Failure to perceive clinical events: An under-recognised source of error

Abstract

Introduction

Attentional focus narrows as individuals concentrate on tasks. Missing an event that would otherwise appear obvious is termed a perceptual error. These forms of perceptual failure are well-recognised in psychological literature, but little attention has been paid to them in medicine. Cognitive workload and expertise modulate risk, although how these factors interplay in practice is unclear. This video-based experiment was designed to explore the hypothesis that perceptual errors affect clinicians.

Methods

142 volunteers with varying levels of experience of adult resuscitation were shown a short video depicting a simulated cardiac arrest. This video included a series of change-events designed to elicit perceptual errors. The experiment was conducted on-line, with participants watching the video and then responding via combinations of open-ended free-text and directed questioning.

Results

141 people experienced at least a single perceptual error. Even the most clinically significant event (disconnection of the patient's oxygen supply) was missed by three in four viewers. Although expertise was associated with increased likelihood of detecting an occurrence, even highly significant events were missed by up to two thirds of the most experienced observers.

Discussion

This study demonstrates, for the first time, that perceptual errors occur during healthcare-relevant scenarios at significant levels. Events such as an oxygen malfunction would meaningfully affect patient outcome and, although expertise conferred some advantages, events were still missed more often than not. Data acquisition is fundamental to good-quality situational awareness. These results suggest perceptual error may be a contributor to adverse events in practice.

Introduction

Situational awareness, a term that describes the ability of an individual to process information about the environment in which they are functioning, is considered to be a “safety-critical skill”,¹ particularly in emergency-care type settings. The development of situational awareness can be modelled as a three-step process:² acquisition of relevant information, integration of that information into a coherent mental model, and then use of the mental model to make decisions. A failure in any one of these steps limits effective care delivery. Data from both healthcare and industry already demonstrate that loss of situational awareness contributes to a high proportion of critical incidents.3, 4, 5

Many factors can cause a loss of situational awareness. Poor training or judgement may limit the synthesis of mental models or the decisions made using them, and these may be the focus of investigations after serious untoward incidents (SUI). Comparatively little attention has been paid to human performance limitation, and specifically the role of perception, in loss of situational awareness.⁶ Introspection, based on day-to-day experience, leads us erroneously to conclude that our ability to perceive objects in our environment is unlimited, richly detailed, and automatic.⁷ There is a general assumption that educated, trained, experienced practitioners are skilled in noticing events in their environment.

This assumption is ill-founded, and it is surprisingly easy to demonstrate that perception is fallible. A number of studies in experimental psychology have demonstrated that individuals frequently miss conspicuous events when placed under relatively trivial stress, and that insight into these failings is universally poor. Consider this situation: you are stopped in the street by a stranger asking for directions. During the conversation you are distracted, and during this distraction the stranger is swapped with a different person. When you return to the conversation you are talking to a completely new individual. How likely are you to notice the change? When asked, 98% of people questioned expected to notice the switch.⁷ This experiment was carried out, but fewer than 50% of people tested were in fact aware that their conversation partner had changed part-way through their interaction.⁸ This phenomenon is termed ‘change-blindness’, and the poor insight into the effect ‘change-blindness-blindness’.

A related phenomenon, ‘inattentional blindness’ occurs when the brain edits out unexpected events before the individual ever becomes aware. A classic demonstration of inattentional blindness involves watching a video of teams passing basketballs.⁹ Part-way through the sequence an unexpected event occurs, as an experimenter crosses the scene dressed in a gorilla costume. If participants simply watch, without a primary task on which to concentrate, all report witnessing the costumed individual. If observers are given a concentration task to perform, for example counting the number of passes made by one team, around half will remain completely unaware of the gorilla's presence. The more taxing the concentration task, the more likely observers are to miss the unexpected event. It is possible that perceptual errors follow the ‘inverted-u’ pattern described by the Yerkes–Dodson Law,¹⁰ which suggests that task performance generally suffers at both high- and low-extremes of arousal. There is some evidence to suggest an optimal level of distraction may exist. Rates of inattentional blindness may paradoxically fall in the presence of low-level distractors in another sensory modality.¹¹

The role of memory in perceptual error has been the subject of debate in the psychological literature.¹² It was thought that the results of the original perceptual error studies might represent a failure to encode the unexpected stimulus into memory – in essence the object was perceived but swiftly forgotten. More recent work has demonstrated that memory traces for such objects do in fact persist even in the event of a perceptual failure,13, 14 suggesting that these phenomena are not primarily the result of memory lapse.

Change-blindness and inattentional blindness are two aspects of ‘visual awareness failure’, an umbrella term that encompasses a range of well-recognised limitations in perceptual ability, including also ‘repetition blindness’ and ‘attentional blink’.¹⁵ Taken together, these phenomena highlight significant gaps in human perceptual abilities. It seems plausible to expect that errors of perception occur in clinical practice, where staff are routinely placed under far higher degrees of stress than described in these studies. Clinical personnel operate in high-stakes, high-pressure environments where accurate decisions must be made, often under significant time constraints.

There are reasons to expect susceptibility to perceptual errors decreases with increasing expertise. When performing a familiar task it is known that experts can derive more relevant information from single glances than do novices,¹⁶ and that experts report lower workload for comparable activities.¹⁷ In the context of perceptual errors however, the protective effects of expertise appear to be highly task-specific – when the gorilla video described above is shown to professional basketball players, they are far more likely to notice the gorilla's presence,¹⁸ but professional athletes from other disciplines perform at levels comparable to the general public. It is unclear how expertise and workload effects interact in clinical practice. If perceptual errors occur at the frequencies suggested by psychologists’ data they may be significant, but largely unrecognised, contributors to loss-of-situational-awareness incidents in healthcare.

Our group, and others, have previously demonstrated that clinical training does not appear to confer any generalisable protection against perceptual error.19, 20 These previous studies have used variations of the Simons “invisible gorilla” paradigm. There are good reasons, however, to expect people not to notice items (such as gorillas) if they are not a feature of their normal working environment,²¹ making it difficult to assess the applicability of these results to normal practice. This study was designed to redress that gap.