International Committee Focuses on Eastern Europe
The International Committee for the Prevention of Anesthesia Mortality and Morbidity (ICPAMM) met in Warsaw, Poland on September 9, 1990, to discuss methodologies and data related to adverse anesthesia outcomes. This was a continuation of earlier meetings to disseminate ideas on how such studies can best be conducted. A special theme of this meeting was promotion of similar studies in Eastern European countries, where previously there appears to have been less openness to connecting these kinds of data.
Randomized Trial of Pulse Oximetry
Dr. Jakob Trier Moeller reported on the Danish-American Randomized Trial of Pulse Oximetry The objective of the study was to test the hypothesis that routine use of pulse oximetry would reduce the rate of complications. The Planning Group estimated that ten thousand patients would be required in a control and experimental group to attain a power of 90 percent for a meaningful reduction in pulmonary and cardiovascular outcomes.
The investigative team had to address the serious question of the ethics of randomizing considering the presumed benefit of pulse oximetry. After much debate, there was unanimous consensus that this was reasonable considering that pulse oximetry was, at the outset of the study, rarely used in Denmark. The decision was approved by the ethics committee of each hospital.
The study included all elective, acute patients receiving either general or regional anesthesia expected to last at least 20 minutes. Excluded were children, cardiac surgical, neurosurgical patients, outpatients, and most patients unable to give consent. Emergency patients seems included and consent sought postoperatively.
Randomization was achieved by drawing envelopes that assigned oximetry to specific operating room after patients had already been scheduled. If a patient was reassigned to another room, the original pulse oximetry assignment was maintained for that patient.
Data on complications was collected intraoperatively and postoperatively up to the seventh hospital day. At the close of the study, 21,600 patients had been randomized, with essentially equal participation in each group. Approximately 20 percent of all operative patients meeting inclusion criteria were not randomized. The vast majority of these were due to either a failure to acquire consent or merely the anesthetist’s inadvertent failure to perform the randomization. There was no difference between patients randomized or not randomized into the study regarding ASA patient status or concurrent risk factors.
Data entry was validated at three different points in the collection sequence. Dr. Trier Moeller noted that, while data collection has been completed, the analysis is preliminary. Them does not, at ” point, appear to be a significant difference in overall outcomes between the two groups, although specific complications have not yet been examined individually. There also does not appear to he a difference in the rate of hypotension or hypovolemia.
During the discussion, Dr. William Runciman (Australia) and Dr. Michael Rosen (U.K.) both questioned the hypothesis and the expectation that differences in serious outcomes would be significant in such a small sample Dr. Cooper responded that the hypothesis was partly based on the unexpected finding in an earlier study. Dr. Trier Moeller also noted that psychological testing of a subset of patients was performed based on the hypothesis that pulse oximetry may prevent hypoxic events which lead to subtle changes in cerebral function.
Mortality and Morbidity Studies in the U.S.A.
Dr. Ellison C. Pierce, Jr. and Dr. Gerald Zeitlin jointly reported on several activities in the United States. Dr. Pierce first noted that there have been dramatic decreases in rates for malpractice insurance for anesthesiologists in recent years. He explained the relativity scale by which physician specialties are assigned a number from one to seven according to the severity of their exposure to financial losses from malpractice suits. Anesthesiologists started the 1980’s with a relativity of five. The St. Paul Insurance Company, which represents 15 percent of all U.S. physicians, has reduced anesthesiologists to relativity of three as has the Massachusetts state insurance underwriting organization. The Harvard affiliated Hospital Insurance Group has reduced anesthesiologists to a relative rate of 2.5, which is expected to be reduced even lower soon. For Harvard-insured anesthesiologists, this represents a decrease of from $21,000 annually to the current rate of $8,000 per physician.
Dr. Pierce also reported on the recently completed pilot study of anesthesia outcomes jointly conducted by the Center for Disease Control, the American Society of Anesthesiologists, and the American Association of Nurse Anesthetists. The objective of the study was to test the feasibility of a larger surveillance system and assessment of adverse anesthesia outcomes in the U.S. Five hospitals of varying characteristics participated. Mortality and serious morbidity, including myocardial infarction, aspiration, and neurologic injury within 48 hours of operation were documented by different means in each hospital. An independent panel including two anesthesiologists and one surgeon reviewed all cases of adverse outcome and assigned anesthesia responsibility according to several different scales. The entire study included 8,000 patients of which 21 suffered some adverse outcome. Only one was assessed as being due to anesthesia alone. Five other adverse outcomes were believed to include anesthesia as one of several causative factors.
