ISMP survey on perceptions of a nonpunitive culture produces some surprising results

Despite a growing awareness of the system-based causes of errors, many in healthcare are still struggling to come to terms with the role of individual accountability in a non-punitive culture. In our June 28, 2001 newsletter, ISMP Medication Safety Alert!, we offered a series of statements about a nonpunitive culture and asked subscribers to tell us about their personal beliefs on this issue. We emphasized that there were no right or wrong answers - only perceptions about a non-punitive culture in healthcare. Responses from 1,255 participants suggest that work is needed on all fronts to fully adopt a nonpunitive culture. Data from the survey showed that frontline staff may have received less education about the basic tenets of a nonpunitive culture.

About 15% of respondents believed that a nonpunitive culture excuses poor performance and absolves staff of personal responsibility for patient safety. Another 21% believed that such a culture might increase carelessness as individuals learn they will not be punished for their mistakes. But significant differences existed between respondents of different backgrounds. Not a single risk manager felt that a nonpunitive culture excuses poor performance, and only 8% agreed that such an approach might absolve staff of responsibility or increase carelessness. On the other hand, 21-26% of pharmacy technicians felt that a nonpunitive culture excuses poor performance, absolves staff of responsibility, and worsens carelessness. Similarly, 26% of staff nurses were worried about increased carelessness if punishment was not an option for mistakes. However, a nonpunitive, system-based approach to error reduction does not diminish accountability; rather it redefines and directs it in a more productive manner so that the workforce and leaders are equally accountable for patient safety, not zero errors (see our August 8, 2001 issue). Further, there is no evidence to support the premise that a nonpunitive culture will increase carelessness. In fact, experience has shown that it increases staff awareness of safety and sparks enthusiasm for changing systems and practices associated with errors.

Frontline staff (22%) were more likely than administrators (16%) or managers (9%) to believe that a nonpunitive culture may be detrimental to an organization. Likewise, nurses (19%), pharmacists (16%), pharmacy technicians (15%), and physicians (13%) agreed more readily than risk managers (5%) and executives (10%) that a nonpunitive culture inhibits their ability to weed out "bad apples." Technically, weeding out "bad apples" should occur during the recruitment process or initial probationary period when competency can be carefully evaluated using a more accurate measure than the presence or absence of errors. A nonpunitive culture does not inhibit this process, it strengthens it by eliminating the use of errors as a performance measure and forcing more accurate means of evaluating basic competency.

Most respondents (64%), especially physicians (78%), nurses (70%), and executives (70%) felt that remedial education was an effective nonpunitive remedy for staff involved in an error. However, remedial education is punitive in nature because it inappropriately singles out individuals who made the error. A more effective, nonpunitive approach to error assumes that there are others who are likely to make the same mistake given similar circumstances. Thus, if a knowledge deficit contributed to an error, educational efforts would be directed more appropriately to all who could make a similar error.

Quality improvement staff and pharmacists often assume a leadership role in carrying out "root cause" analyses of adverse drug events. To effectively use this process, all focus must be taken off individuals and placed on the system-based causes of error. Thus, it makes sense that quality improvement staff (16%) and pharmacists (13%) were more likely than nurses (5%), physicians (5%), and pharmacy technicians (7%) to believe that only system error exists. Yet, even in a nonpunitive culture, we must not lose sight of the fact that human error will occur. We must clearly acknowledge it, but look beyond it to identify and correct the system-based causes that allow human error to reach patients.

About 11% of managers, 14% of administrators, and 18% of staff felt that a nonpunitive culture tolerates failure. Yet, in reality, it is a punitive culture that tolerates failure, and a nonpunitive culture that can remove the fear of failure as a barrier to patient safety. A punitive culture stifles creativity, innovation, and willingness to change because the possibility of failure is greatly feared and perceived to be totally unacceptable. In addition, punitive cultures tend to remember failures for a long time, making it even more difficult for people to speak up about problems or new improvement ideas. Seen from this viewpoint, a nonpunitive culture offers just the opposite - a nurturing environment that is open to innovation, creativity, and change because fear of failing is not a limiting factor.

Overall, respondents were able to offer definitive opinions about many of the statements on the survey. But the jury is still out on the more controversial issues such as using errors as a performance or competency measure; the role of sanctions to improve performance; amnesty reporting policies; "error-prone" individuals; dealing with policy violations; and the public's view of a nonpunitive culture. Over a quarter of all respondents were undecided (could neither agree nor disagree with survey statements) on these more difficult issues that often unravel our best efforts to embrace a nonpunitive culture.

It's one thing to overlook a single lapse in performance, especially if our most qualified staff have been involved and the patient hasn't been harmed. But our tells a different story when it comes to dealing with people who have made frequent or fatal errors, or violated a policy that has resulted in an error.

