A few years ago we set an ambitious task for ourselves of finding ways to improve surgical safety in the United States as a whole and this work has now carried on through the center. Previous studies, including our 1999 study of medical injury in 28 Utah and Colorado hospitals, established that avoidable, serious complications from surgery were a more frequent occurrence than appreciated. (In the Utah/Colorado study, involving 15,000 hospital patients, two-thirds of adverse events resulting in death, disability, or a prolonged hospital stay, were surgical; three percent of surgical patients experienced a medical injury; and these injuries contributed to 12.2 percent of all hospital deaths (1). In all, half of the adverse events were avoidable.) Intraoperative errors, as opposed to errors in the preoperative or postoperative setting, were by far the largest contributor. Most of these were classified as “technical errors.” Little was known about the why these occur or how to reduce them.
In the last two years, we have conducted interviews with surgeons about their errors, studied nearly four hundred malpractice claims against surgeons, and brought human factors experts into operating room to allow a systems analysis of operative care. Again “technical complications” emerged as the most common problem by far. But our interview study of 146 surgical injuries resulting from error at three Boston area hospitals established that three factors were the most common contributors: surgeon inexperience played a role in 53% of the injuries (half of the inexperience was that of attending surgeons; half that of trainees); communications breakdowns (most often from handoff problems) played a role in 43%; and the surgeons cited fatigue or work overload as playing a role in 33%. One or more of these factors contributed to 83% of injuries. (12) Phase I:
The initial study to come out of the malpractice claims review database looked at cases in which patients had instruments or sponges left in them by mistake and was supported in part by the Risk Management Foundation. (13) These are rare complications (occurring in 1 in 9,000 to 19,000 operations), but are dangerous, avoidable, and often seen as a prototypical example of error resulting from negligence. The case-control study established, however, that failure to follow standards for tracking materials during operations was not a significant contributor. Instead, the major risks were emergency surgery (which produced an 880% increase in likelihood of a retained instrument or sponge), an unexpected change in operation (420% increase), and body mass index (obesity doubled the risk). In these situations, our current system of counting everything at the start and end of an operation was itself prone to error. In two-thirds of the cases, a surgical sponge was the item inadvertently left behind. These findings suggested that an improved system for detecting materials left in patients-sponges in particular-is required. RMF also sponsored the observation study of surgical performance, which brought human factors experts into Brigham and Women’s Hospital operating rooms to examine 10 complex cases. Their moment-by-moment analysis revealed, unexpectedly, that avoidable safety-compromising events occurred in every case. (14) Teams managed to act successfully to prevent patient harm in most of these cases, but not all. The counting protocols for tracking instruments and sponges in particular seemed to play a surprisingly negative role. They found that nurses devoted an average of 29.8 minutes after the incision was made just to counting materials. This represented 14.5% of the operative time. Furthermore, a counting discrepancy was observed to occur in two-thirds of the cases. As a result, in several patients, progress of a still-incomplete operation had to be substantially delayed or outright suspended while the nurses attempted to reconcile inconsistencies. (14) This work was considered Phase I of our investigation of surgical safety. It established clear and focused areas for improvement (our system for tracking operative sponges and instruments, in particular) as well as a priority of identifying ways to reduce errors due to inexperience among low-volume surgeons and hospitals. Phase II:
Design of new interventions. RMF was generous in supporting this next phase of research, in which we sought to develop and assess new interventions to improve surgical safety. This work focused primarily on three goals: (a) developing a solution to effectively prevent losing instruments and sponges in surgical patients; (b) developing the Harvard Surgical Safety Score, a simple score for how safe (or unsafe) any given operation was that could permit routine adoption in daily practice and benchmarking for improvements in performance; and (c) finding a way to improve performance of low-volume surgeons. A. Prevention of retained instruments and sponges. We conducted a rigorous search and innovation process which generated five possible technologies to detect when instruments or sponges are left behind. Discussions with engineers and focus groups with operating room nurses identified two of these technologies as most feasible for adoption: (1) Bar-coded sponges; and (2) a computer vision system for tracking instruments. Working with the company in Los Angeles, we have now succeeded in placing bar-coding on operative sponges. FDA approval for the system was secured. A June 22, 2004, simulation study with nurses from four Harvard hospitals demonstrated its safety and practicability. We have