The CDC concluded as a result of this study that it was indeed feasible to successfully design and conduct a larger surveillance involving 300 hospitals in the U.S. The total cost was estimated to be $15,000,000. In February, 1990, agreement to pursue funding for this study was achieved among the participating groups. However, objections were raised by some individuals about the lack of the study’s innovation, its great expense, and its failure to adequately address important methodologic questions such as patient case mix. Also, the U.S. government recently established an agency with the objectives of setting practice standards and collecting outcome data. It is now unlikely that the larger scale study will be conducted as originally planned. Efforts are underway to modify the study objectives and methods so that some national surveillance mechanism, acceptable to all parties can be established. Dr. Pierce also noted that the joint Committee on Accreditation of Hospitals Organizations (JCAHO) is currently entering the second phase of pilot studies to establish a surveillance mechanism for anesthesia quality in U.S. hospitals. It will be collecting data on eight different outcome indicators including death, MI, and re-intubation in the recovery room. These and other efforts represent the interest in the U.S. of establishing some form of quality and outcome measures in anesthesia although substantial obstacles remain in the way of doing that. Nevertheless, Dr. Pierce believes that some mechanism surely will be established.
Dr. Zeitlin then reported on some methodologic aspects of the American Society of Anesthesiologists-sponsored study on anesthesia malpractice claims (closed claim study). Dr. Zeitlin noted that he was representing Dr. Fred Cheney, Dr. Richard Ward and Dr. Robert Caplan, who constitute the committee conducting this study under the auspices of the American Society of Anesthesiologists Committee on Professional Liability. Concentrating on the issue of inter-rater reliability, Dr. Zeitlin first noted that past studies of closed claim did not assess judgments about the anesthesia in relationship to the outcome. Contrary to this, the closed claims study required that reviewers of data describe exactly how they made their judgments about who was responsible for the event. In particular, they describe for every case a possible or probable mechanism by which the injury occurred. They also judged whether the standard of care was equal to, below, or above that which would be given by a prudent anesthesiologist in the U.S. In the precursor to this study by Dr. David Davis, only the standard of care in the local community was used as a basis for judgement.
In an earlier paper arising from the closed claims study, it was shown that patients were given a payment for injury in 40 percent of cases in which the care had been judged satisfactory by the closed claims reviewers. This suggests that the American legal system results in compensation often even in cases where the physician did everything reasonable to prevent an adverse outcome.
Critical to establishing its credibility, the closed claim study carefully assessed the ability of reviewers to come to similar conclusions about the appropriateness of care, the presence of human error and the potential for prevention of the event. Using 48 cases and 42 reviewers, it was found that reviewers reach a more than reasonable degree of agreement, where agreement is defined as by at least 75 percent of reviewers of each case. Thus, it has been shown that a group of peers is an effective way to judge important issues related to adverse anesthesia outcomes after the fact using retrospectively collected data. Dr. Cooper noted the importance of conducting such validity assessments whenever judgments of culpability are made.
Australian AIMS: Update
Dr. Runciman reported on the two experiences in the AIMS study. He reminded that they consider an incident to be any event during anesthesia that could have caused harm to the patient. There are now 70 hospitals and 5,500 anesthesiologists participating, representing an estimated one half million anesthetics annually. Over 1,000 reports have now been reviewed and analysis of some issue conducted. For instance, he noted that approximately 25 percent of incidents occurred during induction and 50 percent during the maintenance phase of anesthesia. Over 50 percent of incidents in this study involve some interaction with equipment. They have focused in particular on the involvement of monitors in detection of incidents. Pulse oximetry, capnometry, and ECG monitors each played a role in approximately 80 of the incidents. Of all incidents, 156 involved some major physiologic change, 42 some morbidity and 21 ended in death.
Of particular interest has been the fact that endobronchial intubations are the most frequent cause of desaturation events intraoperatively. As an example of how AIMS should be used, Dr. Runciman expects that some data should change the algorithm by which people react to desaturation. Surprisingly, there are still numerous incidents of disconnection and hypoxic mixtures occurring which are detected by oximetry He noted that many anecdotes of dose calls are still never reported in the literature.
There are still many breathing circuit incidents occurring among which tracheal tube problems are the most frequent (107 in this set of data). Confirming earlier suggestions, disconnections in particular are frequently not detected by pressure based alarms.
