Secondary Logo

Journal Logo

Special Theme: Medical Errors: SPECIAL THEME ARTICLES

The Roles and Responsibility of Physicians to Improve Patient Safety within Health Care Delivery Systems

Classen, David C. MD, MS; Kilbridge, Peter M. MD

Author Information
  • Free


The release of To Err is Human by the Institute of Medicine (IOM) in 1999 catapulted the issue of patient safety to the forefront of national consciousness with its shocking statistics. The report stated that between 44,000 and 98,000 Americans die annually from medical error, with an associated cost of between $17 billion and $29 billion, making this the leading cause of death in the United States and a major component of the nation's high health care costs.1

Although it was To Err is Human that captured the public's and the nation's attention, the literature is filled with statistical evidence of medical errors. For example:

  • In 1984 a detailed study of 30,000 randomly selected discharges in New York State revealed that 3.7% of patients experienced an adverse event resulting in a longer hospital stay or disability at the time of discharge. Of these events, 13.6% led to death and 58% were preventable.
  • A second study conducted in Utah and Colorado in 1992 reported similar findings. In this instance, 2.9% of the discharges suffered an adverse event, and 8.8% of these led to death. As in New York State, this study revealed that over half of the adverse events, 53%, were preventable.

There has been considerable skepticism among physicians about the overall magnitude of medical-error-related deaths based on these studies. Serious questions about the methodologic validity of the 1984 New York State study have been raised.2 This controversy is likely to continue as a new study of preventable hospital deaths in Veterans Administration Hospital patients raises further questions about the actual number of preventable deaths from medical errors.3

Although debate over the magnitude of the patient safety problem will continue, there is little doubt that health care can be made much safer. Indeed, many industries (aviation, nuclear power, manufacturing) have achieved degrees of safety and reliability far greater than that found in health care. For instance, reliability (safety) as measured by defect rates in manufacturing has achieved rates of fewer than four defects per million opportunities. In health care, comparable defect rates are several thousandfold higher. For example, 79% of eligible heart attack survivors fail to receive beta-blockers, for a defect rate of 790,000 per million heart attacks.4 It is often thought that health care is a far more complex field than other industries and therefore unable to achieve such low defect rates. However, deaths from anesthesia have fallen dramatically over the last two decades, from rates of 25 to 50 deaths per million anesthesia inductions to a rate of only 3.4 per deaths per million inductions.5 Clearly, some parts of the health care system are performing at safety levels on par with the best in other industries; the challenge is to create this level of safety throughout all areas of health care.

The momentum generated by the IOM's first report and the ongoing news coverage of health care safety problems have forced all health care stakeholders to acknowledge that patient safety is an area where there is a tremendous opportunity and need for improvement. Although few patient safety initiatives have emerged from managed care organizations, an increasing number of health care delivery executives now consider patient safety one of the top priorities that they must address within their organizations over the next three years. Health care delivery organizations are responding to the need to improve patient safety in many ways, including incorporating patient safety into the organization's mission, vision, and values and elevating safety to a strategic status throughout the organization; developing organizational awareness of patient safety as an issue; and exploring the use of advanced technologies (e.g., computerized physician order entry, bar coding the administration process, etc.) as enablers to improve clinical processes. Several large health care delivery organizations such as the Veterans Administration hospitals have taken significant leadership roles in enhancing patient safety.6

However, many of these initiatives have been organizationally focused, without clear and consistent physician leadership or engagement. Perhaps this is not unexpected given the ambiguous and vulnerable position in which physicians find themselves with respect to patient safety.7 Whatever the source of industry efforts to promote patient safety, the relationship between physicians, health delivery organizations, and patients lies at the crux of efforts to implement measurable improvements in patient safety. Unfortunately, there has been little research on these relationships and their effects on patient safety. Perhaps this is because most of the focus in safety has been on what Cook refers to as “the sharp end” of care delivery: the frontline practitioner who interacts with patients, rather than on the “blunt end”: organizational and system characteristics that can lead to an unsafe environment.8

We propose new organization-focused models of safety that could guide health care delivery organizations and physicians in shaping their evolving roles and responsibilities in their ongoing efforts to improve the safety of care.


