr. R. Glassenberg and associates from Northwestern University noted a statistically not-significant decline in failed intubations for Caesarean sections from 11300 to 1/500 after the introduction of the fiber optic Laryngoscope. They suggest there may be an irreducible minimum number of failed intubations and note that half the cases occurred in women with normal appearing airways, and, therefore, were not anticipated

Anesthetic Catastrophe

Drs. R. L. Keenan, 1. H. Shapiro, and K. Dawson of the Medical College of Virginia analyzed their institution’s incidence of cardiac arrest in infants. There were 4,343 cases over eight years about evenly divided between anesthesiologists with and without pediatric fellowship training or equivalent. The pediatric anesthesiologists had no cardiac arrests in infants whereas the rate for the other group of anesthesiologists was 19 arrests per 10,000 patients. It was concluded that the attendance of pediatric anesthesiologists significantly decreased anesthetic morbidity in infants.

Drs. H. Schwid and D. O’Donnell of the University of Washington advocate the use of an on-screen computerized anesthesia simulator to give clinicians experience in the early diagnosis and treatment of malignant hyperthermia. Four case scenarios have been developed to allow anesthesiologists working at the simulator to observe the triggers, physiologic changes, and results of therapeutic intervention in MH model cases.

Sleep deprivation in anesthesia personnel was found to have an adverse impact on costive tasks. This study was performed by Drs. J. Zelcer, J. Manton, and J.D. Paull (Royal Women’s Hospital, Melbourne). Seventeen third-year registrars were evaluated during a 24 hour period when each of them stayed awake, whether they were working or not. Each subject also acted as his or her own control, being studied at another time during a normal daytime duty cycle when well rested. As one might expect, measurements of subjective phenomena such as fatigue and sleepiness increased. The results of objective tests of logical reasoning and short term memory worsened after 18 hours of wakefulness and this decrement achieved statistical significance at 24 hours. This work confirms that performance on tasks involving attention span, vigilance, problem-solving ability, short-term memory, information retrieval, and reasoning is negatively affected by 18-24 hours of continual work without sleep. As these are the skills involved in administering anesthesia, the patient safety implications are clear. The authors hypothesized that these measured deteriorations might contribute to increased probability of anesthetic mishaps caused by human error.

Postoperative Drinking and Driving

Dr. J. L. Lichtor and associates from the University of Chicago investigated whether patients having ambulatory surgery obeyed instructions about drinking alcohol or driving within 24 hours of anesthesia or intravenous sedation. They found that 11 of 37 patients either drank, or drove themselves, but none did both, within 24 hours.

In a further study, the same investigators asked the question, “Does it matter?” They performed a placebo-controlled, double blind study on volunteers, giving them either intravenous saline or fentanyl with midazolam. Four hours later the same subjects drank a beverage that either did, or did not contain alcohol. The subjects completed a set of psychomotor performance tests before, during and after this sequence. It was concluded that the effects of short-acting drugs used in ambulatory surgery had probably dissipated by the time the patient arrived home, and so would be unlikely to potentiate or interact with alcohol.

Critical Incident

Intraoperative carbon monoxide toxicity(peak levels of CO Hb of 12-29%) was demonstrated in three patients underpins enflurane anesthesia. While the exact cause or causes of this CO accumulation remain unknown, Dr. R. E. Moon and colleagues (Duke University Medical Center, Durham) noted that this very unusual CO accumulation occurred each time in the risk case on Monday morning. Thorough investigation led to circumstantial evidence suggesting the C02 absorbent or anesthesia circuit as the source, possibly the result of some slow low-grade chemical reaction over the week-end. This report underscores the importance of flushing anesthesia circuits prior to their use

Non-Cognitive Factors Play a Major Causal Role in Critical Incidents

M. F Rhoton, Ph.D. and her colleagues from Case Western Reserve University investigated the role of non-cognitive variables in the genesis of critical incidents involving anesthesia residents. They identified lack of conscientiousness and loss of composure as being most frequently associated with critical incidents. In contrast, an inadequate knowledge base was rarely identified as a causative factor.

Drs. A. DeAnda, D. M. Gaba, and I Lee continued their studies of anesthesiologists’ responses to critical incidents using a comprehensive anesthesia simulator. This time they studied an “experienced” group of private practitioners and faculty members and compared these results to the previous findings in a group of residents. Their subjects were presented with six simulated problems; short circuit hoses, a kink in an IV fine, endobronchial intubation, atrial fibrillation, airway disconnection and a cardiac arrest. They then measured the response times for detection and the beginning of definitive correction of these problems. With some exceptions, there was more rapid detection and correction with increasing physician experience. This was particularly true when first year residents were compared with all others. However, even within the experienced group, there was marked variability, particularly in the appropriateness of the response to the simulated cardiac arrest.

Blinded Study of Pulse Oximetry

Dr. J.T. Moller and colleagues of Herlev Hospital in Copenhagen, Denmark presented a fascinating study that would be difficult to duplicate in the USA Using pulse oximetry, they compared the incidence and duration of hypoxia, in both the recovery room and the operating room, between a group of patients in which the saturation values were available to care-providing personnel and a group in which the values were blinded from the care providers. It was considered feasible to attempt this study m Denmark because pulse oximetry had not yet become a de-facto standard of care them The investigators found the incidence and duration of episodes of hypoxemia were significantly less in both the operating room and the recovery room when oxygen saturation values were available to the caregivers. They also found that hypoxemia was more frequent, more severe, and more prolonged in the recovery room than in the operating room. Their findings strongly support the value of pulse oximetry in risk reduction during anesthesia. Moreover, they indicate that oxygen saturation monitoring should continue during recovery from anesthesia.

Is MAC Safer Than General or Regional for ESWL?

Anesthesiologists often state their belief that for many procedures local anesthesia is not inherently safer than general or regional anesthesia. This belief was supported by the results of a study of post-procedure morbidity in ESWL patients. This study by Drs. A.L. Kovac and J.V. Mangold of the University of Kansas revealed a higher incidence of post-procedure pain, hypertension, and nausea and vomiting with monitored anesthesia care than with general or regional anesthesia.

Drs. Owens (Washington University, St. Louis) and Polk (University of Chicago)moderated ASA scientific sessions. Drs. Aukberg (University of Pennsylvania), Vender [assisted by Dr. Gilbert] (Evanston (IL) Hospital), and Zeitlin (Lahey Clinic, Burlington, MA) are members of the APSF Newsletter Editorial Board and reported on the poster sessions.