Chapter 15. Creating a Culture of Safety

In Chapter 9, I discussed the tragic collision of two 747s on a foggy morning in Tenerife, the crash that vividly illustrated to everyone in the field of commercial aviation the risks associated with steep and unyielding authority gradients. In response to Tenerife and similar accidents, aviation began a series of training programs, generally called “crew resource management” or “cockpit resource management” (CRM) programs, designed to train diverse crews in communication and teamwork. Some of these programs also incorporate communication skills, such as training in Situation, Background, Assessment, and Recommendations (SBAR) and briefing/debriefing techniques (Chapter 9). There is widespread agreement that these programs helped transform the culture of aviation, a transformation that was largely responsible for the remarkable safety record of commercial airlines over the past generation (Figure 9-1).

As the healthcare field began to tackle patient safety, it naturally looked to other organizations that seemed to have addressed their error problems effectively.1 The concept of high reliability organizations (HROs) became shorthand for the relatively mistake-free state enjoyed by airlines, computer chip manufacturers, nuclear power plants, and naval aircraft carriers—but certainly not healthcare organizations.2,3 According to Weick and Sutcliffe, HROs share the following characteristics:2

• Preoccupation with failure: the acknowledgment of the high-risk, error-prone nature of an organization's activities and the determination to achieve consistently safe operations.
• Commitment to resilience: the development of capacities to detect unexpected threats and contain them before they cause harm, or to recover from them when they do occur.
• Sensitivity to operations: an attentiveness to the issues facing workers at the frontline, both when analyzing mistakes and in making decisions about how to do the work. Management units at the frontline are given some autonomy in identifying and responding to threats, rather than being forced to work under a rigid top-down approach.
• A culture of safety: in which individuals feel comfortable drawing attention to potential hazards or actual failures without fear of censure from management.

Over the past few years, the patient safety world has embraced the concept of HROs. Personally, I find the concept intriguing, but wonder whether it is something of a healthcare “MacGuffin.” Legendary film director Alfred Hitchcock used to love to insert a MacGuffin—a plot device, such as a piece of jewelry, a stack of important-looking papers, or a suitcase, that “motivates the characters or advances the story, but the details of which are of little or no importance otherwise”—into his films. Similarly, when I hear a healthcare organization pledge to become an HRO, I wonder whether they are trying to advance their own “patient safety story” without committing themselves to an actionable or measurable target. In this, I find myself agreeing with British safety expert Charles Vincent, who wrote, “Put simply, reading the HRO literature offers a great deal of inspiration, but little idea of what to do in practice to enhance safety.”4,5

This hints at an overarching challenge we face as we enter the crucial but hazy world of safety culture, which can sometimes feel like the weather—everybody talks about it but nobody does anything about it. When it comes to safety culture, everyone seems to have an opinion. As with many complex matters in life, the truth lies somewhere between the commonly polarized viewpoints. Is a safe culture top down or bottom up? Both. Is it “no blame” or dependent on a strong framework for accountability? Both. Is it a local phenomenon (“microculture”) or an attribute of an entire organization? Both. Is it enhanced by following a set of rules and standards or by creating a nimble frontline staff, able to innovate, think on its feet, and even break a rule or two when that's required for safe care? Well, both.

The gossamer nature of safety culture—sometimes defined as “the way we do things around here”—can be frustrating to those entering the field, whether their goal is to improve the safety of their own practice or that of an entire hospital. It seems to be nowhere and everywhere. Yet the past few years have seen real progress in the area of safety culture: our understanding of what it means is growing richer and more nuanced, and we are discovering robust ways of measuring it and, to a degree, improving it. This, thankfully, provides something for even the pragmatist to grab on to.

In a large teaching hospital, an elderly woman (“Joan Morris”*) is waiting to be discharged after a successful neurosurgical procedure. A floor away, another woman with a similar last name (“Jane Morrison”) is scheduled to receive the day's first cardiac electrophysiology study (EPS), a procedure in which a catheter is threaded into the heart to start and stop the heart repetitively to find the cause of a potentially fatal rhythm disturbance. The EPS laboratory calls the floor to send “Morrison” down, but the clerk hears “Morris” and tells that patient's nurse that the EP lab is ready for her patient. That's funny, she thinks, my patient was here for a neurosurgical procedure. Well, she assumes, one of the docs must have ordered the test and not told me. So she sends the patient down.

