Helping Health Care Organizations Help Patients

Sentinel Event

Presentation on Patient Safety Issues

Presented by Paul M. Schyve, M.D., Senior Vice President, at the National Summit on Medical Errors and Patient Safety Research, September 11, 2000

The Joint Commission would like to thank you for the opportunity to present our views on what crucial information is needed to help address patient safety issues of today and tomorrow.

The Joint Commission is the oldest and largest national health care accrediting body in the United States. We accredit approximately 19,000 health care organizations, including hospitals, home care organizations, hospices, behavioral health organizations, nursing homes, clinical laboratories, long term care pharmacies, ambulatory facilities, and managed care entities. The Joint Commission's mission is to improve the quality of care for all Americans, and patient safety is considered by us to be an integral component of health care quality.

The Joint Commission has long been involved in protecting patient safety. In fact, our genesis in the Hospital Standardization Program early in this century arose from efforts by the American College of Surgeons to protect patients undergoing hospital-based surgery from death unrelated to the course of their illness. In the fifty years since its incorporation, the Joint Commission's accreditation programs have contributed to the high quality of care that Americans enjoy today, in areas ranging from control of nosocomial infections to reducing deaths from chemical, electrical and fire-related hazards. An analysis of our current standards demonstrates that over 75% of our accreditation requirements bear a relationship to issues of patient safety. This is because accreditation is, at its heart, a risk-reduction strategy that seeks to maximize good patient care while minimizing untoward events. It is assumed that if an organization complies with accreditation standards, the probability of adverse events is decreased.

Despite the many successes of accreditation and other external and internal oversight activities carried on during the Twentieth Century, patient safety is a major agenda item for health care today. We have reached the current point of increased attention to safety for several prominent reasons. First, we have experienced a significant evolution in health care over the last decade that has had far reaching consequences for the quality of patient care. We have witnessed a virtual revolution in the availability and use of medicines; seen dramatic changes in how health care is configured and delivered; and watched as every type of health care organization has struggled with the need to provide more with fewer resources. At the same time, the rapid explosion of biomedical and health knowledge has made it increasingly challenging for practitioners to stay current while relying on traditional information sources. Technologic advances have made possible complex medical procedures requiring high-level skills and knowledge, and have enabled many of these procedures to be moved outside the hospital setting. In general, workloads have become heavier, creating increased stress and fatigue for health care professionals. Care givers are working in new settings and performing new functions (sometimes with minimal training) -- as a result, skill mixes are shifting. These and other factors have exacerbated the opportunity for serious medical errors.

Second, a mounting concern within health care over ways to address patient safety issues has characterized the last five years. Those of us in the area of health care evaluation were influenced by a spate of nationally publicized serious adverse events in 1995. Some of these events occurred in health care organizations that by both reputation and accreditation status are considered to be outstanding. This latter fact spurred the Joint Commission into looking for new methods that could be employed to reduce the incidence of errors. Consulting with experts within health care and in other high-risk endeavors (e.g., airline and nuclear power industries), the Joint Commission developed, through a series of experience-based iterations, its Sentinel Event Policy, which is described below. In 1996 this policy, which emphasizes the surfacing, analyzing, and sharing of error-related information, was implemented nationally for all of our accredited organizations.

That same year, the Joint Commission partnered with other opinion leaders and expert organizations to highlight the need for action on patient safety within health care. A ground breaking national expert conference - - co-convened with the American Association for the Advancement of Science, the American Medical Association, and the Annenberg Center for Health Sciences at Eisenhower, and co-sponsored by eleven other agencies, associations, foundations, and industry representatives, including the Agency for Health Care Policy and Research -- was held on the subject Examining Errors in Health Care: Developing a Prevention, Education, and Research Agenda. This multidisciplinary conference was an important starting point for health care to develop an agenda for making error reduction a common goal. Many stakeholders in the patient safety arena launched their initiatives at or following this conference or its sequel events. A significant impetus to address errors had been born. For example, the National Patient Safety Foundation, a diverse group of interested parties that includes the Joint Commission, was initiated by the American Medical Association to help health care practitioners and provider organizations meet the challenges of reducing medical errors. The Joint Commission, with other leading advocates, testified on the issue of patient safety during the 1997 President's Advisory Commission on Consumer Protection and Quality in the Health Care Industry. Eventually, in 1999, the publication of the Institute of Medicine Report To Err is Human: Building a Safer Health System would catapult the messages on patient safety from health care leaders into national prominence.

