Other Approaches to Discern the Root Cause of Morbidity and Mortality

The accumulating data clearly demonstrate that risk directly attributable to anesthesia has declined over time. The cause for this reduction in mortality is unclear. Numerous factors have been implicated in the improved outcome, including new monitoring modalities, new anesthetic drugs, and changes in the anesthesia workforce. However, it is difficult to document reduced risk related to any one factor (issues related to manpower are addressed later). Although using newer monitoring modalities, particularly pulse oximetry, would be expected to improve outcomes, no randomized trial has been able to document such a conclusion.^[81] This limitation supports the need for continued monitoring of complications and their root cause with studies such as the ACCS.

Studies similar to the CEPOD have not been performed in the United States, most likely because of the legal system, and information related to perioperative mortality must be obtained from other sources. The Professional Liability Committee of the ASA conducted a nationwide survey of closed insurance claims for major anesthesia mishaps. In the ACCS, fatal and nonfatal outcomes were reviewed. Among the fatal events, unexpected cardiac arrest during spinal anesthesia was observed in 14 healthy patients from the initial 900 claims.^[82] The cases were analyzed in detail to identify patterns of management that might have led to the event. Two patterns were identified: oversedation leading to respiratory insufficiency and inappropriate resuscitation of high spinal sympathetic blockade.

905

Tinker and coworkers^[83] queried the ACCS to determine the role of monitoring devices in the prevention of anesthesia mishaps. They reviewed 1097 anesthesia-related claims and determined that 31.5% of the negative outcomes could have been prevented by the use of additional monitors, primarily pulse oximetry and capnography. Injuries that were deemed preventable with additional monitoring resulted in dramatically more severe injury and cost of settlement than did those judged nonpreventable with additional monitoring. In almost 90% (305 of 346) of the preventable cases, at least one clinical sign of abnormality was identified by the existing monitors.

Caplan and colleagues^[84] reviewed the ACCS for respiratory events ( Table 24-12 ). These claims represented the single largest class of injury (34%), with death or brain damage occurring in 85% of cases. They identified inadequate ventilation, esophageal intubation, and difficult tracheal intubation as the primary causes of respiratory events. Most of the outcomes were thought by the investigators to be preventable with better monitoring ( Fig. 24-4 ). Although no randomized trial has demonstrated the value of pulse oximetry or capnography, an analysis of claims supports the value of such monitoring. This analysis also formed part of the basis for the ASA Task Force on Management of the Difficult Airway guidelines and algorithm.^[85]

Cooper and colleagues^{[86] [87] [88] [89]} approached the problem of sample size for determining perioperative morbidity by identifying "critical incidences," which were defined as those that were preventable but could lead to undesirable outcomes. This definition included events that led to no or only transient effects for the patient. The investigation involved collecting data on anesthesia-related human errors and equipment failures from anesthesiologists, residents, and nurse anesthetists. In a series of reports, the authors identified frequent incidences, such as disconnections in breathing circuits, and causes of discovery of errors, such as intraoperative relief. They confirmed that equipment failure was a small cause of anesthesia mishaps (4%), whereas human error was dominant. They suggested that future studies of anesthesia-related mortality and morbidity should classify events according to a strategy for prevention rather than outcome alone.

Buffington and coworkers^[90] studied the ability of attendees of an anesthesia meeting to identify five faults intentionally created in a standard anesthesia machine. A survey was distributed, and the answer sheets were scored with respect to the number of correct answers. Only 3.4% of the respondents found all five faults. The average number of faults detected was 2.2. The professional background of the participants did not influence the ability to find the faults; distributions were similar between physicians and CRNAs. There was a small improvement in fault-finding ability among those with more than 10 years of experience. Studies such as these highlight the problem of the practitioner's ability to identify conditions that may lead to anesthesia mishaps. Whether improved technology or education can reduce some basal level of mishaps is unknown.