This case study first appeared in the report Committed to Safety: Ten Case Studies on Reducing harm to Patients by Douglas McCarthy and David Blumenthal, M.D.
Organization. The Veterans Health Administration, a component of the Department of Veterans Affairs (VA), operates the nation's largest integrated health care system, providing care to more than 5.1 million veterans and 7.6 million enrollees at more than 1,300 sites of care nationwide. In response to public and congressional concerns about the quality of care in VA facilities, the VA undertook a broad organizational transformation during the 1990s, which included "rationalization of resource allocation, explicit measurement and accountability for quality and value, and development of an information infrastructure supporting the needs of patients, clinicians, and administrators" (Perlin et al. 2004).
Objective and intervention. Former VA Undersecretary for Health, Kenneth Kizer, M.D., stated the VA's objective in this way: "To make health care safe, we need to redesign our systems to make errors difficult to commit and create a culture in which the existence of risk is acknowledged and injury prevention is recognized as everyone's responsibility" (Weeks et al. 2000). The VA established a National Center for Patient Safety to lead organizational culture change and empower local facilities and frontline staff with proven tools, methods, and initiatives for patient safety improvement.
Date of implementation. The VA was a pioneer in instituting a comprehensive patient safety program within a large health care system, beginning its journey toward a culture of safety in 1997⁸two years before the publication of the IOM's To Err Is Human report. These efforts led to the creation of the VA's National Center for Patient Safety in 1999.
Process of change. The VA's culture change program drew upon human factors principles and the experience and lessons from industries such as aviation and nuclear power. To create a culture of safety, the VA adopted a nonpunitive approach to patient safety. "No one commits an error to hurt someone on purpose, but they are punished as if they did," said James Bagian, M.D., director of the VA's National Center for Patient Safety. That kind of "name and blame" approach doesn't correct the underlying system issues. "Now we ask, 'What happened? Why did it happen? What can we do to prevent it from happening in the future?'" he added.
On the other hand, an organization cannot promise a "blame-free" environment because some willful actions are blameworthy. To make this distinction clear, the VA sought the cooperation of Congress, the Joint Council on Accreditation of Healthcare Organizations (JCAHO), and the unions with which it works to define what acts would be subject to blame or punitive action. These were defined as criminal acts; any act involving alcohol, substance abuse, or patient abuse, or a purposefully unsafe act (i.e., the individual knew it was unsafe but did it anyway). The VA clarified that these blameworthy acts should be dealt with administratively, not within the patient safety system. This policy did not lead to any loss of data, Dr. Bagian noted, since such events are unlikely to be reported by those who commit them. Furthermore, the policy gives confidence to those who have not committed blameworthy acts that they will be treated in a nonpunitive manner when they report honest errors.
The National Center clarified that employees should report all adverse events and close calls (i.e., events that could have resulted in patient harm but did not) to their facility's patient safety manager, who would then use a computer system to centrally report these incidents (Figure 2). Because close calls are much more common than adverse events, they can provide valuable information to diagnose latent system weaknesses that may lead to errors. Reports are protected under federal statute. The individual filing the report remains identified until a root cause analysis is completed so that he or she can be notified of and comment on the findings. External reports, described below, are similarly protected.
To discover and learn from a fuller range of patient safety issues, in April 2002, the VA instituted an external patient safety reporting system modeled on NASA's aviation safety reporting system. Operated by the NASA Ames Research Center, the patient safety reporting system is intended to provide a "safety valve" for those who are not comfortable reporting adverse events or close calls to the internal VA patient safety reporting system. Reporters are encouraged to include identifying information so that they can be contacted in case additional details are needed to fully characterize the incident, but the record is stripped of all identifiers once it is deemed complete. Reports are analyzed by a team of NASA patient safety experts.