The Sentinel Event Policy

The Joint Commission's experience developing and progressively revising its Sentinel Event Policy has provided unusual insight into the complexities of successfully attacking the complex issues surrounding patient safety. The Sentinel Event Policy codifies much of the current thinking about how to reduce errors and can be used, therefore, as substrate for advancing scientific inquiry into how to maximize their reduction. The implementation of this policy also presents a context for identifying relevant gaps in our understanding of certain issues. Our experiences from this initiative confirm that the problem of medical errors is endemic to the way health care is carried out, but that we can bring to bear the commitment of health care professionals, lessons learned from other scientific and industrial fields, and certain tools at our disposal to greatly reduce the incidence of errors today.

In brief, the Sentinel Event Policy is a patient safety program that includes the requirements that:

the leaders of health care organizations encourage internal reporting of sentinel events;
the organizations undertake root cause(s) analyses of such events in order to identify system and process redesigns that could reduce the risk of future similar events; and
the organizations develop and implement action plans based on the root cause analyses and assess their effectiveness. Moreover, there are also strong encouragements for organizations to report to the Joint Commission the events, root cause analyses, and action plans to enable the Joint Commission to create a deidentified database that can be analyzed to identify specific risks to patient safety, their underlying causes, and successful preventive actions.

The Sentinel Event Policy was designed to create a protected environment that encourages the systematic surfacing and reporting of serious adverse events. Hoping to foster organizational cultures that promote error reduction efforts, the Joint Commission designed its policies so as to not penalize the accreditation status of an organization that voluntarily surfaces an error and performs the analysis and preventive action required under the policy. Today, the blame-and-punishment orientation of our society drives errors underground. Indeed, we believe that most medical errors never reach the leadership level of the very organizations in which they occur. Therefore, although there is much rumination over the statistics published about medical errors, we believe that no one has a real handle on the actual numbers because all the incentives to report are negative.

A further characteristic of the Sentinel Event Policy -- and I believe its hallmark -- is the production of a credible intensive analysis following the occurrence of a serious event. It is essential that there be a no-holds-barred vetting of all the potential causes underlying the event. We call these responses "root cause" analyses -- a term borrowed from the engineering world's orientation to a systems approach both to solving problems and to producing desired outcomes. A root cause analysis focuses primarily on an organization's systems and processes as opposed to an individual's performance. It is from the root cause analyses that we obtain the insight into how processes must change in order to control undesirable variation that can lead to error and harm. The analyses tell the story of what must be developed to guard against the occurrence of similar error -- thus they are the raw material from which emanate action plans designed to address the specific failures or omissions uncovered during the in-depth evaluation.

Another task of the Sentinel Event Policy is the dissemination of information gleaned from errors and their root cause analyses to all organizations at risk for serious errors. To have a positive national effect on patient safety, information learned from errors - - their occurrence, their characteristics, and their root causes -- must be aggregated, analyzed, and disseminated to the health care community at large. This dissemination can -- and should -- be done at all levels of the system. The Joint Commission's dissemination program periodically issues a Sentinel Event Alert to share the most important lessons learned from our database -- known risks in systems and processes as well as safe practices. To date, there have been 14 such alerts (e.g., on KCl deaths, inpatient suicides, wrong site surgery, restraint deaths, infant abductions, blood transfusion errors, operative and post-operative complications, fatal falls). We also incorporate appropriate information learned from the analysis of sentinel events into accreditation standards. Both statistical and anecdotal information suggest that this sharing of preventative strategies has altered organizations' practices and reduced selected types of errors.

Lastly, the Sentinel Event Policy involves the creation of proactive accreditation standards that apply to the leadership of accredited organizations. The Joint Commission is now vetting proposed new standards that go beyond retrospective analysis of events to focus on creating a culture of safety in health care organizations, on proactive safety engineering of health care processes (e.g., failure mode, effects, and criticality analysis), on safety education of staff, and on disclosure to the patient and (family) of significant unanticipated outcomes in care. All of these proposed new accreditation standards derive from empirical studies on safety and/or the advice of experts in the safety field outside health care.

The Sentinel Event Policy has demonstrated that despite having the tools and commitment to make significant reductions in the incidence of medical errors, the road to success will require hard work and daunting cultural changes. The Sentinel Event Policy provides a framework for honing in on the types of information and research efforts that should be supported at the federal level. We have learned a great deal from this program, including where more investigative study is needed, where incentives must be placed for achieving needed behavior change, and where demonstration projects would be valuable.

Specific Research Needs and Questions

The area of patient safety is broad enough to consume an almost limitless amount of research and demonstration dollars. Further inquiry into "managing" errors must have context for posing and crafting research questions and demonstration projects. New inquiry must assume some existing knowledge and understanding of what is to be achieved. We have derived from our Sentinel Event Policy experience specific critical functions associated with effective error reduction.