Later that morning, the neurosurgery resident enters Joan Morris's room to discharge his patient, and is shocked to learn that his patient is in the EPS laboratory. He sprints there, only to be told by the cardiologists that they are in the middle of a difficult part of the procedure and can't listen to his concerns. Assuming that his attending physician ordered the test without telling him, he returns to his work.

Luckily, the procedure, which was finally aborted when the neurosurgery attending came to discharge his patient and learned she was in the EPS laboratory, caused no lasting harm (in fact, the patient later remarked, “I'm glad my heart checked out OK”).6

The case illustrates several process problems, and it would be easy, even natural, to focus on them. Many hospitals would respond to such a case by mandating that procedural units use complete patient names, hospital numbers, and birthdays when calling for a patient; by requiring a written order from the physician before allowing his or her patient to leave a floor for a procedure; by directing that important information be read back to ensure accuracy; or by implementing a bar coding system that matches a patient's identity with the name in the computerized scheduling system. And these would all be good things to do.

But these changes would leave an important layer of Swiss cheese (Chapter 1)—perhaps the most crucial one of all—unaddressed, the one related to the cultural issues that influence, perhaps even determine, the actions of many of the participants. A closer look at the case, now through a safety culture lens, reveals issues surrounding hierarchies, the “culture of low expectations,” production pressures, and rules.6 After a brief discussion about measuring safety culture, I'll address each of these topics in turn, and then consider solutions such as teamwork training and checklists. Rapid Response Teams and dealing with disruptive behavior are covered in Chapters 16 and 19, respectively.

*Joan Morris and Jane Morrison are fictitious names, but are designed to illustrate the fact that the two patients’ real last names were similar to each other, but not identical.

While one could try to assess safety culture by directly observing provider behavior (and some studies do just that7,8), the usual method is through self-reports of caregivers and managers, collected via a survey instrument. The two most popular instruments are the AHRQ Patient Safety Culture Survey and the Safety Attitudes Questionnaire (SAQ).9–11 A core benefit of using one of these validated instruments is that they have been administered to tens of thousands of healthcare workers and leaders, which means that benchmarks (for various types of clinical units and healthcare organizations) have been established. There has also been considerable research performed using these instruments, with results that have been surprising, sometimes even eye-popping. They include:

• Safety culture is mostly a local phenomenon. Pronovost and Sexton administered the SAQ to clinicians at 100 hospitals.12 While they found considerable variation in safety culture across hospitals (with positive scores ranging from 40% to 80%, Figure 15-1), there was even more variation when they looked at 49 individual units within a single hospital (positive scores ranging from 0% to 100%, Figure 15-2). It is not too much of an oversimplification to say that your hospital doesn't have a safety culture—rather, your fourth floor ICU has one, your step-down unit 50 ft away has a different one, and your emergency department has still another one. This is liberating in a way, since one need not look far afield for the secrets of a strong safety culture—it's likely that some unit in your organization has already figured them out.
• Safety culture varies markedly among different providers, and it is important to query all of them to paint a complete picture of a unit's culture. In a seminal 2000 study, Sexton et al. administered the SAQ to members of operating room teams (Figure 15-3).13 Whereas nearly 80% of the surgeons felt that teamwork was strong in their OR, only one in four nurses and 10% of the residents felt the same way. More recent studies have found that these differences persist.14,15
• Safety culture is perceived differently by caregivers and administrators. Singer et al. found that problematic views of culture were much more common among frontline providers than hospital managers working in the same facilities.16 They argue that this difference may be an important determinant of an organization's safety: if frontline workers perceive poor safety culture but the leaders remain blissfully unaware, the two groups are unlikely to be able to make appropriate decisions regarding how to prioritize safety concerns. Interestingly, another study by the same group found that there were differences even among managers, with senior managers having rosier views of culture than lower-level administrators.17

There is increasingly powerful evidence that safety culture scores correlate with clinical outcomes, including infection and readmission rates.10,18–21 The Joint Commission now requires that accredited organizations administer and analyze safety culture surveys each year, and doing so is one of the National Quality Forum's “safe practices” (Appendix VII). Of course, simply measuring safety culture without a strategy to improve it will be of limited value.