The National Center designed easy-to-use, computer-aided, root-cause analysis tools and cognitive aids for multidisciplinary teams of frontline staff to analyze reported safety incidents. Triage questions and rules of causation help teams identify actionable root causes in six major categories. For example, one of the rules of causation specifies that a violation of procedures cannot be a root cause, since the factors that led to the violation must themselves be investigated. Facility managers were given a standardized methodology (known as the "safety assessment code matrix") to prioritize incidents based on their severity and probability of occurrence. The National Center conducted three-day training programs over a period of nine months to teach frontline staff how to use these tools, and provides ongoing support to facility staff through calls and site visits.
The findings of root cause analyses are presented to the facility's CEO, who either approves recommendations for corrective action or proposes alternatives until there is mutual agreement on the remedies to be taken. The National Center maintains a database of root cause analyses, which it uses to investigate selected issues. Aggregate analyses of similar events are used to identify common systems issues and develop action plans based on findings from multiple cases. Topics have included patient falls, medication errors, missing patients, and suicidal behavior.
The National Center adapted a systems engineering tool for prospective risk assessment, known as Healthcare Failure Modes and Effects Analysis. This tool is used to proactively identify critical system vulnerabilities that might cause patient harm and to design and assess the effectiveness of system improvements. The VA also has been a leader in adopting information technologies, such as a computerized medical record system that provides timely access to clinical information for VA providers whenever and wherever they need it to inform diagnosis and treatment. A medication bar coding system helps improve patient safety by reducing errors in medication ordering and administration.
Results. Within 10 months of enhancing its internal patient safety reporting system, the VA experienced a 30-fold increase in the reporting of incidents, indicating that the promise of confidential, nonpunitive reporting was important to the workforce (Bagian et al. 2001). The relatively small number of reports received by the external patient safety reporting system—fewer than 400 in two years of operation as compared with hundreds of thousands submitted internally in five years—suggests that the VA has achieved a high level of trust in its internal reporting system.
Since the creation of enhanced tools and training, nearly all root cause analyses identify a recommended action, whereas previously about half the reviews of patient safety incidents failed to identify an action. The average number of root causes identified has increased from one to three per incident, while the types of root causes identified have shifted from a focus solely on patient behavior and professional training to human factors and systems issues (Bagian et al. 2002). These trends suggest that teams no longer consider circumstances giving rise to error as nonremediable.
Reported medication administration errors substantially decreased after the introduction of medication bar coding (Johnson et al. 2002). An employee culture survey conducted in 2000 found that inpatient facilities that rated higher on teamwork culture tended to have higher levels of patient satisfaction than those with lower teamwork ratings (Meterko et al. 2004). According to Dr. Bagian, one measure of culture change at the VA is that personnel want to lead the way by exceeding—as opposed to simply meeting—JCAHO requirements.
Lessons learned. The critical success factors identified by Dr. Bagian include:
- creating a culture of safety that emphasizes system learning;
- defining what actions are blameworthy, assuring the confidentiality of safety investigations, and promising nonpunitive responses to reporting;
- providing easy-to-use tools to understand causes of errors and ways to prevent them, and
- understanding and surmounting obstacles to success.
"Creating a culture of safety is the only way to create a sustainable organizational focus on patient safety," said Dr. Bagian. The goal of a patient safety program must be focused on outcomes—preventing patient harm—not just on the process of reducing errors. The aviation industry provides an apt analogy; airline passengers do not want to know how many cockpit errors were averted, they simply want the plane to reach its destination safely and on time.
Because people will never be perfect, systems must be designed to be "fault tolerant." That is, the systems should prevent errors from occurring or from resulting in patient harm when they do occur. If a system relies on people to be perfect and for everything to go right, then harm will certainly continue to occur. Commercial airlines, for example, have at least two engines in case one should fail. "In health care, we are still flying a single-engine aircraft," said Dr. Bagian.
Creating an easy-to-use, effective capability for frontline staff to analyze and take action from incident reports was "the single most noteworthy means of changing the culture," Dr. Bagian said. "Someone shouldn't need a degree in human factors engineering to do root cause analysis," he added. Involving frontline staff created a sense of ownership and promoted new ways of thinking that have transferred to other quality improvements. One physician who was trained to perform root cause analysis said, "I look at my whole job differently now; I'm seeing things that I can fix that I didn't even know I could solve." This kind of positive "buzz" from frontline staff creates demand for training and tools, creating the perception that the patient safety program is useful to care providers and not just a mandate from headquarters.
