Design Decisions to Overcome Clinicians' Fatigue

With Mustafa Ozkaynak, PhD

CU Nursing Associate Professor Mustafa Ozkaynak has received a new grant from the National Institutes of Health to study nurses’ decision-making under conditions of fatigue. Paul is a co-investigator on this project. The study has some cool technical aspects: Besides self-reported measures of fatigue we’re going to use EHR audit logs to see how clinicians interact with the electronic health record when fatigued versus not. But more fundamentally, the study can tell us interesting things about how people who are tired relate to the world. There’s a lot of data showing that people do poorly when fatigued on tasks that require lots of conscious focus and attention. This has been studied extensively in aviation, for example, and is the reason that crews and pilots have mandatory hours of rest time between flights. Clinicians who are fatigued make more medical errors, communicate less effectively with colleagues, and are less able to relate to others. In the language of Two Minds Theory, this occurs because fatigue seems to people’s shift mental processing from the Narrative mind to the Intuitive mind, which is more affected by biases and heuristics.

In the context of emergency department work, a clinician might normally go through her daily routine operating mainly by habit, using mostly the Intuitive mind (aka "System 1"). Most of the time, the Intuitive mind is very successful in making clinical decisions – it works fast, it is good at multitasking, and it gets better as a nurse develops expertise. Sometimes, it can produce the right decision based on recognizing a pattern even though the clinician might not be able to tell you in words what was going on – in predicting mortality, for instance, it’s useful to ask a health care provider “would you be surprised if this patient were to die in the next 12 months?” That question provides information above and beyond what you can get from the patient’s vital signs or the diagnosis in their chart, and shows the power of the intuitive mind. And nurses are able to identify subtle signs that something is "not right" with a hospitalized patient, such that this type of subjective assessment is as beneficial as the patient's vital signs in identifying which patients require a rapid medical response. There’s even some evidence that tired people operating mostly out of their Intuitive mind do better at some tasks, like understanding the gist of things said in a foreign language that they don’t speak. 

The complication is that when something unexpected comes up, critical thinking is needed. At that point a clinician might typically slow down and go into a more deliberate, “System 2” or Narrative mode of thinking. Similarly, stopping to document things in the electronic chart probably involves Narrative thought: it requires reflection on previous intuitive steps taken in order to put them down in language, crossing all the t’s and dotting all the i’s. Sometimes, the electronic record might prompt a clinician to realize that she skipped a step, which she could then go back and deal with. The interplay between Narrative and Intuitive thought is usually a good thing. It helps us respond quickly in the moment but also reflect on our actions, and keeps us from making serious mistakes based on either system's weaknesses.

Although fatigue tends to push people into Intuitive modes of thinking, and Intuitive thought is the norm for clinical practice, fatigue also unfortunately makes the Intuitive mind less efficient. People’s thinking is more likely to be drawn to trivial aspects of a situation, they might skip steps, and they can’t work as fast. People who are fatigued might become more defensive and less able to collaborate in making decisions. Intuitive-mind errors based on heuristics and biases become even more common under conditions of fatigue; the basic theme seems to be that the brain takes even more shortcuts than usual just to keep going. Intuitive expertise is still there, but its performance degrades. Those are the conditions when someone is more likely to make a mistake - for example, a medical error like not following clinical practice guidelines for good antibiotic stewardship. Clinicians are often too eager to hand out antibiotics, because it satisfies social norms and makes the patient happy even if it’s medically inadvisable and creates the risk of antibiotic-resistant bacteria. By leaving the decision up to the Intuitive mind and its social sensitivity, a fatigued brain makes this problem even worse. 

The central idea of Mustafa's grant is that we could make modifications to the electronic health record that use technology to counteract these effects of fatigue. The most common strategy to compensate for weakened Intuitive-level processing is to try kicking the person’s thinking to the Narrative mind instead, to slow things down and engage System 2 with some extra prompts or reminders. That’s often how people prevent errors when things get complex and they aren’t tired. But unfortunately the effects of fatigue on System 2 can be even worse than their effects on System 1. People are less able to concentrate, less able to use logical reasoning, and less able to pay sustained attention when they are tired. A lot of Narrative thinking depends on the ability to focus attention on a problem, and if you think about being very tired you’ll understand that attention is often the first thing to go. So just putting in more reminders might not do the trick. There’s some evidence that under these conditions nurses start to click through the reminders built into their electronic health record system without really seeing them, or develop "workarounds" to avoid them, because the Intuitive mind views the reminders as just one more annoyance. An alternative approach might be to slow down the pace of the interaction with the system, to provide less information rather than more, to cut back to just the essentials based on a person's clinical role so that they have to deal with fewer buttons and prompts. The idea is that reducing the input or level of stimulation might help the nurse to conserve scarce cognitive resources, which would allow for better Intuitive-mind decisions even under conditions of fatigue. 

We aren’t sure what approach is best, so this exploratory study is designed to first describe the ways in which fatigue affects nurses’ decision-making, including the ways in which they interact with the electronic medical system when they are fatigued versus not. The second aim is then to design electronic health record modifications that might help to prevent fatigue-related errors. Antibiotic prescribing is something that happens often, and it’s a situation in which errors are likely to occur, so that clinical scenario gives us an opportunity to measure clinicians’ fatigue in various ways and to see what differences there are in the process when a nurse is fatigued versus not fatigued. 

Up until now, Two Minds Theory has been tested mainly in studies that looked at how patients self-manage their chronic illnesses. Mustafa’s study is particularly exciting to me because this will be the first time that this theory has been used in a study of health care workers’ interactions with technology. Interactions between people and technology are constant in the contemporary health care environment, and if we can gain a better understanding of how those interactions change based on the person’s level of fatigue, it could lead to many potential future applications for improving care.