There has been a substantive increase in the use of monitoring since the start of the Australian study in early 1988. Pulse oximeter use has increased from approximately two-thirds of all cases to 98 percent. Capnometry has increased from approximately 50 to 70 percent of all cases. Dr. Runciman noted that, in 93 percent of all incidents, detectable incidents during induction were detected by either pulse oximetry or capnography. Of the entire data-set, approximately 50 percent of incidents were detected by some form of monitoring.
These data have be-en used to great effect in local and regional quality assurance meetings and as motivation for at least 15 modifications of equipment. Study organizers are seeing the loop being closed as evidenced by feedback from participants who note how AIMS has affected their practice habits.
In response to a question, Dr. Runciman said that there have been exceptionally few incidents reported that were caused by monitors.
Update on Confidential Inquiry on Intraoperative Deaths
Dr. John Lunn reported that CEPOD study continues to promote the joining of forces between anesthesiologists and surgeons seeking causes and recommendations for prevention of deaths. Each year, the focus will be on a different sample group. In 1989, deaths of children less than ten years of age were examined. Each surgeon also provided information of the management of “index” cases which served as controls for the deaths. For the entire national sample, there were 2 1,000 deaths of which 417 were less than ten years of age. Unfortunately, no denominator data is yet available since that is obtained from the government and will not be completed until early 1991. Dr. Lunn report that there continues to be exceptional cooperation on the part of both surgeons and anesthetists in providing information, although the return rate on information about deaths was 70 percent from surgeons and 62 percent from anesthetists.
The most important finding of the analysis of deaths in this year’s sample was a disproportionate number of cases involving physicians with only “occasional” experience with pediatric patients. For children less than six months of age, physicians with less than 20 cases per year experience were involved in 28 percent of the deaths.
Data from the study also points to increasing utilization of monitors during anesthesia. In the 1989 sample, fully 88 percent of all intraoperative patients, were monitored by an ECG and 65 percent with pulse oximetry. The collection of data will likely lead to standards, for instance requiring that patients be seen preoperatively and that some forms of monitoring be required rather than recommended.
In 1990, the focus of CEPOD was on a random sample of deaths in England, North Wales and Ireland, comparing them with index cases. In response to a question from the audience about cost of the study, Dr. Lunn noted that in the first two years, approximately 1.2 million P was allocated by the national government. Unfortunately, the status of funds for future years is not guaranteed.
Continuation of the “Manitoba” Study
Building on her reports at the past two meetings, Dr. Marsha Cohen described the continuing investigations of anesthesia outcomes in Manitoba. Her most recent efforts were detected toward asking the question “Can patient outcomes be compared between hospitals?” A related question was “Can morbidity from anesthesia be measured in some reproducible way to be useful for quality assurance?” Dr. Cohen briefly summarized the anesthesia follow-up program at the Health Sciences Center in Winnipeg between 1975 to 1983. Complications within 72 hours of anesthesia were reported via data collected from charts and patient interviews, including information on 40 major and minor anesthesia outcomes. This study was limited in that it only involved one hospital and spanned a time during which occurred dramatic changes in anesthetic drugs, monitors, techniques, patients and surgical procedures. Also, many definitions of outcome were vague and information was not differentiated to ascertain its source from either the patient interviews or chart reviews.
The latest study was begun in 1987 and involves four hospital departments of anesthesia. It employs an occurrence screening technique in which information is gathered on every patient directly from the anesthesia record, from chart reviews by nurses and from patient interviews. Data has been collected from over 35,000 anesthetics. The disadvantage of this technique is that it is more expensive than sampling only some patients. The advantage is that there are fewer problems with missing data, major and minor events are included and training of staff to complete forms is easier. Spin-offs include the availability of information on resource utilization and practice profiles. Also, information for resident education and performance appraisal is provided. The approach provides good public relations for the department of anesthesia via patient contact after each procedure Also, this is a good approach to getting a “first look” before undertaking the even more expensive randomized control trial.
In the analysis of these data, inpatients and outpatients are being examined separately. There is a case mix adjustment for age (three groups), physical status score (ASA PS I to 2 and ASA PS 3 to 5), sex, length of anesthesia procedure and emergency versus elective status of procedure Hospitals are compared via a multiple logistic regression using hospital A as the reference.