also worked with a company in Cambridge, to design and develop a prototype camera-based system for computer-automated counting of all instruments in the operative field. Initial simulations have demonstrated feasibility, and we are now refining the computer vision system. Both systems are designed to safeguard against human counting errors and reduce the time operating staff spend just counting stuff. B. The Harvard Surgical Safety Score. A core problem in improving surgical safety is that what constitutes an operation well done has never been carefully characterized. At the end of an operation, unlike, say, a baseball game, we cannot confidently say whether it went well or not. We have to wait-usually days to weeks-to see how the patient ultimately does. And even then, because patient illness clouds the picture, it is unclear how much of the outcome is the result of how the team performed. Management expert Peter Drucker has pointed out that “You cannot improve what you cannot measure.” We see this is no less true in the operating room. Will current requirements to reduce working hours, have a team “timeout” at the start of every operation, allow use of scrub technologists instead of registered nurses in the operating room affect the safety of surgical care? We cannot firmly say. Virginia Apgar developed her simple Apgar score for newborns so obstetric teams could quickly identify when a delivery had gone well or badly. It drove rapid changes in obstetric care that have dramatically raised the safety of deliveries. Similarly, we have sought to develop an operative safety score using a mix of easily observed measures. Our initial scorecard uses five indices: the length of the operation, the amount of blood loss, fall in patient temperature, the lowest blood pressure, and the lowest oxygen saturation. The worse anyone of these are, the more unsafe an operation appears to be. We have also identified ten other measures for possible incorporation into such a safety score, including anesthesia time, highest pulse rate, highest pulse pressure, and whether pressors were used. With a validated score of this kind, three major and far-reaching outcomes are sought. First, it could permit routine measurement and awareness for teams and patients of how safe or unsafe their operations are-something that is currently a matter of only speculation and lore. Second, it could allow benchmarking and, simply by existing, drive teams to improve their operative safety. Third, it would enable proper evaluation of interventions for safety, such as simulation training and surgeon work-hour limitations. C. Peer-consultation for low-volume surgeons. With inexperience identified in Phase I as the most important contributing factor in avoidable surgical injuries, we sought to devise a strategy to improve outcomes and safety for low-volume surgeons. Although efforts, like the Leapfrog initiative, have sought to shift patients to high-volume centers, the vast majority of patients still obtain care from low-volume surgeons locally, even for high-risk procedures. (15) We identified a familiar, but infrequently employed, practice for surgeons attempting an operation they are relatively inexperienced with: consulting a more experienced peer for careful operative planning and safety. According to a survey of 1370 doctors from 5 specialties, doctors believe that informal consultation improves quality of care. (16) The three leading reasons physicians cited for obtaining such consultations were to obtain an expert opinion, verify information, and to learn from the consultant. Less experienced physicians were also found to request more informal consultations. With these findings in mind, we devised a protocol for peer consultation with high-volume surgeons. It involves a brief but structured discussion of the patient’s indications for undergoing surgery, the operative plan, and the experienced surgeon’s strategies for averting complications in such situations. If such a protocol can be demonstrated to be effective, it would allow pairing of low-volume surgeons and hospitals with surgeons at high-volume centers for improved operative safety. Phase III:
Testing. Having designed at least three major interventions with potential to substantially improve the safety of surgery, we have now arrived at the point of testing them for their efficacy. Testing is critical for two reasons. First, no matter how intuitively appealing and simple a safety intervention may be, there is nearly always the potential for the approach to have more complex effects in practice than realized in theory. That the benefits are genuine and avoid unacceptable costs in time, expense, complexity, and unanticipated harm must be demonstrated. Second, evidence is critical to persuading clinicians to adopt safety strategies that involve substantial changes in established practices. Because lives are on the line, because many previous strategies have not worked, and because change is hard, clinicians are understandably resistant to modifying their practices without hard evidence of benefit. Evidence itself, however, can spur broad and rapid change. Therefore, in this ultimate phase of our work, we propose to provide careful clinical testing of the efficacy of our three new interventions-a bar-coded sponge system; the Harvard Surgical Safety Index; and peer consultation to reduce errors of inexperience. We will also seek to complete the design and simulation testing of our computer vision system for preventing instruments from being left in patients.
|
Send Feedback to: 
