Human error

Introduction

Human error is often cited as the cause of accidents, when all other factors have been eliminated. However, this does not mean that human error should be considered a final or vague explanation. It can be studied and understood using scientific methods. In fact, today, there is considerable interest in the study of human error^[1]. The aim of this article is to describe human error and its relationship to occupational accidents.

Definition of human error

It is very difficult to provide a clear definition of human error^[2] as it is often the result of a complicated sequence of events and therefore an elusive phenomenon to analyse. However, Reason^[3] has defined “human error” in the following way: "Error will be taken as a generic term to encompass all those occasions in which a planned sequence of mental or physical activities fails to achieve its intended outcome, and when these failures cannot be attributed to the intervention of some chance agency." On the other hand, it has been said that to err (i.e. to make mistakes) is human. Human error is an element that cannot be totally eliminated, but if the typical errors are identified, most of them can also be prevented.

According to the traditional viewpoint, human error s seen as the cause of failures and accidents but other, more modern approaches, consider   human error as a symptom of failure, reflecting deeper problems in a system. Examining human error provides information to delve beneath the simplistic label of 'human error'. Human error is an attribution after the fact, and it is systematically related to people, tools, tasks, and operating environment^[4][5].

Although there is no unanimous definition of human error, the general thinking has changed from attributing guilt to an individual towards a much broader contextual approach.

One classification of human error regards them as 'Action errors' (action not as planned), which can be further categorised as 'slips' or 'lapses'; or as 'thinking errors' (action as planned) - classified as 'mistakes'^[6]. The inadvertent nature of such errors sets them apart from deliberate actions (known as 'violations') where an individual wilfully and knowingly adopts an incorrect course of action. Identification of human error

There are several models for describing human error. The best known is the "Swiss Cheese" model. In addition, various methods exist that can be used to identify the causes of human errors.

Accidents are rare

In the well-known "Swiss cheese" model, Reason^[3] suggested that there are several intrinsic defences and atypical conditions preventing accidents. In an ideal world, each defensive layer would be intact. In reality, however, they are more like slices of Swiss cheese, having many holes. These holes are continually opening, closing, and shifting their location. An accident happens when the holes in many layers momentarily line up to permit a trajectory of accident opportunity^[7]. The main message of the "Swiss cheese" model is that the chance of danger factors finding all in the holes lined up in all of the defences at any one time is very small and that is why accidents are rather rare.

The Swiss cheese model of James Reason

Source^[3]

Cognitive failures

The aim of the Cognitive Failures Questionnaire (CFQ) is to measure self-reported failures in perception, memory, and motor function^[8]. A number of studies have shown that a higher score indicates a greater frequency of cognitive failures and is associated with real-world consequences of these failures, such as occupational accidents^[9]. For instance, the scale was presented to 240 electrical workers in the United States Army. The CFQ predicted both car accidents and work accidents. When the foremen were asked to rate the workplace safety performance of 158 workers the ratings of foremen and workers agreed very well ^[10].

Based on the Cognitive Failure Questionnaire, Wallace and Chen^[11] developed the Workplace Cognitive Failure Scale with 22 items like "Cannot remember whether or not you have turned off work equipment?" Using this scale, the researchers showed that general cognitive failure predicted unsafe behaviours and micro-accidents of American workers. Later with a smaller sample the same scale predicted supervisor safety ratings, injuries and missed workdays.

The Skill–Rule–Knowledge (SRK) model

The Skill–Rule–Knowledge (SRK) model proposed by Rasmussen2 is a framework for understanding how people process information and make decisions in complex systems, especially under varying levels of familiarity and stress. The three levels of behaviour consist of: 

• skill-based behaviour, i.e. automatic, highly practiced actions such as typing on a keyboard, shifting gears while driving. Errors present themselves as slips or lapses, e.g., pressing the wrong button or forgetting a step. These errors often occur when attention is diverted or when a routine is interrupted.
• rule-based behaviour is behaviour guided by stored rules or procedures. Examples are following an emergency shutdown procedure or diagnosing a known machine fault. Rule-based errors, e.g., applying the wrong rule or misinterpreting the situation may lead to an undesired consequence^[12].
• knowledge-based behaviour refers to problem-solving in unfamiliar situations where no rule or routine applies, e.g. improvising during equipment failure. Knowledge-based errors occur especially in complex, high-pressure, or unknown situations and include flawed reasoning and incorrect assumptions. These include situations of incomplete or imprecise knowledge, and inexperience with the task^[12].

