Preventing Medical Errors

Creating a Culture of Safety Reduces Preventable Medical Errors

Tim Vogus Every year, nearly 100,000 Americans die as a result of preventable medical errors. In addition to this staggering toll in human lives, medical errors also cost the U.S. health care system billions of dollars. Since the Institute of Medicine announced these troubling statistics in a landmark 1999 report, patient safety has received renewed attention by researchers and practitioners alike. Preventable errors occur in a variety of forms yet often follow common patterns. Patients, for example, may receive the wrong medication—or the wrong dosage of the correct medication. While the actual number of medical errors and their preventability have been intensely debated since 1999, there is little argument that serious mistakes are both frequent and costly. With the recent announcement that Medicare no longer will reimburse hospitals for costs associated with preventable errors, the need to improve patient safety takes on even greater urgency.

PHOTO: Timothy Vogus, Assistant Professor of Management (Organization Studies and Health Care)

In the search for ways to reduce errors, recent research has concluded that the most fruitful solutions focus on eliminating systemic defects that give rise to errors. Toward that end, many experts are urging health care providers to develop a “culture of safety” that permeates their organizations. A safety culture is more than the product of shared values and attitudes. It also involves patterns of behavior. And it involves an observable, tangible effort with which all members of the organization work together to minimize errors that harm patients. To be effective, moreover, a safety culture requires the ongoing collection, analysis and dissemination of data regarding patient care.

During the first half of this decade, an emerging body of research has linked a safety culture with reductions in medical errors and patient deaths. Some recent studies also have provided evidence that supportive leaders within hospitals enable the development or strengthening of a safety culture. However, noted Professor Tim Vogus of Vanderbilt Owen Graduate School of Management, the research also had left a gap. Although these early studies were suggestive, they offered only indirect evidence of how a safety culture can reduce the number of medical errors. The research mostly had focused on factors that create a context for promoting safety, such as the role of leaders; creation of processes and procedures; and the encouragement of open communication and reporting of errors. But researchers had devoted only limited attention to examining how those practitioners on the front lines of health care delivery—such as nurses—took specific actions to promote patient safety.

There was an understandable reason for this apparent gap of information, explained Vogus and his research colleague, Professor Kathleen Sutcliffe of the University of Michigan. No comprehensive measurement tool existed for directly gauging the safety-promoting behaviors of caregivers. Without an ability to assess these actions, it is impossible to determine precisely which mechanisms and interventions hospitals should use systematically to reduce errors at the point of care.

Vogus and Sutcliffe set out to fill that gap. In the process, they developed and validated a “Safety Organizing Scale” (SOS), a self-report measure for capturing and analyzing behaviors within hospital nursing units that are associated with indicators of patient safety. Their research was published in the journal Medical Care in January of 2007.

For several reasons, they focused their research on nursing units. First, nurses interact with patients more directly and more frequently than anyone else in the hospital. Second, noted Vogus and Sutcliffe, registered nurses serve as a hospital’s early detection system when there are medical errors and complications in care. They are well positioned to head off mistakes and to minimize negative outcomes for patients. Moreover, a safety culture encompasses shared actions and processes at the unit level; by definition, it involves collective behavior rather than the actions of individuals. Finally, documenting behavior at the unit level allows researchers to examine variations among units at different facilities and even within the same hospital.

The researchers also chose to focus on reported medication errors and patient falls–two safety indicators that are primarily influenced by RNs. They controlled for other factors that might affect safety behavior and errors—including the average tenure of RNs on a unit, the percentage of RNs with at least a bachelor’s degree in nursing, and the type of unit surveyed.