Internal reporting systems cannot be used for punishment if they are to obtain the trust of staff. "Reporting errors is not about finding fault. Managers should ask 'who' and 'why' only to fix the problem," Dr. Bagian said. Safety culture surveys conducted at the VA suggest that the biggest obstacle to creating a transparent culture is not the fear of malpractice litigation but the sense of humiliation that professionals say they would feel when admitting to an error.
A voluntary external safety reporting system complements internal reporting systems by providing additional insights into broad system vulnerabilities. The usefulness of such reports does not depend on the volume of data collected, but from the veracity of information obtained. This complements the information obtained through investigations of internal reports, which is more precisely actionable. A combination of internal and external reporting systems helps the VA demonstrate its commitment to a culture of safety and organizational learning, says Dr. Bagian. The volume of reports received by the external system allows managers to gauge how well the internal reporting system is working; a large increase in external reports, for example, might suggest a need to increase the staff's confidence in the trustworthiness of the internal reporting system.
To identify and overcome obstacles, the National Center proactively approaches skeptics within the organization for criticism. This feedback provides a candid assessment of how programs are viewed by frontline staff.
Replication and related results. The VA's experience has been of wide interest both domestically and internationally, with methods and tools adopted in Australia, Canada, Denmark, Hong Kong, Japan, the Netherlands, Singapore, Sweden, and the United Kingdom. As the nation's largest provider of medical education, the VA has developed a patient-safety curriculum for medical residents, medical students, nurses, pharmacists, and other allied health care professionals. "Our goal is to start new health care professionals thinking about patient safety early in their careers, something that generally hasn't been done in the past," Dr. Bagian said in a press release announcing the initiative.
The VA's ability to collect and analyze incident reports across multiple institutions can be compared with the external reporting efforts undertaken by other institutions. Organizations participating in collaborative efforts, such as the Pittsburgh Regional Healthcare Initiative, find that reporting medication errors and health care–associated infections to external databases permits useful analyses across multiple organizations (Sirio et al. 2003). A few states have established nonregulatory patient safety centers to promote statewide patient safety reporting and analysis (Rosenthal and Booth 2004).
Implications. Patient safety incident reporting is not an end in itself, but can benefit the organization when coupled with an effective analytic and management process that drives organizational learning and action. Barriers to reporting include concerns about confidentiality, fear that information will be used punitively, lack of time to make reports, and failure to receive feedback after an error is reported (Jeffe et al. 2004). Management must provide feedback to frontline staff on lessons learned and actions taken in order to promote organizational buy-in.
A crucial element in the VA's success—not only in gaining employee participation in reporting but also in taking effective action from reports—was its ability to guarantee the confidentiality of reports under statute. Violation of this trust can destroy a reporting program; for example, one national aviation safety reporting system failed after an identifiable incident was disclosed (Connell 2000).
The aviation industry's experience suggests that a centralized safety reporting system can improve safety awareness if it is nonpunitive, confidential, independent, and easy to use; and if it produces timely, expert, and systems-oriented feedback (Leape 2002). The newly enacted federal Patient Safety and Quality Improvement Act may contribute to these goals by protecting the voluntary, confidential reporting of safety data to independent, federally certified patient safety organizations.
<back to top>
This study was based on publicly available information and self-reported data provided by the case study institution(s). The aim of Fund-sponsored case studies of this type is to identify institutions that have achieved results indicating high performance in a particular area, have undertaken innovations designed to reach higher performance, or exemplify attributes that can foster high performance. The studies are intended to enable other institutions to draw lessons from the studied organizations' experiences in ways that may aid their own efforts to become high performers. The Commonwealth Fund is not an accreditor of health care organizations or systems, and the inclusion of an institution in the Fund's case studies series is not an endorsement by the Fund for receipt of health care from the institution.