Keywords

1 Introduction

Education lacks a unifying theoretical perspective, but rather derives its principles from a range of associated scientific sources to provide both its philosophy and experimental methodology. This is important, as every education endeavour is inevitably situated within the confines of its chosen sources, which perhaps explains why education remains a subject around which argument frequently rages. This paper particularly references medical education, but the principles are generalizable to other educational domains. While Mental Workload has developed as a theoretical concept with validated measures and proved itself to provide unique insights into human performance [1], it has had little impact on the measurement of educational outcomes. Cognitive load theory has an established literature [2] and uses the principle that excessive mental workload impairs learning. It has been used to ensure that educational materials both present critical information in easily understood formats and to reduce the inclusion of extraneous information. A critical weakness in this approach is that the experimental evidence is based on classroom-based educational activities with outcomes measured by performance in written assessments. While such an approach may be valid within purely academic disciplines, it must be questioned whether such an approach is applicable to educational activities which aim to prepare students for real world activities. An example of the above is the difference between a study of tracheal intubation (a key clinical skill) in the laboratory, which suggested that it is possible to train a complete novice using three attempts in the laboratory using success as the criteria [3] compared to a real world study of experienced staff who required an average of 76 attempts to master the same skill. The role of complexity in educational design will be further explored below.

This paper is an attempt to outline a coherent philosophy which can guide its use in educational activities with practical real world outcomes. Unfortunately, the huge range of educational philosophies makes it impossible to include them all in a single paper and I have, therefore, chosen three philosophies to illustrate what such a philosophy would imply. In each case, the philosophy is described in terms of its theoretical source, its view of learning, the teaching/learning methods adopted, the assessment tools used and its chosen outcomes.

Firstly, the philosophy of Behaviourism [4] is based on laboratory studies of (often animal) behaviour in response to changes in their environment, with the work of Pavlov [5] on dog salivation as perhaps the best known. Learning is viewed as a ‘conditioned response’ to an environmental stimulus with the organism (learner) changing their behaviour in response to a reward or punishment. The educational process is therefore simple; repeatedly present the learner with a problem and if the answer is right, provide a positive stimulus and if the answer is wrong, a negative stimulus. In the past this may have been characterised as a ferocious teacher teaching mathematics using frequent application of the edge of a ruler on the knuckles and a rare sweet. However, the same principles apply to learning high level skills such as sports, where athletes usually repeat the same processes over and over, receiving feedback from their coach. Assessment is therefore a simple matter of providing the correct stimulus and measuring how often the learners pick the correct response. The lab rats press a lever in the cage and the human learners pick the correct answer in multiple choice tests. Success is a simple matter of measuring the % of correct responses and with a direct correlation between % correct and educational success, often published in the media as ‘pass rates’, ‘A grades’ or ‘core indicators’. While such an approach is undoubtedly effective, it is frequently denigrated by higher level educators as mindless ‘training’ which can produce high levels of performance within a narrow domain, but which does not develop understanding, wider knowledge or the ability to transfer skills to other domains. The view of learners as little more than lab rats to be punished or rewarded sits uncomfortably with professional values.

Secondly, the philosophy of ‘competence’ [6] developed from workplace based analysis of tasks splits any complex task into a series of ‘competences’ which can be taught and then assessed in a structured and objective way. For example, during the induction of anaesthesia, the doctor needs to insert a cannula (tube) into a vein to give drugs, needs to communicate effectively with the patient and needs to administer an appropriate dose of drug. These tasks can be taught and tested separately, so for example, the insertion of cannula can be tested in the laboratory on a plastic arm, communication can be tested with an actor in an office and the dosage of drugs tested with a paper based exercise in an examination setting. The abilities of learners can therefore be easily tested in a series of structured assessments with each competency achieved ‘ticked off’ and when all the competencies have been achieved, then the learner is defined as fully competent and fit to practise. Unfortunately, for many practitioners, this logic is a flawed. For example, if you assume that if any musician can demonstrate that they can play each note of their instrument in turn, comment on some recorded music and also identify all the notation marks on a musical score, they are competent to play in an orchestra, you are likely to be disappointed. Complex tasks are far more than the sum of their components and need to be assessed as wholes.