The physician's interaction with the health care delivery organization starts not upon entering practice, but at the beginning of the medical educational experience. A physician's clinical training starts in medical school and occurs within a health care delivery setting. Indeed, it is hard to imagine how changing the education of physicians with respect to patient safety would be successful without corresponding changes in the environment and culture of the health care delivery systems in which they train. Therefore, any changes in the health care delivery system with the goal of creating a culture of safety will have to address not only practicing physicians but also physicians in training.

Physicians are taught to abide by the credo “Primum non nocere” (First, do no harm). This principle, learned in medical school, forms the basis of physicians' views of themselves and the culture in which they work. Perfection is portrayed as the necessary goal of any physician who hopes to care for patients, and its manifestation is the complete absence of error.9 Deeply rooted in the current medical education process and the current culture of delivery systems is the idea that mistakes are not tolerated; identifying and castigating error in the performance of the physician in training is an ingrained part of the educational experience of all physicians. This ritualized aspect of the medical education process is not a celebration of success but a relentless focus on error as a shortcoming to be eliminated.

No forum better illustrates this focus on error and mistakes than the traditional morbidity and mortality conference. This conference is a commonly used teaching venue whose goal all too often seems to be to point out mistakes that a physician has made and to vilify the physician for those mistakes in the presence of peers and colleagues. This approach merely institutionalizes the idea that mistakes are not tolerated, and it hardens the view that mistakes are caused by individuals—usually physicians—rather than by systems. It also goes a step further and suggests that regardless of the true cause of the mistake, the physician is to blame.10

This common educational experience is often so painful and the lessons so recurrently similar that most physicians have little interest in either reporting or investigating serious mistakes. Indeed, resident physicians often do not discuss their errors with attending physicians.11 In part as a result of this focus on personal responsibility, the profession of medicine attracts perfectionists who become more so as they move through the hazing of medical training. Perfectionists do not like to advertise imperfections, and the pain of residency only further supports that view.9 Perhaps this is one reason that physicians focus their ongoing educational efforts on published national studies rather than on learning from local medical mistakes that they or colleagues might be held accountable for. How often do grand-round topics include local cases in which a serious mistake has occurred?12

For most practicing physicians the culture of blame created during their training continues into practice. However, the forum changes from the morbidity and mortality conference to the credentialing and peer-review process. The credentialing process can be a very painful exercise if the physician has any past history of serious mistakes. Local quality assurance committees pick up where credentialing leaves off and handle the ongoing exercise of review of errors and mistakes, and if such errors are serious, they end up in the local peer-review process. These processes are often focused on assigning blame, rather than true learning, and are very painful for both the investigating physicians and the investigated physician. In addition, all of these processes can lead to malpractice cases and licensure actions.9

Both the medical education process and the physician practice experience have created a culture of intimidation in which the expectation is that the physician must achieve perfection and accept responsibility for mistakes as personal failures.13 The emphasis on individual accountability on the part of the physician extends more broadly to all things related to safety. This perspective is hardened by state scope-of-practice laws, which tend to put responsibility on the physician and not the overall care team.

In addition, oversight at the local level and the national level tends to be quite specialty-driven. This leads to great resistance on the part of physicians to fully accept or participate in team-based care, despite much evidence that team-based care can lead to better and safer outcomes. A recent report offers evidence that multidisciplinary team-based care is still far from a reality.14 Indeed, in some health care delivery settings—where more team-oriented care has been implemented and physicians have transferred some care roles to other health care professionals—legal liability has actually increased.14 Other recent reports also suggest that licensure, local peer review, specialty oversight, and legal concerns further emphasize the physician as the sole source of responsibility for mistakes. This only reinforces the lone nature of the physician at the “sharp end” of care, who is blamed if the system provides unsafe care. It is not hard to understand how difficult it will be to engage physicians in leadership roles at health care delivery systems as champions of patient safety, if it only increases their risk.