Frequent or fatal errors: Except for pharmacists (39%), more than half of all respondents at the staff level (60%) and executives (54%) believe that employees who make repeated or fatal mistakes warrant disciplinary action or employment termination to protect the safety of patients. Nurses (65%) were the most likely to feel this way. Yet such harsh action, which is firmly rooted in hindsight bias, i.e., Monday morning quarterbacking, crumbles the very foundation of patient safety because it may encourage some to conceal their mistakes or "redefine" errors as non-errors (ISMP Medication Safety Alert! November 19, 1997). What's more, this punishment is unwarranted. Terminating employment in the wake of a fatal error is an ineffective, emotionally charged, knee-jerk reflex which, quite simply, is easier to do than getting to the bottom of an error and making system changes to make sure it doesn't happen again. In the case of disciplinary action based on the frequency of errors, it's really impossible to determine if one individual is making more errors than another using the typical methods of detecting and reporting errors. Unless research methodologies are employed, perceived patterns of error among staff are invalid, and most likely related to better error detection and reporting. This is especially true if closely monitoring new staff, technology, and processes. Instead of punishing those who appear to fall more frequently into the traps created by our inefficient, complex systems, we should recognize that they are our best lens to examine and strengthen our systems so that even our most experienced staff won't make the same mistakes.

Just as important, while systems and processes can be classified as error prone, individuals cannot. Physicians (38%), pharmacists (26%), nurses (25%), and technicians (24%) were the most likely to believe that people who make frequent errors while performing a specific function are error prone in other tasks as well. Yet, error prone personalities can't be pinpointed. While some people are more adept at detail work, others are better able to manage the "big picture." Although it's important to make sure people are well suited to their jobs, it's impossible to totally divorce detail and "big picture" work. Inevitably, each of us encounters both in a typical day. However, people who are more likely to make mistakes doing detail work are less likely to make mistakes with "big picture" tasks, and vice versa.

Policy violations: About 60% of managers, 54% of staff, and 47% of administrators felt that errors caused by policy violations warranted disciplinary action. Yet, isn't at least one policy violated every time an error happens? The intention is often to bend the rules for what seems like a good reason at the time. Chances are, that same policy has been broken before without a hitch, especially if rules have been used as "Band-Aids" for an ailing system, and following them has become a time-consuming, unrealistic goal. Policies may also be violated if managers fail to provide staff with sound rationale for adhering to them or if they send mixed messages that encourage rule breaking. Is speed or accuracy the real priority? Productivity or safety? The ongoing answers to these questions, sent both in word and deed, help shape staff performance as they make those necessary, moment-to-moment decisions on what can be accomplished within the allotted time and resources. Instead of punishing staff for errors caused by policy violations, it's far more important to determine the underlying causes of the violation and make the changes necessary to facilitate adherence or alter the policy.

Our survey laid bare the difficulties faced in changing the way people view individual and collective responsibility for errors. Recent responses from readers in disagreement with the tenets laid forth in Part II of our survey findings attest to a culture shift that is clearly foreign to the "Spare the rod, spoil the child" doctrine still embedded deeply in our culture. The story of this struggle continues, as survey respondents' views on the role of sanctions, amnesty, and errors as a measure of competence and performance are unfolded below.

Sanctions. In practice, managers and administrators are the most likely staff to impose sanctions on the workforce when an error occurs. Yet our survey showed that managers (22%) and administrators (19%) were less likely than frontline staff (33%) to believe that sanctions produced more careful individuals. Perhaps these findings reveal that many managers and administrators have seen firsthand that sanctions and warnings to "be more careful" have done little to stop errors. But it's equally important to recognize that use of sanctions can perpetuate the vicious cycle of blame. Because these actions have little or no effect on reducing errors, their continued occurrence fuels greater anger and exasperation on the part of managers because the work force has now been warned, yet still makes errors.1

Amnesty. Pharmacists (52%), physicians (50%) and executives (49%) were most likely to agree that an amnesty policy for those who report errors is indicative of a nonpunitive culture. And while amnesty may improve reporting, does it send the right message? If you are serious about changing the culture, do you really want the foundation to be built upon the premise that you should punish people for errors, but if they tell you what they did, you will pardon their offense? When it is safe for staff to report errors, they certainly should be held accountable for doing so. But such a requirement should not be tied to amnesty for making an error. With this in mind, perhaps it's not surprising that nurses (34%), who are often at the sharp end of an error where the caregiver/patient interaction occurs, were the least likely to believe that amnesty equates to a nonpunitive culture.

Performance and competency. About half of all staff, managers, and administrators felt that errors should be used as a performance and competency measure. Despite fairly even distribution between basic professional levels, 63% of executives felt that errors could measure competence and 74% believed that errors could measure performance. In contrast, less than half of all risk managers, quality improvement staff, and pharmacists believed these myths. But despite a growing awareness that even the most competent and experienced staff cannot outperform the systems that bound and constrain them, it's disheartening that so many respondents have not embraced one of the most important tenets of a nonpunitive culture: Errors are not valid measures of competence and performance.

While it's certainly easy to use errors to evaluate performance and competence of individuals, typical means of detecting errors are not comprehensive, and more to the point, errors measure the performance of an organization as a whole, not the individuals who comprise it. Once errors have been removed from individual performance and competency evaluations, healthcare will be forced to identify more accurate, useful, and valid measures that can better ensure that the workforce has the requisite skills, knowledge, and teamwork to provide safe, quality care.

Reference 1: Reason JT. Forward. In: Bogner MS, ed. Human Error in Medicine. Hillsdale, NJ. Lawrence Erlbaum Associates 1994, p. vii-xv.