In 2015, the Association of Perioperative Registered Nurses (AORN) published guidance aiming to reduce surgical site infection (SSI). One stipulation was that surgeons and other practitioners should wear “bouffant” style headwear that covered the hair, ears, nape of the neck, and facial hair, rather than the traditional “skull cap” style hat.1 As a result of this recommendation, the Joint Commission banned traditional operating room (OR) caps in favor of the paper bouffant hats, despite the absence of data supporting any improvement in patient safety. It is not clear where the discussion about bouffant hats began, but it escalated fairly quickly into policy. The policy was immediately met with dismay and dissatisfaction by OR personnel, and particularly by surgeons. As the debate intensified, some authors carried out and published studies which demonstrated either no difference in SSI rates,2,3 or even a potential increased risk of SSI4 with the use of bouffant hats. Despite these data, proponents for the change justified continuing the policy citing “potential benefits”—despite the dearth of supporting evidence.5
Virtually everyone today, when considering how the “bouffant policy” was created, would agree this was a flawed process which has been very difficult to reverse. There are many lessons to learn from how this policy was created, which, unfortunately, followed a pattern that is relatively common in health care. Indeed, readers may consider other examples within their working practice which reflect this style of policy design and implementation. In this case and others, a well-meaning, sensible-sounding policy is proposed and taken action upon to improve outcomes for patients, but without clear evidence to support the change. Then, once implemented, subsequent quality assurance exercises that measure adherence to the policy reassure institutional leadership that they are on the correct path to better patient outcomes, even when what is being measured has nothing to do with outcomes for patients. Such a policy then becomes stubbornly resistant to reversal, even amidst mounting evidence of limited (or no) benefit. This process leads to an illusion of quality improvement (QI), often with unintended consequence. The first, and often significant unintended, consequence is cost. For the bouffant policy, there are personnel costs in implementing and monitoring the policy, and also significant costs associated with the caps themselves. Secondly, this policy has the unintended consequence of environmental damage because these caps are not biodegradable. Thirdly, such a policy disrupts a sociologically-based tradition. This may seem trivial, but loss of agency and denial of tradition may contribute to risk of resentment, dissatisfaction, and burnout. Furthermore, healthcare professionals who disagree with the policy may be perceived as rebellious, obstructive, or menacing by institutional leadership, which may, in turn, erode team dynamics, and lead to apathy and disengagement by clinicians, especially when failure to comply with policy may be met with institutional sanctions.
Why does this happen in health care? The answer is complex; after all, healthcare systems are large, dynamic systems in which there are many tiers of decision-makers. Human decision-making does not always follow the rules of rationality and logic; human judgement can be flawed due to unconscious cognitive biases and heuristics—or mental shortcuts—as described by Kahneman and Tversky in their Nobel Prize winning papers and the former's book Thinking, Fast and Slow.6 A key cognitive bias when considering issues like the bouffant policy is “Substitution Bias”—where the brain is challenged with a difficult problem that may require lengthy thought, but the decision-maker substitutes a simpler and less complicated question. In the arena of improving patient outcomes and, in particular, SSIs, a simple solution such as changing caps may be the result of substitution bias for the more complex questions this situation really calls for. Cognitive biases are further compounded by solutionism—a commonplace desire to “do something,” or to be seen to be doing something, especially when there is pressure from political and institutional stakeholders. Another issue which contributes is the general lack of understanding of how evidence-based medicine is carried out. This is not just limited to laypersons, because healthcare professionals may also lack appreciation for this important facet of patient care.
Because we are all prone to make errors and to interpret data incorrectly, it is no surprise that common heuristics may slip into the day-to-day practice of those who seek to improve health care for patients. When examined from an objective standpoint, cognitive biases, combined with a lack of knowledge of how to design and answer research questions, and a desire for a tangible solution may lead to the instigation of irrational policies. This is human nature—we are probably all guilty of replacing difficult questions with easier questions that require less effort for seemingly meaningful outcomes, even when those questions do not lead to improved care. In the surgical headwear example, the question “How do we reduce the risk of SSI in surgery?” (a potentially difficult and multi-faceted question) was substituted with “How do we ensure that healthcare practitioners are wearing attire to potentially reduce SSI?”, which, in turn, led policy makers to substitute the question “How do we ensure that healthcare professionals are wearing a specific headwear?”. This new question is enticing because it is simple, concentrating on a very practical aspect of work (ie, what people do or do not wear), and it is a visible action (and therefore easy to “measure”). Furthermore, it fulfils the institutional desire for a solution, because it is clear and visible for all to see, and requires a physical act of compliance and display—with an obvious “before” and “after.” All of these aspects of the “solution” demonstrate a seemingly sensible, clear marker of progress, even if none of the actions actually improve patient care or outcomes. Such flawed policy design is particularly important when it originates from a person or organization with respected authority and executive power. Those with such authority have a responsibility to lead the way with true evidence-based policies. Unfortunately, it is likely that there will be similar “illusions of QI” in the future, similar to the bouffant hat debate. Table 1 illustrates some of the characteristics of QI measures that we propose may highlight the risk of this being the case.
One of the many issues in the bouffant hat debate is the complex and sometimes challenging relationship between clinicians (in this case the surgical community) and policy makers. Such difficult relationships may exist at an individual hospital (such as those who determine infection prevention strategies) or higher authorities such as the Joint Commission or Centers for Medicare and Medicaid Services. Each may not necessarily understand the motivations or overall aims of the other, and this can only be improved through open dialog and collaboration. A top-down directive that is unreactive and unresponsive does not seem like the application of effective leadership. Instead, active discussion with multidisciplinary committees that seek and apply best evidence is paramount. Collaborative, active, and positive leadership is required for any new policies to be both justified and successful.7
The bouffant hat debate has been an enduring and uncomfortable one. How can such a thing be avoided in the future? Firstly, policies should be based on evidence from the outset, rather than being subject to cognitive biases and solutionism. This requires insight from both individuals and institutions to identify and correct these human errors. Secondly, QI policies that have the characteristics of an illusionary nature must be addressed in a timely manner, preferably in collaboration with multidisciplinary partners and with effective leadership. We all have a responsibility to continuously review and re-evaluate policies for the overall benefit of patients, rather than for the sake of conformity and visual effect. There is a lot we can learn from the bouffant hat debate; clinicians must not play a passive role in the implementation of policies within their institutions, but instead adopt active, interested, and integrative roles for the benefit of patients.
1. Braswell ML, Spruce L. Implementing AORN recommended practices for surgical attire. AORN J
2012; 95:122–137. [quiz 38-40].
2. Rios-Diaz AJ, Chevrollier G, Witmer H, et al. The art and science of surgery: do the data support the banning of surgical skull caps? Surgery
3. Kothari SN, Anderson MJ, Borgert AJ, et al. Bouffant vs skull cap and impact on surgical site infection: does operating room headwear really matter? J Am Coll Surg
4. Markel TA, Gormley T, Greeley D, et al. Hats off: a study of different operating room headgear assessed by environmental quality indicators. J Am Coll Surg
5. Spruce L. Surgical head coverings: a literature review. AORN J
2017; 106: 306-16.e6.
6. Farrar, Straus and Giroux, Kahneman D. Thinking, Fast and Slow. 2011; 499.
7. Petro CC, Rosen MJ. What surgeons need to know about the Bouffant scandal. JAMA Surg
2019; doi: 10.1001/jamasurg.2019.2107 [Epub ahead of print].