The success of any effort to improve the safety of health care delivery will be directly influenced by the relationship between physicians and health care delivery systems. One of the most difficult tasks in improving the quality and safety of care has been successfully engaging the physician.15 Health care delivery systems have expended much effort in learning how to develop, manage, and govern their relationships with physicians and physician groups. The primary motivator for these efforts to integrate physicians has not been clinical quality improvement, but rather economic concerns related to reimbursement, productivity, and cost containment. This has clearly established an expectation on the part of physicians that their relationships with delivery systems, even when focused on clinical issues such as disease management, are primarily driven by financial concerns.

Another problem with attempts by delivery systems to integrate physicians has been the bewildering complexity of models for this relationship. Physicians often have multiple simultaneous relationships with integrated delivery networks (IDNs), hospitals, management service organizations, physician—hospital organizations, independent practice associations, and IDN-owned insurance companies. Few of these arrangements exist in pure form, and one IDN might offer several different options to meet the different interests of physicians.16

What is clear from all these models is the overwhelming importance physicians place on shared decision making and power sharing. Without these, physicians are much less likely to cooperate in any initiatives, quality or otherwise. With respect to quality, there is evidence that physicians are more likely to respond to quality initiatives if they lead them rather than if they have them imposed from the outside, regardless of the natures of their relationships with the delivery system.17

Unfortunately, physician integration efforts have for the most part failed to yield the expected economic benefits to delivery organizations. As a result, many of these relationships are in the process of unraveling. The uncertain future of the delivery organization—physician relationship makes long-term clinical quality and patient safety initiatives ever more problematic.18


Traditionally in health care delivery organizations the oversight for quality and safety involved the physician only when issues of physician performance, credentialing, medical staff privileging, or regulatory matters—such as a survey by the Joint Commission on Accreditation of Healthcare Organizations (JCAHO)—arose. Quality and safety of care were generally considered the personal province of each individual physician. Prospective responsibility for quality and safety beyond the individual practitioner usually resided with the administrative staff of the delivery organization, often in the quality assurance department and without direct physician oversight.19

As the health care marketplace has become increasingly concerned about the quality and safety of care delivered, health care delivery systems have responded by assigning quality leadership roles to physician executives.18 Despite this trend several reports indicate that hospital leaders continue to confront substantial barriers in cultivating physicians' involvement in quality improvement projects.20 Physicians are often reluctant to participate in hospital quality improvement projects due to distrust of hospital motives, lack of time, and fear that pressure to reduce variations in care in will compromise their flexibility in managing care to meet individual patients' needs.21 In addition, physicians lack training in the principles of quality improvement, team leadership, and general management, making them less comfortable and competent than they would like to be in these roles.

A large study of quality improvement in health delivery systems revealed several factors that were associated with increased involvement of practicing physicians in health system quality improvement:

  • Board involvement and activity in quality improvement
  • Active physician staff involvement in governance, but not “physician at large” involvement in governance
  • Negative correlations of hospital size, membership in a multi-hospital system, and private ownership with practicing physicians' involvement in quality improvement

It has been recognized that successful efforts to improve the quality of care within health care systems requires significant physician leadership around several critical tasks: organizing the system around defined processes of care, developing multidisciplinary teams to manage each of these processes of care, and evolving the autonomous physician into a process-of-care team member.22 Some systems have structured joint governance in quality improvement as a method of creating this necessary organization—physician partnership.

However, as increasing pressure is placed on physician practice revenue, the tension between voluntary committee commitments and the generation of personal income increases. In some markets physicians and hospitals have become direct competitors for revenue.23 In the absence of clear and direct economic benefit to physicians, and particularly in the event of negative economic impact, organizations will be increasingly challenged in their attempts to demonstrate the relevance of process improvement to physicians.24 Simply reducing length of stay or cost to the hospital is not enough to garner ongoing physician support.