Rasmussen’s model shows that human errors are not random but depend on how familiar and practiced the task is. Research has shown that these three types of errors do not all lead to incidents or accidents to the same extent. In the fatal occupational accidents which occurred in Australia, two out of three were due to skill-based errors, one fifth to rule-based errors and the other fifth to knowledge-based errors. Equipment work practices were relatively clearly related to rule-based errors, personal protective equipment to skill-based errors, and management unsafe procedures to knowledge-based errors^[13]. In fatal accidents on British construction sites, skill-based errors and knowledge-based errors both caused nine fatalities, whereas only three fatalities were due to rule-based errors^[14].

In the cardiology ward of a Japanese hospital 181 accidental and incidental events were reported during a six month period. A total of 40 of the reported events were classified as skill-based errors, 52 as rule-based errors, and seven incidents were designated as knowledge-based errors. A total of 12 errors were life threatening^[15]. 

Aircraft mechanics in Australia reported 666 human errors. They spent 65% of their working time correcting skill-based errors, 32% were rule-based errors, and 3% as knowledge-based errors. Based on incident reports, researchers^[16] assessed that the reporting skill-based errors was more reliable than reporting rule- and knowledge-based errors. Subsequently, they^[17] examined a larger data set and revealed that only skill-based errors were related to occupational accidents. In addition, they^[18] reported that memory lapses, rule violations and knowledge-based mistakes were the most commonly identified human errors made by aircraft mechanics.

Skill-based errors were the most common unsafe act encountered in Australian mines. Inadvertent or missed operations were the most general types of skill-based errors. These errors were typically the result of a breakdown in visual monitoring or the inadvertent activation of a control^[19].

Human error and accidents

In everyday life, it is generally believed that human errors can cause injuries. This is confirmed by empirical studies.

It is generally accepted that 80-90% of accidents are due to human error^[20]. For example, approximately 70% of aircraft accidents have been attributed to human error^[21]. In a Finnish study, human errors were involved in 84% of serious accidents and in 94% of fatal accidents^[22]. However, statements on the magnitude of human error contribution to accidents highly depend on differences in taxonomy and approaches used among researchers^[23].

Factors leading to human errors

According to Reason, human errors are shaped by situational and task factors that are part of the environment in which the person is functioning^[24]. For example, in a Japanese train company, drivers who made errors were required to participate in a mandatory training class. In order to avoid this “penalty” – a loss of face - the drivers did not report any mistakes. This practice led to over 100 fatalities in commuter train accidents^[25]. Thus, this organisational measure to criminalise drivers who had made a human error (by forcing them to participate to a training class) resulted in even more fatalities.

A similar effect can sometimes occur when implementing a Zero Accidents Vision approach. While the intention to prevent all accidents is commendable, excessive pressures, whether deliberate or not, can encourage employees and/or middle management not to report some accidents in order to avoid direct or indirect sanctions. This can lead to a failure to address the causes of accidents which can later lead to more serious consequences.

Factors related to the individual, the job, the environment or the organisation can increase or decrease the likelihood of human error or increase or decrease human performance^[26]. These performance influencing (or shaping) factors describe any condition that influences performance and affects the way information is processed, and decisions are made^[27]. The Health and Safety Executive^[28] provides a non-exhaustive list of performance influencing factors divided into three broad categories. Good OSH management can optimise performance influencing factors and reduce the likelihood of all types of human errors.

Source^[28]

Conclusion

Human error in the workplace is a common phenomenon, it can cause disturbances and accidents at work. 

Human errors are typically results of long chains of events, and preventing human error in workplaces requires different types of preventive actions such as establishing clear procedures and guidelines, providing skills training, safe workplace design, managing workload to reduce fatigue, providing appropriate equipment, etc.  These actions help create a resilient system that minimises the likelihood and impact of human error.