In Chapter 9, I introduced the topic of authority gradients—the psychological distance between a worker and his or her supervisor. These gradients are often referred to in terms of hierarchies; an organization is said to have steep hierarchies if these gradients tend to be large. In the wrong patient case described at the start of this chapter, the nurse's reluctance to raise concerns when told that her patient was being called for a procedure she didn't know about, or the neurosurgery resident's decision to acquiesce after getting pushback from the cardiology staff, can be attributed at least partly to steep authority gradients. So too can the tragic events that doomed the KLM 747 in Tenerife (Chapter 9). You may also recall the case that begins Chapter 9, in which a Code Blue is aborted when a physician finds a “no code” order in the wrong patient's chart. There too, a young nurse was reluctant to question the resident's confident proclamation, even though she had received sign-out an hour earlier that portrayed the patient's status as “full code.”

Authority gradients are natural and, to a degree, healthy. But when they prevent caregivers with important knowledge or concerns from sharing them with superiors for fear of being wrong or angering the boss, they become dangerous. This issue pertains not only to direct clinical care but also to whether the ward clerk, pharmacist, or intern will point out hazardous conditions to his or her floor manager, pharmacy director, or department chair. Safe organizations find ways to tamp down hierarchies at all levels, encouraging individuals (including patients, Chapter 21) to speak up when they see unsafe conditions or suspect that something might be awry.

In their masterful discussion of the wrong patient case in our Quality Grand Rounds series in the Annals of Internal Medicine (Appendix I), Chassin and Becher introduced the concept of “the culture of low expectations.”6 In reflecting on the actions—or inactions—of the nurse and the resident, they wrote:

We suspect that these physicians and nurses had become accustomed to poor communication and teamwork. A ‘culture of low expectations’ developed in which participants came to expect a norm of faulty and incomplete exchange of information [which led them to conclude] that these red flags signified not unusual, worrisome harbingers but rather mundane repetitions of the poor communication to which they had become inured.6

Fighting the culture of low expectations is a crucial step in creating a safety culture. Doing this requires a change in default setting of all the providers, from:

1. If you're not sure it is wrong, assume it is right (it's probably just another glitch, we have them all the time around here)

   to

2. If you're not sure it is right, assume that it is wrong—and do whatever it takes to become sure, even at the cost of delaying the first case in the operating room or making someone important a little irritated.

There is no shortcut to fighting the culture of low expectations. Yes, it is important to clean up communications with tools such as SBAR (Chapter 9) and information technology (Chapter 13)—making the instinct to think, “another glitch, but it must be right,” less automatic. (I often tell healthcare leaders that they can improve their organization's safety immediately by simply purging the four words—it must be right—from their staff's vocabulary.)

But healthcare organizations will always have examples of faulty, ambiguous communication or rapid changes that cannot be communicated to everyone. Therefore, changing the dominant mindset from “A” to “B” requires a powerful and consistent message from senior leadership (Chapter 22). As importantly, when a worker does take the time to perform a double check after noticing something that seems amiss—and it turns out that everything was okay—it is vital for senior leaders to vocally and publicly support that person. If praise is reserved for the “great save” but withheld for the clerk who voiced a concern when nothing later proved to be wrong, staff are unlikely to adopt mindset “B,” because they will still fret about the repercussions of being mistaken.22

Another, perhaps more subtle, culprit in the EPS case—and in safety culture more generally—is production pressure.23,24 In a nutshell, every industry, whether it produces hip replacements or SUVs, has to deal with the tension between safety and throughput. The issue is not whether they experience this tension—that would be like asking if they experience the Laws of Gravity. Rather, it is how they balance these twin demands.

