Perspectives on Safety
What Makes a Good Checklist
The use of checklists is a primitive yet remarkably effective strategy for ensuring accuracy in complex tasks. Checklists have long been used in fields such as aviation and space exploration but have only recently made headway in medicine. The reluctance of medical professionals to adopt checklists is often framed as pushback against "more paperwork" and "cookbook medicine," or due to disbelief in their effectiveness.(1) However, a rich literature has helped establish many best practices in checklist design, and health care now stands to benefit.
Much research on checklist design and effectiveness originated in the field of human factors psychology.(2) Human factors research stresses the importance of looking beyond a specific technology to the guiding principles that govern human behavior. This theoretical focus allows for generalizing beyond a specific checklist or environment. Indeed, use of these principles has led to two strikingly successful and widely discussed medical checklists: a surgical safety checklist that resulted in more than a one-third decrease in deaths and complications (3) and a checklist guiding the activities surrounding the insertion of central venous catheters that reduced bloodstream infections up to 66% in approximately 100 Michigan intensive care units.(4) The success of medical checklists even has resulted in government mandates to use them after repeated surgical errors in some hospitals.(5)
Checklists, however, are not a panacea for medical mistakes. Any system has the power to complicate or assist—the design of the checklist determines its utility. A poorly designed or overly lengthy checklist can harm performance just as easily as a good one can improve it. Fortunately, guidelines for checklist design are available. These may be grouped into two general categories: knowing the task and knowing the users. Specific advice concerning the tasks and users is included under these major categories.
Knowing the Task
Knowing a task means much more than being an expert. Indeed, many experts are unable to describe exactly what a task requires because they are experts.(6) One characteristic of experts is that they tend to combine portions of tasks so that discrete steps are chunked together into larger steps. Such chunking tends to result in losing conscious access to each of the specific steps that a novice would need to know.
The human factors technique to "unpack" these essential steps—to make the implicit steps explicit—is known as task decomposition. The first step in task decomposition is task analysis, which takes many forms based on the goal of the analysis.(7,8) For example, cognitive task analysis outlines each step in terms of the abilities and limits of the human performing the task.(6) This type of analysis reveals portions of tasks that are overwhelming or could interfere with each other, such as when an anesthesiologist needs to vigilantly attend a display while remembering to check other systems.
Identify automatized subtasks. To know the task means appreciating which portions of the task are performed automatically and which require attention. The benefit to automatized tasks is high accuracy.(9) However, automatized tasks are no longer open to introspection by the performer. An example would be tying a shoelace—the "expert" lacer can perform the task quickly and accurately but would have a hard time describing the actions step by step. It is this characteristic of automatic tasks that makes them important in a checklist—it is possible to be unaware that an automatic task was performed. A checklist can help support individuals operating in such automatic environments by signaling that a task has been performed, but it generally does not need to address each step along the way. To extend the example, the steps in tying one's shoes do not need to be included in the checklist because once the shoes are tied, it is evident all steps were completed. On the other hand, if a task has automatized subtasks that do not give such evidence of their completion, then those subtasks do belong in the checklist.
Request specific task outcomes. Another way to operationalize the admonition to "know the task" is to make the performer record specific values on the checklist. In poorly designed checklists, the state of the listed items can be marked complete with a check, such as "Blood pressure reading—complete." Some early aviation checklists made this mistake, resulting in operators marking items complete when they were incorrect. Thus, checklists should require the value or status of an item rather than a check of completion (Figure) (e.g., "Blood pressure reading—Pressure is 120 over 80").(10)
Identify conflicting physical demands in the task. If both hands are used in the task, the checklist cannot require the use of hands, or the steps need to be grouped with pauses so that the checklist can be operated between subtasks. Consideration of physical demands is extremely important—for example, if hand hygiene is necessary, a health care worker cannot be prompted to touch the checklist, wash hands, and re-contaminate his or her hands as he or she records the action on the checklist.
Consider all possible task scenarios. Although a lengthy checklist can do more harm than good, checklists need to be complete. For example, in one aviation checklist the emergency procedures covered a two-engine failure at altitude but not failure near the ground when there was little time to react.(11) Including all scenarios requires length—not lengthier checklists, but different checklists for different scenarios. For this reason, modern airplane checklists are catalogued in books, and fast indexing is important.(11) This may be one area where technology can assist by providing fast ways to search the checklist for the appropriate scenario. The search method should be as well designed and thoroughly tested as the checklists.
Be realistic about the task. Know how people actually perform the task, not how the task is prescribed to be performed.(12) Over time, workers often learn to perform complex tasks in ways that are far more efficient than initially conceived, and the checklist should reflect this. Task analyses and observation of procedures can assist with this checklist requirement.