Thirdly, the concept of Situated Learning [7], developed from ethnographic studies and views the learner starting as an outsider and then gradually becoming a full member of a community. The focus is therefore not so much on the acquisition of specific knowledge or skills, but rather on personal change and development. The primary learning activity is therefore working within the community alongside peers/mentors and with success attained when the individual is accepted by the community as one of their own. The medieval apprenticeship with the apprentice working alongside masters and being finally accepted once they produce their own ‘masterpiece’ perhaps illustrates this philosophy best. While such a philosophy has obvious attractions, its lack of objectivity and scope for excluding minority groups means that it is rarely used in professional settings as the sole means of assessment.

Although the above are described as distinct and mutually exclusive philosophies, in practice, they are often combined. While such combinations are understandable in response to external pressures such as cost, the need to provide a hard pass/fail boundary and the need to defend decisions against legal challenge, the results are predictably incoherent. For example, trainee anaesthetists, who are already qualified doctors with at least two years’ work experience are required to join a training programme [8] (apprenticeship) and work alongside senior staff (masters) who provide mentorship and ultimately decide whether the learner has achieved the level of skill required to be accepted to the community of (master) anaesthetists. However, during this period, they are also required to learn vast amounts of abstract knowledge and pass multiple choice examinations, as well as being required to demonstrate specific competencies in real world and laboratory settings. The result is entirely predictable, with some trainees rated as excellent by senior staff who fail the exams and others who are regarded as incompetent who nevertheless pass the exams. The result is frustration, wasted resources and a failure to reliably discriminate between those who will perform to a high standard in the workplace and those who will not.

2 Mental Workload as Educational Theory

The question then, is how Mental Workload should be viewed as an educational philosophy?

Wicken’s Multiple Resource Theory [9] is widely cited within the Mental Workload literature and provides a framework on which to base a wide range of assessment methodology. However, it also underpins a philosophy of perception and cognition which can be brought into the educational domain. These can be expressed as:

  • The world is a highly complex and rapidly changing environment which cannot be fully comprehended by our limited perceptual resources.

  • Our senses can be trained to convert the overwhelming complexity of reality into a manageable and coherent ‘perception’ of our world. However, by necessity that perception is fragmentary and may be inaccurate.

  • Our decisions are largely driven by subconscious processes, with consciousness perhaps seen as a largely retrospective and rationalising process.

  • Our responses to our environment are highly dependent on learned and complex patterns of response which coordinate a wide range of resources to achieve each task.

For example, this can easily be applied to the role of a pilot in combat. The environment is complex and there are a lot of things to do, such as fly, navigate and use weapons. Pilots gradually learn strategies to simplify their world, so that radar contacts are either ‘friend’ or ‘foe’ and firing weapons, at least initially, becomes a simple matter of eliminating a ‘foe’. Pilots also learn that the majority of rapid responses cannot be thought through, but have to become automated responses to specific problems. In the same way, responses cannot be theoretical notions of how to respond, but rather as highly automated psychomotor patterns learned through long practice. However, the same principles can be applied to a doctor in General Practice seeing a patient, with for example, an older lady with back pain. While the symptoms are usually simple, a consultation provides a wealth of verbal and nonverbal cues as to the severity and nature of the pain, so that the way the patient opens the clinic door, their gait across the room, the way they sit down and their facial expression as they describe the pain may be far more informative than the symptoms as described. Further, doctors need years of practice to learn how to notice these clues, but also may miss or misinterpret clinical information. The subconscious nature of these processes is readily evident in the common descriptions of doctors who refer patients for further investigation because they ‘knew something was wrong’ (rather than what) and those who describe the identification of a diagnosis ‘before the patient even sat down’. However, the two examples above provide a key distinction in the work of those investigating the use of Mental Workload as a method to improve performance and that difference is in how we view the complexity of our environment.

In high risk industries, such as aviation, complexity has been seen as a challenge to the limited capacity of the human mind, with the result that it has been systematically designed out of high risk processes [10]. The main application of studies involving mental workload has been to identify processes where cognitive demand are too high so that the workload can be reduced, for example, through simplification, workload sharing or better training. The most widely known example in the UK is perhaps the ban on the use of mobile phones while driving, partly at least based on simulation studies using a secondary task methodology [11]. In the above cases, the theoretical stance taken supports the use of laboratory or simulation based studies in that the task and operator actions can usually be reduced to a small number of defined correct and incorrect options and the environment can be reproduced to a high degree of fidelity. If the mental workload of subjects is measured within a simulator, then interventions which produce a reduction in their workload can be accepted as a successful intervention. For example, if redesigning a computer interface reduced the workload of operators, then we can predict that the performance of operators will increase and their error rates will decline [12].