When my children were small, we often went to a well-known pancake restaurant on weekend mornings for breakfast. I was in awe of the establishment's speed: customers were seated promptly, their orders were taken rapidly, putting a fork down on a cleaned plate immediately cued the server to politely ask “is there anything else I can get you today?,” and the cleanup was brisk. The reason for this pace is obvious: in the fast food business, with large boluses of customers arriving en masse and relatively low profit margins per customer, corporate survival depends on throughput. The food is fine but not great, and there are periodic errors (of the two-forks, no-knife type). But that's what happens when production trumps safety and reliability.

Only a couple of miles from this particular restaurant is San Francisco International Airport (SFO). The tension between production and safety is particularly acute at SFO, since its two main runways are merely 738 ft apart—I often feel like I can read the newspaper headlines of a passenger in an adjacent plane when both runways are operating.

The Federal Aviation Administration (FAA) enforces inviolable rules about throughput at SFO and every other airport. For example, when the cloud cover falls below 3000 ft (which, in foggy San Francisco, happens often), one of the two runways is closed, gumming up the entire U.S. air traffic control system. And whatever the weather, planes cannot land more often than one each minute. The explanation for these rules is as simple as the explanation for the restaurant's speed: when the aviation industry has to choose between production and safety, safety wins. The result: despite 400,000 takeoffs and landings each year, there has not been a fatal crash at SFO in over 50 years.

I've used this metaphor many times in speeches to hospital staff and leaders, usually capping it by asking audiences: “As it balances production and safety, does your hospital look more like the pancake restaurant or SFO?” Although things have improved a bit in the last few years, the answers still run about 10:1 in favor of the restaurant.

Addressing production pressures sometimes involves creating more efficient processes that free up time to do the work safely (just recall the description in Chapter 13 of the maddening waste of time and effort that often occurs when a nurse has to chart a blood pressure measurement). But it also involves choices, such as slowing down the operating rooms when they reach a state of chaos. This is expensive and disruptive, but so is closing a runway. In aviation, our society willingly pays this price to ensure safety, but we have not done so in healthcare.

As always, it is vital to appreciate the limits of analogies between healthcare and other industries. When SFO closes Runway 10R, planes sit on the ground (or circle overhead) and passengers are inconvenienced. But if we close our emergency department to new ambulances (“go on divert”) because of safety concerns, patients, some quite ill, need to be taken to other facilities. This carries its own safety cost, one that needs to be weighed as we address this challenging balance.

Charles Vincent quotes an oil industry executive who captured the tension beautifully; I think of this statement every time I hear a hospital CEO or board member earnestly declare, “safety is our top priority.” The executive said:

Safety is not our top priority. Our top priority is getting oil out of the ground. However, when safety and productivity conflict, then safety takes priority.4

I saw the sign in Figure 15-4 on the gate outside a construction site in London. Can you imagine a similar sign in the entrance to your hospital's operating room or emergency department? Until you can, we need to reflect more on the nexus between production pressures and patient safety.

It should be clear from the above discussion that safety culture involves far more than what happens in the operating room, in the labor and delivery suite, or on the wards. It is influenced, and sometimes determined, by the organizational culture, leadership, context, incentives, and more. For this reason, there has been increasing emphasis on the engagement of healthcare boards and other leaders.25,26 I will return to their crucial role at the end of this section and again in Chapter 22.

But much of the discussion around safety culture does center on the world of caregivers. You have already heard a fair amount about this in our discussions of handoffs and teamwork (Chapters 8 and 9). But let's now turn to efforts to improve culture, most specifically through team training, checklists, and the thoughtful use of rules.

As I've noted, one of the key ways that the commercial aviation industry transformed its culture in the 1980s was through team training (CRM) programs.1 Healthcare was relatively slow to embrace this kind of training for several reasons: it was not required by regulators or accreditors (Chapter 20), it was often challenging to get clinicians (particularly physicians) to participate, many people found the concept unattractive (I've had colleagues ask me whether teamwork training will require them to hug their colleagues and sing “Kumbaya”), and the evidence for its effectiveness was relatively weak.