These functions include:

Surfacing reports of errors so that we can understand their epidemiology.
Analyzing the underlying causes of errors so that we can understand the risk points and system and process failures that allow errors to occur.
Designing safer procedures, equipment, and information systems that minimize human errors and their impact, and latent errors in delivering health care services.
Broadly disseminating knowledge about errors and their causes to maximize preventive strategies throughout the health care system.
Bringing about the necessary cultural changes that support team work, acceptance and integration of decision support systems and clinical practice guidelines into health care practices, and avoidance of a "name and blame" response when errors occur.
Continually monitoring and periodically evaluating health care practices to ensure that safe practices are adopted and appropriately carried out.
Appropriately engaging all responsible stakeholders as part of a learning atmosphere.

These functions can serve as a framework for considering where to pursue additional scientific study. Therefore, we recommend that new research and demonstration projects be federally funded in areas which support the functions listed above. This means supporting both basic and applied research into how to best accomplish these functions, as well as funding demonstration projects to assess the characteristics of effective error reduction activities and the feasibility of various alternative ways to achieving patient safety goals.

Our intensive experience in studying and attempting to improve patient safety has identified a series of questions whose answers the Joint Commission believes could be used to improve our ability to carry out these functions. The answers will come from a diverse set of sciences such as medical anthropology, human factors engineering, operations research, psychology, industrial engineering, informatics, and other fields of investigation.

The first set of questions addresses the fact that we do not have an accurate handle on the statistics underlying errors. The oft-cited studies about the estimates of medical errors are old and have methodologic issues that complicate their usefulness. Furthermore, the Joint Commission's experience is that even the leaders of health care organizations will admit that they do not have an accurate count of the errors within their own facilities. Without better knowledge of the epidemiology of errors, health care will be limited in its ability to appropriately set priorities for resource use in preventing errors, to identify common root causes for multiple types of errors, and to assess the effectiveness of error-reduction activities. Furthermore while it is argued that a proactive approach to error reduction is to analyze data about "near misses," it is not known what relationship near misses have to errors that result in harm, either with respect to similarities and differences in root causes, or with respect to priority setting among near misses in order to reduce actual harm.

We suggest research to answer the following questions:

What is the epidemiology of serious adverse outcomes resulting from errors? What is their distribution, patterns, and numbers within the different segments of health care?
Does their epidemiology differ from that of the "near misses"?

The next group of questions relate to the organizational risk factors for health care processes in general. The more we know about the factors that create latent risks - - especially risks that could create danger by multiple pathways -- the better we will be at designing new systems and/or creating standards to monitor organizational capabilities.

These questions are:
What are the characteristics of organizations or cultural impediments that increase risk or make them less prone to risk? E.g, what are the effects of hierarchical vs. flat organization management structures on patient safety? of communication mechanisms within an organization? of how an organization handles the adoption of new information and technologies? of its ability to promote team work, reduce environmental stress or maintain a positive staff morale?
What is the contribution to safety of staffing numbers, types, and patterns?

We also need information to help us design systems that have appropriately integrated checks and balances, redundancy when necessary, inherent resilience to human errors, and reduced reliability on routine human intervention for producing successful results. While reduction of variation in the "blunt end" of organization systems and processes can decrease risk to patients by enhancing reliability, it is human (i.e., clinician) intervention at the "sharp end" with the patient that creates safety at the intersection between the stable organizational processes and the inherent variation in patients (e.g., with respect to physiological and pathological factors and patient preferences). We also want to make certain that such systems are affordable and practical, and that we can plan for any additional error that they may introduce into processes. An error, for example, in an information system can replicate its adverse result in many patients, whereas the same human error may affect only one patient.

Therefore, we would like to have answers to the following questions:
How can new information technologies be used to reduce specific risks? Which are the most efficacious? The most cost/beneficial? The easiest to implement?
What are the new risks -- and preventive strategies -- associated with these new information technologies? E.g., how do we implement these technologies without introducing new error?
How best can standardization at the blunt end of systems and processes contribute to patient safety?
What specific process designs are associated with reducing specific risks?
What is the contribution that decision support and other knowledge management systems can make to reducing specific medical errors? How can we promote more acceptance and use of these computerized aides in an era of exploding biomedical and health-related information? What databases, information repositories, and clearinghouses are needed to support these systems? And, how can these databases, repositories, and clearinghouses be kept free of error themselves?
What lessons from engineering, human factors research, and other fields of study, both in the United States and other countries, can be brought to bear on the health care system?
How do we design systems that are resistant to chance combinations of latent errors, human fallibility (active failures), and hazards that result in adverse outcomes? How do we achieve more synchronous risk management (continuous risk assessment and prevention) in the design of systems?
What role can integrated, electronic medical record tracking systems that travel with the patient play in reducing specific medical errors?
What is the role of shared decision-making tools in reducing medical errors through increased patient education/compliance and improved ability to track individual patient care progress and processes? How can we make the patient a more informed and better partner in risk reduction?