In contrast, within domains such as medicine, the environment, actions and possible outcomes are in most cases very poorly defined and the mental workload associated with the task can be extremely high [13]. As noted above, a doctor in a clinic setting is required to monitor a wide range of verbal and nonverbal clues, to consider whether each clue is relevant to an existing diagnosis, to consider whether a new diagnosis is possible, to run through the usual questions, to develop new questions depending on previous answers and at the same time, to appear calm, sympathetic and to respond appropriately to whatever they are told. It is perhaps no wonder that patients complain that doctors ‘don’t listen’ as they are engaged in so many different tasks [14]. Although the distinction between the ‘complexity’ of aviation and the ‘complicatedness’ of medicine can be seen as medical hubris, its reality is evidence by the failure of simulation to produce improved clinical performance in many studies [15,16,17]. The key suggestion here is that in complex systems it is still possible to design training and assessment systems which closely link to operator performance; for example, pilots can be tested on a simulator using outputs like approach speed and altitude, with a high certainty that those outputs will predict real world performance. In contrast, in complicated domains like medicine, where there are no high fidelity simulators and few agreed outputs, it is not possible to use the same educational strategies.

As an educational philosophy, therefore, learning is seen as the ability to deal with complexity without excessive mental workload. That is, someone who has learned effectively can make sense of complex information/situations and respond effectively without evidence of mental workload, with the maximum complexity that can be accommodated as a measure of performance. So for example, the performance of a doctor could be measured by the complexity of patients with which they can cope and the performance of a musician with the complexity of the musical score they can play. While the units with which complexity are measured are domain-specific, the principles are universal.

3 Practical Application to Education

What is suggested here is that our view of complexity is the key to the use of mental workload studies in the educational domain. If complexity is viewed as a problem, then it suggests that educational processes should be simplified with material introduced to learners within controlled environments and with the complexity of the material gradually increased once they have mastered the basic elements. In contrast, if we view complexity as the essential characteristic of the domain, then it suggests that exposing learners to that complexity at the earliest possible stage and maximising their exposure to it during their educational is essential. This approach is supported by studies which appear to characterise expertise as ‘knowing where to look’ rather than ‘knowing what to look for’ [18], and those which characterise expertise as highly context dependant [19]. This conclusion is controversial in that most educational processes have relied in the past on the principle that learners must ‘understand the basic principles’ before they engage in realistic practice. This has led to, for example, a medical student being ejected from an operating theatre because they hadn’t studied enough anatomy on the basis that if they didn’t know anatomy, they couldn’t learn from the experience of observing surgery. It’s difficult to determine whether some of this resistance is based on educational belief, on a feeling that anyone without appropriate professional knowledge is not worthy to be admitted to a professional environment or it is a case of ‘I had to go through two years of lectures to get here, so you should too’.

An adherence to existing methods and assessment techniques also obscures the problem that the accepted measures of student performance in medical schools, such as the results of entrance interviews, multiple choice exams, essays and projects do not predict future performance as a doctor and do not even correlate well with each other [20]. What we do know is that ‘knowing’ and ‘understanding’ or even ‘doing in a laboratory setting’ are not adequate to define ‘safe to practice’. The explanation suggested here is that they fail to predict future performance because they provide a challenge which avoids the complexity of the real world.

In other high risk domains such as aviation, this problem has been addressed through the development of simulation [21], so that learners can be developed through a programme of part task simulators which mimic small aspects of a single task in real-time, such as navigation computers, to full scale, high fidelity simulators which allow an entire day’s work to be completed in real time. For many industries, including healthcare, simulators have limited utility, for three reasons [22]. Firstly, people do not come with a manual and individual people do not always behave the way you would expect. This means that people can suddenly behave in unexpected and unusual ways, but also, their physiology can also confound expectation. When learners have to deal with real people, they have to be prepared for the unexpected and it is hard to design simulators that do not behave predictably. Secondly, most industries where simulation has become embedded deal with information provided either verbally or in the form of instrumentation, both of which can be simulated relatively easily. In contrast, the simulation of skin colour, joint movement or nonverbal communication are extremely difficult and unlikely to be simulated effectively in the near future. Thirdly, pilots are few in number and the cost of flying an aircraft for training is very high, making simulation a financially viable option. Healthcare workers are numerous and training on the job is cheap, making simulation training financially impossible. While simulation provides obvious advantages, it can prove to be prohibitively expensive to implement in low income areas [17]. A shift to mental workload as an educational outcome can shift the focus from how a simulator works (anatomical fidelity) and works (functional fidelity) toward more simple devices which just target the effectiveness of training (psychological fidelity) [23, 24].