The last objection has been largely answered by the experience of the U.S. Veterans Affairs (VA) healthcare system. Neily et al. compared surgical outcomes of 74 VA hospitals in which surgical teams had undergone team training with those of 34 facilities that had not yet implemented such training.27 They found that the risk-adjusted mortality rate had fallen 50% more in team training hospitals than in controls. Moreover, there was evidence of a dose–response curve, with additional training yielding even further cuts in mortality. The particular training program used by the VA emphasized four core areas of competency: team leadership, situation monitoring, mutual support, and communication.28 Several other studies have also demonstrated the benefits of team training,29–33 but there have been negative trials.34 Still, after a period of high hopes but substantial uncertainty, the weight of the evidence now supports the premise that teamwork training improves both safety culture and clinical outcomes.

Healthcare CRM programs come in a variety of shapes and sizes, though all focus on improving communication skills, forging a more cohesive environment, encouraging caregivers to speak up when they have concerns, addressing errors in a nonjudgmental fashion, and implementing tools such as time outs and briefs/debriefs (Chapter 5). The general experience is that, even when using an “off the shelf” program, it is best to allow for some local customization. In Table 15-1, Mayer et al. outline a number of implementation factors that help determine success29 and, in Table 15-2, Weaver et al. enumerate 12 best practices for evaluating team training programs.35

From the literature and my own experience in implementing team training at UCSF36 (the outline of our program is shown in Table 15-3), I'd emphasize the following practices and strategies:

1. Employ strong leadership and “champions”: Changing a culture is hard, and one person's “empowerment” might be another's loss of power and status. The case for culture change has to be clearly articulated, and buy-in—particularly from the physicians, nurses, and hospital executives at the top of the authority ladder—is vital. Luckily, thoughtful leaders now appreciate the safety risks associated with poor teamwork and steep authority gradients, and many are welcoming CRM-type programs.

2. Use aviation and other relevant nonhealthcare analogies, but don't overdo them: At UCSF, we used actual airline pilots in some of our training, and this was helpful. Pilots can confess to medical audiences that their culture was previously similar to that of physicians, and that they participated in team training programs reluctantly at first, but that they now have come to believe that such training is essential to improving culture and decreasing errors. The Tenerife story and others like it are powerful ways to energize a healthcare audience, and—particularly for physicians—the pilots' experiences are quite relevant.

   However, healthcare providers will be quick to highlight differences between an operating room and a cockpit. The biggest difference: a cockpit “team” is made up of two or three individuals with similar training, expertise, income, and social status—operating in quiet behind a sealed door. Dampening this hierarchy is easy compared with doing so in a busy operating or delivery room, where the authority ladder might span from the chief of neurosurgery or obstetrics to a ward clerk who is a recent immigrant with a high school degree (Chapter 9). Moreover, the process of flying is now so automated that pilots are mostly in the cockpit (like control room workers at nuclear power plants) to react to emergencies. Medicine is a long way from being that routinized. These differences mean that aviation examples should rapidly give way to medical ones.

3. Consider whether to use simulation: Experts debate the utility of adding simulation to culture and communication training programs (there is less debate on the utility of simulation for improving technical skills; Chapters 5 and 17).37,38 Proponents of simulation in CRM training argue that it “raises the stakes,” allows participants to suspend disbelief more readily, and makes lessons more memorable.39 Others feel that simulation (particularly high-fidelity, realistic simulation, as opposed to role-playing case scenarios) is not worth the added cost and complexity, and might even distract providers from the lessons regarding communication and collaboration.40 Whether simulation is or is not used, it is critical to “get everybody in the room” to work through case-based scenarios, and to provide an opportunity for skillful debriefing and interdisciplinary cross-talk. Didactics can go only so far in transforming deep-seated attitudes and beliefs.