Incentive alignment between physicians and health care delivery organizations is therefore critical to any effort to improve the quality and safety of care. Cost savings from reduced complications often go to the managed care company, not the physician, and depending on the reimbursement approach, often not to the delivery system either.13 Because much of the improvement in quality that these initiatives entail may lead to potentially direct negative economic effects on physicians, physician incentives will of necessity be a major focus of any attempt to link physicians and delivery system around the issue of quality and safety of care. Currently, physician incentives vary tremendously among delivery systems; however, most of these incentives involve financial performance, reimbursement, and productivity, not quality of care.25 Important incentives for physicians to participate in quality improvement initiatives include:

  • Compensation (capitation, bonuses, risk pools, fee for service)
  • Profiling of performance
  • Recognition
  • Leadership opportunities
  • Penalties/sanctions
  • Equity participation

With respect to the safety and quality of care, whichever model of physician and health care delivery system develops, a new and robust model of safety, quality, and clinical outcome measures is needed to allow new methods of measuring the success of such models beyond the usual financially-driven metrics. Without new, standardized, comprehensive models of quality and safety of care, it will be difficult for any delivery system or its associated physicians to adequately demonstrate that it has improved the safety of care.


In the current environment of concern about health care safety and quality, virtually every stakeholder in health care—federal and state governments, employers, patients, physicians, and managed care organizations—is involved in efforts to improve quality. Various health care regulatory agencies have been developing clinical outcome measures that health care delivery organizations will be held accountable for. The JCAHO, the National Committee for Quality Assurance (NCQA), and the American Medical Association (AMA) have cooperated in the creation of disease-specific outcome measurement sets for hospitals, although these efforts did not focus specifically on safety. The Center for Medicare and Medicaid Services (formerly the Health Care Financing Administration) has developed a new scope of work with a great emphasis on clinical outcomes and will issue condition-of-participation rules that will require Medicare participating hospitals to implement medical error reduction programs.26

Congress has considered numerous bills in the area of patient safety; bills pending would improve peer-review protection for safety and quality reporting and address funding for proven technology interventions.19 At the state level, various state legislatures are beginning to pass mandates, as evidenced by California Senate Bill No. 1875 (Minimization of Medication-Related Errors), which requires hospitals and surgical clinics to implement a formal plan to eliminate or substantially reduce medication-related errors by 2005.26

The JCAHO created extensive patient safety standards that went into effect July 2001. The new standards emphasize patient safety programs, safety reporting metrics, patient/family involvement, and proactive redesign of high-risk clinical processes. Strong emphasis is placed on organizational aspects of safety. Indeed, the JCAHO has considered giving organizations a management safety profile as part of the survey process. Ultimately organizations will receive a separate overall accreditation score for patient safety.

Employers have shown significant interest in patient safety, as evidenced by the high profile obtained by The Leapfrog Group, a purchasing coalition made up of over 80 of the largest employers in America. The group is targeting three interventions that they believe will improve patient safety: 1) computerized physician order entry; 2) volume-based referrals for selected procedures, and 3) medical intensivist ICU coverage.26

The AMA has created the National Patient Safety Foundation (NPSF) to sponsor research and education in the area of patient safety. Many of the efforts of the NPSF have built on the experience of the Anesthesia National Patient Safety Foundation. The American Society of Anesthesiology founded the Anesthesia Patient Safety Foundation in 1984 and has since promulgated a number of reforms that have substantially changed the routine practice of anesthesia. Although there has been no systematic evaluation of the impact of these new standards on the safety of care, there has been a substantial decline in anesthesia-related deaths over the last 20 years.26 Although many other physician groups have indicated interest in this area, to date few major patient safety initiatives have been launched by other physician professional societies. So far the leadership for new initiatives in patient safety has resided in health care delivery organizations and their associate regulators rather than in medical professional societies.

Perhaps no group has had a greater public impact in patient safety than the Institute of Medicine (IOM), which recently issued three reports focused on clinical quality and patient safety: 1) To Err is Human,1 discussed above, which focused on medical errors; 2) Crossing the Quality Chasm,15 which examined the current state of the U.S. health care system, delineated numerous shortcomings, and recommended principles for system redesign; 3) Envisioning the National Healthcare Quality Report,27 which defined a framework and measures for clinical quality (Chart 1). All of these reports have called for new ways of thinking about health care quality and new accountability for health care quality; Chart 1 outlines a potential model that health care delivery organizations could use to both integrate patient safety initiatives into an overall quality improvement framework and develop integrated patient safety and quality performance metrics.