While no comprehensive measurement tool existed previously for nursing units, a number of detailed case studies were available of other operations—such as nuclear power plants and aircraft carrier flight decks—that must routinely perform complex, interdependent tasks almost flawlessly while under pressure. These HROs—high-reliability organizations—provided an excellent model for measuring the actions that would define a safety culture in a hospital. Case studies of HROs show that safety cultures are critical in enabling these organizations to achieve their high reliability. Moreover, the safety cultures come to life within HROs among front-line employees who practice “safety organizing”—a set of five interrelated behaviors by the unit as a whole:

Preoccupation with failure: Operating with a chronic wariness of the possibility of unexpected events that may jeopardize safety by engaging in proactive and pre-emptive analysis and discussion.
Reluctance to simplify interpretations: Taking deliberate steps to question assumptions and received wisdom to create a more complex and nuanced picture of ongoing operations.
Sensitivity to operations: Ongoing interaction and information-sharing about the human and organizational factors that determine the safety of a system as a whole.
Commitment to resilience: Developing capabilities to detect, contain and bounce back from errors that have already occurred but before they worsen and cause more serious harm.
Deference to expertise: When attempting to resolve a problem or crisis, decision-making authority migrates to the person or people with the most expertise with the problem at hand, regardless of their rank.

Drawing on their own research on HROs, their fieldwork in hospital nursing units and emergency departments, and on case studies of hospitals attempting to become HROs, Vogus and Sutcliffe translated the five behavior characteristics of high-reliability organizations into nine corresponding, Safety Organizing Scale survey items with which to gather information from hospital nursing units.

Preoccupation with Failure

When giving a report to another nurse who is about to assume the duties of caring for a patient, “we usually discuss what to look out for.”
“We spend time identifying activities we do not want to go wrong.”
“We discuss alternatives as to how to go about our normal work activities.”

Sensitivity to Operations

“We have a good ‘map’ of each other’s talents and skills.”
“We discuss our unique skills with each other so we know who on the unit has relevant specialized skills and knowledge.”

Commitment to Resilience

“We talk about mistakes and ways to learn from them.”
“When errors happen, we discuss how we could have prevented them.”

Deference to Expertise

“When attempting to resolve a problem, we take advantage of the unique skills of our colleagues.”
“When a patient crisis occurs, we rapidly pool our collective expertise to attempt to resolve it.”

To further ensure that their survey items were both theoretically solid and clear to RNs, Vogus and Sutcliffe enlisted seven experts in HROs and nursing to assess the validity of the SOS questionnaire and to pretest it among a sample of RNs. After this preliminary validation, Vogus and Sutcliffe conducted their study of nursing units within 13 different hospitals in seven states. The hospitals varied in size (from 89 to 478 acute-care beds) and setting (five were urban, five were midsized metropolitan facilities, and three were rural).

The results from their survey upheld Vogus and Sutcliffe’s SOS model as a precise measure of specific, self-reported behaviors underlying a safety culture. As they had hypothesized, units where these behaviors prevailed reported substantially fewer medication errors and patient falls. Their findings also rebutted some previous studies, which had predicted that a safety culture would lead to the reporting or more errors rather than fewer. Instead, Vogus and Sutcliffe found that high numbers of reported errors by a nurse unit were associated with lower ratings for quality of care by nurse managers.

More than the Safety Organizing Scale is involved in developing a safety culture. As other research has shown, it also takes leaders who create a supportive context that makes it safe for RNs to discuss and address errors. Research also suggests that safety cultures can take root when RNs trust managers and administrators to treat them fairly. As they had expected, Vogus and Sutcliffe found a strong correlation between nurse units that followed SOS behaviors and those with high levels of trust in their managers. (At the same time, in assessing the validity of the Safety Organizing Scale, Vogus and Sutcliffe were able to demonstrate that their SOS construct was distinct from the related factors of strong leadership and high trust.) Finally, Vogus and Sutcliffe’s research affirmed a link between lower patient-to-RN ratios and lower incidences of errors.

While the researchers acknowledge the limitations of their study—all of the hospitals, for example, were part of the same health system, and the research examined only the behavior of RNs, not doctors—the findings nevertheless offer important and encouraging news for health care providers. By applying the Safety Organizing Scale to nursing units, hospitals can gain a much clearer picture of the ongoing actions and interactions that underlie a safety culture, differentiate between safe and unsafe units, and improve their quality of care.

Send comments to timothy.vogus@owen.vanderbilt.edu

Reported by Randy Horick