4. Programs must live on beyond the initial training: Although a brief (e.g., 4–6 hours) CRM program can sensitize workers to some of the key cultural and communication issues, effective programs outlive their initial training phase. Whether the programs live on in the form of “unit-based safety teams” or other vehicles, the new emphasis on teamwork and collaboration must be supported by structures that endure (see also Chapter 22).41,42

Before leaving the topic of teamwork training, I'd like to loop back to where we began this section. Team training may seem like a convenient solution to poor safety culture, and the evidence now confirms that it does help. But, as Salas et al. point out, such training likely accounts for a small fraction of a team's effectiveness; the rest is determined by the culture and actions of the organization.43 They continue:

The organizational system matters. What the top leadership does, matters. What policies and procedures are in place to support teamwork, matters. The formal and informal signs and symbols of what is important in the organization—as conveyed through the norms, conditions, policies, procedures, metrics in place, and the messages that top leadership sends—make or break transformational culture change. One cannot forget that organizations tend to obtain the behaviors, cognitions, and attitudes that they measure and reinforce. We need to shift from thinking about a “team training intervention” to creating and sustaining an organizational system that supports teamwork. The best team training in the world will not yield the desired outcomes unless the organization is aligned to support it.

I've mentioned checklists several times throughout this book, mostly as a method to ensure adherence to key care processes. It may seem strange to now see a section on checklists in a chapter on safety culture. Yet in his book, The Checklist Manifesto, surgeon and author Atul Gawande reminds us that checklists can address a number of safety targets.44 In leading the multicenter study that established the value of the World Health Organization's surgical checklist (Figure 5-2),45 Gawande came to appreciate that the value of “the lowly checklist” was not simply in ensuring that the team gave the right preoperative antibiotics at the correct time, but in prompting team members in the operating room to introduce themselves to each other before a procedure begins. In other words, it was as much a culture-changing intervention as a cookbook.

Of course, concerns about “cookbook medicine” explain much of the resistance to the use of checklists, particularly among physicians. Gawande writes:

… we have the means to make some of the most complex and dangerous work we do … more effective than we ever thought possible. But the prospect pushes against the traditional culture of medicine, with its central belief that in situations of high risk and complexity what you want is a kind of expert audacity…. Checklists and standard operating procedures feel like exactly the opposite, and that's what rankles many people.44

Reflecting further on the intersection between checklists and physician culture, Gawande writes; “All learned occupations have a definition of professionalism, a code of conduct … [with] at least three common elements”: selflessness, an expectation of skill, and an expectation of trustworthiness. He continues:

Aviators, however, add a fourth expectation, discipline; discipline in following prudent procedure and in functioning with others. This is a concept almost entirely outside the lexicon of most professions, including my own. In medicine, we hold up ‘autonomy’ as a professional lodestar, a principle that stands in direct opposition to discipline…. The closest our professional codes come to articulating the goal [of discipline] is an occasional plea for ‘collegiality. ’ What is needed, however, isn't just that people working together be nice to each other. It is discipline.44

Gawande would doubtless agree that checklists can't solve all our problems, but they—and other safety-oriented activities such as standardization, simplification, forcing functions, and double checks—can help us deliver healthcare that is safer and more reliable. In applying these solutions, though, we need to grapple with the fact that checklists challenge some of physicians' most deeply held beliefs about what it means to be an excellent doctor.

The relationship between checklists and culture is a two-way street: checklists can influence culture, but their success can also be determined by an organization's ambient culture. While today's enthusiasm for checklists is well grounded, it is worth reading descriptions by Bosk et al.46 and Dixon-Woods et al.47 of the co-interventions crucial to the remarkable success of the Michigan central line–associated bloodstream infection study (which is often characterized—too simplistically—as a “checklist intervention”; Chapter 10).48 We are already beginning to see failures of interventions that were organized around the naive assumption that simply embedding a checklist into a complex clinical process, without addressing matters of culture, will enhance safety.