Chart 1. Framework for measuring health care quality and safety.


An overall framework for health care quality and a standard set of performance metrics are prerequisites for physicians and health care delivery organizations to document improvement in the safety and quality of care delivered.26 However, as discussed above, most research has focused on how to improve safety at the “sharp end” of care—the practitioner's role—rather than at the “blunt end,” or organizational level. Therefore, the most immediate priority should be on building safer and more reliable health care organizations. As previously demonstrated with successful improvements in the quality of care, physicians can, will, and must play a large role in this effort to improve organizational aspects of patient safety.

Etiologic models of safety such as that of Reason's “Swiss cheese” model of latent failures28 are not easily adaptable for improvement efforts. What health care lacks is a performance improvement model that is focused on the organizational aspects of safety.8

We propose a model based on high-reliability organizational theory for improving the safety of care within health care organizations.6 This organizational model (Table 1) offers physicians and health care organizations a framework in which to implement best practices in patient safety. The model includes an overall framework for the development of highly reliable, safer health care organizations. It consists of seven essential components, a set of guiding principles, and a model of organizational evolution. The seven essential components are:

Table 1:
Organizational Components of Patient Safety
  • Governance and leadership: Sponsorship of and accountability for the performance of the organization in patient safety
  • Culture of safety: Shared values and beliefs that define the organization's approach and response to issues of safety
  • Learning environment: An environment that allows the organization to openly report, discuss, learn from, and seek optimal resolution to identify safety issues
  • Program objectives: End-state goals or objectives that the organization will achieve from a patient safety program(s)
  • Safe process: Process created by knowledge gained from industry, organizational experience, and internal monitoring/surveillance to actively reduce the probability of error
  • Process implementation: Standard solutions (process and technology) used across the organization, allowing for customization based on unique business and clinical requirements
  • Measure and monitor: A continuous process (proactive and retrospective) of measurement and surveillance to identify emerging trends, new issues, and performance in relationship to patient safety targets
  • Learning environment: Continuous learning from experience, fed back into new or modified program objectives

The type of patient safety program that an organization has in place reflects the interest level and direct involvement of its board and medical staff, the extent to which senior executives have taken ownership of the patient safety agenda, and the clinical culture of the organization. Eight guiding principles provide the foundation for an organization's approach to patient safety:

  • Information sharing
  • Reporting
  • Role of patient and family
  • Education and training
  • Role of the physician and team based clinical leadership
  • Decision making and accountability
  • Threshold of safety
  • Role of technology

The seven components of the safety model can be combined with the eight guiding principles to create an evolutionary model of patient safety in health care organizations that envisions health care organizations as moving through four serial stages of safety (see List 1).

List 1. How Organizational Approaches to Patient Safety Evolve

Based on this model, we can identify significant gaps between the characteristics of most current-day delivery organizations and the high-reliability organization with respect to the roles and responsibilities of physicians in working with delivery organizations. Below we offer recommendations for approaches to bridge these gaps.


The gaps between the stage 4 high-reliability organization and the current state of health care delivery organizations are many. As organizations move from stage 1 toward stage 4 they will have to manage significant changes of a long-term nature vis-à-vis their relationships with physicians. Without a clear organizational mandate, an overall organizational strategy for improving patient safety, and specific measurable goals it is unlikely that patient safety will significantly improve in health care delivery organizations.

Recommendation: Health care delivery organizations should give the measurable improvement of patient safety high-level strategic priority over the next three years.

As organizations move to stage 2 and stage 3, governance and leadership will play critical roles. It is hard to see how any initiative in patient safety and quality of care could be successful without ownership and clear accountability residing at the senior leadership level of the organization. Traditionally, the chief quality officer has not had the same influence as the chief financial officer. Yet a chief patient safety officer might be faced with the need to interrupt a highly profitable clinical service (such as open-heart surgery) that may be temporarily unsafe. Would this individual be able to close a busy intensive care unit because low staffing levels had made it unsafe? These actions might be acceptable and necessary in the stage 4 high-reliability organization, but difficult or impossible in the current financially strapped environment of health care delivery organizations. Without a clear line of accountability in senior management it will be very difficult for organizations to make such tough decisions. For this reason, some health care organizations have appointed a chief patient safety officer in senior management.