The preliminary results suggest that there are substantial differences found for virtually all 60 outcomes across the four hospitals. Dr. Cohn addressed the question of possible causes of these differences. There was some suggestion that anesthetists in one hospital were less compliant in completing data. But, no one hospital was consistently different for all complications. There does not seem to be differences in institutional case-mix. Differences in institutions, e.g., recovery room size and criteria for use, nursing/patient ratio, monitoring equipment availability and differences in physician practice patterns, e.g., anesthetic techniques and drugs for the same surgical procedure are possible explanatory factors for the differences. These are being examined in the analysis.
There were barriers to conducting this research. Anesthetists are not particularly good at completing forms. Despite training, there were problems in interpreting definitions of the variables and coding complications. There were the inevitable unexpected problems. The four hospitals could not agree upon the exact same anesthesia record although they were quite similar. A threatened nursing strike reduced data collection from what was expected.
Despite extensive preparation and attention to detail, Dr. Cohen noted that it is difficult to compare “soft” anesthesia outcomes across institutions. Still, the variations they have identified in anesthesia care are not unexpected considering that similar variations are found in medical and surgical care
A question was asked about the findings of differences in non-life-threatening outcomes, particularly patient satisfaction. Dr. Cohen reported that most patients were satisfied with their care as measured on a one to five scale. There were some interesting findings such as that patients with headaches were less likely to be satisfied and those called at home were more likely to be satisfied. Having trained nurses who asked the same questions of all the patients added to the reliability of the measurements.
Anesthesia Accidents, Near Accidents, Faults and Complications in One Hospital in the Netherlands
Dr. Vimal Chopra reported on the findings of a committee organized at the University of Leiden. AU medical and nursing staff are required to report all accidents, faults and near accidents and complications to the committee. There are specific definitions for each event. The results of ten years of data collection were described. There were 148 reports to the committee for the over 1 13,000 anesthetics during this period. The most common events reported were dental injuries, cardiac arrests, drug related problems, e.g., drug swap, overdose, and nerve injury. fourteen events met the definition of “complication.” In examining causal factors more carefully, the familiar factor of “failure to check” was frequently noted. The overall rate of fatalities related to anesthesia was one in 16,000. Dr. Chopra noted the difficulty in drawing conclusions from such retrospective analysis and noted their expectation of more useful information from their ongoing prospective study.
Dr. Cooper described how accident investigation techniques can be used as still another form of outcome study. In particular, this method is powerful for locating new mechanisms of injury and for altering behavior locally. An example of accident investigation techniques had been published in an article describing a bum from a neuromuscular stimulator. For accident investigation to he most effective, preparations must be made to collect information immediately following a critical event.
Dr. Jan Davies spoke of her work with Dr. Rob Lee, Director of the Bureau of Air Safety Investigation in Canada. They have developed a system for examining anesthetic accidents using investigative techniques developed for air craft accident investigations. They have investigated 12 accidents to date.
Study of Maternal Deaths in Poland
Dr. Teresa Tribus described a study of maternal deaths in Poland. Their approach was retrospective, using a modification of the data collection instrument described by Lunn and Mushin. Using information gathered solely from medical and anesthesia records, the investigators determined if anesthesia was totally, partially, or not all responsible for the death. Contributory factors were also identified including poor judgment, lack of vigilance, incomplete knowledge and experience, etc.
Records from the southeast region of Poland during the period 1976 to 1986 were examined. This represents approximately 15 percent of the Polish population and 1, I 00,000 deliveries. A total of 363 deaths were found of which 17 were determined to be primarily caused by anesthesia and 3 7 of which anesthesia was at least partially contributory. Of the deaths for which anesthesia was primarily responsible, the most frequent cause was aspiration and inability to intubate. For deaths to which anesthesia was partially contributory, inadequate treatment of hypovolemia and wrong choice of anesthetic technique was identified as the predominant causes.
Dr. Tribus noted the recognized difficulty in identifying cause of death from retrospective records. Still, the investigators believed that 76 percent of deaths to be attributable to anesthesia and 54 percent of deaths partly attributable to anesthesia should have been avoided. Overall, the unavailability of adequately trained anesthesiologists was believed to be the most important deficiency in the system.
New South Wales (Australia) Study Committee Report
Dr. John Warden briefly noted the continuing work of the New South Wales (Australia) Committee on Anesthetic Deaths. Since 1983, the Committee has examined 141 anesthesia related deaths during which time there were approximately 500,000 anesthetics per year for a rate of 1125,000 procedures. This includes deaths in which the Committee attributed solely to anesthesia (class 1), to which anesthesia alone partly attributed (class 2), and deaths for which anesthesia and surgery contributed (class 3). Approximately 50 deaths were in the Class I category for a rate of approximately 1/80,000 procedures.
