A Tale of Two Stories
National Patient Safety Foundation

Report from a Workshop on Assembling the Scientific Basis for Progress on Patient Safety

 

Day One Contrasting Cases Uncelebrated Case #1

Uncelebrated Case #1: Bile duct injuries during laparoscopic cholecystectomy The shift to laparoscopic cholecystectomy where surgeons use a small video camera to indirectly observe the process of removing the gallbladder, has been accompanied by an increase in bile duct injuries with significant consequences for patients. Studies of the basis of surgical expertise at this task revealed the need for new critical perceptual and cognitive skills, in particular, a kind of judgment under uncertainty and risk when considering to convert to an open procedure if the anatomy cannot be clearly visualized. The research results identify opportunities to improve performance through new perceptual aids, new techniques for training judgment under uncertainty, and needed changes in organizational behavior. Background Technological change sometimes brings new problems that demand attention. The growth of laparoscopic cholecystectomy provides an excellent example. This form of minimally invasive surgery to remove the gallbladder has largely replaced the older, "open" cholecystectomy. The laparoscopic procedure involves use of a small video camera to provide a view of the gallbladder and surrounding structures as these are manipulated with instruments that penetrate the abdominal wall. The procedure involves a few small incisions rather than one large one and, for this reason, allows faster recovery, shorter hospital stays, and less pain than the older "open" technique. The widespread adoption of laparoscopic cholecystectomy provided obvious benefits to many, but it was accompanied by a significant increase in incidence and severity of injury to the common bile duct (Way, 1992). Injury to this structure, which carries bile from the liver to the intestine, is a catastrophic accident that can lead to protracted hospitalization, multiple surgeries, and even liver transplantation. Bile duct injuries are not unique to laparoscopic cholecystectomy; they can occur even with the open procedure. But decades of experience with open cholecystectomy had reduced the rate of bile duct injury to a very low level. The new rash of bile duct injuries associated with laparoscopic cholecystectomy was troubling, especially because the benefits of laparoscopic surgery were so compelling and the demand for this form of cholecystectomy was intense (Way, 1992). It was also troubling because the severity of the injury increased. Those suffering bile duct injury during laparoscopic cholecystectomy tended to have devastating injuries of the sort that might lead eventually to liver transplantation. Furthermore, bile duct injury was frequently compounded when the original surgeon attempted a repair procedure. The repair was more likely to be successful if it was carried out by a specialist with experience at biliary duct repair (Stewart and Way, 1995). The increase in severity and frequency of injury was detected fairly early during the expansion of laparoscopic cholecystectomy. Epidemiological work suggested that surgeon inexperience played an important role; surgeons less experienced with this technology had a higher rate of injury. Clearly the remote viewing and manipulation of the tissues that goes along with the laparoscopic technique altered the mix of optimal surgical skill in ways that increased the risks for injury. Experts with laparoscopic technique have pointed out that minimizing the risk of bile duct injuries depends on identifying the bile duct anatomy before cutting, clipping, or cauterizing any structure. However, bile duct injuries occur only in the setting where the operating surgeon thinks that the anatomy has been identified -- no one cuts the common duct knowing that it is the common duct! Ideally, practitioners should take steps to definitively identify the anatomy when the structures (ducts and arteries) can not be identified clearly. One option is to convert the procedure from laparoscopic to open to permit direct handling of the tissues and a direct binocular view of the anatomy, but this decision sacrifices the advantages of the laparoscopic procedure. The decision to convert is a new judgment under uncertainty and risk. The research has explored the factors that affect the difficulty of making this judgment and what constitutes expertise at this judgment. The Research Dominguez and colleagues investigated the nature of surgical expertise during laparoscopic cholecystectomy, in particular the judgment to convert to an open procedure. To study this judgment they had to create conditions where visualizing the anatomy is challenging. A challenging case is needed to observe how surgeons evaluate whether to continue laparoscopically or to convert to an open procedure. Surgeons do not confront this decision at a single, well-defined moment. Rather the issue emerges over time as the physician explores the situation and confronts difficulties. Thus, Dominguez et al., used a complicated case that included a number of difficulties that interacted and fed upon each other. Because laparoscopic surgery is done using video displays of the surgical field, they used a videotape of a difficult procedure to present the case to 20 surgeons and surgical residents. For this type of surgery, the videotape record contains precisely the same visual information that was available to the surgeon who actually did the operation, presented in precisely the same way. As each surgeon viewed the video of the surgical field as the case progressed, they commented on the nature of the case, the probable future course and, specifically, the comfort level they were experiencing with continuing the case laparoscopically, as opposed to converting to an open procedure. The investigators also stopped the video at specified points to ask the surgeon questions on his or her assessment of the situation. The study, like any other exploration of people encountering and coping with real problems, produced a data set that was challenging to analyze. The basic results are the step by step assessments of each participating surgeon as the procedure evolved. These are built up from the comments of the participants as they viewed the videotape, linked to the characteristics and difficulties of the case as it unfolds moment by moment. This kind of data analysis, called protocol analysis, examines the process by which someone solves a problem. It is and has been the basic technique used to study problem solving. Dominguez et al.'s protocol analysis was made richer by including data from a variety of practitioners with varying degrees of experience. A variety of results emerged from analysis of the surgeons' commentaries on the videotaped cases. Basically, it reveals that the conversion decision is a difficult tradeoff judgment. The data show some of the visual cues that trigger consideration of whether to convert to an open procedure. The results provide insights about how laparoscopic visualization is limited when compared to direct binocular vision and handling of tissues. The results also provide some insight into the circumstances in which people may make the tradeoff inappropriately, for example, by continuing the procedure laparoscopically even in the face of increasing uncertainty. Significantly, the research goes beyond merely identifing visualization as a key factor. It does much to define what needs to be enhanced, where the critical visual cues reside, and how the cooperative work of the