As strong physician leadership is essential to engage physicians in any efforts to improve patient safety as well as the overall quality of care,23 the high-reliability organization will have to address various barriers to practicing physicians' involvement in improving the safety of care. These issues include legal risks, medical training, credentialing, peer review, team-based approach to care, and creating aligned incentives for improved physician performance in safety and quality. It will be very difficult for any individual to break through these barriers, and especially arduous for a non-physician.

Recommendation: Health care delivery organizations should create a new position in senior leadership for the chief patient safety officer, with an established budget and clear lines of accountability. This role should be filled by a physician with experience and training in quality improvement.

Among the challenges facing health care delivery organizations in improving system aspects of safety, creating a culture of safety is perhaps the most difficult. For physicians it requires a complete change in perspective from focus on personal perfection to a broader view of system reliability. External issues will inevitably have impacts on efforts to create a culture of safety. For example, the new JCAHO requirement that patients must be informed about unanticipated outcomes outlines the urgent and compelling need for integration of physicians into safety initiatives.

Physicians will be very leery of sharing information about their own errors or those of peers if they know this information may be divulged (without their consent) to their patients, potentially placing them at risk for litigation or disciplinary action. It has been suggested that the physician has an ethical obligation to disclose to patients all errors and mistakes29; in any case, managing this debate and defining policies for an organization clearly are areas where physician leadership is essential.

The issue is further complicated by differential malpractice coverage between physicians and health care delivery organizations. Differential liability between a health delivery organization and physicians will serve only to emphasize a culture of blame rather than a culture of learning.30 Some form of shared liability between physicians and health care delivery organizations will be necessary before physicians and health care delivery organizations can effectively cooperate and create a true culture of safety. As organizations move to stage 3, where risk is prospectively identified with better adverse-event detection systems and prospective risk prediction models, better measurement of both risk and performance in safety would allow liability to be prorated for documented safety performance and ultimately also for safety risk. For this approach to be feasible, some form of shared liability will be a prerequisite.

Recommendation: Health care delivery organizations and their associated physicians should pursue joint malpractice liability arrangements that are prorated for measured performance in safety.

Creating a learning environment will require health care delivery organizations and their physicians increasingly to focus educational efforts on learning from local errors and mistakes, in addition to taking lessons from the literature. A promising venue for this change in focus is the continuing medical education apparatus, which often combines continuing medical education for practitioners with primary education for residents and interns. Examples of inserting patient safety teaching into medical education do exist.31

Recommendation: Physicians' professional organizations should lobby for the addition of patient safety education and training to the core curricula of medical schools and residency programs, in fellowship programs, and in all continuing medical educational programs.

The traditional approach to reviewing medical mistakes—the morbidity and mortality conference—needs to be evolved into a forum where physicians can openly and honestly discuss medical errors, and begin to address issues such as the balance between physician autonomy and collaboration and teamwork, the risk-prone nature of the complex medical environment, and the need to continually learn from mistakes.

The current hospital peer-review and credentialing processes also enhance the separation between physicians and the health delivery system regarding accountability and liability. In addition, the oversight of specialty boards and state licensing boards tends to emphasize a culture of blame, focusing on individual accountability rather than health system accountability. Approaches to team licensure for high-risk processes of care would be one possible response to these issues.

Recommendation: Health care delivery organizations, in concert with associated physicians, should overhaul their peer-review and credentialing processes to support a learning environment for safety for all clinical practitioners.

Designing safe processes of care is an essential part of improving the safety of care delivered as well as meeting JCAHO patient safety standards. These standards require the redesign of one high-risk process per year. This requirement obligates organizations to shift from a reactive to a proactive mode to improve patient safety; organizations must identify, prevent, and mitigate risk as part of the process of care.