Mental workload provides us with a new way of looking at the problem of assessing the performance of individuals dealing with complex environments in that the ability of an individual to deal with a complex situation without becoming cognitively overloaded can be used to define expertise [1]. This is characterised by sensory systems which effectively reduce the complexity to comprehensible concepts, an ability to make sense of those concepts as a whole as well as the ability to rapidly produce a coordinated and effective response. The difference between the novice and expert is perhaps most graphically described by those involved in large scale conflicts, where those new to combat are described as becoming overwhelmed by the situation and becoming completely paralysed and unable to respond at all. In contrast, with even a few days experience, soldiers become able to deal with the experience and start to respond appropriately [25].

4 Links to Other Education Theories

This concept of expertise links closely to the concept of the four stages of competence [26], with novices moving from Unconscious incompetence, conscious incompetence, conscious competence and finally unconscious competence. However, while the label of ‘unconscious competence’ is correct in that the primary reason for the improved performance is the development of subconscious cognitive processes which allow the operator to complete the task without conscious effort (low mental workload), the suggestion here is that the process of developing expertise does not pass through four different developmental stages. Rather, the subconscious processes are the key elements of expertise and are developed from the beginning of the training process. The conscious monitoring of action and the decision to practice are obviously important to make sure an individual continues to train, but is not an essential part of the development of expertise. For example, those learning to play the piano usually go through a process of being told about musical notation, learning how to convert the notes to key strokes and eventually, at the highest level, being told to ‘forget the notes and just play the music’. However, there are many people who have just sat down at a piano and learned to play music without any instruction or theory. That is, they develop unconscious competence without going through a stage of unconscious incompetence or conscious competence. In this context, the use of the term ‘competence’ is also problematic as noted above.

Often linked to the concept of unconscious incompetence is Miller’s pyramid of competence [27], which describes an increase in performance from Knows, Knows How, Shows How and Does. That is a learner progressing from knowing theory, to being able to describe how something would be done in practice, to being able to demonstrate how and finally being able to complete the task in the real world. However, the problem with this description of learning is that it is a greater reflection of the teaching methods commonly used than the development of expertise. That is, learners are usually taught the theoretical principles first, the practicalities next and then taken through simple practical teaching before being allowed to move into the workplace. The example of the self-taught pianist readily demonstrates that it is not necessary to acquire theoretical knowledge before engaging in practice. However, an important distinction must be made here between what is necessary for effective learning and what is possible in the real world. For example, while it might be acceptable to allow a novice free rein on a piano to learn through experimentation, the same would not apply to those learning how to control a nuclear reactor. The suggestion is that Mental Workload can provide the basis for a new educational concept of expertise in that learning is a process by which we learn to respond to complexity without developing cognitive overload. The level of educational attainment achieved is therefore defined by the complexity/speed of the task which can be completed successfully. ‘Command’ appears an appropriate term to describe an individual who has achieved a level of expertise at which they have acquired the ability to cope with the cognitive demands of the most difficult predicted situation and yet still maintain a cognitive workload within their capacity so that they retain spare capacity to monitor their own performance. Such an individual would therefore be able to explain and justify their decisions after the event. While this concept is perhaps most easily applied to pure psychomotor tasks, such as typing, where expertise can be defined in terms of ‘words per minute’, it can also be applied to much more complex tasks. An objection to this definition could be that professional occupations perceive their own occupations as far more than the successful completion of tasks, but this does not negate the concept of mental workload as an educational outcome.

5 Practical Implementation

The follow will describe previous studies using mental workload within an educational setting to demonstrate these principles. As noted above, a medical consultation provides a complex cognitive challenge for even experienced healthcare workers due to the need to ask questions, monitor the answers, watch for nonverbal clues, make diagnoses, empathise etc. There are few studies of mental workload during consultations, but they appear to confirm that the consultation is a high workload task [28]. An interesting observation is that the observed levels of workload appear consistent with periods of overload that could be associated with significant levels of error or poor performance.

