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

The conversation that took place over two days at the workshop was a wide ranging discussion about safety, accidents, and research in health care. The discussion was mainly about contrasts:

• between celebrated and uncelebrated cases;
• between success and failure; and
• between naïve attributions of failure to human error and detailed investigations of the strengths and weaknesses of expert human performance in context.

This report presents one synthesis of the materials from the workshop-it follows a few threads through the two days of talk.

Just like any real conversation, there were sometimes several topics discussed at once. Many issues were raised, questions put, and subjects left hanging without conclusion. The workshop did not attempt to develop a consensus. Rather it was an effort to obtain a variety of perspectives, to engage in an exploration of safety in new ways.

The researchers offered no solutions, nor did they identify easy paths to success. Rather they pointed out how myopic our present approaches to safety actually are. In polite and sometimes not so polite terms they indicated that fascination with the celebrated cases of failure is unlikely to yield any real progress towards safety. They encouraged research into the basis for success as a means for understanding failure. They pointed out how careful examination of seemingly peripheral questions about how practitioners work provided the new insights. They showed how research on understanding of the real tasks of real practitioners can lead to new technology that actually improves performance.

The researchers warned against narrow focus on practitioners at the sharp end, pointing out that the lessons from other industries are that accidents reflect systemic factors and not individual ones. They warned, too, against trying to treat safety in isolation from the other aspects of health care. Rather, safety is an embedded feature of a complex, dynamic system. They were optimistic and encouraging about prospects for research that bears on safety in health care but, as seasoned researchers with long experience on these topics from other industries, they were also cautious about the recent flurry of interest in safety as a goal. Several times the conversation turned from potential for progress to warnings against efforts to improve safety directly by programs that look attractive but are disconnected to the larger research base on human and system performance. Their experience with other industries indicated that the need for the appearance of a commitment to safety can sometimes take precedence over the long, painstaking efforts required to make real progress. It is much easier to talk about a "safety culture" than it is to create one.

Charles Dickens' book A Tale of Two Cities begins with a famous litany contrasting the time as both the best and the worst. The title of this report is meant to evoke that same sense of contrast.

At a superficial level-at the level of the celebrated cases as they are usually presented-the story of safety in health care is about repeated, unrelated, isolated, incomprehensible accidents that dog the heels of the vaunted successes of modern technological health care.

At a deeper level, the story is about the ways in which success and failure are derived from the same sources. It is about the ways in which exposure to hazards is indivisibly connected to the pursuit of success. In this contrasting view, the bad events are not separate phenomena that can be eliminated by the use of some managerial or technological tool. Safety is not a separate entity that can be manipulated or achieved in isolation. Rather it is an emergent property of the ways in which the technical, individual, organizational, regulatory, and economic factors of health care join together to create the settings in which events-the best ones and the worst ones-occur.

Although the researchers had much to say about safety, none of them were researchers on safety in itself. Their research is primarily about human performance, technology, organizational behavior, and even philosophy. These are all fields that bear on safety and describe the ways in which the factors interact to cause safety to emerge. John Flach spoke for many of the researchers when he said, "researchers pursue interesting questions." His point was that this work may lead to new views and applications that advance safety, but that the really effective research does not start out that way. His statement can be taken as a warning about the need to establish and sustain a variety of lines of research on human and system performance in health care in order to make progress on safety. The uncelebrated cases all involved long-term efforts focused on apparently small questions. This much is clear-gaining more insight requires sustained, detailed efforts.

The situation confronting those who want to increase safety today is not unlike that confronting those who wished to eradicate cancer a generation ago or more recently those who wished to find a cure for AIDS. At first glance, these were simply applied problems that needed bigger, better, more powerful treatments of the sort that were already being applied. But the real improvement in treatment of these diseases came not from direct applications. The real improvements came from study of the mechanisms of disease, often in areas that appeared only superficially related to the problem at hand. The discovery of genetic causes of cancer, the development of the protease inhibitors that now offer HIV-infected people a chance for long life, came out of efforts that looked scientifically into the complex mechanisms that underlie these diseases. It is ironic that the research pathways that led to these successes do not immediately suggest to us similar approaches to learning about safety by studying the complex mechanisms that lead to success and failure.

The organizers of the conference thought to challenge the assembled researchers to describe how their research could be used to gain new insights into safety in health care. The researchers provided a host of pointers and engaged in a wide-ranging conversation about the opportunities and obstacles to research on safety. Several contacts between individual researchers and potential users of their work happened at the workshop. The transfusion medicine work was particularly interesting to some of the workshop sponsors. The lessons about incident reporting and analysis captured the attention of other people and organizations. But in the end, the researchers challenged the organizers and by extension, the health care community to take a new look at safety, to change long-held views about the sources of success and failure, and to look more closely at the ways in which our fascination with the celebrated cases limits our ability to see the larger world in which safety is created and nurtured.