Dr. John Lunn commented that such estimates should he interpreted with caution considering that the denominator is not carefully documented.
Study of Anesthetic Cardiac Arrests in West Berlin
Dr. Jurgen Link reported on data that he has been collecting in his hospital in West Berlin since 1973. Information has been documented on all patients undergoing anesthesia. He has recently focused on the incidence of cardiac arrest, comparing the incidence and contributory factors for the periods 1973-1980 and 19W 1 989. The rates of cardiac arrest for the two periods were respectively 1:2079 and 1:2439 procedures. Not unexpectedly, the rate is much higher for the very young, the very old and patients with pre-existing diseases. He has identified respiratory insufficiency, error during emergence, faulty apparatus monitoring, insufficient fluid replacement, inadequate treatment of hemorrhage, and insufficient handling of complications as factors exhibiting notable changes between the two time periods. There were reductions in all of the above with the exception of the last category.
How should mortality and morbidity and quality assurance be studied with minimal resources?
The group discussion began with a report from Dr. Gribomont about the system of event reporting and mortality and morbidity conferences as practiced in his hospital in Brussels. That system is believed to be a very inexpensive and effective method for study and quality assurance The group agreed that the term “incident” was preferable to I ‘critical incident.” The term “critical” could imply “criticism” and therefore, is less desirable. Many such systems have been devised to facilitate collection and analysis of incidents. The prime virtue of such systems was that individual groups could use the information to improve their own performance. It was noted that measuring quality means identifying good things as well as bad. Working to identify such events and promote them is believed to be a good objective.
Patient Safety in the 1990’s
Dr. Davies identified four major areas covered by the group. 1.) Education the need for uniformity in training including the problem that will be faced in Europe with the consolidation of multiple standards. There was controversy about the impact of continuing education on patient safety. The use of simulators is seen as a potential value with the caveat that recognition of problems will still be problematic. 2.) The need for feedback to anaesthetists and & public about what are minimal standards and about anesthetic outcomes. 3.) Equipment: greater efforts in developing uniformity among manufacturers. The use of automated record keeping is likely to increase and should be tied to quality assurance. Monitoring of depth of anesthesia and EEG monitoring is expected to increase as well. 4.) Well-being of the anaesthetist great differences exist between countries regarding the work load of anaesthetists. In some cases, vacations or any time off is rarely granted while in other countries it has become a common practice to allow the anaesthetist to be relieved from responsibilities the day after a night on call. Greater emphasis is expected to be placed on the impaired anaesthetists, e.g., abuse of alcohol or narcotics.
What strategies for reducing risk are most effective given limited resources?
Despite the fact that this group consisted of only six individuals, there was quite a spread of opinion on this question. Dr. Marx from Uruguay reported on an excess of anesthesiologists to the point of often having two anesthesiologists in one operating theatre. The opposite situation exists in countries such as Poland and Romania where there are very few trained anesthesiologists. Yet, there was agreement that the most pressing issue was improved training for anesthesiologists, which would generally be cheaper than purchasing technology because of the problem of foreign currency exchanw. Dr. Jastrzebski personally emphasized the need for more training on regional techniques. The monitoring standard felt to be most important was the presence of an anesthetist at all times. The technology believed to be of most importance is pulse oximetry. Legal protection will be needed to permit the collection of outcome data related to errors given the concerns of people about retribution. Dr. Cooper noted the broad contrast between the group discussions. In one group the emphasis was on the need for training individuals in the most rudimentary concepts of anesthesia while in another group discussion focused on technologies for monitoring depth of anesthesia. Perhaps, developing countries need purchase only the most highly leveraged monitor and put other resources toward basic education. To this end, the more developed countries could provide such educational materials and methods, e.g., video cassettes and video cassette players. The developing countries now have the unique opportunity of studying the impact of these various safety efforts as they are introduced and have an effect.
The next meeting of ICPAMM is planned for June 1992 in the Hague, Netherlands, coincident with the next World Congress of Anaesthesiology. Attendance is by invitation, but suggestions for presentations from new countries and investigators are welcome.
Dr. Cooper, Massachusetts General Hospital, is a member of the APSF Executive Committee and co-founder and organizer of ICPAMM.