A recent study [unpublished] in medical students measured the effect of combining a listening task with the physical task of taking a blood sample. While the simple task of listening to a recording of clinical information appeared to be a low workload task, students taking a blood sample showed evidence of cognitive overload. The more interesting finding was that when the two tasks were combined, there was no increase in mental workload, which was unexpected. However, the performance of the students on the listening task deteriorated. The explanation appears to be that in response to the excessive cognitive workload, the students ‘shed’ the listening task in order to be able to complete the physical task of taking blood. If confirmed in future studies, the implication is that during a consultation, healthcare workers are routinely overloaded and are likely to ignore patients in order to complete tasks such as filling in forms or entering data into a computer. This would provide an explanation of the problem that patients frequently complain that healthcare workers don’t listen to them and suggest that rather than staff being uncaring, they are simply overwhelmed with the number of tasks they have to perform.

Studies in medicine have largely used subjective reporting of workload using the NASA-TLX scale [29], and secondary task methods using mental arithmetic/visual change/vibrotactile stimulus as outcomes. Although the methodology varies, all studies have used deterioration in second task performance as evidence of cognitive overload. These studies confirm that the mental workload of staff correlates well with their perception of the difficulty of the tasks, with, for example, higher levels of mental workload during the induction of anaesthesia and emergence from anaesthesia (often compared to takeoff and landing for pilots) and lower during the maintenance (cruise) phase. Although the average levels of workload appear closely linked to the task, the workload of individuals appears to vary widely, which was an unexpected finding and one at odds with studies in other fields. The explanation may be that while admission to many high risk occupations depends on a selection process which includes a range of psychomotor and cognitive tasks, admission to medicine is almost exclusively based on written examinations and interviews. It is therefore possible for an individual with cognitive problems with psychomotor coordination (dyspraxia) or other limitations to qualify as a doctor who would never qualify, for example, as a pilot. This suggests that the inclusion of mental workload as an educational outcome could provide learners with valuable feedback on their aptitude for a variety of medical specialities at an early stage of their career and avoid them investing years in a speciality for which they are unsuited.

A study looking at the mental workload of anaesthetists during their seven year training programme confirmed that delivering anaesthesia appears to be a high workload task and confirmed that training reduces mental workload, with a near linear decrease in workload in relation to the number of years training completed [30]. Interestingly, those tested in their final year of training, who should have completed all the compulsory aspects of their training and were engaged in training for sub-specialist tasks did not show evidence of cognitive overload, suggesting that a seven year programme successfully trains subjects to an appropriate level of expertise and that this expertise can be reliably determined by measuring mental workload. The interesting finding from this research was that the mental workload of qualified, permanent staff was also measured and was shown to be higher than that of final year trainees. This unexpected finding indicates that either the expertise of qualified staff had deteriorated once they had completed their training or could be associated with a decrease in cognitive capacity with age, as qualified staff were inevitably older than trainees. This suggests that healthcare workers in high cognitive load tasks could face significant and different challenges as they get older.

In addition to the challenges of healthcare working, we have observed that the assessment process can also be challenging for observers. For example, the most common assessment of clinical practice involves the observation of a learner by a trained observer, who then ticks off each aspect of the performance that has been completed. A study of observers during a practice exam suggested that the mental workload of the observers was indeed very high, which is explained by the need to observe for up to 28 different actions, to rate the completeness of each action and then to record the action on a chart [31]. Importantly, a follow-up study showed that a short training period had no effect on either the accuracy or mental workload of observers which supports the concept that the performance of complex tasks depends on subconscious processes developed through long periods of practice rather than simply understanding or knowing what to do [32].

6 Conclusion

In conclusion, this paper suggests that human performance should be viewed as three tightly-knit and largely subconscious processes: a highly sophisticated analysis of sensory input to derive comprehensible wholes, the synthesis of those wholes into situational awareness and the coordinated response to the situation. Expertise is defined as the ability to complete those three processes in response to a complex situation without cognitive overload and professional expertise requiring the ability to consciously justify the actions taken. This level of performance is characterised by not just the ability to perform at an expert level, but to also retain enough spare cognitive capacity to monitor the performance and adapt as necessary and therefore should be regarded as achieving ‘Command’. Mental workload therefore provides a unifying theory of educational activity with well-developed links to appropriate assessment methodologies.