One other point about checklists has emerged from recent research. After participating in the WHO surgical safety intervention, 21% of clinicians disagreed with or gave neutral responses to the statement, “The checklist prevents errors in the operating room.” Yet 93% of the same respondents stated that they would want the checklist to be used if they themselves were having an operation.49 This is Lake Wobegon Meets Patient Safety: humans naturally believe they are above average in their carefulness.50 Accordingly, tools such as checklists might seem useful for “the other guy”—you know, the one who isn't quite as careful as I am. But life doesn't work that way. Inserting checklists—to be used by everyone—into error-prone processes as part of a comprehensive strategy to improve culture represents one of the major recent advances in patient safety.

Developing a great checklist is not a trivial undertaking. Like many good checklists, the WHO surgical checklist went through scores of iterations before the final version was produced.44 Through experiences like this and collaboration with other industries that have been using checklists for decades, we are beginning to understand some of the design principles for effective healthcare checklists. Figure 15-5, from McLaughlin, depicts a poor and an improved surgical checklist, illustrating some of these tenets.51

Gawande's comments on discipline naturally lead us into a discussion of rules and standard procedures. Healthcare providers are, of course, familiar with clinical rules: treat hypovolemia with fluids, don't give gadolinium to patients in renal failure, abscesses need to be drained. Although such rules were developed informally in the past and handed down by experts to trainees, our modern paradigm is to base clinical guidelines on firm scientific evidence (Chapter 3).

The field of patient safety is not agnostic about such clinical guidance. Providing high-quality, safe care requires that we pay attention to evidence-based rules and guidelines while allowing clinicians the flexibility to deviate when their patients do not fit the template (such as in patients with multiple and overlapping diseases).52 In Chapter 13, we discussed the role of computerized clinical decision support systems, which should make it easier for clinicians to follow such rules by providing up-to-date guidance at the point of care.

But these clinical rules fall largely in the domain of quality of care (Chapter 3). In the safety field, we are more concerned with rules and policies enacted to prevent patient harm. Examples might include: perform a “time out” prior to the first incision (Chapter 5), read back a critical lab value (Table 13-3), use a checklist before inserting a central line (Chapter 10), and check the patient's identity before sending her off to a potentially dangerous procedure (this chapter's case). Such rules (e.g., “flight attendants, please arm doors and cross check for takeoff”) are the mainstays of many industries' safety programs, but healthcare's attention to them has been sporadic, at best.

Before we criticize healthcare workers too harshly for breaking safety rules, it is worth noting that rule breaking is part of the human condition, and not always a bad thing. History, of course, is filled with stories of gadflies who were responsible for major breakthroughs. And who among us never breaks rules, such as the ones concerning jaywalking and speed limits?53 In a fascinating study quoted by Vincent,4 a social scientist interviewed British railway shunters, the individuals responsible for coupling and uncoupling trains, depositing them into sidings, and generally ensuring their safe movement.54 The researcher found that, despite the risky nature of the work, the shunters frequently violated established safety procedures, sometimes for reasons of inexperience, sometimes because they felt a rule was unnecessary or silly, and sometimes because they believed their action did not pose any serious risk (Table 15-4). James Reason calls the last category a routine rule violation, and most times they are harmless enough.55

The problem is that an unblinking acceptance of rule violations has diffused throughout medicine, particularly among physicians, whose status atop the hierarchy and, at least in the United States, as independent entrepreneurs (rather than employees of the hospital or healthcare system) has traditionally given them license to ignore rules. But, just as with the broken windows theory of policing (which holds that letting people get away with minor violations, such as breaking windows in abandoned houses, tacitly encourages them to try their hand at larger crimes),56 it can be argued that our permissiveness toward healthcare rule violators is responsible for the high rate of violations of some very important rules, such as hand hygiene. As Henry Kissinger famously said, “History teaches that weakness is provocative….”57