Another set of questions involve the best ways to get information about high risk situations and risk-reducing practices into the hands of individuals who can best respond. Too often, the wrong person may receive information, or it may be in the wrong format to catch his/her attention or to be usable. The problems associated with dissemination, assimilation, and implementation of clinical practice guidelines are likely to be repeated in our patient safety efforts - - but will be tolerated much less by the public. Further, information must be actionable and reach all parts of the system that have responsibility.

These concerns lead to the following questions:
What are the best methods for disseminating error-related information, on a national basis, to ensure that it reaches appropriate organizations and individuals? Are some information networks/channels optimal for transmitting safety information?
How should the information be formatted to receive necessary attention and immediate and effective responses?

There are significant cultural, educational and behavioral barriers to adopting safety practices and maintaining a positive culture for safety. The current environment of blame, shame, and liability is a strong disincentive for reporting and analyzing errors, despite the ethical imperative to "first, do no harm."

Therefore, we pose the following questions for investigation:
How do we create a more constructive environment that rewards the voluntary reporting of errors?
What are the special skills and knowledge required of clinicians to create safety at the sharp end interaction with the patient? How can we best support acquisition of these skills/knowledge?
What model(s) of safety reporting systems work best? How can we construct a model of safety reporting that brings the strengths of all responsible stakeholders to bear on the issues while creating the appropriate learning environment? How can such a model best handle the issue of public accountability?
What methods for assessing an individual's competence would least interfere with the disclosure and analysis of errors and serious adverse outcomes?
How can we use the educational curricula and training in the health professions to achieve our goals in patient safety? Are there impediments to teamwork and acceptance of information technologies built into the way we educate/train practitioners? Does the value we place on professional autonomy interfere with patient safety goals? That is, are the more well qualified and highly motivated professionals less receptive to a systems approach to delivering care? If so, how do we maximize these desirable personal values and patient safety at the same time?
What are the post incident effects on those practitioners who were most proximal to a serious event?

Further study of the foregoing questions is important because in the rush to solve the errors problem, we can easily make new mistakes and introduce new sources of error. Moreover, it is often tempting to focus safety management on ways to bring down the count of negative incidents during a period of time, rather than focus on the hard work of designing and redesigning systems so that they possess an inherent resilience to even a chance accumulation of errors.

A review of these questions leads to a critical conclusion: improving patient safety will ultimately require action by many stakeholders, each of whom has both an interest and a role in reducing errors and improving safety. These stakeholders include clinicians and their associations, health care organizations and their associations, purchasers, accrediting bodies, state and federal regulatory agencies, consumers, patients and their families, lawyers and their associations, legislatures, institutions of professional education, and the media. It is likely that the most effective methods for improving patient safety will be based on collaboration and coordination among multiple stakeholders. In the Joint Commission's experience, collaborative projects in patient safety have been especially rewarding. Therefore, we would urge the agency to support funding of collaborative demonstration projects, in which the collaboration itself (e.g., among state agencies, accrediting bodies, state professional and health care associations, and federally-sponsored Peer Review Organizations) serves as an ingredient of the patient safety effort.

The Joint Commission is anxious to collaborate with others in seeking answers to these and other questions, and to use those answers to improve patient safety. We would be pleased to further mine our database and experiences in keeping with federal research goals, and we also extend the Joint Commission as a resource to the Agency for Healthcare Research and Quality in the broad area of quality improvement.

Summary

It is important to note that patient safety will never disappear as a concern, because safety issues will track advances in medicine and technology, and changes in the delivery of care. More to the point, human beings will continue to be fallible, and human errors cannot always be predicted. However, we have a unique opportunity to make considerable progress on how we design resilient systems that minimize both the likelihood of untoward events and the impact of their adverse effects when they occur. Future levels of error in the health care arena will depend in large part upon how we apply what we know today and what questions we seek to answer for tomorrow. Fortunately, we are better equipped now than ever before to understand how to take advantage of research in many scientific fields and apply it to reducing error and harm in the health care system.