Physicians do not typically closely attend to the details of process design; they focus on medical decision making rather than on the details of execution. At the same time, physicians are loath to accept detailed care plans that they have not played a major role in developing.21 There is no doubt that building safe and reliable high-risk processes requires keen attention to the details of execution; Norman and others have shown the importance of designing these new processes with error in mind, using lessons from human-factors research.1

In systems that have successfully implemented redesigned care processes, physicians have led many aspects of such processes.24 Such a process focus will of necessity involve a team-based approach, with a division of responsibilities and tasks. It will inevitably lead to loss of autonomy on the part of physicians; if this loss of autonomy is not coupled with a sharing of accountability by the provider organization, it will fail: physicians have already experienced the schism resulting from separation of responsibility from accountability in previous dealings with managed care organizations.

Recommendation: Physicians' professional organizations should lobby for changes in professional scope-of-care laws to foster team-based process-of-care initiatives to improve the safety of care for the organization, rather than the practitioner.

Measuring and monitoring the quality of care is the only way to know whether organizational initiatives actually achieve the desired goals. Without an overall framework for quality and specific metrics for patient safety it will be impossible to determine whether care is actually becoming safer. Such metrics should be focused on processes of care, rather than practitioners of care. In the past, managed care and health care delivery organizations have developed and measured individual physicians' performance metrics, only perpetuating the emphasis on individual—not team or organizational—accountability. It will be necessary to build system-level measures of performance as the primary metrics if organizations hope to evolve highly reliable processes.26 In addition, if these metrics are not aligned with financial performance incentives for both physicians and the organization, any resulting financial disincentives for either will seriously limit the success of safety initiatives.

Recommendation: Health care delivery organizations in concert with their associated physicians should identify specific and relevant patient safety metrics based on specific patient-related events, to measure the impacts of organizational safety initiatives on patients.

Recommendation: Health care delivery organizations in concert with their medical staffs should create an aligned financial incentive structure such that improvements in safety and quality of care do not lead to significant financial disincentives for either group.

Implementation of redesigned care processes is always a difficult task for health care delivery organizations. Key to successful implementation is strong clinical leadership throughout project conception, design, implementation, and evaluation. Physicians will need to lead this effort, will need to be educated in these approaches, and will need to be assured that implementation will not lead to loss of control of their patients or loss of revenue.

Recommendation: Practicing physicians should take an active role in frankly assessing the impact of organizational patient safety and other quality improvement initiatives on the safety of care delivered to their patients at the sharp end of care.


Physicians' participation is the keystone of any health care delivery organization's efforts to improve the safety of care. Acknowledging the significant improvement in safety that is possible is only the first step on a long road to achieving safety in patient care at a level that is as high as that in high-reliability industries. To achieve this, health care delivery organizations must commit themselves to this serious improvement, and find a way to successfully integrate physicians into their efforts to improve the overall safety of care. At the same time, physicians and their professional organizations must take a clear and unambiguous leadership role in this area. Physicians need physician leaders, effectively aligned incentives, and tools for improving and documenting improvements in quality and safety. Delivery organizations need the same things. How long will it take for delivery organizations to achieve high reliability in safety and quality improvement? The answer lies in the relationships they are able to develop with their physicians, and the physicians' willingness to take leadership roles in improving the safety of care.