Knowing the Users
The users of medical checklists are both homogeneous and heterogeneous. They are homogeneous in that they have received similar training concerning some components of the task. They are heterogeneous in that every individual has a different history and background. It is likely that the health care field differs from the aviation field in this area, as there can be more variability in surgical team experience. For example, in a cockpit, the checklist may only address the actions of pilot and co-pilot—individuals with similar training, expertise, and status—whereas a surgical checklist may need to influence the actions of both a new nursing assistant and a senior surgeon. The following guidelines capitalize on knowing the users to ensure good communication via the checklist.
Include pauses. Most aviation checklist errors occurred immediately after an interruption to the checklist or when the user was fatigued or overwhelmed.(10) The operators then lost their place in the checklist (as they tried to remember where they left off) and sometimes skipped vital actions. In the implementation of the WHO surgical checklist, pauses were built into the checklist.(3) These "time outs" contain checklist items but are a review of past action and a plan for future action—no task progress occurs during the pause. An example item from the WHO surgical checklist is "Nursing team reviews: has sterility (including indicator results) been confirmed?"
Adhere to basic usability guidelines. Basic usability guidelines improve human perception and expectancy of the task steps. Although the developers of the WHO surgical checklist embraced human factors principles, which helped user perception, they later learned that basic usability still needed to be improved.(12) Recommended changes included use of readable fonts, avoiding ALL CAPS, and increasing the contrast of the words against the background.(12) Last, use familiar and accessible language. For example, a checklist should refer to "each day," not the more jargony "per diem." Standardize these terms as much as possible across clinical units and sites of care.
Indicate task possession. A crucial component of checklist design is to allocate tasks to individuals.(13) This prevents diffusion of responsibility (i.e., no one ends up doing a task assigned to everyone) and avoids assumptions. For example, an aviation checklist that addressed actions to be taken during an in-flight fire instructed the flight attendant to take one action if the plane was landing and another if it was not. Unfortunately, the checklist did not specifically instruct the attendant to contact the captain or co-pilot to determine flight status. Lack of task possession resulted in the flight attendant taking the wrong action, and the fire destroyed the airplane.(13)
Test the checklist. Every checklist should be tested with real users in realistic scenarios. Realism does not necessarily mean convincing surroundings, but it does mean a time-sensitive, high-pressure test with each tester multi-tasking and dealing with distractions. Testing is as much a part of making a good checklist as any principle of design.
The use of checklists is an integral part of many complex industries, from aviation to space flight to marine travel and nuclear reactors. The field of medicine is no less complex. Though the benefits of checklists are large, these benefits depend on proper design, testing, and use. Researchers and caregivers should welcome the advent of checklists in medicine, customizing the tools for their particular purposes while taking advantage of the literature from other domains to create successful, usable, and well-adopted checklists.
Anne Collins McLaughlin, PhDAssistant Professor
Department of Psychology, North Carolina State University
1. Gawande A. The Checklist Manifesto: How to Get Things Right. New York, NY: Metropolitan Books; 2009. ISBN: 9780805091748.
2. Drury CG. Human factors and ergonomics audits. In: Salvendy G, ed. Handbook of Human Factors and Ergonomics. 3rd ed. Hoboken, NJ: John Wiley & Sons, Inc; 2006:1106-1132. ISBN: 9780471449171.
6. Crandall B, Klein G, Hoffman RR. Working Minds: A Practitioner's Guide to Cognitive Task Analysis. Cambridge, MA: The MIT Press; 2006. ISBN: 9780262532815.
7. Hollnagel E. Task analysis: why, what, and how. In: Salvendy G, ed. Handbook of Human Factors and Ergonomics. 3rd ed. Hoboken, NJ: John Wiley & Sons, Inc; 2006:243-268. ISBN: 9780471449171.
8. Kirwan B, Ainsworth LK. A Guide to Task Analysis. Boca Raton, FL: CRC Press; 1992. ISBN: 9780748400584.
10. Degani A, Wiener EL. Human factors of flight-deck checklists: The normal checklist. Moffett Field, CA: NASA Center: NASA (non Center Specific); 1991. Accession Number: 91N27144; Document ID: 19910017830; Report number: A-90183, NAS 1.26:177549, NASA-CR-177549.
11. Sullenberger CB III, Zaslow J. Highest Duty: My Search for What Really Matters. New York, NY; William Morrow: 2009. ISBN: 9780061924699.
12. Davies JM, Caird JK. Surgical safety checklist: a redesign using human factors guidelines background to the surgical safety checklist. University of Calgary, Canadian Patient Safety Institute, Calgary, Canada; 2009. [Available at]
13. Burian BK. Emergency and abnormal checklist design factors influencing flight crew response: A case study. In: Proceedings of the International Conference on Human-Computer Interaction in Aeronautics, 2004. Toulouse, France: EURISCO International; 2004.