surgical team is organized around visualization. These kinds of results are important because they suggest different interventions that can improve performance. Information about difficulties in visualizing the anatomy, combined with knowledge of human perception, suggests that perceptual aids to enhance surgeons' ability to visualize the anatomy through a two-dimensional video view of the surgical field would be valuable. The method used by Dominguez et al. also provides insight into how training for surgeons should be modified to prepare them for the new judgments that new technology demands. By confronting videos of cases specifically chosen to display the variety of factors that play into the decision to convert to an open procedure, surgeons can expand their expertise. This technique, sometimes called exploratory learning, is now being used with simulation technology to train high-performance skills (e.g., Feltovich, Spiro and Coulson, 1989; Howard, Gaba, Fish, Yang, and Sarnquist, 1992). While Dominguez's study is closely focused on perceptual and cognitive factors, the results have much broader implications. When patients, referring physicians, administrators, and colleagues discover that a case has been converted from laparoscopic to open, their response influences future surgical decisions. This is a specific example of the way reactions of larger organizational and professional groups to practitioner decisions in specific cases has a strong influence on the way people make judgements in the face of uncertainty and risk. Organizational responses to cases are an important part of the practitioners' world and encourage them to adjust their approach to risk. Indeed, this organizational response becomes one point of certainty in a world where future outcomes are inherently uncertain. The core issues of bile duct injury in laparoscopic surgery are mainly how individuals and groups cope with uncertainty. While they did not study these factors directly, Dominguez et al.'s results show that it is impossible to formulate narrow, rule-based approaches to the problem of bile duct injury in laparoscopic surgery. They show that injuries arise from the same sources that produce the (usual) success of this method. Finally, the technological change represented by laparoscopic techniques raises new questions about skill that have organizational and professional implications. As a new generation of surgeons emerges, will they have experience only with laparoscopic techniques? Will they be reluctant to convert even in cases where uncertainty is high? The skill mix changes as technology changes. This has profound implications for training, especially for more difficult or complex situations. These are the same issues raised by Way in his editorial. They are especially important in an environment where there are substantial pressures to reduce training time and costs, to reduce the skills required of practitioners, and to increase production. Dominguez et al. provide a model for exploration of these issues that can be extended and reused. Their research is not simply a study but also the model for a host of studies that can explore the complexity and uncertainty of surgical decision making, expertise, and injury. Implications of the Research This uncelebrated but researched case is interesting not simply as a specific area in the landscape of patient safety, but also as a model that illuminates broad generic issues. This is an excellent example of the way periods of significant or rapid technological change create demands for new skills and judgments. These demands can contribute to new kinds of failures with new consequences for failure. In the case of laparoscopic cholecystectomy, there is a path toward failure that did not exist before the new technology (i.e., failing to convert to an open procedure when uncertainty is high). The consequences of failure are changed, too: bile duct injuries, when they occur, are more likely to be severe. This same pattern of new technology, leading to new demands, leading to new forms of failure with altered consequences, has also been predicted for infusion-based total intravenous anesthesia (Cook and Woods, 1996). The case of bile duct injury during laparoscopic cholecystectomy also shows the importance of pursuing the second story that lies behind the first, superficial story of isolated practitioner failure as the source of accidents. This deeper look identifies a set of factors that combine to produce both success and failure. Perceptual, cognitive, and organizational factors all play roles in this case. Reducing the rate of failure involves improving the system which depends on a coordinated approach that develops and evaluates (1) perceptual aids, (2) exploratory learning techniques to enhance high performance skills and expertise, and (3) changes to organizational behavior. This requires investments of time, energy and financial resources. Grappling with an area like laparoscopic cholecystectomy means developing an understanding of how practitioners handle uncertainty, risk, and hazard. In this, researchers are confounded by hindsight bias. Uncertainty exists only so long as the outcome is undetermined. Hindsight bias tends to make it hard for us to appreciate the uncertainty practitioners confront. It is easy, when there has been no bile duct injury, to see the decision to convert to open cholecystectomy as too conservative, as sacrificing important goals when no such sacrifice was necessary. In hindsight, we readily identify practitioners as risk-averse or risk-seeking. But in actuality, risk is an inherent part of their world, a fluid and changing characteristic that can be difficult to localize and is impossible to quantify. Given the very high consequences of bile duct injury, handling the conversion tradeoff well means that sometimes surgeons will convert even though hindsight will reveal it was probably unnecessary. Some patients will be harmed (suffer the undesirable effects of an open cholecystectomy) so that others may benefit (avoid bile duct injury). Although Dominguez et al. began with a relatively narrow study of perception and visualization in laparoscopic cholecystectomy, their results provoke consideration of much larger issues, making a rich network of connections with other research results. The Dominguez et al. study illustrates how one can proceed to enhance safety in other areas of health care. First, they looked at the sources of both success and failure. They began by studying what makes problems more or less difficult. This helped them identify the human performance issues relevant to expertise (e.g., perceptual factors and judgment under uncertainty and risk). Significantly, they recognized that practitioner performance depended on the larger organizational context. Getting the results depended on tracing the process of how practitioners handle different kinds of situations. Getting this story, in the form of a problem-solving protocol, is necessary in order to learn about human contributions to risk and safety. Researchers then can look for and tabulate patterns across these problem-solving protocols. When these methods are used, investigators begin to escape from hindsight bias to find the set of multiple interacting factors that contribute to accidents.   Uncelebrated Case #2 Antibody misidentification and transfusion reactions Uncelebrated Case #3 Drug misadministrations via computerized infusion devices in the operating room

Copyright 1998 National Patient Safety Foundation at the AMA Prepared for Web publication by Annenberg Center for Health Sciences