1. Kohn LT, Corrigan JM, Donaldson MS. To Err is Human: Building a Safer Health System. Washington, DC: Committee on Quality of Health Care in America, Institute of Medicine, 2000.
2. McDonald CJ, Weiner M, Hui SL. Deaths due to medical errors are exaggerated in Institute of Medicine report. JAMA. 2001;284:93–5.
3. Hayward RA, Hofer TP. Estimating hospital deaths due to medical errors: preventability is in the eye of the reviewer. JAMA. 2001;286:415–20.
4. Chassin MR. Is health care ready for Six Sigma quality? Milbank Q. 1998;76:565–91, 510.
5. Ross AF, Tinker JH. Anesthesia risk. In: Miller RD (ed). Anesthesia. New York: Churchill—Livingstone, 1994.
6. Pizzi LT, Goldfarb NI, Nash DB. Promoting a culture of safety. In: Markowitz AJ (ed). Making Healthcare Safer: A Critical Analysis of Patient Safety Practices. Washington, DC: AHRQ Publications, 2001.
7. Sandroni S. Iatrogenic error and physician responsibility. Conn Med. 2001;65:167–8.
8. Cook RI, Woods DD. Operating at the sharp end: the complexity of human error. In Bogner MS (ed). Human Error in Medicine. Hillsdale, NJ: Lawrence Erlbaum; 1994:255–310.
9. Prager LO. Physicians can help implement “culture of safety.” Minn Med. 2000;83:14–5.
10. Greenwald L. Medical error in the physician office: an insurer's perspective. Medicine and Health Rhode Island. 1983;83:312–5.
11. Ely JW, Levinson W, Elder N, Mainous AG, Vinson DC. Perceived causes of family physicians' errors. J Fam Pract. 1995;40:337–44.
12. Parker JB. The effects of fatigue on physician performance—an underestimated cause of physician impairment and increased patient risk. Can J Anaesth. 1987;34:489–95.
13. Becher EC. Improving quality, minimizing error: making it happen. Health Aff. 2001;20(3):68–81.
14. Clarke J. What residents are not learning: observations in an NICU. Acad Med. 2001;76:419–24.
15. Kohn LT, Corrigan JM, Donaldson MS. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: Committee on Quality of Health Care in America, Institute of Medicine, 2001.
16. Shortell SM, Waters TM, Clarke KW, Budetti PP. Physicians as double agents: maintaining trust in an era of multiple accountabilities. JAMA. 1998;280:1102–8.
17. Hughes JM. Total quality management in a 300 bed community hospital: the quality improvement process translated to healthcare. Qual Rev Bull. 1992;18:293–300.
18. Leider HL. Influencing physicians: the three critical elements of a successful strategy. Am J Managed Care. 1998;4:583–8.
19. Shipon DM, Nash DB. Quality in health care: what are the problems and what are the solutions? Tex Med. 1996;96:61–5.
20. Weiner BJ, Shortell SM, Alexander J. Promoting clinical involvement in hospital quality improvement efforts: the effects of top management, board, and physician leadership. Health Ser Res. 1997;32:491–510.
21. Blumenthal D, Edwards JN. Involving physicians in total quality management: results of a study. In: Blumenthal D, Scheck AC (eds). Improving Clinical Practice: Total Quality Management and the Physician. San Francisco, CA: Jossey—Bass, 1995.
22. Berwick DM, Godfrey AB, Roessner J. Curing Healthcare: New Strategies for Quality Improvement. San Francisco, CA: Jossey—Bass, 1990.
23. Goodroe JH, Murphy DA. The algebra of health care reform: hospital—physician economic alignment. J Cardiovasc Management. 1999;10:16–20.
24. Hosler FW, Nadle PA. Physician—hospital partnerships: incentive alignment through shared governance within a performance improvement structure. Joint Commission Journal on Quality Improvement. 2000;26:59–73.
25. Weber V, Joshi MS. Effecting and leading change in health care organizations. Joint Commission Journal on Quality Improvement. 2001;26:388–99.
26. Trowbridge R, Wachter RM. Legislation, accreditation, and market driven and other approaches to improving patient safety. In: Markowitz AJ (ed). Making Healthcare Safer: A Critical Analysis of Patient Safety Practices. Washington, DC: AHRQ Publications, 2001.
27. Hurtado MP, Swift EK, Corrigan JM. Envisioning the National Healthcare Quality Report., Washington, DC: Committee on the National Quality Report on Healthcare Delivery, Institute of Medicine, 2001.
28. Reason JT. Human error: models and management. BMJ. 2000;320:768–70.
29. Hebert PC, Levin AV, Robertson G. Bioethics for clinicians: 23. Disclosure of medical error. Can Med Assoc J. 2001;164:509–13.
30. Brennan TA. Hospital peer review and clinical privileges actions: to report or not report. JAMA. 1999;282:381–2.
31. Estrada CA. Medical errors must be discussed during medical education. BMJ. 2000;321:507.
© 2002 Association of American Medical Colleges