Yet before we embrace a more punitive approach to rule-breakers (which I discuss more fully in Chapter 19), it is worth recognizing the extraordinary complexity swirling around this area. Vincent4 highlights the contributions of Diana Vaughan (whose study of the Challenger shuttle disaster introduced the concept of normalization of deviance) and Rene Amalberti (whose studies have illuminated the frequent migration of workers from the “legal” to the “illegal-normal,” sometimes driven by differences in risk perception by individuals).58,59 Vincent and coworkers also describe a fascinating study by French anesthesiologists documenting the brief lifespan of many safety rules.60 “They are born,” Vincent writes, “they have a vigorous youth where most people take note of them, they begin to get old and less notice is taken of them, and finally they die.”4

As our quest to improve safety drives us to pay more attention to rules and procedures, we need to be sure that the rules we promulgate are important enough to be worth paying attention to, that we distinguish “must follow” from “nice to follow” rules (some organizations call the inviolable ones “Red Rules”),61 and that we enlist teams and peers to help enforce crucial rules. One aspect of high functioning teams is that they reach consensus that deviations from certain rules will not be tolerated, and then enforce such decisions with everything from gentle reminders to outright sanctions. You'll know you've achieved a safe culture when you see someone low in the hierarchy—say, a new nurse—reminding a senior physician to wash his or her hands, and the physician responds by simply saying, “thank you,” then turns to the sink or gel dispenser.

I end the chapter with a few general reflections on the remarkably rich and challenging issue of safety culture. Rather than emphasizing the issues above (promoting teamwork and collaboration, dampening down authority gradients, fighting the culture of low expectations, staying attuned to production pressures, mandating adherence to key rules and policies), some institutions define a safety culture in terms of the following questions:

• Do workers report errors?
• Is there a “no blame” culture?
• Is “improving patient safety” listed in the strategic plan?

Let's consider these points one by one.

As I discussed in Chapter 14, I don't believe that pushing workers to report, report, and report some more (particularly to an institutional or governmental reporting system) is likely to generate a safety culture. That said, local reporting (especially when there is a supportive unit leader who listens to problems and helps fix them, and when reports fuel engaging M&M conferences and unit-based problem solving) is important, and reporting can be a useful part of a broader effort to discover problems and engage frontline workers.

The issue of “no blame” is one of the most complex in the safety field. Throughout the book, I have discussed the importance of a systems approach, fighting the instinct to blame and replacing it with a focus on identifying system flaws that allow our inevitable human errors to cause harm. I believe that this new mental model has been responsible for much of the progress that we have made in the first decade of the patient safety movement.

That said, the “no blame” mantra has been taken by some to absurd levels. Blame is appropriate for individuals who make frequent, careless errors, who fail to keep up with their specialty, who come to work intoxicated, or who choose to ignore reasonable safety rules. David Marx and others have promoted the concept of a “Just Culture” (rather than a “no blame” culture) as a way to reconcile an appropriate focus on “no blame” when caring, competent workers make errors with the need to hold individuals (and institutions) accountable for blameworthy errors or conditions.62 This concept is explored more fully in Chapter 19.

Finally, placing patient safety in, or even atop, an institution's strategic plan is no guarantee of a safe culture (Chapter 22). Senior leadership and hospital boards make an important statement when they prioritize patient safety among their values and missions, but too many organizations and leaders offer lip service without promoting the effort and allocating the resources to make this commitment real.25,26 Creating a safe culture is hard work. But once it is created, it is easy to identify its presence, as well as its absence.

• Safe cultures are ones in which people willingly speak up when they see risky situations and behaviors, there are relatively flat hierarchies, workers follow critical safety rules, and the need for throughput is balanced against the need for safety.
• Many institutions are employing teamwork training programs, usually modeled on aviation's CRM programs, and emerging research on the results of well-implemented programs is promising.
• Validated survey instruments are available to measure safety culture and to benchmark against like clinical units or organizations.
• One aspect of a safe culture is fighting “the culture of low expectations,” in which workers assume faulty communication and therefore fail to perform double checks despite clear warning signs.
• Checklists can not only encourage compliance with safe processes of care but also facilitate culture change by improving teamwork and communication. Their success, in turn, can also be influenced by the ambient culture.
• The concept of a “Just Culture” has been advanced to emphasize the importance of blending a systems focus with appropriate individual and institutional accountability.